DRAFT - DO NOT CITE OR QUOTE                  EPA/635/R-08/011A
                                          External Review Draft
f/EPA
          TOXICOLOGICAL REVIEW

                            OF

              Tetrachloroethylene
              (Perchloroethylene)
                       (CAS No. 127-18-4)
           In Support of Summary Information on the
           Integrated Risk Information System (IRIS)
                          June 2008
                           NOTICE

This document is an External Review Draft. This information is distributed solely for the
purpose of pre-dissemination peer review under applicable information quality guidelines. It has
not been formally disseminated by EPA. It does not represent and should not be construed to
represent any Agency determination or policy. It is being circulated for review of its technical
accuracy and science policy implications.
                  U.S. Environmental Protection Agency
                         Washington, DC

-------
                                     DISCLAIMER

       This document is a preliminary draft for review purposes only.  This information is
distributed solely for the purpose of pre-dissemination peer review under applicable information
quality guidelines.  It has not been formally disseminated by EPA. It does not represent and
should not be construed to represent any Agency determination or policy.  Mention of trade
names or commercial products does not constitute endorsement or recommendation for use.
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    ii       DRAFT—DO NOT CITE OR QUOTE

-------
CONTENTSCTOXICOLOGICAL REVIEW for TETRACHLOROETHYLENE
(PERCHLOROETHYLENE) (CAS No. 127-18-4)
LIST OF TABLES	ix
LIST OF FIGURES	xii
LIST OF ABBREVIATIONS AND ACRONYMS	xvi
FOREWORD	xix
AUTHORS, CONTRIBUTORS, AND REVIEWERS	xx

1.  INTRODUCTION	1
   REFERENCES FOR CHAPTER 1	3

2.  BACKGROUND	1
    2.1. USES AND PHYSICAL/CHEMICAL PROPERTIES	1
    2.2. OCCURRENCE AND EXPOSURE	1
        2.2.1. Air	1
        2.2.2. Water	4
        2.2.3. Food	5
        2.2.4. Breast Milk	6
        2.2.5. Direct Ingestion	7
   REFERENCES FOR CHAPTER 2	8

3.  TOXICOKINETICS	1
    3.1. ABSORPTION	1
        3.1.1. Inhalation	1
        3.1.2. Oral	2
        3.1.3. Dermal	2
    3.2. DISTRIBUTION AND BODY BURDEN	2
    3.3. METABOLISM	4
        3.3.1. Introduction	4
        3.3.2. Extent of Metabolism	4
        3.3.3. Pathways of Metabolism	5
              3.3.3.1. Cytochrome P450-Dependent Oxidation	6
              3.3.3.2. Glutathione (GSH) Conjugation Pathway	12
              3.3.3.3. Relative Roles of the Cytochrome P450 (CYP) and Glutathione
                    (GSH) Pathways	18
        3.3.4. Susceptibility	19
        3.3.5. Comparison of Tetrachloroethylene Metabolism with Trichloroethylene
              Metabolism	20
    3.4. EXCRETION	20
    3.5. PHYSIOLOGICALLY BASED AND OTHER TOXICOKINETIC MODELING	23
        3.5.1. Various Physiologically Based Pharmacokinetic (PBPK) Models	23
        3.5.2. Variability and Uncertainty	28


          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                iii     DRAFT—DO NOT CITE OR QUOTE

-------
                              CONTENTS (continued)

         3.5.3. Animal-to-Human Extrapolation Using a Physiologically Based
               Pharmacokinetic (PBPK) Model	33
              3.5.3.1.  Choice of Physiologically Based Pharmacokinetic (PBPK) Model	33
              3.5.3.2.  Implementation of Physiologically Based Pharmacokinetic
                      (PBPK) Models	36
         3.5.4. Comparison of Physiologically Based Pharmacokinetic (PBPK)
               Simulations With Experimental Data	39
         3.5.5. Physiologically Based Pharmacokinetic (PBPK) Model Comparisons
               and Interspecies Differences	47
         3.5.6. Metabolic Interactions With Other Chemicals	56
   APPENDIX FOR CHAPTER 3: COMPARISONS OF TETRACHLOROETHYLENE
     METABOLISM WITH TRICHLOROETHYLENE METABOLISM	58
   REFERENCES FOR CHAPTER 3	62

4.  HAZARD IDENTIFICATION	1
    4.1.  OVERALL APPROACH	1
    4.2.  OVERVIEW OF TETRACHLOROETHYLENE METABOLISM	1
    4.3.  GENOTOXICITY	3
    4.4.  LIVER TOXICITY	7
         4.4.1. Human Effects	7
              4.4.1.1.  Liver Damage	7
              4.4.1.2.  Liver Cancer	11
         4.4.2. Animal  Studies	11
              4.4.2.1.  Liver Toxicity	11
              4.4.2.2.  Liver Cancer	13
         4.4.3. Summary of Liver Effects in Humans and Animals	15
         4.4.4. Mode of Action for Liver Toxicity	16
              4.4.4.1.  Background	17
              4.4.4.2.  Relationship of Metabolism to Potential Mode of Action and
                      Organ Toxicity	18
              4.4.4.3.  Description of a Hypothesized Mode of Action (MOA):
                      Peroxisome Proliferator-Activated Receptor (PPAR) Mediated
                      Hepatocarcinogenesis	19
              4.4.4.4.  Effects That Could Be Related to Other Potential Modes of Action. ...32
              4.4.4.5.  Summary and Conclusions	34
    4.5.  KIDNEY TOXICITY	35
         4.5.1. Human  Studies	35
              4.5.1.1.  Kidney Toxicity in Humans	35
              4.5.1.2.  Kidney Cancer	37
         4.5.2. Animal  Studies	39
              4.5.2.1.  Kidney Toxicity in Animals	39
              4.5.2.2.  Kidney Cancer in Animals	40
         4.5.3. Summary of Kidney Effects in Humans and Animals	41

          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                 iv       DRAFT—DO NOT CITE OR QUOTE

-------
                               CONTENTS (continued)

         4.5.4. Mode of Action for Kidney Toxicity and Carcinogenicity	42
               4.5.4.1.  Background	42
               4.5.4.2.  Summary Description of a Postulated Mode of Action—
                      alpha-2u-globulin Accumulation	43
               4.5.4.3.  Other Modes of Action for Tetrachloroethylene-Induced Renal
                      Tumors in Rats	48
               4.5.4.4.  Summary	51
    4.6. NEUROTOXICITY	52
         4.6.1. Human Studies	52
               4.6.1.1.  Environmental Chamber Studies	58
               4.6.1.2.  Chronic Exposure Studies	63
               4.6.1.3.  Summary of Neuropsychological Effects in Low- and Moderate-
                      Exposure Studies	96
         4.6.2. Animal Studies	104
               4.6.2.1.  Inhalation Studies	104
               4.6.2.2.  Oral andIntraperitoneal Studies	Ill
         4.6.3. Summary of Neurotoxic Effects in Humans and Animals	114
         4.6.4. Mode of Action for Neurotoxic Effects	117
    4.7. DEVELOPMENTAL/REPRODUCTIVE STUDIES	118
         4.7.1. Human Studies	118
         4.7.2. Animal Studies	124
               4.7.2.1.  Summary of Animal Studies	134
         4.7.3. Summary of Human and Animal Developmental/Reproductive Studies	134
         4.7.4. Mode of Action for Developmental Effects	137
    4.8. TOXIC EFFECTS IN OTHER ORGAN SYSTEMS	138
         4.8.1. Human Studies	138
               4.8.1.1.  Noncancer Effects	138
               4.8.1.2.  Cancer	146
         4.8.2. Animal Studies	156
               4.8.2.1.  Noncancer Effects	156
               4.8.2.2.  Cancer Effects	158
         4.8.3. Summary of Immunotoxicologic Effects in Humans and Animals and
               Potential Mode of Action	163
    4.9. SUSCEPTIBLE POPULATIONS	164
         4.9.1. Life Stages	165
               4.9.1.1.  Life Stage-Specific Exposures	165
               4.9.1.2.  Early Life Stage Effects	166
               4.9.1.3.  Later Life Stages	173
         4.9.2. Other Susceptibility Factors	173
               4.9.2.1.  Health and Nutritional Status	173
               4.9.2.2.  Gender	174
               4.9.2.3.  Race/Ethnicity	175
               4.9.2.4.  Genetics	175

           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   v        DRAFT—DO NOT CITE OR QUOTE

-------
                              CONTENTS (continued)

         4.9.3. Multiple Exposures and Cumulative Risks	176
         4.9.4. Uncertainty of Database for Susceptible Populations	178
              4.9.4.1.  Uncertainties of Exposure	178
              4.9.4.2.  Uncertainties of Effects	180
         4.9.5. Conclusions on Susceptibility	180
    4.10. SUMMARY OF HAZARD IDENTIFICATION	182
         4.10.1.  Description of Effects and Exposure Levels at Which They Occur	182
              4.10.1.1. Summary of Effects in Humans	182
              4.10.1.2. Summary of Effects in Animals	187
              4.10.1.3. Summary of Effect Levels	189
         4.10.2.  Characterization of Cancer Hazard	190
              4.10.2.1. Background	190
              4.10.2.2. Hazard Characterization for Tetrachloroethylene	191
         4.10.3.  Mode-of-Action Summary	193
         4.10.4.  Rationale for Select!on of Dose Metric	201
              4.10.4.1. Liver	201
              4.10.4.2. Kidney	201
              4.10.4.3. Hematopoietic Target Organ	202
              4.10.4.4. Central Nervous System	203
   APPENDIX 4A: CONSISTENCY OF TETRACHLOROETHYLENE AND
      TRICHLOROACETIC ACID HEPATOCARCINOGENICITY	204
   APPENDIX 4B: HUMAN STUDIES OF CANCER	215
   REFERENCES FOR CHAPTER 4	219

5.  DOSE-RESPONSE EVALUATION	1
    5.1. INHALATION REFERENCE CONCENTRATION (RFC)	1
         5.1.1. Choice of Principal Study and Critical Effect	1
         5.1.2. Method of Analysis	10
         5.1.3. Reference Concentration (RfC) Derivation, Including Application of
               Uncertainty Factors	11
         5.1.4. Supporting Studies	15
         5.1.5. Previous Inhalation Assessment	19
    5.2. ORAL REFERENCE DOSE (RFD)	20
         5.2.1. Choice of Principal Study and Critical Effects	20
         5.2.2.  Methods of Analysis, Including Models	21
         5.2.3.  Reference Dose (RfD) Derivation, Including Application of Uncertainty
               Factors	24
         5.2.4. Supporting Studies	27
         5.2.5. Previous Oral Assessment	29
    5.3. UNCERTAINTIES IN INHALATION REFERENCE CONCENTRATION
         (RFC) AND ORAL REFERENCE DOSE (RFD)	30
         5.3.1. Point of Departure	32
         5.3.2. Extrapolation from Laboratory Animal Studies to Humans	32

          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                 vi       DRAFT—DO NOT CITE OR QUOTE

-------
                              CONTENTS (continued)

         5.3.3.  Human Variation	33
         5.3.4.  Database Uncertainties	33
    5.4. CANCER DOSE-RESPONSE ASSESSMENT	34
         5.4.1.  Choice of Study/Data with Rationale and Justification	35
         5.4.2.  Dose-Response Data	36
               5.4.2.1. Liver Tumors in Mice	36
               5.4.2.2. Mononuclear Cell Leukemia in Rats	41
               5.4.2.3. Other Tumor Sites in Male Rats	45
         5.4.3.  Estimation of Dose Metrics for Dose-Response Modeling	46
               5.4.3.1. Dose Metric for Hepatocellular Carcinogenicity	46
               5.4.3.2. Dose Metric for Rat Leukemias and Kidney Tumors	48
               5.4.3.3. Dose Metric for Sites Not Addressed by Physiologically Based
                      Pharmacokinetic (PBPK) Modeling	49
         5.4.4.  Extrapolation Methods	49
               5.4.4.1. Dose-Response Models and Extrapolation to Low Doses	49
               5.4.4.2. Extrapolation to Human Equivalent Environmental Exposure	51
         5.4.5.  Cancer Risk Values	55
               5.4.5.1. Dose-Response Modeling Results	55
               5.4.5.2. Recommended Inhalation Unit Risk	71
               5.4.5.3. Recommended Oral Slope Factor	73
               5.4.5.4. Quantitative Adjustment for Sensitive Populations	73
         5.4.6.  Discussion of Uncertainties in Cancer Risk Values	75
               5.4.6.1. Sources of Uncertainty	75
               5.4.6.2. Summary and Conclusions	85
   APPENDIX 5A: BENCHMARK DOSE MODEL RESULTS	86
   APPENDIX 5B: PROBABILITY DISTRIBUTIONS OF CANCER RISK ESTIMATES	96
   REFERENCES FOR CHAPTER 5	102

6.  CHARACTERIZATION OF HAZARD AND DOSE-RESPONSE	1
    6.1. SUMMARY OF HUMAN HAZARD POTENTIAL	1
         6.1.1.  Exposure	1
         6.1.2.  Absorption, Metabolism, Distribution, and Excretion	1
         6.1.3.  Noncancer Toxicity in Humans and Laboratory Animals	4
         6.1.4.  Carcinogenicity in Humans and Laboratory Animals	5
         6.1.5.  Mode-of-Action Information	7
               6.1.5.1. Liver Mode-of-Action Information	7
               6.1.5.2. Kidney Mode-of-Action Information	8
               6.1.5.3. Mode-of-Action Information for Other Targets of Toxicity	8
               6.1.5.4. Mode-of-Action Conclusions and Implications for Dose-Response
                      Analyses	9
         6.1.6.  Weight-of-Evidence Descriptor for Cancer Hazard	9
          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  vii      DRAFT—DO NOT CITE OR QUOTE

-------
                              CONTENTS (continued)

    6.2. DOSE-RESPONSE CHARACTERIZATION	11
         6.2.1.  Noncancer Toxicity (Reference Concentration [RfC]/Reference Dose
               [RfD])	11
               6.2.1.1. Assessment Approach Employed	12
               6.2.1.2. Impact of Assumptions, Uncertainties and Alternatives on
                      Reference Concentration and Reference Dose	15
         6.2.2.  Cancer Risk Estimates	21
               6.2.2.1. Assessment Approach Employed	22
               6.2.2.2. Impact of Assumptions, Uncertainties and Alternatives on Unit
                      Risk Estimates	24
               6.2.2.3. Quantitative Analysis of Multiple Uncertainties on Cancer Unit
                      Risk	32
               6.2.2.4. Conclusions	35
   REFERENCES FOR CHAPTER 6	40

APPENDIX A:  SUMMARY OF EARLIER ASSESSMENTS	A-l
          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  viii      DRAFT—DO NOT CITE OR QUOTE

-------
                                   LIST OF TABLES


2-1.       Physical and chemical properties of tetrachloroethylene	2-2

3-1.       Comparison of Vmax/Km ratios	3-25

3-2.       Variation in values of metabolic parameters for tetrachloroethylene, as seen
          in the literature	3-30

3-3.       Ratio of average daily dose at various life stages to the average daily dose
          for a 25-year-old adult: PBPK simulations	3-31

3-4.       Parameters for tetrachloroethylene PBPK modeling	3-40

3-5.       Comparison of venous blood tetrachloroethylene concentrations: PBPK
          simulations and Altmann et al. (1990) study	3-46

4-1.       Summary of studies of human liver toxicity	4-10

4-2.       Summary of rodent liver toxicity studies	14

4-3.       Summary of human kidney toxicity marker studies in dry cleaners	4-38

4-4.       Summary of human neurotoxicology studies	4-54

4-5.       Summary of neuropsychological effects of tetrachloroethylene in humans	4-97

4-6.       Summary of animal inhalation neurotoxicology studies	4-109

4-7.       Summary of oral neurotoxicity animal studies	4-115

4-8.       Developmental/reproductive studies in humans	4-125

4-9.       Exposure concentrations (ppm) at which effects occurred in a two-
          generation study	4-132

4-10.      Summary of animal developmental/reproductive studies for
          tetrachloroethylene, in chronological order	4-135

4-11.      Immune-related conditions in studies of dry cleaning or tetrachloroethylene
          exposure in humans	4-142

4-12.      Summary of low-effect levels of exposure to tetrachloroethylene	4-189

4-13.      Summary of potential modes of action for cancer	4-196
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   ix       DRAFT—DO NOT CITE OR QUOTE

-------
                             LIST OF TABLES (continued)

4-14.      Quantitative implications of different modes of action: candidate modeling
          approaches	4-199

5-1.       Summary of rationale for principal study selection	5-3

5-2.       Inhalation studies considered in the development of an RfC	5-5

5-3.       Oral studies considered in analysis of the oral RfD	5-25

5-4.       Oral RfV: point of departure and uncertainty factors	5-26

5-5.       Tumor incidence and estimated metabolized doses in mice exposed to
          tetrachloroethylene	5-37

5-6.       Historical control data of the Japan Bioassay Research Center, Crj/BDFl
          mouse, 104-week studies	5-39

5-7.       Incidence of mononuclear cell leukemia, kidney tumors, and brain gliomas
          in rats exposed to tetrachloroethylene by inhalation	5-42

5-8.       Historical control data of the Japan Bioassay Research Center, F344/DuCrj
          (Fischer) rat, 104-week studies	5-43

5-9.       Dose-response modeling summary for tumor sites using total
          tetrachloroethylene metabolites as the dosimeter; tumor incidence data
          from JISA (1993) and NTP (1986)	5-62

5-10.      Human equivalent risk per unit concentration, in terms of continuous
          environmental exposure, derived using total tetrachloroethylene
          metabolites as the dosimeter; tumor incidence data from JISA (1993)
          and NTP (1986)	5-63

5-11.      Dose-response summary and cancer risk estimates using continuous
          equivalent administered tetrachloroethylene levels as dosimeter, from
          NTP (1986) and JISA (1993)	5-64

5-12.      Summary of tetrachloroethylene oral slope factors, estimated from dose-
          response modeling of inhalation-exposed animals and by extrapolation to
          oral exposure using pharmacokinetic models	5-74

5-13.      Summary of uncertainties in tetrachloroethylene cancer unit risk estimate	5-76

5-14.      Summary of considerations for each rodent tumor type	5-82
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   x        DRAFT—DO NOT CITE OR QUOTE

-------
                             LIST OF TABLES (continued)

6-1.       Summary of rationale for principal study selection	6-13

6-2.       Summary of dose-specific extra risks (means and 95% confidence limits)
          for four dose-response models fit to incidence of leukemias in male rats
          exposed to tetrachloroethylene via inhalation	6-27

6-3.       Summary of considerations for each rodent tumor type	6-30

6-4.       Combined impact on tetrachloroethylene cancer risk estimates (per |ig/m3)
          of statistical uncertainty, PBPK model and tumor site(s), using multistage
          model in observed range and linear low-dose extrapolation	6-34

6-5.       Considerations leading to the determination of a reasonable upper bound on
          risk	6-37

6-6.       Considerations that impact uncertainty in reasonable upper bound risk
          estimates	6-38
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xi       DRAFT—DO NOT CITE OR QUOTE

-------
                                  LIST OF FIGURES

3-1.       Postulated scheme for the metabolism of tetrachloroethylene by the P450
          oxidative pathway	3-7

3-2.       Metabolism of tetrachloroethylene by the glutathione conjugation pathway	3-13

3-3.       PBPK simulations of variations with age and gender in blood concentrations
          of tetrachloroethylene and its main metabolite trichloroacetic acid (TCA)	3-32

3-4.       Comparison of model predictions for blood concentration with inhalation
          experiment	3-41

3-5.       Comparison of model predictions for alveolar concentration of
          tetrachloroethylene with experimental data on humans	3-42

3-6a.      Comparison of model predictions for alveolar concentration as a fraction of
          inhaled tetrachloroethylene concentration with experimental Opdam and
          Smolders (1986) data on male human subjects	3-44

3-6b.      Comparison of model predictions for alveolar concentration as a fraction of
          inhaled tetrachloroethylene concentration with experimental data on female
          human subjects	3-45

3-7.       Comparison of tetrachloroethylene concentrations in blood in rats and humans.... 3-48

3-8.       Comparison of various model predictions of tetrachloroethylene blood
          concentration in humans and rats following steady state	3-49

3-9.       Model predictions of total tetrachloroethylene metabolites produced
          following a 6-hr inhalation exposure in rats, mice, and humans	3-50

3-10.      Model predictions of rate of total metabolism in humans at steady state	3-51

3-1 la.     Rate of metabolism in rat and human models:  time course for low exposure	3-53

3-1 Ib.     Rate of metabolism in rat and human models:  time course for high exposure	3-54

3-12.      Oral ingestion of tetrachloroethylene: blood concentration in humans
          versus time	3-55

3-13.      Rate of metabolism of tetrachloroethylene in humans: oral exposure	3-56
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xii       DRAFT—DO NOT CITE OR QUOTE

-------
                            LIST OF FIGURES (continuted)

4-1.       Visual contrast sensitivity functions for control and exposed children (top),
          adults that were identified as having impaired function (i.e., 5 of the total 11)
          and their matched controls (middle), and the control and exposed individuals
          over 60 years of age	4-85

4-2.       Summary of the relationship between LOAEL concentrations (ppm) and
          treatment duration (hours)	4-111

5-1.       Array of PODs and reference values for a subset of neurotoxic effects of
          studies in Table 5-2	5-16

5-2.       Organ-specific reference values for inhalation exposure to tetrachloroethylene .... 5-19

5-3.       Time course of venous blood concentration in humans as predicted by the
          Bois et al. (1996), Rao and Brown (1993), and Reitz et al. (1996) PBPK
          models for ingested tetrachloroethylene	5-23

5-4.       Oral organ-specific reference values for exposure to tetrachloroethylene	5-28

5-5.       Mouse liver tumor responses (hepatocellular adenomas  and carcinomas) for
          three chronic bioassays (Table 5-5), plotted against continuous equivalent
          concentration (ppm) and total tetrachloroethylene metabolism
          (mg-equivalents/kg-day), for male and female mice	5-40

5-6.       Rat mononuclear cell leukemia responses (minus control) in two chronic
          bioassays (Table 5-7), plotted against continuous equivalent exposure
          (ppm) and total tetrachloroethylene metabolites, in mg-equivalents/kg-day,
          for male and female rats	5-44

5-7.       Sequence of steps for extrapolating from tetrachloroethylene bioassays  in
          animals to human-equivalent exposures expected to be associated with
          comparable cancer risk	5-47

5-8a.      Incidence of hepatocellular adenomas and carcinomas in male mice (JISA,
          1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
          and multistage model fit showing BMC and BMCL at 10% extra risk	5-56

5-8b.      Incidence of hepatocellular adenomas and carcinomas in male mice (JISA,
          1993) corresponding to human equivalent continuous tetrachloroethylene
          exposure (ppm) and multistage model fit showing BMC and BMCL at
          10% extra risk	5-56
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xiii      DRAFT—DO NOT CITE OR QUOTE

-------
                           LIST OF FIGURES (continuted)

5-9a.      Incidence of hepatocellular adenomas and carcinomas in female mice
          (JISA, 1993) corresponding to total tetrachloroethylene metabolism
          (mg-eq/kg-day) and multistage model fit showing BMC and BMCL at
          10% extra risk	5-57

5-9b.      Incidence of hepatocellular adenomas and carcinomas in female mice
          (JISA, 1993) corresponding to human equivalent continuous
          tetrachloroethylene exposure (ppm) and multistage model fit showing
          BMC and BMCL at 5% extra risk	5-57

5-10a.     Incidence of malignant hemangiosarcomas in male mice (JISA, 1993)
          corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
          and multistage model fit showing BMC and BMCL at 10% extra risk	5-58

5-10b.     Incidence of malignant hemangiosarcomas in male mice (JISA, 1993)
          corresponding to human equivalent continuous tetrachloroethylene
          exposure (ppm) and multistage model fit showing BMC and BMCL at
          5% extra risk	5-58

5-1 la.     Incidence of mononuclear cell leukemia in male rats (JISA, 1993)
          corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
          and multistage model fit showing BMC and BMCL at 10% extra risk	5-59

5-1 Ib.     Incidence of mononuclear cell leukemia in male rats (JISA, 1993)
          corresponding to human equivalent continuous tetrachloroethylene
          exposure (ppm) and multistage model fit showing BMC and BMCL at
          5% extra risk	5-59

5-12a.     Incidence of mononuclear cell leukemia in female rats (JISA, 1993)
          corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
          and multistage model fit showing BMC and BMCL at 10% extra risk	5-60

5-12b.     Incidence of mononuclear cell leukemia in female rats (JISA, 1993)
          corresponding to human equivalent continuous tetrachloroethylene
          exposure (ppm) and multistage model fit showing BMC and BMCL at
          5% extra risk	5-60

5-13a.     Incidence of kidney adenomas and adenocarcinomas in male rats (NTP,
          1986) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
          and multistage model fit showing BMC and BMCL at 5% extra risk	5-61
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  xiv      DRAFT—DO NOT CITE OR QUOTE

-------
                            LIST OF FIGURES (continuted)

5-13b.    Incidence of kidney adenomas and adenocarcinomas in male rats (NTP,
          1986) corresponding to human equivalent continuous tetrachloroethylene
          exposure (ppm) and multistage model fit showing BMC and BMCL at
          5% extra risk	5-61

5-14.      Comparison of inhalation risks per unit concentration for tetrachloroethylene
          derived from rodent bioassays using four different dose metrics	5-72

5-15.      Illustration of sensitivity to model selection for low-dose extrapolation	5-80

6-1.       Array of PODs and reference values for a subset of neurotoxic effects in
          inhalation studies	6-16

6-2.       Organ-specific RfVs for inhalation exposure to tetrachloroethylene	6-17

6-3.       Oral organ-specific reference values for exposure to tetrachloroethylene	6-18

6-4.       Cancer risk estimates for tumor sites associated with tetrachloroethylene
          exposure in rodent bioassays, using the multistage model	6-33
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    xv      DRAFT—DO NOT CITE OR QUOTE

-------
                   LIST OF ABBREVIATIONS AND ACRONYMS
8-OHdG     8-hydroxydeoxyguanosine
AAP         alanine aminopeptidase
ALT         alanine transferase
AST         aspartase amino transaminase
ATSDR      Agency for Toxic Substances and Disease Registry
AUC         area-under-the-curve
BMC         benchmark concentration
BMCL       95% lower bound benchmark dose
BMD         benchmark dose
BMDS       Benchmark Dose Software
BMDU       95% upper bound benchmark dose
BUN         blood urea nitrogen
BW          body weight
CARB       California  Air Resources Board
CASRN      Chemical Abstracts Service Registry Number
CCI          Color Confusion Index
CI           confidence interval
CLL         chronic lymphocytic leukemia
CNS         central nervous system
CC>2          carbon dioxide
CT          carbon tetrachloride
CYP         cytochrome P450
CYP P450    cytochrome P450
DCA         dichloroacetic acid
DEHP       di(2-ethylhexyl)phthalate
EEGs        electroencephalograms
EPA         U.S. Environmental Protection Agency
FDA         Food and Drug Administration
FMO3       flavin-containing monooxygenase 3
GOT         gamma-glutamyltransferase
GSH         glutathione
GST         glutathione S-transferase
GSTx        glutathione S-transferase isoform, where x denotes different isoforms (such as M,
             T, P, S, Z)
FIEC         human equivalent concentration
HSIA        Halogenated  Solvents Industry Alliance
i.p.          intraperitoneal
IAP          intestinal alkaline phosphatase
IARC        International  Agency for Research on Cancer
IOM         Institute of Medicine
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xvi      DRAFT—DO NOT CITE OR QUOTE

-------
              LIST OF ABBREVIATIONS AND ACRONYMS (continued)

IPCS        International Programme on Chemical Safety
IRIS         Integrated Risk Information System
IUGR        intrauterine growth restriction
JISA        Japan Industrial Safety Association
Km          Michaelis-Menten constant
LECioS       95% lower confidence limits on the air concentrations associated with a 10%
             extra risk of cancer incidence
LGL         Large granular lymphocyte
LOAEL      lowest-observed-adverse-effect level
MLE        maximum likelihood estimate (please verify inserted correctly on pg 5-69, MLE
             was there but not the exact definition)
MCA        monochloroacetic acid
MCL-5       microsomal epoxide hydrolase
MCL        mononuclear cell leukemia
MOA        mode of action
MRL        minimal risk level
NAG        N-acetyl-p-D-glucosaminidase
NCI         National Cancer Institute
NHL        non-Hodgkin's lymphoma
NIOSH      National Institutes of Occupational Safety and Health
NK          natural killer
NOAEL      no-observed-adverse-effect level
NRC        National Research Council
NTP         National Toxicology Program
NYS DOH   New York State Department of Health
NYS O AG   New York State Office of Attorney General
OR          odds ratio
P450        cytochrome P450
PBPK        physiologically based pharmacokinetic
PCO         palmitoyl CoA oxidation
PHG        public health goal
POD         point of departure
PPAR        peroxisome proliferater activated receptor
PPAR-a      peroxisome proliferater activated receptor, alpha isoform
PPAR-5      peroxisome proliferater activated receptor, delta isoform
RBP         retinol binding protein
REAL        revised European-American Lymphoma
RfC         reference concentration
RfD         reference dose
RfV         reference value
RR          relative risk
SAP         Scientific Advisory Panel
SCE         sister chromatid exchange

           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xvii      DRAFT—DO NOT CITE OR QUOTE

-------
               LIST OF ABBREVIATIONS AND ACRONYMS (continued)

SES          socio-economic status
SGA         small for gestational age
SIR          standardized incidence ratio
SMR         standardized mortality ratio
SSB          single-strand breaks
TCA         trichloroacetic acid
TCE          trichloroethylene
TCOH        Trichloroethanol
TCVC        S-(l,2,2,-trichlorovinyl)-L-cysteine
TCVCSO     S-(l,2,2,-trichlorovinyl)-L-cysteine sulfoxide
TCVG        S-(l,2,2-trichlorovinyl) glutathione
TNAP        tissue non-specific alkaline phosphatase
TWA         time-weighted average
U/L          international units per liter
UDS          unscheduled DNA synthesis
UF           uncertainty factor
VCS          visual contrast sensitivity
VE           ventilation rate
VEP          visually evoked potential
Vmax          maximum velocity
WHO         World Health Organization
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    xviii     DRAFT—DO NOT CITE OR QUOTE

-------
                                      FOREWORD

       The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to
tetrachloroethylene. It is not intended to be a comprehensive treatise on the chemical or
toxicological nature of tetrachloroethylene.
       In Chapter 6, Characterization of Hazard and Dose-Response, the United States
Environmental Protection Agency (EPA) has characterized its overall confidence in the
quantitative and qualitative aspects of hazard and dose-response by addressing knowledge gaps,
uncertainties, quality of data, and scientific controversies. The discussion is intended to convey
the limitations of the assessment and to aid and guide the risk assessor in the ensuing steps of the
risk assessment process.
       For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    xix      DRAFT—DO NOT CITE OR QUOTE

-------
                  AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER

KathrynZ. Guy ton1
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Karen A. Hogan1
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Robert E. McGaughy2
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
AUTHORS

Stanley Bar one
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Rebecca Brown
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Glinda Cooper
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Nagalakshmi Keshava
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
       1 Chemical Manager since September 2007.
       2 Retired prior to final revisions to document.
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  xxi      DRAFT—DO NOT CITE OR QUOTE

-------
            AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)


Leonid Kopylev
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Susan Makris
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Jean Parker2
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Cheryl Scott
National Center for Environmental Assessment
Office of Research and Development
Washington, DC

Ravi Subramaniam
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Larry Valcovic2
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC


CONTRIBUTORS

Nancy Beck
AAAS Fellow
U.S. Environmental Protection Agency
Washington, DC

David Bussard
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  xxi      DRAFT—DO NOT CITE OR QUOTE

-------
            AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)


Jane Caldwell
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Weihsueh Chiu
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Deborah Rice
Formerly with National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Marc Rigas
Formerly with Environmental Sciences Division
National Exposure Research Laboratory
U.S. Environmental Protection Agency
Las Vegas, NV

Bob Sonawane
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC

Paul White
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC


REVIEWERS

       This document and the accompanying IRIS Summary have been peer reviewed by EPA
scientists and independent scientists external to the EPA. Comments from all peer reviewers
have been evaluated carefully and considered by the EPA during the preparation of this external
review draft. During the preparation of this draft, the IRIS Program Director achieved common
understanding of the assessment among the Office of Research and Development; the Office of
Air and Radiation; the Office of Prevention, Pesticides, and Toxic Substances; the Office of
Solid Waste and Emergency Response; the Office of Water; the Office of Policy, Economics,
and Innovation; the Office of Children's Health Protection; the Office of Environmental
Information, and the EPA's regional offices.

          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    xxii     DRAFT—DO NOT CITE OR QUOTE

-------
            AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
INTERNAL PEER REVIEWERS

Hugh Barton, U.S. Environmental Protection Agency, National Health and Environmental
Effects Research Laboratory, Research Triangle Park, NC

Robert Benson, U.S. Environmental Protection Agency, Office of Partnerships and Regulatory
Assistance, Region 8, Denver, CO

William Boyes, U.S. Environmental Protection Agency, National Health and Environmental
Effects Research Laboratory, Research Triangle Park, NC

Jane Caldwell, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Research Triangle Park, NC

Jim Cogliano, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Washington, DC

Herman Gibb, formerly with U.S. Environmental Protection Agency, National Center for
Environmental Assessment, Washington, DC

John Lipscomb, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Cincinnati, OH

Elizabeth Margosches, U.S. Environmental Protection Agency, Office of Pollution, Prevention,
and Toxics, Washington, DC

Dierdre Murphy, U.S. Environmental Protection Agency, Office of Air Quality and Planning and
Standards, Research Triangle Park, NC

Onyemaechi Nweke, U.S. Environmental Protection Agency, Office of Policy, Economics, and
Innovation, Washington, DC

Robert Park, National Institute for Occupational Safety and Health, Education and Information
Division, Cincinnati, OH

Brenda Perkovitch, U.S. Environmental Protection Agency, Office of the Administrator, Office
of Children=s Health Protection, Washington,  DC

Bruce Rodan, U.S. Environmental Protection Agency, National Center for Environmental
Assessment, Washington, DC
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xxiv    DRAFT—DO NOT CITE OR QUOTE

-------
             AUTHORS, CONTRIBUTORS, AND REVIEWERS (continued)
Val Schaeffer, Occupational Safety and Health Administration, Directorate for Health Standards,
Washington, DC

Vanessa Vu, U.S. Environmental Protection Agency, Office of Prevention, Pesticides, and Toxic
Substances, Washington, DC

Diana M. Wong, U.S. Environmental Protection Agency, Office of Science and Technology
Office of Water, Washington, DC

Tracey Woodruff, formerly with U.S. Environmental Protection Agency, Office of Policy,
Economics, and Innovation, Washington, DC


CONSULTANTS

Anne Aschengrau, Department of Epidemiology, Boston University School of Public Health,
Boston, MA

Matt Bogdanffy, Lincoln University, PA

George Lucier, formerly with National Institute of Environmental Health Sciences, Research
Triangle Park, NC

Neurotoxicity Expert Review Panel Reviewers at public workshop on February 25, 2004
Kent Anger (Chair), Center for Research on Occupational and Environmental Toxicology,
Oregon Health and Science University, Portland, OR

Rosmarie Bowler, San Francisco State University, San Francisco, CA

Diana Echeverria, Battelle Center for Public Health Research and Evaluation, Seattle, WA

Fabriziomaria Gobba, Dipartimento di Scienze Igienistiche, Universita di Modena e Reggio
Emilia, Modena, Italy

William Merigan, Department of Ophthalmology and Center for Visual Science,
University of Rochester School of Medicine and Dentistry, Rochester, NY


ACKNOWLEDGMENTS

       The authors would like to acknowledge the contributions of the following individuals:
Terri Konoza of NCEA who managed the document production activities; Cristopher Broyles of
Intellitech Systems, Inc. who provided editing support; Lana Wood of Intellitech  Systems, Inc.

           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   xxv     DRAFT—DO NOT CITE OR QUOTE

-------
who provided word processing support; Patricia von Brook of KBM Group who provided editing
support for the previous draft; and Christine Chang of KBM Group who provided word
processing support for the previous draft.
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    xxvi     DRAFT—DO NOT CITE OR QUOTE

-------
 1                                     1. INTRODUCTION
 2
 O
 4          This document presents background information and justification for the Integrated Risk
 5    Information System (IRIS) Summary of the hazard and dose-response assessment of
 6    tetrachloroethylene. IRIS Summaries may include oral reference dose (RfD) and inhalation
 7    reference concentration (RfC) values for chronic and other exposure durations, and a
 8    carcinogenicity assessment.
 9          The RfD and RfC, if derived, provide quantitative information for use in risk assessments
10    for health effects known or assumed to be produced through a nonlinear (presumed threshold)
11    mode of action.  The RfD (expressed in units of mg/kg-day) is defined as  an estimate (with
12    uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
13    population (including sensitive subgroups) that is likely to be without an appreciable risk of
14    deleterious effects during  a lifetime. The RfC is defined  as an estimate, with uncertainty
15    spanning perhaps an order of magnitude, of a continuous inhalation exposure to the human
16    population (including sensitive subgroups) that is likely to be without an appreciable risk of
17    deleterious non-cancer effects during a lifetime. The RfC considers toxic effects for both the
18    respiratory system (portal-of-entry) and for effects peripheral to the respiratory system
19    (extrarespiratory or systemic effects). Reference values are generally derived for chronic
20    exposures (up to a lifetime), but may also be derived for acute (#24 hrs), short-term (>24 hrs up
21    to 30 days), and subchronic (>30 days up to 10% of lifetime)  exposure durations, all of which are
22    derived based on an assumption of continuous exposure throughout the duration specified.
23    Unless specified otherwise, the RfD and RfC are derived for chronic exposure duration.
24          The carcinogenicity assessment provides information  on the carcinogenic hazard
25    potential of the substance  in question and quantitative estimates of risk from oral and inhalation
26    exposure. The information includes a weight-of-evidence judgment of the likelihood that the
27    agent is a human carcinogen and the conditions under which the carcinogenic effects may be
28    expressed. Quantitative risk estimates may be derived from the application of a low-dose
29    extrapolation procedure. If derived, the oral slope factor  is an upper bound on the estimate of
30    risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is an upper bound on the
31    estimate of risk per ug/m3 air breathed.
32          Development of these hazard identification and dose-response assessments for
33    tetrachloroethylene has  followed the general guidelines for risk assessment set forth by the
34    National Research Council (NRC, 1983, 1994).  U.S. Environmental Protection Agency (EPA)
35    Guidelines and Risk Assessment Forum Technical Panel  Reports that may have been used in the
36    development of this assessment include the following:  Guidelines for the Health Risk
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    1 -1        DRAFT—DO  NOT CITE OR QUOTE

-------
 1   Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines for Mutagenicity Risk
 2   Assessment (U.S. EPA, 1986b), Recommendations for and Documentation of Biological Values
 3   for Use in Risk Assessment (U.S. EPA, 1988), Guidelines for Developmental Toxicity Risk
 4   Assessment (U.S. EPA, 1991), Interim Policy for Particle Size and Limit Concentration Issues in
 5   Inhalation Toxicity (U.S. EPA, 1994a), Methods for Derivation of Inhalation Reference
 6   Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b), Use of the
 1   Benchmark Dose Approach in Health Risk Assessment (U. S. EPA, 1995), Guidelines for
 8   Reproductive Toxicity Risk Assessment (U.S. EPA,  1996), Guidelines for Neurotoxicity Risk
 9   Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk Characterization (U.S.
10   EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S. EPA, 2000b),
11   Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures (U.S.
12   EPA, 2000c), A Review of the Reference Dose and Reference Concentration Processes (U.S.
13   EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), Supplemental
14   Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens (U. S. EPA,
15   2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A Framework
16   for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA, 2006b).
17          The literature search strategy employed for tetrachloroethylene was based on the
18   Chemical Abstracts Service Registry Number (CASRN) and at least one common name. Any
19   pertinent scientific information submitted by the public to the IRIS Submission Desk was also
20   considered in the development of this document. A comprehensive literature review was carried
21   out through July 2004. In addition, a number of relevant publications since that time have been
22   considered and incorporated in the document.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  1-2        DRAFT—DO NOT CITE OR QUOTE

-------
  1                                    REFERENCES FOR CHAPTER 1
  2
  O

  4    NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
  5    (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3F1 mice. 311, 1-190. U.S. Department of Health
  6    and Human Services. Technical Report Series.
  7
  8    U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
  9    mixtures.  Fed. Register 51(185):34014-34025. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
10
11    U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk assessment. Federal
12    Register 51(185):34006-34012. Available online at http://www.epa.gov/ncea/raf.
13
14    U.S. EPA (Environmental Protection Agency). (1988) Recommendations for and documentation of biological values
15    for use in risk assessment. Environmental Criteria and Assessment Office, Office of Health and Environmental
16    Assessment, Cincinnati, OH; EPA 600/6-87/008. Available from: National Technical Information Service,
17    Springfield, VA; NTIS PB88-179874/AS.
18
19    U.S. EPA (Environmental Protection Agency). (1991) Guidelines for developmental toxicity risk assessment.
20    Federal Register 56(234):63798-63826.
21
22    U.S. EPA (Environmental Protection Agency). (1994a) Methods for derivation of inhalation reference
23    concentrations and application of inhalation dosimetry. Environmental Criteria and Assessment Office, Office of
24    Health and Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical
25    Information Service, Springfield, VA; NTIS PB88-179874/AS.
26
27    U.S. EPA (Environmental Protection Agency). (1994b) Peer review and peer involvement at U.S. Environmental
28    Protection Agency. Signed by the U.S. EPA Administrator Carol M. Browner, dated June 7, 1994.
29
30    U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose approach in health risk
31    assessment. Risk Assessment Forum, Washington, DC; EPA/630/R-94/007. Available from: National Technical
32    Information Service, Springfield, VA, PB95-213765, and online at http://www.epa.gov/raf.
33
34    U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive toxicity risk assessment. Federal
35    Register 61(212):56274-56322. Available online at http://www.epa.gov/ncea/raf.
36
37    U.S. EPA (Environmental Protection Agency). (1998a) Guidelines for neurotoxicity risk assessment. Federal
38    Register 63(93):26926-26954. Available online at http://www.epa.gov/ncea/raf.
39
40    U.S. EPA (Environmental Protection Agency). (1998b) Science policy council handbook: peer review. Prepared by
41    the Office of Science Policy, Office of Research and Development, Washington, DC; EPA 100-B-98-001.
42    Available from: National Technical Information Service,  Springfield, VA, PB 98-140726, and online at
43    http://www.epa.gov/OSA/spc.
44
45    U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment.  Federal Register
46    70(66)17765-18717.  Available online at http://www.epa.gov/cancerguidelines.
47
48    U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing cancer susceptibility
49    from early-life exposure to carcinogens.  Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
50    online at http://www.epa.gov/cancerguidelines.
51
52    U.S. EPA (Environmental Protection Agency). (2006a) Peer review handbook, 3rd edition. Science Policy Council,
53    Washington, DC; EPA/100/B-06/002. Available online at
54    http://www.epa.gov/peerreview/pdfs/Peer%20Review%20HandbookMav06.pdf.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        1-3         DRAFT—DO NOT CITE OR QUOTE

-------
1    U.S. EPA (Environmental Protection Agency). (2006b) A framework for assessing health risks of environmental
2    exposures to children. National Center for Environmental Assessment, Washington, DC; EPA/600/R-05/093 A.
3    Available online at http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid= 158363.
                  This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                        1-4         DRAFT—DO NOT CITE OR QUOTE

-------
 1                                      2. BACKGROUND
 2
 O
 4    2.1. USES AND PHYSICAL/CHEMICAL PROPERTIES
 5          Tetrachloroethylene is a widely used solvent that is produced commercially for use in dry
 6    cleaning, textile processing, and metal-cleaning operations. It has the following use pattern:
 7    55% as a chemical intermediate, 25% for metal cleaning and vapor degreasing, 15% for dry
 8    cleaning and textile processing, and 5% for other unspecified uses (ATSDR, 1997).
 9          Table 2-1 lists the physical and chemical properties of tetrachloroethylene (ATSDR,
10    1997). The reference citations can be found in the Agency for Toxic Substances and Disease
11    Registry (ATSDR) document and are not included in the reference list for this document.
12
13    2.2. OCCURRENCE AND EXPOSURE
14          Tetrachloroethylene has been detected in ground water and surface water as well as in air,
15    soil, food, and breast milk.  The primary exposure routes of concern are inhalation of vapor and
16    ingestion of contaminated water. Although dermal exposure is possible via contaminated tap
17    water during showering, bathing, or swimming, this is generally not considered a major route of
18    exposure.
19
20    2.2.1.  Air
21          Because of its high volatility, there is considerable potential for release of
22    tetrachloroethylene into the atmosphere. Once in the air, it is not susceptible to wet deposition
23    because of its hydrophobicity.  The primary method for removal is photooxidation to
24    trichloroacetyl chloride, trichloroacetic acid (TCA), carbon monoxide, ozone,  and phosgene
25    (U.S. EPA, 1982). However, this reaction is very slow, so tetrachloroethylene is not implicated
26    in the buildup of any of the reaction products in the troposphere.  Though the half-life  of
27    perchloroethylene can vary based on season and environmental conditions, it has been estimated
28    at 96 days under typical conditions (ATSDR, 1997).
29          Ambient tetrachloroethylene concentrations vary from source to source and with
30    proximity to the source. It should be noted that outdoor concentrations can vary widely within a
31    period of a few hours as a function of wind velocity and direction, precipitation, humidity, and
32    sunlight. ATSDR (1997) reported mean tetrachloroethylene concentrations of 8.8 |ig/m3 in areas
33    close to points of release.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   2-1        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Table 2-1. Physical and chemical properties of tetrachloroethylene
Property
Molecular weight
Color
Physical state
Melting point
Boiling point
Density at 20EC
Density at 25EC
Odor
Odor threshold: water
Odor threshold: air
Solubility: water at 25EC
Solubility: organic solvent(s)
Partition coefficients:
Log KQW
Partition coefficients:
Log KOC
Vapor pressure at 25EC
Henry's law constant at 25EC
Autoignition temperature
Flashpoint
Flammability limits
Conversion factors, air
Explosive limits
Information
165.83
Colorless
Liquid (at room temperature)
-19EC
121EC
1.6227g/mL
No data
Ethereal
0.3 ppm
1 ppm
150mg/L
Miscible with alcohol, ether,
chloroform, benzene, solvent
hexane, and most of the fixed
and volatile oils
3.4
2.2B2.7
18.47mmHg
1.8H 10'2 atm-m3/mol
No data
None
Nonflammable
1 mg/L = 141.4 ppm
1 ppm = 6.78 mg/m3
No data
Reference
Lide (1990)
Sax and Lewis (1987)
Sax and Lewis (1987)
Lide (1990)
Lide (1990)
Lide (1990)

HSDB (1996)
U.S. EPA(1987b)
U.S. EPA(1987b)
HSDB (1996)
HSDB (1996)
HSDB (1996)
Seipetal. (1986)
Zytneretal. (1989a)
HSDB (1996)
Gossett(1987)

HSDB (1996)
HSDB (1996)
HSDB (1996)

     Source: ATSDR(1997).
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   2-2        DRAFT-DO NOT CITE OR QUOTE

-------
 1          EPA has carried out modeling to characterize the geographic distribution of
 2    tetrachloroethylene for its National-Scale Air Toxics Assessment database (U.S. EPA, 1996).
 3    Median census tract-based tetrachloroethylene concentrations across the United States were
 4    estimated at about 0.3 |ig/m3 for urban areas and 0.1 |ig/m3 for rural areas (75% upper percentiles
 5    of 0.4 and 0.2 |ig/m3, respectively).  The California Air Resources Board (CARB,  1998) reported
 6    a statewide median air concentration of 0.3  |ig/m3 in 2001, which represents the lowest value in
 7    what has been a decreasing trend since 1990. Note that these averages, which are based on
 8    geographic areas, only characterize the likely exposure of individuals who spend an equal
 9    amount of time in all parts of the defined area, and they may, therefore, significantly
10    underestimate the exposure of individuals who consistently spend time in subareas that have
11    higher tetrachloroethylene concentrations.
12          Near points of use, such as dry cleaners or industrial facilities,  indoor exposure to
13    tetrachloroethylene is more significant than outdoor exposure (U.S.  EPA, 2001). Indoor air
14    concentrations in an apartment above a dry  cleaning shop have been measured at up to 4.9 mg/m3
15    (Verbek and Scheffers, 1980), whereas mean concentrations inside dry cleaning facilities have
16    been found to vary from 48 mg/m3 to 200 mg/m3, depending on type of facility (Solet et al.,
17    1990). Concentrations in facilities with post-1990 equipment are likely to be lower (U.S. EPA,
18    1998).
19          The off-gassing of garments that have recently been dry-cleaned may be of concern
20    (Tichenor et al., 1990).  In the home, tetrachloroethylene vapors may off-gas from the clothes of
21    occupationally exposed individuals, or they may come directly from the exhaled breath of
22    exposed workers (ATSDR, 1997). Relatively high tetrachloroethylene air concentrations have
23    been measured in the proximity of freshly dry-cleaned clothing stored in small, close spaces.  A
24    residential closet storing newly dry-cleaned clothing had an air concentration of 2.9 mg/m3 after
25    1 day, which rapidly declined to 0.5 mg/m3  and  persisted for several days (Tichenor et al., 1990).
26    There is one documented mortality case:  a 2-year-old boy was found dead after being put to
27    sleep in a room with curtains that had been incorrectly dry-cleaned (Gamier et al.,  1996).
28          Dry-cleaned garments transported in an automobile may  also lead to unexpectedly high
29    levels of exposure.  Park et al. (1998) used simulated driving cycles to estimate the
30    concentrations of several contaminants emitted from in-vehicle sources.  Using dry-cleaned
31    clothes as a source, tetrachloroethylene levels inside a stationary vehicle after 30 minutes reached
32    0.230 mg/m3.  Approximating these exposures is not easy because specific exposure levels would
33    depend on many factors: car velocity, wind speed, ventilation, and time spent in the automobile.
34    Another study demonstrating exposure in a  car found that transporting a freshly dry-cleaned
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   2-3         DRAFT-DO NOT CITE OR QUOTE

-------
 1    down jacket in a car resulted in a cabin air concentration of 24.8 mg/m3 after 108 minutes (Chien,
 2    1997).
 3          Air exposure may also occur during showering or bathing as dissolved
 4    tetrachloroethylene in the warm tap water is volatilized. Rao and Brown (1993) used an adult
 5    physiologically based pharmacokinetic (PBPK) model combined with a microenvironmental
 6    exposure model to estimate the dose received by inhalation exposure during showering and
 7    bathing as well as by dermal exposure to the water. The tap water concentration of
 8    tetrachloroethylene was 1 mg/L, which is probably a higher concentration than exists in most
 9    water supplies. They also demonstrated that a majority  of the tetrachloroethylene in the blood, as
10    a result of their bathing scenario, resulted from inhalation exposure, while about 15% resulted
11    from dermal absorption.
12
13    2.2.2. Water
14          Because of its relatively low aqueous solubility (see Table 2-1), it is not likely that
15    volatilized tetrachloroethylene will enter surface or rain water. However, it has been detected in
16    drinking water, ground water, and surface water (U.S. EPA,  2001; AT SDR, 1997).  Most of this
17    contamination is probably due to release in water following industrial use or by public use of
18    consumer products.
19          Unless a surface water body is in the vicinity of a highly contaminated site, surface waters
20    are expected to have a lower concentration of tetrachloroethylene than ground water. In an
21    estimate of drinking water contamination in California, McKone and Bogen (1992) assumed that
22    surface water would have a negligible contribution to the concentration of tetrachloroethylene
23    measured in drinking water. Based on data from wells in California, they estimated an average
24    drinking water concentration of 0.3 |ig/L, with  a standard deviation of 0.35 |ig/L.
25          In areas near sources of contamination,  ground water, and surface water concentrations
26    can be considerably higher than average.  Because the density of tetrachloroethylene is about
27    60% higher than that of water, tetrachloroethylene is expected to accumulate near the bottom of a
28    stagnant receiving water body after a large-volume point discharge.  Water samples collected
29    near the bottom of the St. Clair River near Sarnia, Ontario, downstream from several petroleum-
30    based production facilities, contained tetrachloroethylene concentrations ranging from 0.002 to
31    34.6 |ig/L (EC, 1993).  The concentrations in 17 samples of surface water from the lower Niagara
32    River in New York State in 1981 averaged 0.036 |ig/L (with a maximum of 0.134 |ig/L; EC,
33    1993).
34
35
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    2-4        DRAFT-DO NOT CITE OR QUOTE

-------
 1          Exposure models have been developed to predict the fate and transport of organic
 2    compounds such as tetrachloroethylene in environmental media, including air, water, and soil.
 3    The outputs from two similar but independently developed environmental exposure models,
 4    CalTOX and FugSONT, were compared for a scenario designed to reproduce a residential area
 5    near an industrial contamination site (Maddalena et al., 1995), in which 75 moles/day are
 6    released into the air and 0.7 moles/day are released into surface water. Although the soil
 7    predictions differed, the predictions of tetrachloroethylene in air and ground water were similar,
 8    with the concentration of air predicted by CalTOX approximately 6 |ig/m3 and the surface water
 9    concentration 82 |ig/L.  It should be noted that agreement of the models does not confirm the
10    validity of either one, but lends some support to the usefulness of the results.
11          The off-gassing of tetrachloroethylene from a drinking water supply can result in
12    exposure.  In 1976, EPA measured tetrachloroethylene levels ranging from 800 to 2,000 |ig/L in
13    drinking water samples in Massachusetts (Paulu et al.,  1999). Similar levels were reported
14    elsewhere in New England.  These concentrations  were attributed to the vinyl-lined asbestos-
15    cement pipes that were used to carry water in this area (Webler and Brown, 1993). Letkiewicz et
16    al. (1982) estimated that 53% of newborn infants are formula-fed from drinking water sources
17    and the other 47% receive all of their fluid from breast milk.  Taking into account volatilization
18    during boiling of water, they indicate that the uptake of tetrachloroethylene in formula-fed infants
19    on a mg/kg-day basis is 10 times higher than in adults with the same level of drinking water
20    contamination.
21          Although dermal exposure is possible via contaminated tap water during showering,
22    bathing, or swimming, this is generally not considered a major route of exposure. Rao and
23    Brown (1993) demonstrated that only 15% of the tetrachloroethylene in the blood resulted from
24    dermal exposure as compared to inhalation of vapors.
25
26    2.2.3. Food
27          Certain foods have been found to be contaminated with tetrachloroethylene (U.S. EPA,
28    2001). Because of the lipophilic nature of tetrachloroethylene, it may bind to lipid molecules in
29    such foods as margarine, oils, meats,  and other fatty foods stored in areas where there is
30    tetrachloroethylene in the air. In 1988, elevated tetrachloroethylene levels were seen in
31    margarine and butter samples obtained from grocery stores located near dry cleaning facilities
32    (Entz and Diachenko, 1988).  Further studies confirmed that close proximity to a dry cleaning
33    facility was associated with elevated tetrachloroethylene levels in butter  samples (Kacew and
34    Lambert,  1997).  Nonetheless, food is not considered to be a major exposure pathway. Other
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   2-5         DRAFT-DO NOT CITE OR QUOTE

-------
 1    sources of information about tetrachloroethylene in foods are the Food and Drug Administration
 2    (FDA, 2003) and Fleming-Jones and Smith (2003).
 3
 4    2.2.4. Breast Milk
 5          Due to its lipid solubility, tetrachloroethylene can concentrate in milk (NYS DOH, 2000;
 6    Schreiber, 1993; Sheldon et al., 1985).  Breast milk can contain high concentrations of
 7    tetrachloroethylene and some of its toxic metabolites. Reported levels of tetrachloroethylene in
 8    breast milk have ranged up to 43 |ig/L in the general population (U.S. EPA, 2001).
 9          Schreiber (1993) used a PBPK model to estimate the dose a nursing infant might receive
10    from an exposed mother's breast milk.  This study showed that it is possible for the dose an
11    infant receives through breast milk to approach levels that could result in adverse health effects
12    and exceed the 1988 EPA RfD of 0.01 mg/kg-day (U.S. EPA, 1988).  Actual indoor air
13    concentrations (24-hr average), as measured in apartments in New York  State, were used to
14    predict potential levels in breast milk in these modeling scenarios. The apartments included one
15    located above a dry cleaning facility that used an old dry-to-dry machine (average concentration,
16    45.8mg/m3), three located above facilities that used transfer machines (average concentration,
17    7.7mg/m3), and two located above facilities that used newer dry-to-dry machines (average
18    concentration, 0.25 mg/m3; Schreiber, 1993).  The predicted breast milk concentrations in these
19    scenarios ranged from 16 to 3,000 |ig/L. Assuming that a 7.2 kg infant ingests 700 mL of breast
20    milk per day,  Schreiber (1993) determined that the infant dose from milk could range from
21    0.0015 to 0.3 mg/kg-day.
22          Using the same exposure conditions as Schreiber (1993), Byczkowski et al. (1994)
23    predicted lower doses to the infant (0.0009-0.202 mg/kg-day), although these doses approached
24    levels that could result in adverse health effects. Exceedances of the RfD were seen only in those
25    apartments above old dry-to-dry machines (0.202 mg/kg-day) or above transfer machines (0.029
26    mg/kg-day).  Ingestion through breast milk and infant exposures is discussed further in
27    Section 4.8. However, Schreiber (1997) has suggested that if infants live adjacent to or in close
28    proximity to dry cleaning facilitates, the dose received through breast milk ingestion will be
29    insignificant when compared with that from their inhalation exposure.
30          In one case study, the breast milk of a woman was found to contain 10 mg/L of
31    tetrachloroethylene 1 hr following a visit to her husband at his work in a dry cleaning
32    establishment. This concentration dropped to 3 mg/L after 24 hrs. Her child suffered from
33    obstructive jaundice and hepatomegaly, but these conditions improved when breastfeeding was
34    discontinued (Bagnell and Ellenberger, 1977).
35
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   2-6        DRAFT-DO NOT CITE OR QUOTE

-------
1    2.2.5. Direct Ingestion
2          In rare circumstances, direct ingestion of tetrachloroethylene has been documented. A
3    6-year-old boy who directly ingested 12-16 g tetrachloroethylene experienced drowsiness,
4    vertigo, agitation, and hallucinations.  He then lost consciousness and went into a coma, and later
5    recovered (Koppel  et al., 1985). Follow-up testing on the boy was not reported, so any potential
6    long-term effects of the exposure are unknown.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   2-7        DRAFT-DO NOT CITE OR QUOTE

-------
  1                                    REFERENCES FOR CHAPTER 2

  2
  3     ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
  4     (update). Prepared by Sciences, International, under subcontract to Research Triangle Institute.
  5
  6     Bagnell, PC; Ellenberger, HA. (1977) Obstructive jaundice due to a chlorinated hydrocarbon in breast milk. Can
  7     MedAssocJ 117:104761048.
  8
  9     Byczkowski, JZ; Kinkead, ER; Leahy, HF; et al. (1994) Computer simulation of the lactational transfer of
10     tetrachloroethylene in rats using a physiologically based model.  Toxicol Appl Pharmacol 125:228B236.
11
12     CARD (California Air Resources Board). (1998)  1990B1996 Statewide perchloroethylene summary, ppb. California
13     Environmental Protection Agency.  Available online at http://www.arb.ca/gov/aqd/perc/pcstate.htm.
14
15     Chien, Y-C. (1997) The influences of exposure pattern and duration on elimination kinetics and exposure assessment
16     of tetrachloroethylene in humans. PhD thesis, Rutgers University, New Brunswick, NJ.  Available from: IRIS
17     Information Desk, U.S. Environmental Protection Agency, Washington, DC.
18
19     EC (Environment Canada). (1993) Priority substances list assessment report: tetrachloroethylene. En-40-215/28E.
20     Canadian Environmental Protection Agency, Ottawa, Canada.
21
22     Entz, RC; Diachenko, GW. (1988) Residues of volatile halocarbons in margarines. Food Addit Contam 5:267B276.
23
24     FDA (U.S.  Food and Drug Administration). (2003) Food and Drug Administration total diet study: summary of
25     residues found, ordered by pesticide. 91-3-01-4. Center for Food Safety and Nutrition. Washington, DC. Available
26     online at http://www.cfsan.fda.gov/~acrobat/tdslbvps.pdf.
27
28     Fleming-Jones, ME; Smith, RE. (2003) Volatile organic compounds in foods: A five year study. J. Agric and Food
29     Chem 51:8120-8127
30
31     Gamier, R; Bedouin, J; Pepin, G; Gaillard, Y. (1996) Coin-operated dry cleaning machines may be responsible for
32     acute tetrachloroethylene poisoning: report of 26 cases including one death. J Toxicol Clin Toxicol 34:19lBl97.
33
34     Kacew, S; Lambert, GH. (1997) Environmental toxicology and pharmacology of human development. Washington,
3 5     DC: Taylor and Francis.
36
37     Koppel, C; Arndt, I; Arendt, U; et al. (1985) Acute tetrachloroethylene poisoning-blood elimination kinetics during
38     hyperventilation therapy. J Toxicol Clin Toxicol 23:103B115.
39
40     Letkiewicz, F; Johnston, P; Macaluso, C; et al. (1982) Occurrence of tetrachloroethylene in drinking water, food, and
41     air.  Prepared by JRB Associates JRB Project No. 2-613-03-852-29, for EPA contract 86-01-6388, task 29.
42
43     Maddalena, RL; McKone, TE; Layton, DW; et al. (1995) Comparison of multi-media transport and transformation
44     models: regional fugacity model vs. CalTOX. Chemosphere 30:869B889.
45
46     McKone, TE; Bogen, KT. (1992) Uncertainties in health-risk assessment: an integrated case study based on
47     tetrachloroethylene in California groundwater.  Regul Toxicol Pharmacol 15:86Bl03.
48
49     McKone, T; Bogen, K. (1993) CaLTOX: A multimedia total-exposure model for hazardous waste sites. Part II:
50     multimedia transport and transformation model. Prepared for the State of California, Department of Toxic
51     Substances Control. UCRLBCRB111456Ptl.
52
53     McKone, TE; Daniels, JI. (1991) Estimating human exposure through multiple pathways from air, water, and soil.
54     Regul Toxicol Pharmacol 13:36B61.
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         2-8          DRAFT-DO NOT CITE OR QUOTE

-------
  1     NYS DOH. (2000) Evaluation of residential exposure to tetrachloroethene using biomarkers of dose and neurological
  2     tests (non peer-reviewed draft). Albany, NY.
  O

  4     Park, JH; Spengler, JD; Yoon, DW; et al. (1998) Measurement of air exchange rate of stationary vehicles and
  5     estimation of in-vehicle exposure. J Expo Anal Environ Epidemiol 8:65-78.
  6
  7     Paulu, C; Aschengrau, A; Ozonoff, D. (1999) Tetrachloroethylene-contaminated drinking water in Massachusetts and
  8     the risk of colon-rectum, lung, and other cancers.  Environ Health Perspect 107:2656271.
  9
10     Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
11     groundwater for a bathing and showering determination. Risk Anal  13:37B49.
12
13     Schreiber, JS. (1993) Predicted infant exposure to tetrachloroethene  in human breastmilk. Risk Anal 13:515B524.
14
15     Schreiber. JS. (1997) Transport of organic chemicals to breast milk:  Tetrachloroethene case study. In Kacew S,
16     Lambert G (eds): Environmental Toxicology and Pharmacology of Human Development. Washington, DC: Taylor
17     and Francis.
18
19     Sheldon, L; Handy, R; Hartwell, W; et al. (1985) Human exposure assessment to environmental chemicals: nursing
20     mothers study.  Final report. Research Triangle Institute, Research Triangle Park, NC.
21
22     Solet, D; Robins, TG; Sampaio, C. (1990) Perchloroethylene exposure assessment among dry cleaning workers.  Am
23     Ind Hyg Assoc J 51(10):566-574.
24
25     Tichenor, BA; Sparks, LE; Jackson, MD; et al. (1990) Emissions of perchloroethylene from dry cleaned fabrics.
26     Atmospheric Environment 24A: 1219B1229.
27
28     U.S. EPA (Environmental Protection Agency). (1982) An exposure and risk assessment for tetrachloroethylene.
29     Office of Water, Regulations, and Standards, Washington, DC; EPA-4404-85-015.
30
31     U.S. EPA (Environmental Protection Agency). (1988) IRIS summary of tetrachloroethylene RID.  Available online at
32     http://www.epa. gov/iris/subst/0106.htm.
33
34     U.S. EPA (Environmental Protection Agency). (1996) Modeled ambient concentration for perchloroethylene
3 5     (CAS#127184). National Air Toxics Assessment, Technology Transfer Network, Office of Air and Radiation.
36     Available online at http://www.epa.gov/ttn/atw/nata/pdf/perc/conc.pdf.
37
38     U.S. EPA (Environmental Protection Agency). (2001) Sources, emission and exposure for trichloroethylene (TCE)
39     and related chemicals. National Center for Environmental Assessment, Washington, DC; EPA/600/R-00/099.
40     Available from: National Technical Information Service, Springfield, VA, and online at http://www.epa. gov/ncea.
41
42     Verberk, MM; Scheffers, TM. (1980) Tetrachloroethylene in exhaled air of residents near dry-cleaning shops.
43     Environ Res 21(2):432^37.
44
45     Webler, T;  Brown, HS. (1993) Exposure to tetrachloroethylene via contaminated drinking water pipes in
46     Massachusetts:  a predictive model. Arch Environ Health 48:293B297.
                    This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        2-9         DRAFT-DO NOT CITE OR QUOTE

-------
 1                                    3. TOXICOKINETICS
 2
 O
 4    3.1. ABSORPTION
 5           Tetrachloroethylene is rapidly absorbed into the bloodstream following oral and
 6    inhalation exposures. It can also be absorbed across the skin following dermal exposure to either
 7    pure or diluted solvent or vapors (Stewart and Dodd, 1964; Nakai et al., 1999; Poet et al., 2002).
 8
 9    3.1.1.  Inhalation
10           The major exposure route for tetrachloroethylene is considered to be inhalation (U.S.
11    EPA, 1985; IARC,  1995).  Pulmonary uptake of tetrachloroethylene is rapid; however, complete
12    tissue equilibrium occurs only after several hours. Absorption into the systemic circulation
13    through pulmonary uptake is proportional to the ventilation rate, the duration of exposure, and, at
14    lower ambient concentrations to which humans are likely to be exposed, the concentration in the
15    inspired air (Hake and Stewart, 1977; Monster et al., 1979).
16           Chiu et al. (2007) reported that peak levels of tetrachloroethylene in venous blood and air
17    occurred near the end of a 6-hr inhalation exposure to 1 ppm and declined thereafter.  In the
18    Monster et al. (1979) study, uptake after 4 hrs was 75% of its value at the onset of exposure.
19    Increased physical activity increases uptake but lowers the alveolar partial pressure, thus
20    removing more tetrachloroethylene from the alveoli, resulting in a longer time to reach tissue
21    equilibrium (Pezzagno et al., 1988).
22           The blood/gas partition coefficient for tetrachloroethylene describes how the chemical
23    will partition itself between the two phases.  Specifically, it is the ratio of concentrations at
24    steady state; i.e.,  when all rates are constant after equilibrium has been reached. Reported values
25    for the coefficient in humans range from  around 10 to 20 (e.g., Byczkowski and Fisher, 1994;
26    Reitz et al., 1996; Droz and Guillemin, 1986; Ward et al., 1988; Gearhart et al., 1993; Hattis et
27    al., 1990), meaning that if tetrachloroethylene is in equilibrium, the concentration in blood will
28    be 10 to 20 times higher than the concentration in the alveoli.
29           Opdam and Smolders (1986) determined concentrations of tetrachloroethylene in alveolar
30    air for 1- to 60-second residence times (the time interval from the beginning of an inhalation to
31    the end of the next inhalation) for six volunteers exposed to 0.5 to 9.8 ppm of chemical for 1 to
32    60 minutes.  These investigators found the concentrations of tetrachloroethylene in alveolar air to
33    decrease with residence times for breaths during exposure periods but to increase during post-
34    exposure for residence times less than 10 seconds.  Alveolar air tetrachloroethylene
35    concentration correlated with the concentrations in pulmonary artery mixed venous blood.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    3-1         DRAFT-DO NOT CITE OR QUOTE

-------
 1          Like the studies in humans, inhalation studies in laboratory animals provide clear
 2    evidence that tetrachloroethylene is readily absorbed via the lungs into the systemic circulation
 3    (e.g., Pegg et al., 1979; Dallas et al., 1994a).
 4
 5    3.1.2. Oral
 6          Gastric absorption of tetrachloroethylene occurs at a relatively rapid rate and is
 7    essentially complete.  Close to 100% of oral doses are absorbed from the gut, according to
 8    reports of several studies conducted in mice, rats, and dogs (Dallas et al., 1994a, 1995; Frantz
 9    and Watanabe, 1983; Pegg et al., 1979; Schumann et al., 1980). Absorption into the systemic
10    circulation was indicated by blood tetrachloroethylene levels of 21.5 |ig/mL following accidental
11    ingestion of the chemical by a 6-year-old boy (Koppel et al., 1985).
12
13    3.1.3. Dermal
14          Absorption of tetrachloroethylene by humans following dermal exposure to vapors of the
15    chemical has been reported to be relatively insignificant (only  1%) when compared with
16    absorption via inhalation of vapors (Riihimaki and Pfaffli, 1978; Nakai et al., 1999).  The
17    amount of chemical absorbed during the immersion of one thumb  in liquid tetrachloroethylene is
18    equivalent to the uptake during inhalation of 10 to 15 ppm of the compound for the same time
19    period (Stewart and Dodd, 1964).
20          Studies in animals confirm that dermal uptake of tetrachloroethylene following vapor
21    exposure is minimal when compared with pulmonary uptake (Tsuruta, 1989; McDougal et al.,
22    1990), whereas dermal uptake is greater following direct skin application (Jakobson et al., 1982).
23    Notably, the  conclusions of Bogen et al. (1992), based on the results of their study in hairless
24    guinea pigs, indicate that dermal absorption of tetrachloroethylene from contaminated water
25    supplies could be an important route of exposure for humans.  These investigators estimated that
26    a standard 70 kg man with 80% of his body immersed in water would completely absorb the
27    amount of tetrachloroethylene in 2 L of that water.
28
29    3.2. DISTRIBUTION AND BODY BURDEN
30          Once absorbed, tetrachloroethylene is distributed by first-order diffusion processes to all
31    tissues in the mammalian body.  The highest  concentrations of tetrachloroethylene are found in
32    adipose tissue due to the lipophilicity of the compound (U.S. EPA, 1985). Concentrations of
33    tetrachloroethylene reach higher levels in brain and liver than in many other tissues (Gamier et
34    al., 1996; Levine et al., 1981; Lukaszewski, 1979).  Absolute tissue concentrations are directly
35    proportional to the body burden or exposure dose.  Due to its lipid solubility, tetrachloroethylene
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-2        DRAFT-DO NOT CITE OR QUOTE

-------
 1    is also concentrated in milk, and it has been measured in human breast milk (Schreiber, 1993,
 2    1997; Schreiber et al., 2002; NYS DOH, 2000).  Higher concentrations occur in milk having
 3    higher fat content; e.g., a noticeable difference exists between the milk/blood partition
 4    coefficients for rats (12) and for humans (2.8; Byczkowski and Fisher, 1994), reflecting the
 5    higher fat content of rat milk.  Tetrachloroethylene readily crosses both the blood-brain barrier
 6    and the placenta. Partition coefficients for various tissues, relative to blood or air, have been
 7    reported by several investigators (Ward et al., 1988; Dallas et al., 1994a, b; Gearhart et al., 1993;
 8    Byczkowski and Fisher, 1994).  Section 3.5 presents examples of these.
 9          Repeated daily inhalation exposures of human volunteers to tetrachloroethylene indicate
10    accumulation of the compound in the body, which is thought to be due to its high lipid solubility.
11    Because of its long residence time in adipose tissue, repeated daily exposure results in an
12    accumulated concentration; tetrachloroethylene from new exposures adds to the residual
13    concentration from previous exposures until steady state is reached.  Blood levels of
14    tetrachloroethylene increase over several days with continued daily exposures.  Following
15    cessation of these exposures, it is still present in the blood. Exhalation of the compound
16    continues over a number of days due to its slow release from the adipose tissue (Stewart et al.,
17    1977; Altmann et al., 1990; Skender et al., 1991). For a given concentration in blood or air, the
18    half-time—the time necessary to equilibrate the adipose tissue to 50% of its final
19    concentration—is about 25 hrs (Monster, 1979; Fernandez et al., 1976).  Therefore, during a
20    single 8-hr exposure, adipose tissue does not reach steady-state equilibrium.
21          Tetrachloroethylene uptake by fatty tissue during the working hours of the week is
22    countered by the elimination that occurs  during nonexposure times of nights and weekends; thus,
23    for persons exposed to tetrachloroethylene on a five-day-a-week work schedule, an equilibrium
24    is eventually established, but it requires a time period of 3 to 4 weeks of exposure for adipose
25    tissue to reach plateau concentrations (U.S. EPA, 1985).
26          Animal studies provide clear evidence that tetrachloroethylene distributes widely to all
27    tissues of the body, readily crossing the blood-brain barrier and the placenta (Schumann et al.,
28    1980; Ghantous et al., 1986; Savolainen et al., 1977; Dallas et al., 1994b). Following exposure
29    of rats to tetrachloroethylene, the compound has been measured in blood, fat, brain, lungs, liver,
30    kidneys, heart, and skeletal muscle (Savolainen et al., 1977; Dallas et al., 1994b). Highest tissue
31    concentrations were found in adipose tissue (60 or more times blood level) and in brain and liver
32    (4 and 5 times blood level, respectively), as can be calculated from the rat tissue distribution data
33    of Savolainen et  al. (1977). Dallas et al.  (1994b) found the concentration in fat to be 9 to 18
34    times the concentrations found in nonfat tissues. Skeletal muscle contained the lowest
35    concentration. In one human fatality case, the concentration of tetrachloroethylene in the brain

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-3        DRAFT-DO NOT CITE OR QUOTE

-------
 1    was 120 times higher than concentrations measured in the lung. In another case the
 2    concentration in the liver was 8, 3.4, and 3.5 times higher, respectively, than concentrations
 3    measured in the lung, kidney, and brain (Levine et al., 1981).
 4
 5    3.3. METABOLISM
 6          This section describes the metabolism of tetrachloroethylene, identifying metabolites
 7    thought to be causally associated with toxic responses as well as those used to evaluate the flux
 8    of parent compound through the known metabolic pathways.  Sex- and species-dependent
 9    differences in the metabolism of tetrachloroethylene and potential contributors to interindividual
10    differences are identified.  Factors that influence metabolism in humans are mentioned. See
11    Section 4.9 for further discussion of how these factors affect variability and susceptibility.
12
13    3.3.1.  Introduction
14          The metabolism of tetrachloroethylene has been studied mostly in mice, rats, and humans
15    (for reviews, see Dekant et al., 1987,  1989; Anders et al.,  1988; IARC, 1995; U.S. EPA, 1985,
16    1986,  1991; Lash and Parker, 2001).  Tetrachloroethylene is metabolized in laboratory animals
17    and in humans through at least two distinct pathways:  oxidative metabolism via the cytochrome
18    P450 (CYP [also abbreviated as P450 and CYP 450]) mixed-function oxidase system and
19    glutathione (GSH) conjugation followed by subsequent further biotransformation and processing,
20    either through the cysteine conjugate  beta lyase pathway or by other enzymes including flavin-
21    containing monooxygenase 3 (FMO3) and CYP3A (Daniel, 1963; Filser and Bolt, 1979; Pegg et
22    al., 1979; Costa and Ivanetich, 1980;  Dekant et al., 1987,  1989; Anders et al., 1988; U.S. EPA,
23    1985,  1991; IARC, 1995; Birner et al., 1996; Lash et al., 1998; Volkel et al., 1998; Lash and
24    Parker, 2001).  The conjugative pathway, although the minor route quantitatively, is
25    lexicologically significant because it  yields relatively potent toxic metabolites (Vamvakas et al.,
26    1987,  1989a, b, c; Dekant et al., 1986a, b, 1989; Werner et al., 1996; Anders et al., 1988; Lash
27    and Parker, 2001).
28
29    3.3.2.  Extent of Metabolism
30          Studies in both animals and humans indicate that overall metabolism of
31    tetrachloroethylene is relatively limited (reviewed in U.S. EPA, 1985, 1991; Lash and Parker,
32    2001), as evidenced by the high percentage of absorbed dose excreted in the breath as the parent
33    molecule (Stewart et al., 1961, 1970;  Monster et al., 1979, 1983; Boettner and Muranko, 1969;
34    Ikeda and Otsuji, 1972;  Essing et al.,  1973; Fernandez et al., 1976; May, 1976; Ohtsuki et al.,
35    1983; Yllner, 1961; Daniel, 1963; Filser and Bolt, 1979; Pegg et al., 1979; Frantz and Watanabe,
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-4        DRAFT-DO NOT CITE OR QUOTE

-------
 1    1983; Schumann et al., 1980; Buben and O'Flaherty, 1985; Volkel et al., 1998). Because of its
 2    high lipid solubility, tetrachloroethylene can be sequestered in fat and, thus, not all metabolism is
 3    evident in short sampling time periods.
 4          The extent of metabolism after inhalation exposure in humans has been estimated by
 5    measuring trichloro-compounds excreted in the urine and exhalation of tetrachloroethylene in
 6    expired air (Bolanowska and Golacka, 1972; Fernandez et al., 1976; Monster et al., 1979, 1983;
 7    Monster and Houtkooper, 1979; Ikeda et al., 1972; Boettner and Muranko, 1969; Essing et al.,
 8    1973; May, 1976; Stewart et al., 1961, 1970).  Several studies reported only about 1-3% of the
 9    estimated amounts inhaled were metabolized to TCA and other chlorinated metabolites, although
10    additional tetrachloroethylene—as much as 20% or more of the dose—may be metabolized over
11    a longer period (Monster etal., 1979; U.S. EPA, 1985, 1991;  Bois et al., 1996; Bogen et al.,
12    1992). For example, Chiu et al. (2007) noted that although an average of 0.4% of
13    tetrachloroethylene intake (1 ppm for 6 hrs) was recovered in urine as TCA, total recovery in
14    urine and exhaled air accounted for on average only 82% of intake.  This would imply 18%
15    metabolized, but Chiu et al. (2007) noted substantial uncertainty and variability in these
16    calculations and concluded they were consistent with previous studies at higher exposures.
17    Interestingly, Chiu et al. (2007) also noted significant variability among the seven subjects and
18    among the four occasions, contributing to the uncertainty in measurements.
19          The extent of metabolism in animals has been estimated by conducting excretion-balance
20    studies using isotopically labeled tetrachloroethylene.  In rodents, 2-88% of the dose was
21    metabolized, depending on dose level and species: the higher the dose the smaller the percent
22    metabolized. Rats metabolized a lower percent of a given tetrachloroethylene body burden than
23    did mice (Yllner,  1961; Daniel, 1963; Filser and Bolt,  1979; Pegg et al., 1979; Frantz and
24    Watanabe, 1983; Schumann et  al., 1980).  As an example, using data from the Pegg et al. (1979)
25    and Schumann et al. (1980) studies in rats, EPA calculated that the percent of body burdens
26    excreted were unchanged following exposure to 10 and 600 ppm for 6 hrs, were 68 and 99%,
27    respectively (U.S. EPA, 1985). For comparison, studies  in mice exposed to 10 ppm for 6 hrs
28    found pulmonary excretion of only 12%, whereas 83% of the tetrachloroethylene was excreted
29    by the pulmonary route for a body burden of about 11 mg from oral administration (U.S. EPA,
30    1985). As body burden is increased, the proportion of tetrachloroethylene excreted unchanged
31    increases and the percent metabolized decreases.
32
33    3.3.3. Pathways of Metabolism
34          The two known biotransformation pathways for tetrachloroethylene metabolism are (1)
35    oxidation by cytochrome P450  (CYP) enzymes and (2) conjugation with GSH followed by

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-5       DRAFT-DO NOT CITE OR QUOTE

-------
 1    further processing of the conjugate through various pathway bifurcation branches. The initial
 2    step in the metabolism of tetrachloroethylene may be either epoxidation or chlorine migration for
 3    the oxidative pathway or conjugation with GSH for the secondary pathway (Costa and Ivanetich,
 4    1980; Miller and Guengerich, 1982,  1983; Dekant et al., 1986b, 1987, 1998; Lash et al., 1998;
 5    Lash and Parker, 2001). It is possible that other as yet unrecognized pathways for
 6    tetrachloroethylene exist in humans (Sakamoto, 1976; Monster et al., 1979; U.S. EPA, 1985,
 7    1991; Boisetal., 1996).
 8
 9    3.3.3.1.  Cytochrome P450-Dependent Oxidation
10          Oxidative metabolism by the cytochrome P450, or CYP-dependent, pathway is
11    quantitatively the major route of tetrachloroethylene biotransformation (U.S. EPA, 1991; IARC,
12    1995; Lash and Parker, 2001). This pathway was initially proposed by Powell (1945) for
13    trichloroethylene and was subsequently supported for tetrachloroethylene by the results of Yllner
14    (1961), Daniel (1963), Leibman and Ortiz (1970, 1977), Costa and Ivanetich (1980), and others.
15    The pathway is operative in humans and rodents and leads to several metabolites, some of which
16    are known toxins and carcinogens (U.S. EPA, 1991; IARC, 1995). Figure 3-1 depicts the overall
17    scheme of tetrachloroethylene P450 metabolism. Known metabolites presented in this figure are
18    identified by an asterisk.
19          The major excretory metabolite of the oxidative pathway, TCA, is excreted in the urine of
20    all species tested. Figure 3-1 identifies many common urinary metabolites, including
21    dichloroacetic acid (DCA), trichloroacetylethanolamide, oxalylethanalamide, and oxalic acid.
22    Trichloroethanol (TCOH) has been measured in some, but not all, studies (Bonse et al., 1975;
23    Bonse and Henschler, 1976; Yllner,  1961; Dmitrieva, 1967; Pegg et al., 1979; Ogata et al., 1962,
24    1971; Tanaka and Ikeda, 1968; Ikeda and Otsuji, 1972; Ikeda  et al., 1972; Monster et al., 1983;
25    Weichardt and Lindner, 1975; Dekant et al., 1986b,  1987; Birner et al., 1996; U.S. EPA,  1985,
26    1986,  1991). Oxalic acid is a relatively major urinary metabolite in rats (Dmitrieva,  1967; Pegg
27    et al., 1979). Pulmonary excretion of carbon dioxide (CO2) has been identified in exhaled breath
28    from rodents exposed to 14C-labeled tetrachloroethylene (Pegg et al., 1979; Schumann et al.,
29    1980). Oxalic acid and formic acid plus CO2 are hypothesized to arise from  action of
30    microsomal epoxide hydrase on the initial epoxide intermediate to yield tetrachloroethylene
31    glycol, which may then be further processed via two routes to these aforementioned end
32    products.
33          Oxidative metabolism of tetrachloroethylene, irrespective of the route of administration,
34    occurs predominantly in the liver but also occurs at other sites. For example, the kidneys exhibit
35    cytochrome P450 enzyme activities, mostly in the proximal tubules,  although total activity is

                This document is a draft for review purposes only and does  not constitute Agency policy
      06/06/08                                  3-6        DRAFT-DO NOT CITE OR QUOTE

-------
                                               \
              I
                          ,o
                                       n    o    ci
                                        \/\/
                                          c — c
                                        /  z   \
                                            and
                                     ITelrachloroclhylcnc -
                                       P450 intermediate)
                                             17
C1,C 	 C
n
•s
on •«
o
*
"/"Ii
                                                                xnc!K-rn,-on
                                                                                              ,
                                                                                  c '  OH   on ( '
     II

UN —C

     H
             — C — Oil
                                                                                        I
                                                                                   0
                                                                                             0
                                                                                           \
                                                                                             t)II
                                                                                             OH
                                                                                            J_6
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
      Figure 3-1. Postulated scheme for the metabolism of tetrachloroethylene by
      the P450 oxidative pathway. Tetrachloroethylene and identified (*) metabolites:
      1 tetrachloroethylene, 2 tetrachloroethylene oxide, 1 trichloroacetyl chloride, 4
      trichloroacetic acid, 5 dichloroacetic acid, 6 monochloroacetic acid, 7 chloral, 8
      trichloroethanol, 9 trichloroethanol glucuronide, K) trichloroacetyl ethanolamide,
      11 tetrachloroethylene glycol, J_2 ethandioyl dichloride, H oxalic acid, 14
      glyoxylic acid chloride, 1_5 carbon dioxide, 16_ formic acid, j/7 P450 intermediate,
      JJ5 chloral hydrate, 19 glyoxylic acid, and 20 glycine.

      Sources: Adapted from Pegg et al. (1979), Costa and Ivanetich (1980), U.S. EPA
      (1985), Dekant et al. (1986a), Lash and Parker (2001).
                 TTz/'s1 document /'s a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    3-7        DRAFT-DO NOT CITE OR QUOTE

-------
 1    markedly less than in the liver (Lash and Parker, 2001; Lash et al., 2001).  CYP enzymes
 2    occurring in other extrahepatic tissues—brain and lungs, for example—may also contribute to
 3    oxidative metabolism of tetrachloroethylene.
 4
 5    3.3.3.1.1. Formation of tetrachloroethylene oxide.  The first step in the oxidation of
 6    tetrachloroethylene is hypothesized to yield 1,1,2,2-tetrachloroethylene oxide, a relatively
 7    unstable epoxide (Costa and Ivanetich, 1980; Miller and Guengerich, 1982, 1983). Although an
 8    initial epoxide metabolite has not been unequivocally demonstrated for tetrachloroethylene,
 9    evidence for this epoxide does exist. The epoxide has been chemically synthesized (Frankel et
10    al., 1957; Bonse et al., 1975; Kline et al., 1978). The several potential fates of
11    tetrachloroethylene epoxide include trichloroacetyl chloride, oxalate dichloride through
12    tetrachloroethylene glycol, trichloroacetyl aminoethanol, and possibly chloral hydrate (in
13    equilibrium with chloral; Bonse and Henschler, 1976; Henschler and Bonse, 1977; Pegg et al.,
14    1979; U.S. EPA, 1985, 1986). Formation of trichloroacetyl chloride directly from
15    tetrachloroethylene, without the formation of the epoxide intermediate, via the mechanism of
16    CYP-mediated olefin oxidation has also been postulated (Guengerich and Macdonald,  1984).
17
18    3.3.3.1.2. Metabolism to Trichloroacetic Acid (TCA) and possibly Trichloroethanol (TCOH).
19    Measurement of urinary TCA has been used as a biomarker for tetrachloroethylene exposure
20    (U.S. EPA, 1985; IARC, 1995), although TCA can be a by-product of metabolism of other
21    chemical compounds.  TCA, a major tetrachloroethylene urinary metabolite in both humans and
22    laboratory rodents (Yllner, 1961; Daniel, 1963; Leibman and Ortiz, 1970,  1977; Birner et al.,
23    1996; Dekant et al., 1987; Ohtsuki et al., 1983; Volkel et al.,  1998), is believed to result
24    primarily from the oxidation of tetrachloroethylene to trichloroacetyl chloride. This oxidation
25    may occur through the epoxide intermediate, with chloride migration leading to the reactive
26    trichloroacetyl chloride, which can then react with amino groups of cellular proteins or undergo
27    hydrolysis to produce the TCA. N-(di- and trichloroacetylated)-L-lysines, formed by interaction
28    of tetrachloroethylene reactive metabolites with protein, have been identified in liver and kidney
29    tissue of rats exposed to tetrachloroethylene (Birner  et al., 1994; Pahler et al., 1999a).
30           The proposed chloral hydrate intermediate is another potential source of TCA, but chloral
31    hydrate can also be further metabolized to TCOH (Sellers et al., 1972; Birner et al., 1996). This
32    latter pathway to TCOH would be the favored  reaction, and it is thought to be catalyzed by both
33    alcohol dehydrogenase (Larson and Bull, 1989) and  CYP2E1 (Schultz and Weiner, 1979; Ni et
34    al., 1996). The resulting TCOH is then conjugated with glucuronide, a reversible reaction, and
35    both the alcohol and its glucuronide conjugate  have been reportedly detected as urinary excretion
36    products following tetrachloroethylene exposures. TCOH has been detected in the urine of
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-8         DRAFT-DO NOT CITE OR QUOTE

-------
 1    subjects exposed to tetrachloroethylene in some studies (Birner et al., 1996; Ikeda and Otsuji,
 2    1972; Ikeda et al.,  1972; Ogata et al., 1962, 1971; Tanaka and Ikeda, 1968; Monster et al., 1983;
 3    Weichardt and Lindner, 1975; Schreiber et al., 2002), but could not be identified by others
 4    (Fernandez et al., 1976; Hake and Stewart, 1977; Monster et al., 1979; Volkel et al., 1998;
 5    Yllner, 1961; Daniel, 1963; Buben and O'Flaherty, 1985; Costa and Ivanetich, 1980).  TCOH is
 6    thought to be an artifact of the methodology used or could arise due to unknown exposures to
 7    other chemicals. Thus, because TCOH is clearly not a significant metabolite for
 8    tetrachloroethylene, very little, if any, TCA produced from tetrachloroethylene metabolism is
 9    likely to come through chloral, either directly or indirectly through TCOH (Lash and Parker,
10    2001).
11
12    3.3.3.1.3. Formation of dichloroacetic acid (DCA) and other products. TCA is the major
13    source of DCA from the tetrachloroethylene P450 oxidation pathway. Although DCA has been
14    identified as a tetrachloroethylene urinary metabolite (Yllner,  1961; Dekant et al., 1987; Volkel
15    et al., 1998), it is not clear whether the DCA is a product of further metabolism of TCA, of
16    another pathway originating with GSH conjugation, or both. The major organ site of DCA
17    production is likely to differ for each pathway, with DCA arising from oxidative metabolism
18    primarily in the liver and from GSH-dependent metabolism products mostly in the kidney. The
19    amount of DCA produced from tetrachloroethylene oxidative  metabolism may vary across
20    species and is likely to be less than TCA.  This is because DCA derived from P450 oxidation
21    comes only from dechlorination of TCA, which is not extensively metabolized, but rather, is
22    mostly excreted unchanged in urine.
23           The lack of a role for DCA in tetrachloroethylene liver toxicity is supported by the
24    limited findings of Maronpot and his coworkers (Anna et al., 1994;  Maronpot et al., 1995),
25    which showed no similarities in mutation  spectra between tetrachloroethylene-induced liver
26    tumors and DCA-induced liver tumors. It is interesting to note, however, that the kinetics of
27    metabolism and the sensitivity of target tissue to TCA and DCA and their precursors are likely of
28    key importance to  understanding species differences in responsiveness to tetrachloroethylene.
29           Dechlorination of TCA to DCA is catalyzed by gut contents (ingested food and bacteria)
30    of the rat  and mouse (Moghaddam et al., 1996); isolated mouse microflora have been shown to
31    convert TCA to DCA (Moghaddam et al., 1997).  DCA can be rather quickly processed to other
32    chemical  species, such as monochloroacetic acid (MCA), glycolic acid, glyoxylic acid, and
33    oxalic acid (Abbas and Fisher, 1997; Lash et al., 2000; Bull, 2000; Board et al., 1997; Tong et
34    al., 1998a, b; Lash and Parker, 2001).  Conversion to glyoxylic acid is thought to occur by action
35    of the GST zeta (GSTZ in humans) isoform of glutathione S-transferase (GST; Lash et al., 2000).
36    DCA is a mechanism-based inactivator of GSTZ, of which five polymorphic variants exist
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-9        DRAFT-DO NOT CITE OR QUOTE

-------
 1    (Tzeng et al., 2000). Potent and irreversible inhibition of GSTZ activity by DCA occurs, and the
 2    substrate inhibition of the enzyme in vitro differs between rats and humans, with the enzyme
 3    being relatively more sensitive to inhibition by DCA in rats. Further degradation of DCA in the
 4    liver occurs primarily in hepatic cytosol (Lipscomb et al., 1995). Human liver cytosol is less
 5    efficient than either rat or mouse liver cytosol in processing DCA (Lipscomb et al., 1995).
 6           Trichloroacetyl ethanolamide also may be formed from the tetrachloroethylene oxide
 7    intermediate or from the alternative chlorine migration in an oxygenated tetrachloroethylene
 8    transition state. 14CO2 has been recovered from laboratory animals administered 14C-labeled
 9    tetrachloroethylene (Frantz and Watanabe, 1983; Pegg et al., 1979; Schumann et al., 1980).  A
10    measurable portion of tetrachloroethylene is completely metabolized in a dose-dependent manner
11    to CO2.  The oxalate metabolite excretory product may be derived from DCA or MCA (long et
12    al., 1998a, b), although oxalic acid is also produced from the epoxide through tetrachlorodiacetyl
13    chloride and oxalic acid dichloride intermediates (Pegg et al., 1979; Costa and Ivanetich, 1980).
14    The occurrence of oxalic acid and of CO2 as major metabolites of tetrachloroethylene, at least in
15    rodents, indicates the existence of pathway(s) of metabolism other than the primary TCA
16    pathway.
17
18    3.3.3.1.4. Species-dependent differences. Although thought to be qualitatively similar, there are
19    clear differences among species in the quantitative aspects of tetrachloroethylene metabolism
20    (Schumann et al., 1980; Ikeda and  Otsuji, 1972; Volkel et al., 1998; U.S. EPA, 1991; Lash and
21    Parker, 2001).  These differences are in the relative yields and  kinetic behavior of metabolites
22    (Volkel et al., 1998; Ohtsuki et al., 1983;  Green et al., 1990; U.S. EPA, 1985, 1991).  Rodents
23    and humans differ in relative rates  of tetrachloroethylene metabolism in key target organs, in the
24    doses at which saturation of metabolism occurs, and in the half-times in the body.
25           The rate of metabolism  of tetrachloroethylene is faster in rodents than in humans and
26    higher blood levels of metabolites are obtained in rodents as compared to humans. The higher
27    blood levels of metabolites in rodents are particularly noticeable at the higher tetrachloroethylene
28    exposure levels because saturation is approached at lower exposure levels in humans than in
29    rodents. The half-time in the body of these metabolites is, however, noticeably longer for
30    humans than for rodents (144 hrs in humans vs. approximately 10 hrs or less in rodents; see  U.S.
31    EPA, 1985). It is for this reason that examinations of tetrachloroethylene concentration and
32    toxicity associations must reflect both blood concentration and time-integrated dose metrics such
33    as area-under-the-curve.
34           A study of species differences in tetrachloroethylene metabolism conducted by Dekant
35    and colleagues is presented in Volkel et al. (1998).  These investigators compared both oxidative
36    and GSH-dependent metabolism in rats and humans exposed for 6 hrs to 10, 20, or 40 ppm
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-10       DRAFT-DO NOT CITE OR QUOTE

-------
 1   tetrachloroethylene by inhalation. Rats were also exposed to 400 ppm concentrations.  TCA was
 2   the major urinary excretion product in both species; however, the elimination half-time was more
 3   than four times slower in humans than in rats. Blood plasma concentrations of the metabolite
 4   were higher (three- to eight-fold, depending on the dose) in rats than in humans exposed to
 5   identical air concentration levels.  These observations are in agreement with metabolic rates in
 6   general, which are higher in mice than in rats; rats, in turn, have higher metabolic rates than do
 7   larger animals, including humans. Dekant and his coworkers also reported urinary excretion of
 8   DCA by rats but not humans. They concluded most of the DC A resulted from GSH-dependent
 9   metabolism. DCA, however, is further metabolized by P450 enzymes,  which, in turn, limits its
10   detectability in urine.
11
12   3.3.3.1.5. Cytochrome P450 (CYP) isoforms and genetic polymorphisms. Relatively few
13   studies provide information about which specific CYP isoforms play a role in tetrachloroethylene
14   oxidative metabolism. CYP2E1 is presumed to have an important role  in tetrachloroethylene
15   metabolism (Lash and Parker, 2001); however, the chemical-specific related  data are too sparse
16   to provide strong support for this assumption (Doherty et al., 1996). CYP2B1/2 may also be
17   important for the metabolism of tetrachloroethylene. CYP3A isoenzymes may contribute to the
18   generation  of reactive sulfoxides from metabolites of the GSH pathway (see below). Costa and
19   Ivanetich (1980) showed increased hepatic metabolism following treatment with agents now
20   known to induce these isoenzymes specifically.
21          Genetic polymorphisms are DNA sequence variations that result in changes in protein
22   sequence of an enzyme that can alter the enzyme's ability to catalyze a  reaction or alter the
23   expression  of an allele. Polymorphisms are known for most of the CYP enzymes including
24   CYP2E1 (McCarver et al., 1998; Hu  et al., 1999) and CYP3A4 (Sata et al., 2000; Westlind et al.,
25   1999).
26          Metabolism of tetrachloroethylene to its putative epoxide is likely affected by CYP
27   enzymes. The metabolism of the putative metabolite chloral hydrate to TCOH and TCA may be
28   catalyzed by both alcohol dehydrogenase and CYP2E1.  Oxidation of TCOH is catalyzed by
29   P450 enzymes, with CYP2E1 the likely predominant isoform involved, although other
30   isoenzymes may also play a role, even substituting for CYP2E1 in processing
31   tetrachloroethylene. Rat kidney expresses  CYP2E1 (Cummings et al., 1999;  Speerschneider and
32   Dekant, 1995), although human kidney has not been shown to do so (Amet et al.,  1997;
33   Cummings et al., 2000a).  Therefore, renal CYP metabolism by this isoform in rat kidney would
34   be relevant only  insofar as the involvement of other isoenzymes in metabolizing
3 5   tetrachl oroethy 1 ene vi a thi s route.
36
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  3-11      DRAFT-DO NOT CITE OR QUOTE

-------
 1    3.3.3.2.  Glutathione (GSH) Conjugation Pathway
 2          Figure 3-2 shows the second metabolic pathway for tetrachloroethylene (Dekant et al.,
 3    1987, 1988; Sausen and Elfarra, 1990; Volkel et al., 1998). This GSH pathway was
 4    subsequently shown to exist in both rodents and humans (Volkel et al., 1998).
 5          The GSH pathway is initiated by the conjugation of the parent tetrachloroethylene
 6    molecule with GSH to form S-(l,2,2-trichlorovinyl)glutathione (trichlorovinyl glutathione, or
 7    TCVG). This reaction, which is catalyzed by the GSH-S-transferase enzymes (GSTs), a group
 8    of enzyme isoforms, was traditionally considered to be a detoxification reaction, leading to more
 9    water-soluble compounds that are more readily excreted.  In many cases, however, as with
10    certain halogenated alkanes and alkenes such as tetrachloroethylene, GSH conjugation can be
11    important for bioactivation. The critical step for the alkenes would occur after the enzymatic
12    removal of the glutamyl and glycine residues from the GSH conjugate to yield the corresponding
13    cysteine S-conjugate, which in the case of tetrachloroethylene would be S-(l,2,2-trichlorovinyl)
14    cysteine (trichlorovinyl cysteine, or TCVC; Dekant et al., 1987, 1989; Anders et al., 1988; Green
15    et al., 1990; Vamvakas et al.,  1987, 1989a, b, c).
16          Tetrachloroethylene conjugation with GSH is thought to primarily occur through an
17    interorgan process.  GSH conjugation occurs predominantly in the liver to form TCVG, which is
18    then further metabolized to the corresponding cysteine conjugate, TCVC, by the enzymes
19    gamma-glutamyltransferase (GGT) and cysteinylglycine dipeptidase. TCVC acts as a substrate
20    for several enzymes. Beta lyase cleaves TCVC to yield an unstable thiol, giving rise to cytotoxic
21    and mutagenic products, particularly the reactive thioketene. TCVC may also be activated by
22    cysteine conjugate S-oxidase  activity, which also can rearrange to form the reactive thioketene.
23    Although Green et al. (1990)  hypothesized that GSH conjugation and subsequent activation of
24    tetrachloroethylene did not occur in humans, the N-acetyl urinary metabolite has subsequently
25    been clearly identified in humans exposed to tetrachloroethylene in occupational settings, in
26    laboratory studies, and in residential buildings (Birner et al., 1996; Volkel et al., 1998; Schreiber
27    et al., 2002). Therefore, this pathway is now known to operate in humans as well as in rodents.
28          This GSH conjugation pathway was recognized much later than was the oxidative
29    pathway, probably because it is relatively minor quantitatively compared with the CYP pathway,
30    yet it may be lexicologically influential (U.S. EPA, 1991; IARC, 1995; Lash and Parker, 2001).
31    The evidence for this is based on in vitro kinetics and the relatively low recovery of urinary
32    mercapturates as compared with urinary TCA and other CYP-derived metabolites (Green et al.,
33    1990; Birner et al., 1996). Urinary mercapturates comprise from 1% to as little as 0.03% of total
34    recovered urinary metabolites, but this does not reflect the total flux through the GSH pathway
35    but rather only the portion that is excreted.  In particular, the amount of the mercapturate product

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-12       DRAFT-DO NOT CITE OR QUOTE

-------
                      Cl          Cl
                          GST
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
              c     c
          Cl          Cl
                 1
                                                   Cl        O    COO

                                           C    C    CH2    C    CH2
                                                                  £T)
                                       Cl          S    CH   NH^NH,

                                               GGT      HN   CH2  CH

                                                     V                     e
                                                             C    CH2  COO

                                                            O
                                          DP
                  P-Lyase
ci          c:

   c    c

Cl       6  S
                                      Cl          ci  V

                                         C     C     CH,  COO"

                                      Cl          S    CH     3

                                                     tl7 K i l_j
                             ci
                                           FM03
                                            or
                                           P450
                                              Cl
                                                 CCNAT
                                                             Acylase
                                                            C     C    CH,   COO"
           CCS
        Cl
                                C    C     R

                             Cl          S
                                     P-Lyase
                                       or
                                   Spontaneous
                       CljOHCCXT
                          9*
                                                              Cl          S    CH
                                                              C         O    CH,
                                                             C   NH

                                                           e°
                                                          CYP3A1/2 (rat)
                                                          CYP3A4 (human)
                                        Cl          Cl  V

                                            C    C    CH,    COO®

                                        Cl     .     S    CH   O
                                               5
                                                    O    HN   C

                                                            CH,    0°
       Figure 3-2. Metabolism of tetrachloroethylene by the glutathione
       conjugation pathway.  Tetrachloroethylene and identified (*) metabolites:  1
       tetrachloroethylene, 2 S-l,2,2-trichlorovinyl) glutathione (TCVG), 3
       S-(l,2,2-trichlorovinyl)-L-cysteine (TCVC), 4 N-acetyl trichlorovinyl cysteine
       (NAcTCVC), 5 NAcTCVC sulfoxide, 6 1,2,2-trichlorovinylthiol, 7 TCVCSO, 8
       2,2-dichlorothioketene,  and 9 dichloroacetate. Enzymes: glutathione-
       S-transferase (GST), gamma-glutamyltransferase (GGT), dipeptidase (DP), beta
       lyase, FMO3, CCNAT,  acylase, CYP3A1/2, and CYP3A4. Unstable reactive
       metabolites are shown in brackets.

       Source:  Lash and Parker (2001).
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   3-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1    excreted in the urine also does not reflect the amount of the more important portion that is
 2    converted to toxic by-products through further metabolism.
 3          For tetrachloroethylene, the GSH pathway is associated with renal toxicity (Anders et al.,
 4    1988; Dekant et al., 1989; U.S. EPA, 1991; IARC, 1995; Lash et al., 2000; Lash and Parker,
 5    2001). The initial conjugation with GSH occurs mainly in the liver (Dekant et al., 1987; Green
 6    et al., 1990; Vamvakas et al.,  1987, 1989a), with transport of the conjugate and its cysteine
 7    counterpart to the kidney target organ for further processing.  This first step also occurs within
 8    the kidney (Lash et al., 1998). As shown in Figure 3-2, tetrachloroethylene is initially
 9    conjugated with GSH to form TCVG. This reaction is catalyzed by cytosolic and microsomal
10    GSTs. TCVG is then processed through the cysteinylglycine conjugate S-(l,2,2,trichlorovinyl)-
11    L-cysteinylglycine to TCVC by the enzymatic removal of glutamyl and glycine residues by GGT
12    and various membrane-bound dipeptidases known as cysteinylglycine dipeptidase (reviewed by
13    Anders et al., 1988; Dekant et al., 1989; U.S. EPA, 1991; Lash and Parker, 2001).  These
14    enzymes reside in tissues other than the kidneys (e.g., the brain), indicating a potential for toxic
15    reactive metabolite formation in those tissues as well. Conversion of TCVG to TCVC by these
16    cleavage enzymes leads to a critical bifurcation point of the GSH pathway because the TCVC
17    may be processed by certain enzymes to yield reactive, toxic chemical species, although it may
18    be metabolized via a different route to yield an excretory product (Lash and Parker, 2001).
19          Importantly, the TCVC metabolite may also act as a substrate for renal beta lyases
20    (Dekant et al., 1988 reviewed by Anders et al., 1988; Dekant et al., 1989; U.S. EPA, 1991; Lash
21    et al., 2000; Lash and Parker,  2001). Renal beta lyases are known to cleave TCVC to yield an
22    unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a reactive chemical species that  can
23    form covalent adducts with cellular nucleophiles, including DNA and proteins (Volkel et al.,
24    1999). Beta lyases are a family of pyridoxal phosphate-containing enzymes that are located in
25    several tissues besides the kidneys, including liver and brain, and in intestinal flora, although
26    their substrate specificities may vary. Hepatic beta lyase is distinct from renal beta lyase and  has
27    not been found to have a role in TCVC metabolism. Beta lyase activity is higher in rat kidney
28    than in human kidney (Cooper, 1994; Lash et al., 1990), which is consistent with overall
29    metabolic rates being higher in smaller versus larger mammalian species.
30          In addition to activation by beta lyases, TCVC may be metabolized by a flavin-containing
31    monooxygenase, FMO3, or CYP enzymes to TCVC sulfoxide (TCVCSO), another reactive
32    metabolite (Ripp et al., 1997). TCVSO is a more potent nephrotoxicant than TCVC (Elfarra and
33    Krause 2007). These TCVC sulfoxide and beta lyase cleavage products rearrange, forming a
34    thioketene (Dekant et al., 1988; Ripp et al., 1997), which is a potent acylating agent capable of
35    binding to cellular macromolecules, including DNA (Birner et al.,  1996; Pahler et  al., 1999a,  b;
36    Volkel et al., 1999).  Interestingly, the thioketene can degrade to form DC A, potentially making
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-14       DRAFT-DO NOT CITE OR QUOTE

-------
 1   this metabolite a product of both tetrachloroethylene metabolism pathways (Dekant et al., 1987;
 2   Volkeletal., 1998).
 3          In addition to beta lyase and FMO3/CYP activation of TCVC, reactive sulfoxides can
 4   also be produced by further CYP3A metabolism of N-acetyl-S-(l,2,2 -trichlorovinyl)-L-cysteine
 5   (NAcTCVC; Werner et al., 1996). This tetrachloroethylene-derived mercapturate metabolite
 6   results from TCVC being acetylated via a reversible reaction (Bartels, 1994; Birner et al., 1996;
 7   Duffel and Jakoby, 1982).  N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine may be excreted in the
 8   urine.  However, in addition to its activation to sulfoxides via CYP3 A metabolism, it can also be
 9   transported to other organs and deactylated intracellularly, regenerating the cysteine conjugate
10   TCVC and thus making it available to other enzymes for activation (Uttamsingh et al., 1998).  It
11   should be noted that the N-acetylation reaction is catalyzed by an enzyme located in the
12   endoplasmic reticulum that is distinct from the cytosolic enzymes that are polymorphic in
13   humans (Lash and Parker, 2001).
14
15   3.3.3.2.1. Glutathione S-transferase (GST) isoenzymes/polymorphisms. GSTs are a family of
16   isoenzymes (Mannervik, 1985) found in cytoplasm.  A distinct microsomal GST isoenzyme also
17   exists in most mammalian tissues (Otieno et al., 1997). Although GST activity occurs in most
18   cell types, the liver is by far the predominant site of GSH conjugation.  GST alpha, designated as
19   GSTA in humans, is the predominate isoenzyme expressed in normal kidney from rodents and
20   humans (Campbell et al., 1991; Overby et al., 1994; Mitchell et al., 1997; Rodilla et al., 1998;
21   Cummings et al., 2000b). Available data thus far do not indicate that variability in activity of
22   this isoenzyme is important to differences in individual susceptibility to toxicity. GSTZ
23   catalyzes the oxidative metabolism of DCA to glyoxylate (Board et al.,  1997; Tong et al., 1998a,
24   b), however, the tetrachloroethylene metabolite DCA has been shown to be a potent, irreversible
25   inhibitor of GSTZ activity (Tzeng et al., 2000).
26          There are five human polymorphic variants of this GSTZ isoenzyme (Tzeng et al., 2000;
27   U.S. EPA, 1998). These genetic polymorphisms may influence tetrachloroethylene metabolism
28   although human data regarding this hypothesis are lacking. There are some species differences
29   in the other three cytoplasmic GSTs relevant to liver and kidney. GSTP expression is the most
30   variable and appears to be polymorphic in humans (Rodilla et al., 1998). It has been found in rat
31   liver (Cummings et al., 1999), but only in biliary ducts in humans (Terrier et al., 1990; Campbell
32   et al., 1991). GSTP has been detected within the human kidney in various cell types (Terrier et
33   al., 1990) but has not been isolated from rat kidney cells (Cummings et al., 1999), although
34   GSTP has also been detected in rabbit kidney (Cummings et al., 1999).
35          Two homodimeric GST theta (GSTT) isoenzymes have been identified in human kidney
36   (Veitch et al., 1997; Cummings et al., 2000a).  GSTT has been detected in rat and mouse liver
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   3-15       DRAFT-DO NOT CITE OR QUOTE

-------
 1    and in mouse but not rat kidney (Cummings et al., 1999; Quondamatteo et al., 1998).  GST mu
 2    (GSTM) has been detected in rat kidney distal tubule cells (Cummings et al., 2000b) and in
 3    mouse and rabbit liver and kidney (Overby et al., 1994; Mitchell et al., 1997), but it was not
 4    detected in human kidney (Cummings et al., 2000a). It is not clear just how the differences in
 5    these isoenzymes are related to species differences in tetrachloroethylene toxicity because the
 6    isoenzyme specificity and reaction rates have not yet been studied with regard to
 7    tetrachloroethylene (Lash and Parker, 2001).
 8          Some controversy surrounds the importance of the GSH conjugation pathway with regard
 9    to tetrachloroethylene metabolism in humans. As noted above, the GSH pathway for
10    tetrachloroethylene was originally demonstrated only in rodents, and interpretation of the then-
11    existing data led some scientists to conclude that the pathway was not operative in humans
12    (Green et al., 1990; U.S. EPA, 1991).  More recent data clearly demonstrate the existence of the
13    pathway in humans (Birner et al., 1996; Volkel et al., 1998; Schreiber et al., 2002).  There are
14    discrepancies regarding reported rates of tetrachloroethylene GSH metabolism, however (Green
15    et al., 1990; Dekant et al., 1987; Lash et al., 1998; Lash and Parker, 2001). These differences
16    may be due, in part, to different chemical assay methodology or to problems resulting from the
17    stability of the chemical product being measured or both (Lash and Parker, 2001).  Some of the
18    published findings concerning TCVG production would not predict any less susceptibility for
19    humans than for rodents with regard to renal toxicity (Lash et al., 1998).
20
21    3.3.3.2.2. Gamma-glutamyltransferase (GGT).  Species-dependent differences in GGT
22    (Hinchman and Ballatori, 1990) also are not thought to be limiting, because renal activity is
23    present at high enough levels even in humans so that GGT activity is not the rate-limiting step in
24    the metabolism. Species-dependent differences in this enzyme (described below) would have
25    only a very small quantitative effect on the overall metabolism of TCVG and other similar GSH
26    conjugates.  Species differences in GGT activities, therefore, would not have a major role in
27    species differences in renal toxicity (Lash and Parker, 2001) in affecting transformation of
28    TCVG to TCVC, and thus, should not be important to differences in susceptibility to
29    tetrachloroethylene-induced renal toxicity.
30          GGT is the only enzyme that can split the gamma-glutamyl bond in the GSH conjugates
31    to form cysteine conjugates (Lash and Parker, 2001). It is this reaction that creates TCVC, the
32    substrate for the enzymes that generate the toxic metabolites.  Therefore, the distribution of GGT
33    is important. Renal proximal tubular cells have the highest activities of GGT of all tissues,
34    although GGT activity also occurs in the liver, and the kidney-to-liver ratio of this enzyme varies
35    among species. In the rat, the specific activity ratio is 875 (Hinchman and Ballatori, 1990). The
36    ratio is lower in other species that have been studied. The tissue distribution and relative activity
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-16       DRAFT-DO NOT CITE OR QUOTE

-------
 1    have not been fully studied in humans, but it is known that GOT activity is considerably higher
 2    in human liver than in rodent liver (Lash and Parker, 2001). The kidney-to-liver ratio of GGT
 3    for humans is thought to be closer to those of pigs (2) and Macaques (5) than to those of rats or
 4    mice (423). For this reason, use of a rodent model for the processing of the tetrachloroethylene
 5    GSH conjugate to the corresponding cysteine conjugate would overestimate the contribution of
 6    the kidneys and underestimate the contribution of the liver in cleaving TCVG to TCVC. Even
 7    so, the liver excretes most of the cysteine conjugates such as TCVC into the bile or plasma,
 8    where it is cycled to the kidneys and taken up into renal epithelial cells. So, the TCVC will still
 9    end up in the kidneys.
10
11    3.3.3.2.3. Betalyase. The beta lyase enzyme is among the most important activator of toxic
12    products in the conjugation pathway, a fact particularly well documented in the kidney. There
13    are some data, however, that indicate that renal beta lyase-dependent metabolism is greater in
14    rats than in mice or in humans and greater in male than in female rats (Green et al., 1990; Lash et
15    al., 1990; Volkel et al., 1998).  This is not entirely in keeping with metabolic rates in general,
16    which are higher in mice than in rats, and rats, in turn have higher metabolic rates than do larger
17    animals, including humans.  Studies that measured only cytoplasmic beta lyase activity did not
18    consider the importance of mitochondrial beta lyase activity, which may be key to
19    tetrachloroethylene metabolite toxicity (Lash et al., 2001).
20          In contrast, it must also be noted that species comparisons of tetrachloroethylene
21    metabolism in chronic exposures on a surface area- or metabolic-rate basis rather than on a direct
22    body-weight basis, particularly when including the total area-under-the-curve (AUC) for amount
23    metabolized, indicate that metabolite production in rats and humans may not differ significantly
24    (U.S. EPA, 1986; Rhomberg, 1992; Calabrese, 1983).  The fact is that metabolic rates and the
25    amounts metabolized are not the same thing. Metabolic rates are always faster in smaller
26    species.  Total AUC may or may not be similar among species. Even if AUC is the same, the
27    peak blood levels may differ greatly from species to species. In other words, the
28    pharmacokinetics are not the same.
29          The higher percentage of mercapturate found in rat versus human urine does not indicate
30    a higher level of production of toxic products in the rat, because excreted mercapturate allows no
31    estimate of the amount of TCVC or N-acetyl TCVC being processed through alternate routes
32    (Lash and Parker, 2001).  The relatively higher percentage of DCA in the urine may, however,
33    indicate a relatively higher beta lyase enzyme activity  and a higher thioketene production in rats
34    if the DCA is indeed largely the product  of the GST pathway rather than the oxidative pathway
35    (Volkel et al., 1998). It is not known whether sex-dependent variation  of beta lyase activity
36    exists in humans as it does in rats (Volkel et al., 1998).
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1          And finally, it is important to note that because the enzymes involved in this activation
 2    pathway are also present in other tissues (Dohn and Anders, 1982; Stevens, 1985; Stevens and
 3    Jacoby, 1983; Tateishi et al., 1978; Tomisawa et al., 1984, 1986; Larsen, 1985; Larsen and
 4    Stevens, 1986; Alberati-Giani et al., 1995; Malherbe et al., 1995), there exists a potential for
 5    formation of the reactive metabolites at sites other than the kidney, e.g.,  in the brain.  In one
 6    carcinogenicity bioassay of tetrachloroethylene, a biologically significant elevation of gliomas in
 7    the rat brain was reported (NTP, 1986).  Whether or not toxic metabolites resulting from beta
 8    lyase activity in the brain play a role in the development of the gliomas in the rat has not been
 9    studied.  The possibility that such tetrachloroethylene metabolites could be involved in the mode
10    of tumorigenic action producing gliomas is not unrealistic.
11
12    3.3.3.3. Relative Roles of the Cytochrome P450 (CYP) and Glutathione (GSH) Pathways
13          Although it is clear that the oxidative CYP pathway is quantitatively more important than
14    the GSH conjugation pathway, the interorgan patterns for some of the intermediate metabolites,
15    as well as the relative toxicity of certain key metabolites generated from these pathways,
16    influence the relative importance of the two pathways in determining toxicity.  It is still not
17    certain which metabolites, alone or in combination, are explicitly responsible for specific
18    tetrachloroethylene toxicities, and it is likely that different metabolites contribute to toxicity at
19    different target sites. In general, CYP metabolism is associated with tetrachloroethylene-induced
20    liver toxicity, whereas GSH conjugation followed by further processing  by beta lyase and other
21    enzymes is associated with tetrachloroethylene-induced renal toxicity. There is a possibility that
22    beta lyase products could contribute to toxicity in the brain,  for example, and be a factor in the
23    gliomas observed in rats.  The parent compound itself is also likely to be a contributing factor to
24    tetrachloroethylene neurotoxicity, particularly central nervous system (CNS) effects.
25          Data from experiments designed to assess the effects of enzyme modulation suggest
26    competition between the two pathways (Dekant et al., 1987; Lash et al.,  1999; Volkel et al.,
27    1998; Lash et al., 2001).  Other data show relatively low urinary excretion of mercapturates  as
28    compared to CYP-derived products.  On the basis of these findings, some scientists have
29    concluded that there is a lack of toxicological significance for the low-affinity, low-activity GSH
30    pathway except when the high-affinity CYP pathway approaches saturation (Green et al., 1990,
31    1997; Volkel et al., 1998).  However, this conclusion does not consider the relative toxicological
32    potency or chemical reactivity of the metabolites from the two pathways or the fact that the
33    amount of mercapturate excreted is not a valid quantitative indicator of the extent of conjugative
34    pathway metabolism (Lash and Parker, 2001).
35          Specific tetrachloroethylene metabolites are known to be associated with certain
36    toxicities when they are administered directly. Exactly how these same  compounds—as
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    3-18       DRAFT-DO NOT CITE OR QUOTE

-------
 1    metabolites of tetrachloroethylene—contribute to the various toxicities associated with exposure
 2    to the parent compound is not yet well understood.
 O
 4    3.3.4.  Susceptibility
 5          Differences in enzyme activity may lead to variations among individuals in their
 6    sensitivity to tetrachloroethylene toxi cities. A 10-fold difference in CYP enzyme metabolic
 7    capacity among humans is a generally accepted norm.  Although individual variations in the
 8    CYP2E1 enzymatic activity as high as 20- to 50-fold have been reported (Stephens et al., 1994;
 9    Yoo et al., 1988; Lieber, 1997), these in vitro measurements would be taken out of physiological
10    context if used to estimate in vivo interindividual variations.  Measurable and obvious
11    differences in CYP enzymatic activity are observed among various ethnic groups and age groups
12    (Goldstein et al., 1969; Raunio et al.,  1995). No  chemical-specific data regarding the manner in
13    which CYP enzyme isoforms might affect susceptibility to adverse effects  are available for
14    tetrachloroethylene.
15          Diagnosis of polymorphisms in carcinogen-activating and -inactivating enzymes and
16    cancer susceptibility have been noted (Stephens et al.,  1994; Yoo et al., 1988; Raucy, 1995).
17    Potential strain-dependent differences among rodents and human genetic polymorphisms in
18    metabolizing enzymes involved in biotransformation of tetrachloroethylene are now known to
19    exist.  Whether CYP polymorphisms could account for interindividual variation in
20    tetrachloroethylene metabolism among humans-and thus differences in susceptibility to
21    tetrachloroethylene-induced toxi cities-is not known.
22          The GSTs involved in tetrachloroethylene metabolism are described in Section 3.3.2. A
23    potential exists for interindividual variation to occur in tetrachloroethylene metabolism as a
24    result of variability in GST enzyme expression. It is important to note that GST polymorphism
25    has been associated with increased risk of kidney cancer in people exposed to trichloroethylene.
26    This information is discussed in EPA's draft health assessment report on trichloroethylene (U.S.
27    EPA, 2001). There are no direct, chemical-distinctive data with regard to the specific isoenzyme
28    family responsible for TCVG formation in metabolism of tetrachloroethylene.  There are
29    species-dependent differences as to which isozymes occur in liver and kidney, although it is
30    unknown how the various enzymes are related to differences in metabolism of
31    tetrachloroethylene.  The compound is likely a good substrate for GSTA (Lash and Parker,
32    2001). GSTT and GSTP occur in human kidney, as does GSTA, the primary isozyme in human
33    kidney, meaning that there is a potential for differences in the ability to produce TCVG. GSTZ
34    transforms the tetrachloroethylene metabolite DCA. DCA has also been shown to have a potent
35    irreversible inhibitory effect on the GSTZ isoenzyme, which is known to have at least four
36    polymorphic variations.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-19       DRAFT-DO NOT CITE  OR QUOTE

-------
 1          Inhibition or induction of the enzymes responsible for tetrachloroethylene metabolism
 2    can, and likely does, alter susceptibility to toxicity (U.S. EPA, 1985; IARC, 1995; Lash and
 3    Parker, 2001).  Numerous environmental pollutants and therapeutic agents alike have the
 4    potential to induce or inhibit tetrachloroethylene-metabolizing enzymes. For example,
 5    tetrachloroethylene metabolism is increased by inducers of cytochrome CYP enzymes such as
 6    toluene, phenobarbital, and pregnenolone-16 alpha-carbonitrile, whereas CYP inhibitors such as
 7    SKF 525A, metyrapone, and carbon monoxide decrease tetrachloroethylene metabolism (Moslen
 8    et al., 1977; Ikeda and Imanura, 1973; Costa and Ivanetich, 1980).  Chronic exposure to
 9    tetrachloroethylene has been shown to cause self-induction of metabolism (Kaemmerer et al.,
10    1982; Savolainen et al.,  1977; Vainio et al., 1976).  Other factors, such as health status or disease
11    state, activity patterns, or concomitant exposure to other chemicals, can potentially influence
12    tetrachloroethylene metabolism and its resulting toxicity.  Section 4.9 addresses issues
13    coexposures and cumulative risk in greater detail.
14
15    3.3.5.  Comparison of Tetrachloroethylene Metabolism with Trichloroethylene Metabolism
16          Tetrachloroethylene is structurally related to trichloroethylene, and these two compounds
17    cause similar adverse health effects. The toxic effects, with the possible exception of
18    neurotoxicity, are attributed to metabolites.  TCA, DCA,  chloral, and TCOH are reported P450
19    biotransformation products of tetrachloroethylene and trichloroethylene; however, both the
20    relative amounts of these metabolites produced and the precursor intermediates in the oxidative
21    pathways are different for the two compounds. Interestingly, although tetrachloroethylene is not
22    as extensively biotransformed as trichloroethylene,  it is slightly more toxic. Differences in
23    pharmacodynamics of precursors to P450 metabolic products as well as pharmacokinetic
24    differences between the two parent compounds may be related to their pharmacologic potencies
25    (Buben and O'Flaherty, 1985).  Excretion of urinary mercapturates indicates that, relative to
26    P450 oxidation, tetrachloroethylene is more extensively metabolized via GSH conjugation than
27    is trichloroethylene. However, these urinary excretion products do not reflect the total flux
28    through the GSH pathway since the resulting glutathione and cysteine conjugates have been
29    shown to undergo further processing to products that  are  highly reactive. The Appendix for
30    Chapter 3 provides  additional discussion of tetrachloroethylene/trichloroethylene comparative
31    metabolism.
32
33    3.4. EXCRETION
34          Tetrachloroethylene is eliminated from the body by two major processes: pulmonary
35    excretion and rate-limited metabolism.  Tetrachloroethylene that is not metabolized is exhaled
36    unchanged, and this elimination process  is the primary pathway of tetrachloroethylene  excretion
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-20      DRAFT-DO NOT CITE OR QUOTE

-------
 1    in humans for all routes of administration (Monster et al., 1979; Stewart et al., 1961, 1970, 1974,
 2    1977; Guberan and Fernandez, 1974; Opdam and Smolders, 1986; Koppel et al., 1985; Stewart
 3    and Dodd, 1964).  Pulmonary elimination of (unchanged) parent compound is also important to
 4    tetrachloroethylene excretion by animals (Pegg et al.,  1979; Yllner, 1961; Frantz and Watanabe,
 5    1983; Schumann et al., 1980; Bogen et al., 1992).  A very small amount of tetrachloroethylene
 6    has been shown to be eliminated through the skin (Bolanowska and Golacka, 1972); however, it
 7    represents an insignificant percent of total tetrachloroethylene elimination.
 8          Pulmonary elimination of unchanged tetrachloroethylene and other volatile compounds is
 9    related to ventilation rate, cardiac output, and the solubility of the compound in blood  and tissue.
10    The lung clearance of tetrachloroethylene in six adults exposed at rest to 72 ppm and 144 ppm of
11    tetrachloroethylene averaged 6.1 L/min initially and decreased to 3.8 L/min after 4 hrs (Monster
12    et al., 1979). Lung clearance represents the volume of air from which all tetrachloroethylene can
13    be removed per unit time. Normal ventilation rates in adults range from 5-8 L air/min at rest.
14    Pulmonary elimination of unchanged tetrachloroethylene at the end of exposure is a first-order
15    diffusion process across the lungs from blood into  alveolar air, and it can be thought of as the
16    inverted equivalent of its uptake from the lungs. Pulmonary excretion occurs in three  first-order
17    phases of desaturation of blood vessel-rich tissues, muscle tissue, and adipose tissues (Monster et
18    al., 1979; Guberan and Fernandez, 1974). For humans, the half-times of elimination from these
19    three tissue groups are  12 to 16 hrs, 30 to 40 hrs, and 55 to 65 hrs, respectively (Monster et al.,
20    1979).
21          The long half-time of tetrachloroethylene elimination from adipose tissue, due to the high
22    adipose tissue/blood partition coefficient and the low rate of blood perfusion  of the fat tissue
23    (Eger, 1963), is independent of the body burden of tetrachloroethylene, indicated by parallel
24    blood and exhaled air concentration decay curves (U.S. EPA, 1985). However, the exhaled air or
25    end alveolar air concentrations and the blood concentrations after exposure and throughout
26    desaturation are proportional to the acquired body burden or exposure concentration and
27    duration, and they can serve as a means of estimating body burdens.  The half-life of
28    tetrachloroethylene in the human body, measured as the inverse of the slope of the log-
29    concentration versus time curve of the exhaled chemical, varies from 5 to 20  minutes for the first
30    phase of elimination up to approximately 50 hrs during its extended phase (Chien,  1997; Monster
31    et al., 1979). The long half-time of tetrachloroethylene pulmonary excretion  indicates that a
32    considerable time is necessary to completely eliminate the compound.  This time is greater than
33    five times the half-life, or about 2 weeks for humans.  For the rat, the half-time of pulmonary
34    elimination is about 7 hrs.
35          Metabolism of tetrachloroethylene provides another means of elimination of the parent
36    compound.  Metabolism in humans is not considered to be as important to tetrachloroethylene
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-21       DRAFT-DO NOT CITE OR QUOTE

-------
 1    excretion as is pulmonary excretion; however, at low ambient exposure concentrations, this may
 2    not be the case.  The rationale for assigning greater importance to elimination by metabolism in
 3    humans is discussed later in this section. The biotransformation process is well accepted as
 4    being important to elimination of tetrachloroethylene in rodents (see Metabolism, Section 3.3).
 5          The mean half-time of elimination for total trichloro-compounds for 13 subjects exposed
 6    to tetrachloroethylene was determined to be 144 hrs (Ikeda and Imamura,  1973). When TCA is
 7    administered directly, however, the half-life is not that long.  The longer half-life of TCA from
 8    tetrachloroethylene metabolism is likely due to constant metabolic conversion of the parent
 9    compound to TCA as tetrachloroethylene is cycled to the liver over the period of time it is
10    released from adipose tissue.
11          The urinary excretion of tetrachloroethylene biotransformation products, primarily TCA,
12    has been thought to represent only a small percent of the total absorbed dose of
13    tetrachloroethylene in humans  (U.S. EPA, 1985; ATSDR, 1997). Urinary excretion of TCA (or
14    total trichloro-compounds) was estimated to be only 1 to 3% in balance studies conducted in
15    humans (Stewart et al., 1961, 1970; Monster et al., 1979, 1983; Monster and Houtkooper, 1979;
16    Boettner and Muranko, 1969; Ikeda et al., 1972; Essing et al., 1973; Fernandez et al., 1976; May,
17    1976). The shortcomings of the human balance studies include the lack of follow-up of the
18    subjects over a long time period.  It is highly  likely that a larger percent of the
19    tetrachloroethylene dose was eventually metabolized.  Not all of the dose was accounted for in
20    these studies, indicating that more of the dose may be metabolized. Part of the dose may be
21    metabolized to biotransformation products, such as oxalic acid, that were not measured.  It is
22    important to note that estimates of risk calculated directly from the data from such studies would
23    seriously underestimate risk of exposure, because the tetrachloroethylene dose in some of these
24    studies does not likely reflect low-dose  exposure metabolism (U.S. EPA, 1985,  1991; Bois et al.,
25    1996).
26          A literature review published by Hattis et al. (1990) reported estimates of the fraction of
27    tetrachloroethylene metabolized at a low dose of 1 ppm to range from 2 to 86%.  Based on data
28    from the 1979 Monster et al. study, Bois and his colleagues (Bois et al., 1996) determined that at
29    exposure levels above the current occupational  standards, a median of approximately 1.5% of
30    inhaled tetrachloroethylene would be metabolized, whereas at ambient air levels (0.001 ppm) the
31    median estimate of the fraction of inhaled dose that would be metabolized is 36%, a considerably
32    higher fraction of the dose.
33          Metabolism of tetrachloroethylene has generally been reported to contribute more to its
34    elimination in rats and mice than in humans.  The relative importance of metabolism elimination
35    of tetrachloroethylene in rodents  depends on the species and  the dose (Pegg et al., 1979;
36    Schumann et al., 1980; Dallas et  al., 1994a; Bogen et al., 1992). As the body burden of
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-22       DRAFT-DO NOT CITE OR QUOTE

-------
 1    tetrachloroethylene is increased in the rat or mouse, the percentage excreted as unchanged parent
 2    compound increases. Conversely, as metabolism is the other principal route of elimination of
 3    tetrachloroethylene, when the body burden increases, the percentage of the burden metabolized
 4    decreases, although the absolute amount metabolized increases (Pegg et al., 1979; Schumann et
 5    al., 1980). These observations  suggest that, in the rodent, metabolism of tetrachloroethylene and
 6    urinary excretion of its metabolites are rate limited and dose dependent, whereas pulmonary
 7    excretion is a first-order process and is dose independent, with half-time and rate constant being
 8    independent of the dose.  Data from studies by Filser and Bolt (1979) and Buben and O'Flaherty
 9    (1985) suggest that elimination of tetrachloroethylene by metabolism is greater in mice than in
10    rats.
11
12    3.5. PHYSIOLOGICALLY BASED AND OTHER TOXICOKINETIC MODELING
13          Most of the understanding of the pharmacokinetics of tetrachloroethylene in humans is
14    based on a limited number of human data sets (Monster et al., 1979; Fernandez et al., 1976;
15    Volkel et al., 1998) and on extrapolations from animal data to humans using PBPK modeling.
16    PBPK models can provide estimates of tissue concentration as well as total metabolism of
17    tetrachloroethylene.  Models that incorporate transfer into milk and subsequent infant exposure
18    have also been validated using measured human milk concentrations of tetrachloroethylene
19    (Byczkowski and Fisher, 1994). In addition, researchers have looked at the variability in the
20    measured and estimated PBPK model parameters and the implications for applying the models to
21    risk assessment. The critical difference among the various models is in their different
22    approaches to estimating the metabolic parameters.
23
24    3.5.1. Various Physiologically Based Pharmacokinetic (PBPK) Models
25          Chen and Blancato (1987) developed a PBPK model for rats, mice, and humans. The
26    metabolic parameters maximum velocity (Vmax) and Michaelis-Menten constant (Km) were
27    derived by fitting the model to the total amount of metabolized tetrachloroethylene.
28    Experimental data on total metabolite were available for  rodents.  However, for humans, it was
29    assumed that the urinary metabolite TCA,  as measured by Monster et al. (1979), accounted for
30    30% of the total metabolite.  This percentage was chosen because it resulted in a better fit. The
31    model consisted of five compartments: lung, fat tissue, richly perfused tissue, poorly perfused
32    tissue, and liver. The model was used to estimate cancer risk from inhalation and drinking water
33    exposures, based on total daily  absorbed tetrachloroethylene.
34          Reitz et al. (1996) developed a PBPK model for rats, mice, and humans that  describes the
35    total metabolism of tetrachloroethylene using Michaelis-Menten kinetics. The partition
36    coefficients for the five tissue compartments were measured independently and were similar to
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1    those used by Chen and Blancato (1987), giving confidence in the reasonableness of both sets of
 2    numbers. For rats and mice, the metabolic parameters Vmax and Km, as well as the volume and
 3    blood flow rates of the fat compartment, were obtained by simultaneously optimizing the fit to
 4    three sets of in vivo data gathered in 6-hr inhalation radiolabeled tetrachloroethylene exposure
 5    studies.  These data were (a) concentration of tetrachloroethylene in exhaled breath, (b)
 6    radioactive body burden present in animals at end of exposure, and (c) total post-exposure
 7    radioactive metabolites recovered from all excreta and carcass homogenates.
 8          The metabolic parameters for humans were estimated as follows using a "parallelogram
 9    approach" (Reitz et al., 1989). First-order constants for the rate of metabolism were measured in
10    vitro using isolated liver microsomes of all three species. The ratio of these in vivo and in vitro
11    metabolic rates was assumed to be nearly constant across species,  as was found to be the case for
12    rats and mice. Using this constant ratio, the human  in vivo metabolic rate constant per gram of
13    liver could be determined from the human in vitro value. Km was  assumed to be invariant across
14    species because it is  derived solely from the reaction rate constants for the enzyme-catalyzed
15    metabolic reactions.  In contrast, Chen and Blancato (1987) found very different values of Km for
16    each species because this was a fitted parameter in their model.  Vmax, on the other hand, depends
17    on the concentration of the enzyme (substrate) present and is likely to exhibit large inter- and
18    intraspecies variability. As also noted by Reitz et al. (1996), there are inherent uncertainties in
19    estimates from in vitro studies.
20          Reitz et al. (1996) also used a second method for estimating Vmax, which was based on
21    extrapolation from in vivo animal studies of other chemicals metabolized by cytochrome P450
22    enzymes.  Vmax, so estimated, was allometrically scaled to humans. The values obtained by Reitz
23    et al. (1996) through both these independent methods were comparable. The overall average
24    value of 32.9 mg/hr was then used in the PBPK model.  This value compares with the value of
25    42.2 mg/hr used by Chen and Blancato  (1987).  The Chen and Blancato (1987) and Reitz et al.
26    (1996) models differ considerably in the values of Km for humans:  4.66 mg/L and 32.04 mg/L,
27    respectively. Chen and Blancato (1987) also demonstrated that because tetrachloroethylene is so
28    poorly metabolized, the levels of tetrachloroethylene in blood and tissues are not extremely
29    sensitive to the values of Vmax and Km.
30          A human PBPK model was developed for the purpose of investigating neurotoxicological
31    endpoints (Rao and Brown, 1993).  In this case, tetrachloroethylene, not its metabolites, is of
32    toxicological interest. This model was  similar to the others previously discussed except that it
33    included a skin compartment to allow for dermal absorption of tetrachloroethylene from shower
34    water and a brain compartment so that the researchers could evaluate tetrachloroethylene
35    concentrations in this organ. The model was coupled with an exposure model that predicted the

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-24       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
amount of tetrachloroethylene a human would be exposed to from water during showering and
bathing.
       The values for Vmax and Km in Rao and Brown (1993) were estimated using the method
of Reitz and Nolan (1986). The predictions of the model were fit to total metabolite levels
measured in rats and mice (Schumann et al., 1980; Pegg et al.,  1979) to obtain the maximum rate
of metabolism, Vmax (which varied across species and was allometrically related to body weight
raised to 0.74 power), and Km (considered invariant across species). Other parameters for
tetrachloroethylene were derived from various experimental data reported in the literature.  The
value of Vmax for humans was determined by fitting the predicted total metabolite level to that
estimated from urinary metabolite measurements in humans (Monster et al., 1979, and Fernandez
et al., 1976, combined), assuming that the ratio of urinary to total metabolites would be the same
in humans as that observed in rats (equal to 0.71). Although the value of Km for humans  in the
Rao and Brown (1993) and Reitz et al. (1996)  models were similar, their values for Vmax  differed
significantly (see Table 3-1). Rao and Brown  (1993) also provided parameter estimates for 6-
and 10-year-old children.

       Table 3-1. Comparison of Vmax/Km ratios3
Model
Rao and Brown
Reitz
Bois et al.
Gearhart et al.
» max
(mg/hr)b
6.77
32.9
1.86
5.48
Km
(mg/L)
4.56
4.66
0.133C
7.7
» max' ^-m
(L/hr)
1.48
7.06
14
0.712
19
20
21
22
23
24
25
26
27
28
29
30
31
32
a Ratios are for the human models used for animal-to-human extrapolations in this assessment and those of Gearhart
  et al. (1993), which accurately predicted production of major metabolite, TCA, in urine.
 b 70 kg human.
 0 The "posterior" value for Km in Bois et al. (1996) was multiplied by the liver volume and the liver tissue/blood
  partition coefficient in order to conform to the format of the pharmacokinetic equations in this document.
       Bois et al. (1996) used a Bayesian analysis in conjunction with a PBPK model that was
structurally similar to that used by Reitz et al. (1996). The analysis used "prior" empirically
determined distributions for the parameters in the model that were based on values in the
literature and other previously conducted PBPK analyses.  The Markov Chain Monte Carlo
method was used with the PBPK model to compute updated "posterior" population distributions
of these parameters that provided optimal fits to the individual blood and exhaled
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   3-25       DRAFT-DO NOT CITE OR QUOTE

-------
 1    tetrachloroethylene concentrations of subjects in Monster et al. (1979).  Due to lack of good prior
 2    knowledge about the population variability of physiological parameters, standard reference
 3    values were assumed and standard deviations were selected by "reasonable guess" to generate a
 4    relatively diffuse prior.
 5          The Michaelis-Menten parameter, Vmax, was obtained first for rats and mice by fitting the
 6    model to in vivo data on the rate of total metabolite formation (Bois et al., 1990). The
 7    investigators then estimated human values by using the ratio of human and animal values of
 8    Vmax, as determined by Reitz (1992). The geometric mean of these values was in agreement with
 9    that obtained by extrapolation from the animal values, based on allometric scaling by body
10    weight. This value was used as the "prior" estimate of Vmax. Km was treated as invariant across
11    species, and the geometric mean of values determined for rats and mice was used as the "prior"
12    estimate.  The Monte Carlo simulation was run for 10,000 iterations, after which time the
13    parameter distributions converged. Tetrachloroethylene concentrations in blood and exhaled air
14    were fit extremely well to the Monster et al. (1979) data.  The shape of the prior distribution was
15    seen to have little impact on final results. Model predictions were compared against alveolar
16    concentrations of subjects in the Opdam and Smolders (1986) study, and all data points were
17    seen to fall within the 95th percentile envelope of predictions. The exposure concentrations in
18    this study were 5 to 100 times lower than those used  in the Monster et al. (1979) study; thus, this
19    comparison provides further weight to the strength of the model.
20          The mean value for the posterior estimate of Vmax was 20 times lower than the prior
21    estimate.  Thus, the Bois et al.  (1990) results imply that the maximum rate of tetrachloroethylene
22    metabolism in humans is much lower than that extrapolated by body weight raised to 3/4 power
23    allometric scaling from rodents.
24          Other authors have developed models for tetrachloroethylene that specifically describe
25    the kinetics of its major metabolite, TCA. Gearhart et al. (1993) developed a model for
26    tetrachloroethylene that also included the kinetics of TCA, assuming that TCA comprised 60%
27    of the total tetrachloroethylene metabolized in the rodent and using similar parameters for TCA
28    as in a model for trichloroethylene.  Tetrachloroethylene metabolism parameters for mice were
29    estimated by fitting the model to the time course of decrease in chamber concentration of
30    tetrachloroethylene in gas uptake studies. The model was independently validated at low oral
31    doses (acute oral gavage of tetrachloroethylene in corn oil) by comparing the time course of
32    blood concentrations of tetrachloroethylene and TCA in mice.  Details pertaining to the
33    derivation of parameters for metabolism in humans are not provided in the original paper but are
34    available in the review by Clewell et al. (2005).
35          The parameters for describing tetrachloroethylene metabolism in humans were derived by
36    fitting the model to urinary excretion of TCA in two  subjects in a study by Fernandez et al.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-26       DRAFT-DO NOT CITE OR QUOTE

-------
 1    (1976), assuming the same ratio of TCA to total metabolite as in the rodent. This value was set
 2    to 0.6 and attributed to Dekant et al. (1986b). The validity of using this value for humans has not
 3    been evaluated.  Reitz et al. (1996), in their radiolabeled tetrachloroethylene studies, determined
 4    the fraction of urinary to total metabolites to range from 0.49 to 0.59 in rats and from 0.56 to
 5    0.66 in mice for exposure concentrations that varied by two orders of magnitude.
 6          Clewell et al. (2005) evaluated the Gearhart et al. (1993) model further, comparing its
 7    predictions against the more recently available urinary and blood TCA data gathered by Volkel
 8    et al. (1998) on human subjects exposed to tetrachloroethylene concentrations of 10 to 40 ppm
 9    for 6 hrs.  The predicted blood TCA concentrations were in general agreement with the
10    experimental data, but the rate of urinary excretion of TCA was overpredicted by roughly a
11    factor of 2. Clewell et al. (2005) extended the Gearhart et al. (1993) model to include
12    metabolism of tetrachloroethylene in the kidney, allowing for excretion directly into urine.
13    Assuming metabolism in this organ to be at 10% of the capacity of the liver, substantial
14    improvement was noted in the agreement with experimental data. An advantage in using the
15    Volkel et al. (1998) data is that they pertain to exposure concentrations that are lower than those
16    in other studies (e.g., 72 to 144 ppm in the Monster et al., 1979, study). In addition to
17    developing this refined model, the Clewell et al. (2005) work provides an extensive review and
18    evaluation of available PBPK models for tetrachloroethylene.
19          Loizou (2001) used a PBPK model that was structurally similar to that of Gearhart et al.
20    (1993). The model assumes a 15% stoichiometric yield for the total metabolite produced across
21    various dose levels (i.e., 15% of the parent compound in the liver is metabolized), but the basis
22    for these assumptions is not substantiated. The above yield is also assumed to hold for the
23    production of TCA because it is the major metabolite (E-mail dated June 26, 2002, from G.
24    Loizou, Health and Safety Laboratory, UK, to R. Subramaniam, U.S. EPA). Elimination rates of
25    TCA through blood and urine were chosen by calibrating the model to fit blood and urinary TCA
26    kinetics and exhaled tetrachloroethylene TCA concentration levels obtained from Monster et al.
27    (1979).
28          Other compartments have been added to human PBPK models to answer specific
29    questions.  One model was developed for the purpose of predicting cancer risk in breastfed
30    infants (Byczkowski and Fisher, 1995). This model did not include a brain compartment but
31    instead included a milk compartment for the mother.  Hence, milk concentrations were predicted
32    on a real-time basis, and the daily dose to  a nursing infant could be computed.  To assess  cancer
33    risk, the authors used a standard method based on intake dose of the parent compound (U.S.
34    EPA, 1989).
35          This document has described three human PBPK models—those of Rao and Brown
36    (1993), Reitz et al. (1996), and Bois et al.  (1996)—in order to extrapolate health risk from
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-27       DRAFT-DO NOT CITE OR QUOTE

-------
 1    laboratory animals to human. The rationale for the selection of these models is discussed in
 2    Section 3.5.3.
 O
 4    3.5.2. Variability and Uncertainty
 5          A number of models and studies have been discussed in the preceding sections on animal
 6    and human pharmacokinetics, and other models have appeared in older literature (see Hattis et
 7    al., 1990; Clewell et al., 2005).  These models can be shown to adequately predict human data on
 8    concentrations of the parent tetrachloroethylene compound in blood and exhaled air and have
 9    been used to varying degrees along with cancer risk models to attempt to predict the risk of
10    tetrachloroethylene exposure. There is likely to be considerable biological variability in many of
11    these parameters, and the uncertainties about the values and their interpretations are significant.
12          Variability in pharmacokinetic measurements can exist within an individual over repeated
13    measurements (intra-individual variability) and between different individuals in a population
14    (inter-individual variability). Although this variability can introduce uncertainty into risk
15    assessments that are based on single-point estimates, it is also a factor that can be explored using
16    physiologically based mathematical modeling (including PBPK models) along with statistical
17    techniques to estimate parameter distributions. Some studies have attempted to examine how
18    variability among individuals affects risk (Bois et al., 1990, 1996; Gearhart et al., 1993;
19    Isukapalli et al., 1998).
20          Bois et al. (1996) used Markov Chain Monte Carlo analysis to investigate the  sensitivity
21    of model output to changes in parameters and to determine the parameter distributions needed to
22    explain the intersubject variability in humans.  In simple Monte Carlo analysis, parameter values
23    are selected from predetermined empirical or experimental distributions to investigate variability
24    in model output.  The Markov chain technique takes into account prior knowledge and collected
25    data to modify the parameter distributions. In this case, population distributions were developed
26    using knowledge of the six subjects in the Monster et al. (1979) data set for which a fit was
27    desired.  The new (posterior) parameter distributions were then used to estimate the amount of
28    tetrachloroethylene metabolized during ambient exposures of approximately 1 ppb inhaled dose.
29    The investigators estimated that a median value of 36% was metabolized, with 95% confidence
30    bounds of 15% and 58% at low inhalation exposure concentrations of 0.001 ppm, in contrast to a
31    median value of 1.7% metabolized at a 50 ppm concentration.
32          In an earlier study, Bois et al. (1990) used the conventional Monte Carlo method in
33    conjunction with PBPK modeling to consider the effect of pharmacokinetic parameter
34    uncertainties on the precision in model predictions and, subsequently, on cancer risk estimates.
35    Empirically reasonable probability distributions were assigned to the scaling coefficients
36    associated with each parameter.  It was acknowledged that certain parameters would co-vary (for
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-28        DRAFT-DO NOT CITE OR QUOTE

-------
 1    example, body weight and liver weight), so these parameters would not be altered independently
 2    in the Monte Carlo simulation.  The metabolic parameters were seen to be the most important
 3    determinants in the sensitivity of the results, as also observed by Reitz et al. (1996).  The
 4    investigators calculated a median rate of metabolism in humans of 58 ng/day/kg2/3, with 5th and
 5    95th percentiles of 34 and 104 ng/day/kg2/3, upon continuous exposure to 1 ng/L of
 6    tetrachloroethylene. The variability in the rate of metabolism was estimated to be much lower
 7    than that expected from the variability in Vmax and Km due to the covariance of these two
 8    parameters.
 9           To assess variability in uptake and elimination within a single individual over multiple
10    exposures under different exposure patterns, Chi en (1997) collected exhaled breath
11    measurements on a single individual following four different exposure scenarios in a controlled
12    environmental facility (three replicates per scenario for a total of 12 exposures) and following
13    tetrachloroethylene exposure in 22 dry cleaning facilities, where ambient levels of
14    tetrachloroethylene were recorded  and exposures were carefully timed. Hence, the variability in
15    exhaled breath as a biomarker measurement and surrogate for internal dose could be evaluated in
16    the same individual under clinical conditions in  which exposure magnitude, duration, and pattern
17    were carefully controlled and in a field environment where only the duration of exposure was
18    controlled.  The controlled exposures occurred for either 30 minutes or 90 minutes, with
19    exposure concentrations ranging from 0.5 to 3 ppm.  The experiments were designed to result in
20    potential inhalation exposures of 297 |ig/L-min. Differences in percent uptake and elimination
21    half-life between exposure sessions at the same  environmental concentration were statistically
22    insignificant.  However, percent uptake was dependent on environmental concentration.
23           Gearhart et al. (1993) performed 600 runs of a PBPK model in Monte Carlo fashion to
24    produce a distribution of output results and attempted to look at the effect of the variation in the
25    values of partition coefficients on the prediction of different dose surrogates such as area under
26    the blood time curve for metabolites in the liver. For this dose surrogate, the investigators
27    determined that the coefficient of variation was  25% and that the maximum was less than twice
28    the mean.  They concluded that parameter uncertainty in the models does not constitute a
29    significant source of variability in using PBPK models for risk assessment. It must be noted that
30    variation of the metabolic parameters was not included in their exercise.
31           The quantity of metabolite produced was observed to be very sensitive to Vmax and Km,
32    parameters that vary significantly across models. On the other hand, the concentration of
33    tetrachloroethylene in the blood is  relatively insensitive to these parameters. Tables 3-1 and 3-2
34    indicate the range in the values of these parameters reported in the literature for humans and
35    laboratory animals.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-29        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
            Table 3-2.  Variation in values of metabolic parameters for
            tetrachloroethylene, as seen in the literature
Subject
Reference
'max
(mg/hr)
Km
(mg/L)
Comment
Animal
Sprague-
Dawley rat
B6C3F1
mouse
F344 rat
B6C3F1
mouse
Sprague-
Dawley rat
Chen and Blancato
(1987)
Chen and Blancato
(1987)
Reitzetal. (1996)
Reitzetal. (1996)
Dallas et al.
(1994a,b)
Byczkowski and
Fisher (1994)
0.35
0.18
0.325
0.355
0.009
0.012
2.94
1.47
5.62
3.69
0.019
0.32
Fit model to 10 and 600 ppm inhalation exposure
from Peggetal. (1979).
Least squares fit to total metabolized from a gavage
and inhalation study of Pegg et al. (1979).
Optimization to fit entire data set collected and
reported in this study for 6 hr inhalation exposures.
Metabolic parameters and size/perfusion rate of fat
compartment were fit.
Same as F344 rats in this paper (above).
Metabolic parameters and blood/air partition
coefficient estimated via nonlinear regression to an
intra-arterial injection of 10 mg/kg in the rats.
Fit model to exhaled tetrachloroethylene from
closed chamber inhalation studies.
Humans

10-year-old
child
3 -year-old
child
Chen and Blancato
(1987)
Reitzetal. (1996)
Rao and Brown
(1993)
Boisetal. (1996)
Rao and Brown
(1993)
Rao and Brown
(1993)
Byczkowski and
Fisher (1995)
42.2
32.9
6.77
1.86
4.25
2.64
2.94
32
4.66
4.56
0.1333
4.56
4.56
0.32
Least squares fit to urinary TCA data from Monster
et al. (1979) and Fernandez et al. (1976), assuming
that TCA represents 30% of overall metabolism.
Determined from in vitro studies with human liver
cells and in vitro and in vivo studies in animals.
Published literature.
Central estimate, Markov Chain Monte Carlo
posterior distribution fit to Monster et al. (1979)
exhaled and venous blood concentrations.
Published literature.
Published literature.
Allometric scaling from model optimizations based
on the exhaled breath of rats during closed chamber
inhalation exposure.
4
5
      See text and footnote in Table 3-1.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   3-30       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
       Age and gender-specific differences in pharmacokinetics can have a significant impact on
tissue dosimetry. The immaturity of metabolic enzyme systems in the perinatal period may lead
to decreased clearance of toxic chemicals as well as decreased production of reactive
metabolites.  Clewell et al. (2004) examined these differences for various stages in life using
PBPK modeling for tetrachloroethylene and five other chemicals that differed considerably in
their physicochemical (lipophilicity, solubility, and volatility) and metabolic characteristics.
Parameters describing growth of various tissues were taken from the literature, and blood flow
changes with age were assumed to change proportionally with tissue volumes. For
tetrachloroethylene, only oxidative metabolism—specifically the production of TCA—was
considered. Data on age-dependent development of CYP2E1 was used for this purpose (Vieira
et al., 1996). The parameters for tetrachloroethylene were taken from the Gearhart et al. (1993)
model, and the age-dependence of metabolism was based on the CYP2E1 data. The Gearhart et
al. (1993) model describes the amount of TCA produced as 60% of the total metabolized
tetrachloroethylene; this was fixed in the life-stage model.
       The dose metrics examined were blood concentrations of the parent compound and the
metabolite TCA. Continuous lifetime oral exposure was simulated at a daily dose rate of 1
|ig/kg/day.  Table 3-3 provides the average daily dose during different life stages  of a male
expressed relative to that of a 25-year-old adult male. The gender and age differences in
tetrachloroethylene and TCA blood concentrations are detailed further in Figure 3-3.

       Table 3-3. Ratio of average daily dose at various life stages to the average
       daily dose for a 25-year-old adult: PBPK simulations
Dose metric
Tetrachloroethylene blood
concentration
TCA blood concentration
Life stage
0-6 months
0.33
0.057
0.5-5 years
0.42
0.16
5-25 years
0.76
0.59
25-75 years
1.2
1.4
24
25
26
27
Source: Clewell et al. (2004).
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-31       DRAFT-DO NOT CITE OR QUOTE

-------
                1.5E-4
 2
 3
 4
 5
 6
 1
 8
 9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
       o>
      O
      f£
      UJ
      Q-
      "5
       d
       §
      O
      -Q
       8
      CO
                1.0E-4--
                5.0E-5-
O.OE+0
                                                               - mates
                                                        PERC - females
                                                        TCA - males
                                                             - females
                                                                       5.0E-8
                                                                     --4.0E-8
                                                                          f 3.0E-8  P
                                                           4-2.0E-8  g
                                                                     P
                                                                     "D
                                                                      O
                                                              1.0E-8 £
                      n | I 11 I jI11 I 111 11 I IiI IiIt I iI I I M I I  (
                                                                             O.OE+0
                  0     10    20    30    40    50    60    70    80
                                      Age (years)

       Figure 3-3. PBPK simulations of variations with age and gender in blood
       concentrations of tetrachloroethylene and its main metabolite trichloroacetic
       acid (TCA). Simulations are for continuous lifetime oral exposure at a constant
       daily intake of 1 |ig/kg/day.

       Source:  Clewell et al. (2004).

       Considerable gender differences in blood concentrations of TCA and tetrachloroethylene
were seen in these predictions. Internal dose during infancy differed most from the
corresponding dose in a 25-year-old.  Tetrachloroethylene and TCA blood concentrations
increased with age, which the authors attributed to the lower metabolic and pulmonary clearance
of tetrachloroethylene when compared with other volatiles as well as its higher lipophilicity, both
resulting in storage of the compound in fat and other tissues.  These age and gender differences
in pharmacokinetic sensitivity are significant, but they need to be considered together with
pharmacodynamic considerations in determining the contribution of exposure at a life stage to
lifetime risk.
       The same group of authors (Gentry et al., 2003) developed a PBPK model for
tetrachloroethylene that compared maternal and fetal/neonatal blood and tissue dose metrics
during pregnancy and lactation. The manuscript contains the details on the structure  of the
model.  Oxidative metabolism (TCA) in the mother and lactating infant was modeled using data
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  3-32      DRAFT-DO NOT CITE OR QUOTE

-------
 1    for CYP2E1 (Vieira et al., 1996); metabolism in the fetus was not included due to lack of
 2    information pertaining to the development of this pathway during gestation. The dose metrics
 3    were the fetal  and infant blood concentrations of tetrachloroethylene and TCA. Changes in fetal
 4    blood concentrations were not pronounced because changes in tissue composition occurred in
 5    both mother and fetus during pregnancy (Gentry et al., 2003). A decrease of nearly three orders
 6    of magnitude of blood concentrations in the lactating infant when compared with that of the fetus
 7    was calculated. This decrease was attributed to the lower exposure rate during lactation as
 8    compared with placental exposure. Concentrations in the lactating infant were considerably
 9    lower, by more than two orders of magnitude, than in the mother. The largest variation in blood
10    concentration  occurred in the early postnatal period.
11          As the authors indicated, validation of the results in the Clewell et al. (2004) and Gentry
12    et al. (2003) work and further refinement of the parameters in the models are necessary.  It would
13    therefore be premature to consider the results of such analyses for use in risk assessment.
14    Further investigation of variability in the parameters used in the Clewell et al.  (2004) analysis is
15    needed before the results from Table 3-3 can be used to weigh upon considerations of a
16    pharmacokinetic uncertainty factor for age and gender variability. Nonetheless, these models
17    will enable future studies to focus on the key factors that are likely to influence pharmacokinetic
18    susceptibility.
19
20    3.5.3. Animal-to-Human Extrapolation Using a Physiologically Based Pharmacokinetic
21          (PBPK) Model
22    3.5.3.1.  Choice of Physiologically Based Pharmacokinetic (PBPK) Model
23          As explained above, the evidence suggests that by-products of tetrachloroethylene
24    metabolism are implicated in carcinogenesis in both rodent species. Inhaled concentration of the
25    parent compound is therefore not an appropriate dosimeter. The use of pharmacokinetic
26    modeling is expected to be useful in this regard.  Various dose metrics are explored in detail in
27    Chapter 4.  Because the choice of the most appropriate dose metric has bearing on our selection
28    of PBPK models, the issues are briefly summariuzed here.
29          Both the oxidative and GSH-dependent pathways of tetrachloroethylene metabolism are
30    known to be involved significantly in the various tumors. Tetrachloroethylene hepatotoxicity is
31    associated with cytochrome P450 metabolism occurring in the liver. TCA is considered to be the
32    predominant metabolite associated with this P450 oxidation pathway.  However, TCA may not
33    be the sole contributory metabolite to tetrachloroethylene-induced hepatotoxicity and cancer, and
34    reactive intermediates  such as tetrachloroethylene oxide and trichloroacetyl chloride may also be
35    involved. In the case of renal toxicity, GSH conjugates formed in the liver and transported to the

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-33       DRAFT-DO NOT CITE OR QUOTE

-------
 1    kidney are thought to be the primary agents. The GSH pathway is also implicated in the mode of
 2    action for leukemia (see Chapter 4).
 3          Tetrachloroethylene is a chemical that has generated prolific pharmacokinetic modeling
 4    endeavors. A consideration in determining appropriate PBPK model structures for use in risk
 5    assessment is the ability to use the same model to predict dose metrics for all the endpoints.
 6    Although many models have been developed to predict concentrations of the parent compound
 7    and total metabolite levels, only the model by Gearhart et al. (1993) and its variations, developed
 8    by Clewell et al. (2005) and Loizou (2001), predict both  tetrachloroethylene and TCA
 9    concentrations. These models were reviewed in previous sections. As noted in Chapter 4, there
10    is no reliable quantitative data on GSH conjugates formed by tetrachloroethylene;  accordingly,
11    there are no models that can specifically predict these metabolite levels.
12          Various uncertainties are associated with the use  of PBPK models developed to predict
13    the kinetics of TCA produced as a result of tetrachloroethylene metabolism.  One assumption
14    pertains to the fraction of tetrachloroethylene metabolized to TCA in humans. Loizou (2001)
15    made the assumption that 15% of tetrachloroethylene reaching the liver is metabolized.  In
16    related models, Gearhart et al. (1993) and Clewell et al. (2005) estimated their parameters on this
17    assumption that urinary TCA in humans accounts for 60% of the total metabolism of
18    tetrachloroethylene.  This percentage was assumed to be independent of dose and was based on a
19    range around the value seen in rodents (Reitz et al., 1996; Clewell et al., 2005); however, its
20    reliability for humans is not known.
21          At an exposure concentration of 72 ppm in humans, Monster et al. (1979) determined that
22    95% of inhaled tetrachloroethylene was exhaled unmetabolized, 2% was excreted  as TCA in
23    urine, and 1% of TCA remained in systemic circulation.  Thus, these data suggest  that urinary
24    TCA may comprise roughly 40% of the total metabolite  in humans. It may be noted that other
25    tetrachloroethylene metabolites are also known to be excreted in the urine of exposed humans
26    (Ikeda and Ohtsuji, 1972).
27          The measurement of urine levels of TCA using the photometric Fujiwara reaction
28    method, which was used in the Fernandez et al. (1976) and Monster et al. (1979) studies, is
29    hindered by analytical or methodological problems, such as providing information only on the
30    total trichloro content and blank levels being significantly high in unexposed subjects (Reitz et
31    al., 1996). The TCA measurements in the Fernandez et al. (1976) study were used in estimating
32    the metabolic parameters in the Gearhart et al. (1993) model. Other problems include the half-
33    life of TCA in humans being long—in the neighborhood of 100 hrs (Muller et al.,  1974).
34          TCA alone may not be an adequate dose metric for liver tumors. Buben  and O'Flaherty
35    (1985) compared various indices of liver toxicity for tetrachloroethylene and trichloroethylene,
36    finding tetrachloroethylene to be at least twice as potent  as trichloroethylene  on  a molar basis for
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-34      DRAFT-DO NOT CITE OR QUOTE

-------
 1    equivalent amounts of total metabolite generated. They concluded that the toxic metabolite in
 2    tetrachloroethylene is considerably more toxic than that in trichloroethylene with regard to liver
 3    toxicity. In Appendix 4A, we compare the potency of liver tumors in the TCA and
 4    tetrachloroethylene bioassays and conclude that TCA (produced in the metabolism of
 5    tetrachloroethylene) alone does not appear to be sufficient to account for the tumorigenicity of
 6    tetrachloroethylene for the exposures that were examined.  Clewell et al. (2005) had similar
 7    conclusions and suggest a combination of metabolites as the responsible agents.
 8          All the above factors combined led to the use of the rate of overall metabolism (the total
 9    amount of tetrachloroethylene metabolized per day) as a surrogate for the toxic dose in the route-
10    to-route and animal-to-human extrapolations for liver tumors, leukemia, and kidney tumors.
11    This is more reasonable than using the parent chemical, even though total metabolized dose is
12    not a perfect dose metric in that it does not actually estimate the tissue concentration of toxic
13    metabolites. Inhaled concentration of the parent chemical was used as the basis for
14    extrapolations for other cancers.
15          In this assessment, three of the  most recently developed human PBPK models that predict
16    total metabolite levels were considered: those of Rao and Brown (1993; the Rao and Brown
17    model), Reitz et al. (1996; the Reitz model), and Bois et al. (1996; the Bois model).  These three
18    models were chosen to allow cancer risk estimates to reflect uncertainties that arise from using
19    different data and methods to calibrate human  PBPK models. This enables provision of a range
20    of values for extrapolation from laboratory animals to human. In later sections, these models are
21    compared with each other and with experimental data. The three models were chosen on the
22    basis of their different approaches to estimating metabolic parameters, as summarized in the
23    previous section. Although the models describe the overall metabolism of the parent compound,
24    they do not describe the kinetics of the metabolites.
25          The Reitz model used in vivo rodent data on total metabolism and parent compound
26    concentrations in blood and exhaled breath. The development of a human model used a
27    "parallelogram" approach wherein in vivo metabolic rate constants were related to
28    experimentally determined in vitro values by assuming the relationship of in vivo to in vitro
29    metabolic rates to be invariant across species.  The Bois et al. model, on the other hand, used
30    Bayesian inference methods to fit model predictions to laboratory data on exhaled air and blood
31    concentrations of tetrachloroethylene in human volunteers. The Rao and Brown model used the
32    same human study but assumed the ratio of urinary TCA to total metabolite levels to be equal to
33    0.71 in order to derive metabolic  parameters. The Rao and Brown model was included to permit
34    examination of the range in risk values that would arise if metabolic parameters are derived from
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-35       DRAFT-DO NOT CITE OR QUOTE

-------
 1    urinary TCA data extrapolated to total metabolites with this assumption. l  These and other PBPK
 2    models generally seem to predict parent concentrations well; however, they differ considerably
 3    in their predictions of the amount metabolized. As shown in Tables 3-1 and 3-2, there are large
 4    differences in the metabolic parameters used by various authors.
 5           The human PBPK models chosen suffer from the limitation that their predictions of total
 6    metabolite cannot be accurately evaluated because such data are not available. The models
 7    predict only total metabolite levels, so it is not possible to validate them against specific
 8    metabolites, such as TCA, that have been measured in experiments. These models have been
 9    validated against concentrations of the parent compound in blood and exhaled breath. However,
10    because the metabolism of tetrachloroethylene is slow (particularly in humans, with roughly 95%
1 1    of tetrachloroethylene being exhaled unmetabolized), the concentrations of the parent compound
12    are not sensitive to the values of the metabolic parameters. A similar argument applies to our use
13    of the Bois model. The posterior distributions of parameters in this model were obtained by
14    fitting to the parent compound concentrations  in the Monster et al. (1979) study. Further, as
15    explained above, these three models are limited in different ways.
16
17    3.5.3.2. Implementation of Physiologically Based Pharmacokinetic (PBPK) Models
18           Implementation of the Rao and Brown, Reitz, and Bois models follows the PBPK model
19    structure of Ramsey and Andersen (1984).  The Reitz and Bois models are composed of four
20    compartments: poorly perfused tissues, well-perfused tissues,  fat, and liver. The Rao and Brown
21    model contains, in addition, a separate compartment for the brain. In the implementation of the
22    Rao and Brown model herein, there is no separate skin compartment. The compartments are
23    assumed to be homogeneous, and distribution  is limited by blood flow.  The metabolism of
24    tetrachloroethylene is modeled by a Michaelis-Menten term in the differential equation for the
25    liver compartment.  The simulation is represented by the following equations:
26
                                                ,
                                     at
28                                  — = Q,(Cart -CJ -- ^— Cv/
                                     dt   ^l^  art    vl)  Km+Cvl  vl
29
30
             Another approach is to adjust the urinary TCA predicted by Clewell et al. (2005) by the inverse of this
      factor to derive total metabolite levels.  A preprint of this in-press manuscript was not received in time to be able to
      exercise the Clewell et al. (2005) model for the purposes of this document.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-36       DRAFT-DO NOT CITE OR QUOTE

-------
 1    where:
 21= compartments other than liver
 3           /          = liver
 4           M;        = mass of tetrachloroethylene in the ith compartment
 5           Cvi        = venous concentration of tetrachloroethylene at the exit from compartment I
 6           Cart       = arterial concentration of tetrachloroethylene
 7           Q;        =  blood flow rate into the ith compartment
 8           Vmax, Km  = Michaelis-Menten constants.
 9
10    Pulmonary exchange is represented by:
11
12                                Qalv (Cinh - Calv ) = Qtot (Cart - Cven )
13                                Cart = "ba L-a/v
14                                Cexh = 0.67Ca/v % 0.33C,nh
15
16    where:
17           Qaiv    = alveolar ventilation rate (which is different from the inspiratory flow rate
18                   because of the respiratory dead space). The alveolar ventilation rate and cardiac
19                   output (the ratio that is referred to as the ventilation-to-perfusion ratio) increase
20                   with activity but at different rates
21
22           Qnh    = inhaled concentrations of the chemical
23
24           Caiv    = alveolar concentrations of the chemical
25
26           Qtot    = total blood flow rate (equal to the cardiac output)
27
28           Cart    = arterial concentrations of the chemical
29
30           Cven    = venous concentrations of the chemical
31
32           hba    = blood/air partition coefficient
33
34           Cexh    = exhaled concentrations of the chemical
35
36           For oral exposures, the gastric route was added by assuming "first-pass" metabolism—by
37    assuming that all tetrachloroethylene ingested is transported directly to the liver, the
38    metabolizing organ. A separate PBPK compartment for the stomach was, therefore, not
39    necessary. The absorption of tetrachloroethylene in the stomach was modeled as a first-order

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-37       DRAFT-DO NOT CITE OR QUOTE

-------
 1    process with an absorption rate constant, ka. Then the mass balance equation for the liver may be
 2    modified to have an additional source term as follows:
 3
                                                        C  +*
                                      at   v
                            "'                  Km+Lvl
 5
 6    MO is amount of tetrachloroethylene ingested and is itself a function of time.  In these
 7    simulations, tetrachloroethylene was administered via drinking water as a series of boluses.
 8           These model equations were solved using the Simulink Module of the MATLAB
 9    computational software package (The Mathworks, Natick, MA) and the single-point estimation
10    module in the software package MCSIM (Bois et al., 1996).  It was verified that both packages
1 1    produced the same results when applied to the same set of equations and parameters.
12           A note is in order regarding this implementation of the Bois model. Bois's Bayesian
13    approach produces a (posterior) distribution of parameters and, therefore, a distribution rather
14    than a point estimate of dose. However, this assessment is not  carried out within such a
15    statistical framework.  The central estimate of the parameters in the Bois et al. posterior
16    distribution was used to provide point PBPK estimates of internal dose of tetrachloroethylene
17    and of its overall metabolic rate.  The point estimates obtained  in this manner reproduce
18    (coincide with) the median population estimated by Bois et al. for the amount of
19    tetrachloroethylene metabolized for a large range of exposure concentration,  0.001 to 50 ppm.  It
20    is therefore reasonable to use the central estimates of Bois's posterior distribution of parameters
21    to provide point estimates of dose for extrapolation purposes.
22           Most human PBPK models have been implemented to investigate inhalation exposure
23    and do not incorporate gastric absorption rate constants.  Values in the literature for the gastric
24    absorption rate vary widely. Ward et al. (1988) reported a gastric absorption rate constant in
25    mice of 0.5 L/hr. Dallas et al. (1995) reported oral absorption rate constants in rats and dogs as
26    1.5 and 20.4 L/hr, respectively, obtained by fitting blood concentrations following oral gavage.
27    For modeling purposes, a gastric absorption rate constant of 1 .6 L/hr was chosen. This predicts a
28    reasonably rapid gastric absorption consistent with the data.  It was determined that the resulting
29    blood  concentrations of tetrachloroethylene are not particularly sensitive to larger values of this
30    parameter.  Simulations of gastric absorption of tetrachloroethylene were carried out for humans
3 1    for use in route-to-route extrapolation. Because these simulations were at low exposures, and
32    because of first-pass metabolism effects, the uncertainty in the  gastric absorption rate constant is
33    not likely to significantly affect the results of the extrapolation.  Increasing the gastric absorption
34    rate constant to 20 L/hr results in an approximately twofold increase in peak blood concentration.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-38       DRAFT-DO NOT CITE OR QUOTE

-------
 1    Changing this parameter does not substantially impact the elimination profile. Table 3-4 shows
 2    the parameter sets used in this modeling effort.
 3          For inhalation exposures, ventilation rate is a key parameter. In rodents, ventilation rate
 4    (VE) was calculated as a function of body weight using the following equations (U.S. EPA,
 5    1994):
 6                              For mice:  VE(L/min) = ea326+L05 ln^
 1                              For rats:  VE(L/min) = e°-578+0-821 ln^
 8
 9    where w is body weight in kilograms and In represents the natural log operation. These
10    equations provide total ventilation rate. The alveolar ventilation rate is the total ventilation rate
11    less the volume of air that is inhaled through the physiological dead space (total effective volume
12    not involved in gas exchange) in a given time. For the rats and mice and for resting inspiratory
13    rates (7.5 L/min) in humans, Qaiv . 0.67 VE (Brown et al., 1997). For the exercising individual
14    (24 to 49 L/min), Qaiv increases up to 0.8 VE (Brown et al., 1997). For the ventilation rates
15    covered in this document, it was considered reasonable to use the relationship Qaiv . 0.67 VE
16    throughout. These values represent reasonable physiological values, recognizing that there is
17    substantial variation.  The alveolar ventilation rate corresponding to the resting inhaled minute
18    volume is 5.5 L/min.  However, the EPA typically assumes a total ventilation rate of 13.8 L/min
19    for a 70 kg human.  Thus, unless otherwise stated, the calculations presented  in this assessment
20    assume an alveolar ventilation rate of 9.3 L/min.
21          In order to extrapolate between equivalent metabolized doses in animals and humans, the
22    PBPK structure of the Reitz model was used for rats and mice, and all three PBPK models (Rao
23    and Brown, Reitz, and Bois) were used for humans.  The animal PBPK models were run to
24    simulate the exposure conditions of the animal bioassay studies. During the human equivalent
25    exposures, the model was run to simulate continuous low-level chronic exposure at steady-state
26    conditions. Chapter 5 discusses the details and results of the extrapolation.
27
28    3.5.4.  Comparison of Physiologically Based Pharmacokinetic (PBPK) Simulations With
29          Experimental Data
30          The models were run to  simulate various experimental and clinical exposure scenarios
31    from the literature.  Simulated concentration levels of tetrachloroethylene in the blood and in
32    exhaled air were compared with measured values. As discussed in a previous section, it was not
33    feasible within the constraints of these models to make credible quantitative comparisons with
34    data on urinary or blood levels of major tetrachloroethylene metabolites.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-39       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
        Table 3-4.  Parameters for tetrachloroethylene PBPK modeling
Parameter
BW(kg)
Cardiac output (L/hr)b
Alveolar ventilation
(L/hr)b
Mouse
variable3
0.275 x 60 x
(BW)0-75
0.67 x VEC
Rat
variable3
0.275 x 60 x
(BW)0-75
0.67 x VEd
Human model
Rao and
Brown
70
430
558
Reitz
70
430
558
Bois
70
430
558
Tissue volumes6 (%)
Rapidly perfused
Slowly perfused
Fat
Brain
Liver
4.5
79.5
9.18
1
5.86
4.5
75.7
6.8
1
2.53
1.7
57
23.1
2
3.4
3.71
57
23.1

3.14
15.1
49.4
23.1

2.54
Blood flow (% cardiac output)
Rapidly perfused
Slowly perfused
Fat
Brain
Liver
44
19
1.98
10
25
39.2
19
6.84
10
25
41
19
5
11
24
52
19
5

24
63.7
12.9
4.88

17.9
Partition (tissue/blood)
Rapidly perfused
Slowly perfused
Fat
Brain
Liver
Blood/air
3
2.59
48.3
3
3
16.9
3.73
1.06
86.9
3.73
3.73
18.9
3.72
1.06
86.6
3.72
3.72
10.3
5.88
3.1
119.13

5.88
10.3
1.92
2.9
84.1

3.08
16
Metabolic parameters
Vmax (mg/hr)
Km (mg/L)
0.355
3.69
0.325
5.62
6.77
4.56
32.9
4.66
1.86
0.133f
Gastric absorption rate
Ka (1/hr)
1.6
1.6
1.6
1.6
1.6
 4
 5
 6
 7
 8
 9
10
11
  The simulations in this document use 0.03 kg and 0.3 kg for the mouse and rat, respectively.
                 o 326+1 osinfBW)^ where VE is the minute ventiiation.
b VE(L/hr) = 60 x

C VE(L/hT) = 60 X
                        1 ln(BW).
 d Values used in the animal-to-human extrapolation.
 e A density of 0.92 and 1 g/cc was used for fat and for other compartments, respectively.
 f The "posterior" value for Km in Bois et al. (1996) was multiplied by the liver volume and the liver tissue/blood
  partition coefficient in order to conform to the format of the pharmacokinetic equations in this document.
BW = body weight.
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                       3-40        DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
             Figure 3-4 shows a comparison of the blood concentration levels predicted by the three
      models considered in this assessment with clinical data from the human inhalation study by
      Monster et al. (1979).  In that study, six male volunteers breathed 72 ppm or 144 ppm
      tetrachloroethylene at rest. In a third session, they breathed 142 ppm tetrachloroethylene at rest
      with two intermittent 30-minute exercise excursions. All exposures lasted a total of 4 hrs. The
      researchers measured tetrachloroethylene in blood and exhaled air after exposure until almost no
      tetrachloroethylene remained. They monitored TCA in blood and urine for up to 100 hrs after
      exposure.  Their data set is widely cited and is perhaps the most complete in terms of human
      distribution and in vivo metabolism data. The concentration of tetrachloroethylene in exhaled air
      and blood as measured in the Monster experiments have been used to validate several human
      PBPK models. In the comparison presented here, only tetrachloroethylene exposure at 72 ppm
      was considered, and simulations were carried out at two different ventilation-to-perfusion ratios
      (ratio of alveolar ventilation rate to cardiac output),  corresponding to occupational activity levels.
               10 -,
          O)
          _§
          c
          O
           CD
           O
           O
          O
          T3
           O
           O
          CO.
           O
           CD
                1 -
              0.1 -
             0.01
                                                           Bois model, VPR=1.2
                                                           Rao, Brown model, VPR=1.2
                                                           Reitz model, VPR=1.2
                                                           Bois model, VPR=1.6
                                                           Rao, Brown model, VPR=1.6
                                                           Reitz model, VPR=1.6
                                                           Monster Experiment
                 -20
                            20    40    60    80    100   120
                                    Post Exposure Time (hrs)
                                                              140
                                                                    160   180
            Figure 3-4. Comparison of model predictions for blood concentration with
            inhalation experiment.  Tetrachloroethylene was inhaled at a concentration of 72
            ppm. Simulations were performed at different ventilation-to-perfusion ratios
            (VPR) and at an alveolar ventilation rate of 7 L/min (the geometric mean of
            values in the Monster experiment). Standard deviations on the experimental data
            were very small (e.g., 0.025 mg/L and 0.003 mg/L at 20 and 140 hrs,
            respectively).

            Source: Adapted from Monster et al. (1979).

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    3-41       DRAFT-DO NOT CITE OR QUOTE

-------
 1          For any particular model, the increase in ventilation-to-perfusion ratio from 1.2 to 1.6
 2    does not appear to make much difference, as shown in Figure 3-4. The much closer
 3    correspondence of the Bois model predictions to the Monster data is to be expected because the
 4    model's posterior distribution of parameters was arrived at by fitting to the Monster data. The
 5    Rao and Brown and Reitz model predictions are less than the experimental values, generally
 6    within a factor of 2 and 3, respectively.  These two models do not differ much in their predictions
 7    of tetrachloroethylene blood concentrations.
 8          Stewart et al. (1970) analyzed the expired breath of subjects repeatedly exposed to 100
 9    ppm tetrachloroethylene (7 hrs/day for 5 days) using gas chromatography and infrared
10    spectrometry. Figure 3-5 compares the mean alveolar concentration of tetrachloroethylene in
11    these subjects with the results of model simulations. The subjects were  assumed to be at rest
12    (alveolar ventilation rate of 5.02 L/min and a ventilation-to-perfusion ratio of 1). All three
13    models agree reasonably well with the experimental data; the Bois et al. model differs the least
14    from the experimental result.
          o
          o
         O
         JS
          o
          (D
            10 -
             1 -
                                  O   Bois model
                                  V   Rao & Brown model
                                  D   Reitz model
                                  ^   Stewart Expt
3*7  O
°v  •
 %     v  0*
                                             0
                                                 o
                                                    El
                                                           O
                                                                 i
                                                                       D
                                                                       ¥
                       50
              100
150
200
250
300
350
                                  Post-Exposure Time (hrs)
15       Figure 3-5. Comparison of model predictions for alveolar concentration of
16       tetrachloroethylene with experimental data on humans. Tetrachloroethylene was
17       inhaled at a concentration of 100 ppm, 7 hrs/day for 5 days.  The experimental data
18       Stewart et al. (1970) show the mean alveolar concentration of tetrachloroethylene in these
19       subjects.  Resting breathing conditions (alveolar ventilation rate of 5.02 L/min and a
20       ventilation-to-perfusion ratio of 1.0) were assumed. Some points early in the time course
21       were deleted because of difficulty in obtaining numerical values from the author's plot.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    3-42       DRAFT-DO NOT CITE OR QUOTE

-------
 1          Opdam and Smolders (1986) exposed six human subjects to constant levels of
 2    tetrachloroethylene ranging from 0.5 to 9 ppm and measured the concentration of
 3    tetrachloroethylene in their exhaled breath (from which their alveolar concentrations could be
 4    deduced) during exposure up to 50 to 60 minutes. Separate data were gathered on males and
 5    females. Figures 3-6a and 3-6b compare their results for the ratio of alveolar to inhaled
 6    concentrations of tetrachloroethylene with predictions from the three models.  The experiments
 7    were performed for different breathing scenarios that included normal breathing (with no breath
 8    holding) as well  as paused breathing with different durations of breath holding. However, the
 9    simulations were carried out only for the normal breathing scenario at resting inspiratory rates
10    and with the ventilation-to-perfusion ratio set equal to 1. Body weights and lean body weights
11    were considered differently for males and females (as given in ICRP, 1975). While running the
12    Bois model, other parameters remained unchanged across gender.  On the other hand, blood flow
13    distributions to various compartments differed across gender in the other two models (as given in
14    Brown etal., 1997).
15          Although simulations were carried out for a range of tetrachloroethylene exposure
16    concentrations from 0.5 to 9 ppm (only a range has been provided by the authors), the plots in
17    Figures 3-6a and 3-6b show only the simulations for 5 ppm, as there were no substantial
18    differences across this range of exposure concentrations.  The agreement with this experimental
19    data is particularly noteworthy, considering that the Bois model was parameterized on the basis
20    of the Monster et al. (1979) data, in which exposures were 5-100 times higher than in the Opdam
21    and Smolders (1986) measurements.
22          The comparison shows the Reitz and Bois models' predictions to be generally closer to
23    the experimental results.  The disagreement of the Rao and Brown model appears to  be greatest
24    at the longer time durations:  at 40 hrs, it overpredicts by roughly a factor of 1.5, whereas the
25    Reitz and Bois models are in close agreement. Alveolar concentrations in male subjects are
26    generally slightly less than those in females in both simulations and experiment.
27          Comparisons were also performed for the Altmann et al. (1990) study, in which subjects
28    were exposed to 10 ppm and 50 ppm tetrachloroethylene by inhalation for 4 hrs repeatedly on 4
29    days. Table 3-5  shows the values corresponding to measurements  at the end of exposure on the
30    first day of exposure.  Relative to the Reitz model, the Rao and Brown and Bois models appear
31    to compare well  with the Altmann et al. (1990) experiment at lower exposure, but they predict
32    nearly twice the  experimental levels at higher exposure.  The comparison with the Altmann et al.
33    (1990) data may be less conclusive because the measurement was made immediately after
34    exposure, when variation in both experiment and simulation is large.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-43       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
              0.45 -i
              0.40 -

           o
           o  0.35 -
           O
           T3
           -^  0.30 H
           CD
           .c
           _c
           -Q  °-25-
           o

           fe  °'2CH
           "o
           j>  0.15 -
              0.10 -
              0.05
                                                      	 Bois model
                                                      	 Rao & Brown model
                                                      	Reitz model
                                                       •  Opdam Expt: male a
                                                       V  Opdam Expt: male b
                                                       •  Opdam Expt: male c
                           10
                        20
30        40

Time (hours)
50
60
70
Figure 3-6a. Comparison of model predictions for alveolar concentration as
a fraction of inhaled tetrachloroethylene concentration with experimental
Opdam and Smolders (1986) data on male human subjects. Some
physiological parameters specific for males were used (see text for details).
Experiment exposure concentrations ranged from 0.5 to 9 ppm; plots for
simulations depict only results for 5 ppm (no significant difference at other
exposures in this range). Breathing conditions at rest assumed for the
simulations: alveolar ventilation rate of 5.02 L/min and ventilation-to-perfusion
ratio of 1).  Simulations and experimental data shown here are with no pause in
breathing.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   3-44       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
     o
     c
     o
    O
    _CD
     CD
     O
     O
    O
    _
     o
     CD
    I
             0.45 -


             0.40 -


             0.35 -

             0.30 -


             0.25 -


             0.20 -


             0.15 -


             0.10 -
        0.05
                                                                   Bois model
                                                                   Rao & Brown model
                                                                   Reitz model
                                                                   Opdam Expt: female a
                                                                   Opdam Expt: female b
                                                                   Opdam Expt: female c
.-' V
                    10
                                  20
                 30       40

                 Time (hours)
50
60
70
       Figure 3-6b.  Comparison of model predictions for alveolar concentration as
       a fraction of inhaled tetrachloroethylene concentration with experimental
       data on female human subjects.  Some physiological parameters specific for
       females were used (see text for details). Experiment exposure concentrations
       ranged from 0.5 to 9 ppm; plots for simulations depict only results for 5 ppm (no
       significant difference at other exposures in this range). Resting breathing
       conditions (alveolar ventilation rate of 5.02 L/min and a ventilation-to-perfusion
       ratio of 1.0). Simulations and experimental data shown here are with no pause in
       breathing.

       Source: Opdam and Smolders (1986).
           TTz/s document /'s a draft for review purposes only and does not constitute Agency policy
06/06/08                                   3-45       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
             Table 3-5.  Comparison of venous blood tetrachloroethylene concentrations:
             PBPK simulations and Altmann et al. (1990) studya
Inhaled exposure
concentration
(ppm)
10
50
Blood concentration (g/L)
Altmann
333
1,106
PBPK simulations
Bois et al.
350
1,855
Rao and Brown
385
1,940
Reitz
262
1,332
 4
 5
 6
 7
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
     a Values correspond to measurements at the end of the first day of study in the Altmann et al. (1993) experiment.


           It is concluded that these four comparisons provide no particular basis for preferring one
     model over another. The comparisons in Figures 3-4, 3-5, 3-6a, and 3-6b indicate that, with
     regard to alveolar and blood concentrations, all three models provide reasonably good
     predictions and are not markedly different from each other.
           The three models differ most in their values for the metabolic parameters (see Tables 3-1
     and 3-2) and consequently, as shown later, in their predictions of the rate of total metabolite
     production. The PBPK models presented in this document predict the rate of production of the
     total amount of metabolites but do not describe their transformation and clearance.  It is therefore
     not possible to compare their predictions on metabolite produced with experimental data without
     making major assumptions.  Such a comparison was attempted with the experimental data of
     Fernandez et al. (1976) on the amount of TCA excreted in urine.2 In this experiment, two
     individuals were exposed to 150 ppm  of tetrachloroethylene by inhalation for 8 hrs and followed
     for 72 hrs.  The data indicate that approximately 30 mg of TCA were eliminated through urine in
     these subjects during the post-exposure period.  In order to make a rough comparison, we
     assumed that urinary excreted TCA constituted the bulk  of overall metabolites formed.
     Simulations using the Rao and Brown and Bois models predict approximately 60 mg and 320 mg
     (tetrachloroethylene equivalent, determined by multiplying the amount of the metabolite by the
     ratio of tetrachloroethylene and TCA molecular weights), respectively, of total metabolite
     produced during the post-exposure period.
              Other studies that have reported data on the concentrations of TCA in blood or urine include Stewart et al.
      (1961, 1970), Monster et al. (1979, 1983), Boettner and Muranko (1969), Ikeda et al. (1972), Essing et al. (1973),
      Guberan and Fernandez (1974), Volkel et al. (1998), and the New York State Department of Health (NYS DOH,
      2000).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-46       DRAFT-DO NOT CITE OR QUOTE

-------
 1          In contrast to this large difference, when the two models are applied to the exposure
 2    scenario of the Monster et al. (1979) experiment, they predict approximately the same rate of
 3    metabolite produced immediately following a short 4-hr exposure to 72 ppm tetrachloroethylene.
 4          In addition to the problems in making the comparison, the discrepancies between the
 5    models and with experimental data on TCA may point to large uncertainties in the parameters
 6    used in these models. Because the accuracy of the models has been evaluated only against blood
 7    and breath concentrations of the parent compound, their reliability for predicting the production
 8    of overall metabolites is an unknown. The use of all three of these models to provide a range of
 9    risk estimates is intended to capture some of this uncertainty.
10          Furthermore, there are many difficulties associated with estimates of the extent of
11    metabolism in humans based on TCA excretion reported in the experimental studies of
12    Fernandez et al. and Monster et al. Some of the problems encountered are the accurate
13    measurement of the retained dose  of tetrachloroethylene from inhalation exposure, the
14    imprecision of the older methodologies using the Fujiwara reaction for metabolite quantification,
15    the possibility of an important contribution for metabolites other than TCA (e.g., oxalic acid,
16    CC>2, TCVC, or as yet unrecognized products) that may be excreted, and the relatively long half-
17    life of certain urinary metabolites, which necessitates extended collection of samples.
18          The fraction of TCA in urine relative to that in blood or that stored in body organs is not
19    known.  Furthermore, TCA is only one component of metabolism. Although it is  considered to
20    be the major  metabolite (the conclusions from Monster et al. (1979) indicate that it may comprise
21    60% of the metabolites), some tetrachloroethylene is converted to other compounds, and not all
22    potential metabolites have been adequately evaluated. In addition, TCA, itself, might be further
23    metabolized, reducing the amount of TCA available for urinary excretion.
24          Gearhart et al. (1993) used a model similar to the one used by Rao and Brown to predict
25    the same urinary TCA data set and included a parameter for urinary excretion of TCA. A
26    comparison of the metabolic parameters is shown in Table 3-1. The ratio Vmax/Km is directly
27    proportional to the rate of metabolism at low doses of tetrachloroethylene. The values shown in
28    Table 3-1 indicate that the rate of metabolism predicted by the model of Gearhart  et al. (1993),
29    which directly fit the urinary TCA data, is slightly lower than the one used by Rao and Brown.
30    The Reitz and Bois models use values of Vmax/Km that are greater than those of Gearhart et al. by
31    an order of magnitude or more.
32
33    3.5.5.  Physiologically Based Pharmacokinetic (PBPK) Model Comparisons and
34          Interspecies Differences
35          For an example of tetrachloroethylene tissue concentrations in different species, blood
36    concentrations resulting from a 1 ppm inhalation exposure for a duration of 16 hrs were
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-47       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
 9
10
11
12
13
14
15
16
17
18
19
20
21
      simulated. Figure 3-7 indicates blood concentrations in both rats (using the Reitz model) and
      humans (using the Rao and Brown model). The blood concentrations of the two species differ
      by a factor of about 2.  It will be shown later that the difference in metabolism is significantly
      greater on a body-weight basis. Evidence of this exists in these plots; in particular, the shape of
      the decay curve between the two species is different. The blood concentrations are higher in rats
      up to approximately 20 hrs post exposure; the curves cross at this point in time. In Figure 3-7 the
      decay appears faster in humans than in rats, which is likely a consequence of steady-state
      behavior not having been attained.
                             10
                                  20
30     40      50
   Time (hours)
60
70
BO
           Figure 3-7. Comparison of tetrachloroethylene concentrations in blood in
           rats and humans. Blood concentrations in humans (solid line) and rats (dashed
           line) from a 16-hr inhalation exposure to 1 ppm tetrachloroethylene. The PBPK
           models used were those of Rao and Brown (1993) for humans and Reitz et al.
           (1996) for rats.
           In contrast to Figure 3-7, long-time exposure (12 days at 1 ppm) is simulated in
    Figure 3-8 so as to ensure that steady state has been attained in both species. The blood
    concentration shows at least two modes of decay. In the initial phase, the concentrations in the
    Rao and Brown and Reitz human models decay faster than those in the model for the rat,
    whereas those in the Bois et al. model decay more slowly.  In the second phase, the
               This document is a draft for review purposes only and does not constitute Agency policy
    06/06/08                                   3-48       DRAFT-DO NOT CITE OR QUOTE

-------
                 0.12 -i
                 0.10 -
             £  0.08 H
             CD
              0)
              o
              8
                 0.06 -
              o  0.04 H
              O

             I
                 0.00 -
                -0.02
                      	  Bois (Human)
                      	  Rao, Brown (Human)
                      	  Reitz (Human)
                      	  Reitz (Rat)
                                100
200         300
   Time (hours)
400
500
 2          Figure 3-8. Comparison of various model predictions of tetrachloroethylene
 3          blood concentration in humans and rats following steady state.  Blood
 4          concentrations in humans and rats due to 12-day inhalation exposure to 1 ppm
 5          tetrachloroethylene.  Inspiratory rate is 13.9 L/min, ventilation-to-perfusion ratio
 6          is 1.3.
 7
 8
 9    concentrations in all three models for the human decay more slowly than those for the rat.  The
10    decay as predicted by the Bois model is much slower than the decay for other models.
11          Figure 3-9 shows the daily rate of total amount of tetrachloroethylene metabolized
12    following a 6-hr exposure. It illustrates differences in metabolism among the three species
13    (mouse, rat, and human).  The human model is that of Rao and Brown; the  animal models  are
14    those of Reitz. Steady state was not attained—at least not in the rat and human.
15          The human models differ most in the metabolic parameters Vmax and Km (see Table 3-2).
16    The effect of these differences is reflected graphically in Figure 3-10, which shows the rate of
17    metabolism after steady state has been attained as a function of inhaled exposure concentration in
18    units of milligrams per day per kilogram of body weight. At low exposures, the rate of
19    metabolism is nearly equal to the ratio Vmax/Km.  The values of this ratio in the Rao and Brown
20    and Reitz models are one-tenth and one-half of the value in the Bois model. Km in the Rao and
21    Brown and Reitz models is greater than in the Bois model by a factor of 35. Therefore,
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-49       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
                 10
             03
             5
             "S>
             .*:
             O)
             E
             "o
             .Q
             03
             O
             0)
                 0.1
               0.01
                                                           — Rat(BW= 0.3 kg)
                                                           .--- Human (BW = 70 kg)
                                                           . — Mouse (BW=0.025 kg)
                                   50             100            150
                                      Inhaled Concentration (ppm)
                                                                     200
Figure 3-9. Model predictions of total tetrachloroethylene metabolites
produced following a 6-hr inhalation exposure in rats, mice, and humans.
The PBPK model parameters were from Reitz et al. (1996) for rodents and from
Rao and Brown (1993) for humans.

BW = body weight
               This document is a draft for review purposes only and does not constitute Agency policy
    06/06/08                                  3-50       DRAFT-DO NOT CITE OR QUOTE

-------
            CD
           o:
10e1



















^
^



































X



















x






































^x
s


















^
•^

































0.0001 0.001


/














^



























F
1


^












3ois
?ao, Brown
Reitz

•'
^













^





































—

















^



















_^»-
X



































^



















^*


















^^



















0.01 0.1 1
                                       Inhaled Cone (ppm)
                  10
             en
               I5
               13>
                 0.1
                0.01
                                            Bois
                                            Rao, Brown
                                 	Reitz
                                     10                100
                                       Inhaled Cone (ppm)
                                                                        1000
I
2
3
4
5
6
7
       Figure 3-10. Model predictions of rate of total metabolism in humans at
       steady state. Rate of metabolism in humans, normalized to body weight, with
       continuous exposure after steady-state conditions have been reached, as predicted
       by the Bois et al., Rao and Brown, and Reitz models.  Inspiratory minute volume
       is 13.9 L/min, ventilation-to-perfusion ratio is 1.3, and body weight is 70 kg.
           77?/'s document /'s a draft for review purposes only and does not constitute Agency policy
06/06/08                                   3-51        DRAFT-DO NOT CITE OR QUOTE

-------
 1    saturation occurs at correspondingly higher concentrations in these models (between 1,000 and
 2    2,000 ppm) than in the Bois model (at about 50 ppm). At saturation, the rate of metabolism is
 3    lowest in the Bois  et al. model, as reflected in the relative values for Vmax. Ohtsuki et al. (1983)
 4    monitored all the trichloro metabolite compounds excreted in the urine of 36 male and 25 female
 5    workers exposed to tetrachloroethylene.  Their line of regression of the urinary metabolite level
 6    versus exposure concentration of tetrachloroethylene indicated saturation of metabolism
 7    occurring at roughly 600 ppm of exposure concentration. On the other hand, in an earlier study
 8    of 85 male workers, Ikeda et al. (1972) determined this saturation to occur at between 50 and
 9    100 ppm.
10          In Figures 3-1 la and 3-1 Ib, the simulated rate of metabolism is shown as a function of
11    time for two  extremes in exposures (0.001 ppm and 50 ppm) for the rat and humans. The
12    exposure time was considered to be 12 days (as in Figure 3-8). At a 50 ppm exposure
13    concentration, the rate of tetrachloroethylene metabolism decreases very slowly with time,  post-
14    exposure.
15          Table 3-4 lists all the parameters used in the models. Vmax differs considerably between
16    the three models. Km of the Bois model is the least and is less than the value in the other human
17    models by a factor of 35. There are also  other significant differences between the human
18    models.  The volume of the rapidly perfused compartment with the associated blood flow and the
19    blood air partition  coefficient in the Bois model are considerably different from those in the Rao
20    and Brown and the Reitz models.  The perfusion per unit volume of rapidly perfused tissue is
21    much less in  the Bois model (equal to 4.2) than in the Reitz (equal to 14) and Rao and Brown
22    (equal to 24) models.  The partition coefficient for the slowly perfused compartment in the Rao
23    and Brown model is only about one-third that of the other two human models.  The partition
24    coefficient for fat in the Reitz model is considerably higher.
25          A sensitivity analysis was carried out in order to determine the dominant parameters
26    underlying the differences between the results of the Bois and the Rao and Brown models.  The
27    differences in the blood concentrations and in the amount metabolized were found to be largely
28    accounted for by the metabolic parameters (Vmax and Km), with the blood/air partition coefficient
29    playing a secondary role. Results were found to be insensitive to the other parameters in the
30    model.
31          The results of PBPK simulations of oral exposure to tetrachloroethylene are shown  in
32    Figures 3-12 and 3-13. In these simulations tetrachloroethylene was orally  delivered via
33    drinking water in nine bolus doses spaced 2 hrs apart during an 18-hr time period, followed by
34    6 hrs of no dosing. Because tetrachloroethylene concentrations and the rate of metabolism were
35    found to be negligible at the end of the 24-hr period, the simulation was terminated after 24 hrs.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-52       DRAFT-DO NOT CITE OR QUOTE

-------
             5e-4 -i
ro   4e-4 -
T3
         E   3e-4 -
         *«<•
         E

         "o
             2e-4 -
         OT
         0   1e-4 -
         o>
                                                                      Bois Human
                                                                      Rao, Brown Human
                                                                      Reitz Human
                                                                      Reitz Rat
                                100      200      300     400

                                             Time (hours)
                                                         500
600
700
1
2
3
4
5
6
7
  Figure 3-lla. Rate of metabolism in rat and human models: time course for
  low exposure. Rate normalized by body weight as predicted by the rat model
  (Reitz model) and various human models (Bois, Rao and Brown, and Reitz) for
  low-exposure concentration, 0.001 ppm. Human inspiratory minute volume is
  13.9 L/min, ventilation-to-perfusion ratio is 1.3; for the rat, these parameters are
  as given in Table 3-4.
               This document is a draft for review purposes only and does not constitute Agency policy
    06/06/08                                  3-53       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
1
     ro
          CTi
          ^
     E
     tn
          TO
          4—1
          QJ
          E
          4—
          O
          QJ


          QL
         14  -,
         12 -
         10  -
          4  -
          2  -
          0  -
         -2
                         /'
                                                                Bois Human
                                                                Rao, Brown Human
                                                                  jitz Human
                                                    ---------   Reitz Rat
                    0       100      200      300      400     500      600     700

                                        Time  (hours)
       Figure 3-llb. Rate of metabolism in rat and human models: time course for
       high exposure. Rate normalized by body weight as predicted by the rat model
       (Reitz model) and various human models (Bois et al., Rao and Brown, and Reitz)
       for high-exposure concentration, 50 ppm.  Human inspiratory minute volume is
       13.9 L/min, ventilation-to-perfusion ratio is 1.3; for the rat, these parameters are
       as given in Table  3-2.
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  3-54       DRAFT-DO NOT CITE OR QUOTE

-------
             0.10 -i
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
                                           10           15

                                           Time (hours)
                                                         20
Figure 3-12. Oral ingestion of tetrachloroethylene: blood concentration in
humans versus time. Time course of venous blood concentration in humans as
predicted by the Rao and Brown model for ingested tetrachloroethylene.  A total
of 76 mg tetrachloroethylene was delivered orally via drinking water in 9 bolus
doses spaced 2 hrs apart during an 18-hr time period, followed by 6 hrs of no
dosing.  The Bois and Reitz models result in nearly the same blood concentrations
at this exposure concentration.  The dashed line shows the corresponding steady-
state blood concentration due to inhaled tetrachloroethylene of 0.7 ppm exposure
concentration that results in the same area under the curve as the above curve
integrated over a 24-hr period.  The inspiratory rate is 13.9 L/min and the
ventilation-to-perfusion ratio is 1.3.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   3-55       DRAFT-DO NOT CITE OR QUOTE

-------

^c
~3>
£,
E
"o
.a
CD
-i— »
0
M—
0
CD
-I-J
CD

0.0012 -
0
0

0

0

0

0
0010 -
0008 -

0006 -

0004 -

0002 -

0000 -

£ \
n
i *.
\
1 1
• i
i \
i \
\
V
! V
I v
\
\
\
\
i
? A
1 11
I M
1 '!
1 1
\ ! \
i i \
\ i \
\
\
\
\
\
\
\ \l \ \
\( \ •
ii 't t *
A A A |\ Raog
1 1 1 , l : ? '. Kao c
I |> |1 I 	 Reitz
1 h M M
1 M M 1
1 M M
\ 1 h i \
i ' \ !
\ i \\ \\ \
\i \i \j \
* * • "X
^
'•>^
i
1
''•••^r— •— •_

                                           10
15
20
25
                                           Time (hours)
 1          Figure 3-13. Rate of metabolism of tetrachloroethylene in humans: oral
 2          exposure. Rate of metabolism (mg/min) versus time of tetrachloroethylene
 3          ingested orally as predicted by two PBPK models (Rao and Brown; Reitz).
 4          Tetrachloroethylene was delivered orally via drinking water in 9 bolus doses
 5          spaced 2 hrs apart during an 18-hr time period, followed by 6 hrs of no dosing.
 6          The oral exposures that resulted in production of 0.01  mg/kg/day of overall
 7          metabolite were 21,5.1, and 2.25 mg of total ingested tetrachloroethylene for the
 8          Rao and Brown, Reitz, and Bois models, respectively. The Bois model is not
 9          shown here to avoid a congested draft.
10
11
12    3.5.6.  Metabolic Interactions With Other Chemicals
13          Fisher et al. (2004) used PBPK modeling and complementary studies in mice to
14    investigate the effect of co-exposures of orally administered carbon tetrachloride (CT) and
15    tetrachloroethylene on metabolic interactions between the two chemicals. CT is known to inhibit
16    its own metabolism (referred to as suicide inhibition). TCA was used as a biomarker to assess
17    the inhibition of the cytochrome P450 system by CT. Oral bolus intubation in the dose range of
18    1-100 mg/kg of CT was followed by a dose of 100 mg/kg of tetrachloroethylene an hour later.  It
19    was concluded that dose additivity could not be used to predict interactions between the
20    compounds in this dose range because the metabolic interactions were found to be highly
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-56       DRAFT-DO NOT CITE OR QUOTE

-------
1    nonlinear.  The inhibition in metabolic capacity of tetrachloroethylene 2 hrs after administration
2    of CT and  1 hr after administration of tetrachloroethylene was found to be 5, 52, and 90% at CT
3    doses of 1.5, 10, and 19 mg/kg, respectively.
4          Dobrev et al. (2002) performed gas uptake studies in F344 rats and developed a mixture
5    PBPK model for humans to study interaction effects during co-exposure to mixtures of
6    trichloroethylene (TCE), tetrachloroethylene, and methylchloroform. Corresponding to a 10%
7    increase in TCE blood concentration, the production rates of toxic conjugative metabolites
8    exceeded 17%, pointing to  a nonlinear interaction effect due to co-exposure to TCE.
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   3-57       DRAFT-DO NOT CITE OR QUOTE

-------
 1                                APPENDIX FOR CHAPTER 3:
 2              COMPARISONS OF TETRACHLOROETHYLENE METABOLISM
 3                      WITH TRICHLOROETHYLENE METABOLISM
 4
 5
 6    3.A.I. EXTENT OF METABOLISM
 7          The available data indicate that, overall, tetrachloroethylene is less extensively
 8    metabolized than is the closely related chemical, trichloroethylene. The difference is due to the
 9    fact that a lower fraction of a tetrachloroethylene dose is metabolized via the major oxidative
10    CYP pathway when compared with an equivalent dose of the trichloroethylene congener
11    (Ohtsuki et al., 1983; Volkel et al., 1998; Lash and Parker, 2001). For example, in balance
12    studies of humans, only about 1-3% of the estimated amounts of tetrachloroethylene inhaled
13    were shown to be metabolized to TCA and other chlorinated metabolites, although these studies
14    fail to account for total dose (see Section 3.3.2 for further discussion). These amounts can be
15    compared to the 40-75% of trichloroethylene shown to be metabolized in various human balance
16    studies similar to the ones conducted for tetrachloroethylene (U.S. EPA, 1985).
17          Because of its higher lipid solubility, tetrachloroethylene may appear to be less well
18    metabolized than trichloroethylene, at least to a certain degree, simply because it is more slowly
19    metabolized due to fat sequestration.  However, the animal data from studies of the two
20    compounds provide results similar to those of the human studies regarding the relative extent of
21    metabolism.  For example, using data from laboratory animal studies of tetrachloroethylene
22    (Pegg et al., 1979; Schumann et al., 1980), EPA reported the percent of tetrachloroethylene body
23    burdens excreted as unchanged parent compound following exposure to 10 and 600 ppm for
24    6 hrs to be 68 and 99%, respectively (U.S. EPA, 1985). By comparison, rats and mice exposed
25    to equivalent 10 and 600 ppm trichloroethylene doses (Stott et al., 1982) metabolized a higher
26    percentage of this compound, with mice metabolizing essentially all of the dose and rats
27    metabolizing 98 and 79% of the low and high doses, respectively.
28          Saturation of metabolism occurs at a higher dose for trichloroethylene than for
29    tetrachloroethylene; thus, at certain dose levels, the differences in the amounts of the two
30    compounds metabolized is relatively greater than at other dose levels. Tetrachloroethylene
31    appears to be a lower-affinity substrate for CYP enzymes than trichloroethylene (Ohtsuki et al.,
32    1983; Volkel et al., 1998). The Km value for tetrachloroethylene is certainly higher than the Km
33    value for trichloroethylene (Lipscomb et al., 1998).
34          Both tetrachloroethylene and trichloroethylene are liver toxicants and cause liver
35    hepatocellular carcinomas in mice. The liver toxicity, including carcinogenicity, of these
36    compounds is thought to be due to metabolites. It is interesting to note that although

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-58      DRAFT-DO NOT CITE OR QUOTE

-------
 1    trichloroethylene appears to be more extensively metabolized—due to greater CYP metabolism
 2    in the liver—the relative cancer potency for liver tumors is similar for the two compounds.
 O
 4    3.A.2. DIFFERENCES IN CYTOCHROME P450 (CYP) METABOLITES
 5          TCA, DC A, chloral, and TCOH are reported biotransformation products of both
 6    tetrachloroethylene and trichloroethylene; however, the relative amounts produced and the
 7    precursor intermediates are different for the two compounds. TCA is the major urinary
 8    metabolite for tetrachloroethylene and is also an excretion product of trichloroethylene, whereas
 9    TCOH is the major trichloroethylene urinary excretion product. The formation of chloral in
10    metabolism of tetrachloroethylene has been called into question, and the measurements of TCOH
11    in urine following tetrachloroethylene exposures have also been challenged.
12          Regardless,  TCOH clearly is not the significant metabolite for tetrachloroethylene that it
13    is for trichloroethylene (Lash and Parker, 2001). The fact that the major urinary metabolite for
14    tetrachloroethylene is TCA (with little, if any, TCOH being formed), whereas the major urinary
15    metabolite for trichloroethylene is TCOH in the form of its glucuronide, clearly indicates
16    qualitative and quantitative differences in precursor intermediates. Very little, if any, TCA
17    produced from tetrachloroethylene metabolism comes through chloral, either directly or
18    indirectly through TCOH (Lash and Parker, 2001). The TCA from tetrachloroethylene comes
19    through trichloroacetyl chloride, possibly via the epoxide. On the other hand, the TCA produced
20    from trichloroethylene metabolism is thought to come through chloral, both directly and through
21    TCOH enterohepatic circulation (Lash et al., 2000).
22          DCA is a biotransformation product of both tetrachloroethylene and trichloroethylene,
23    although it is believed that a greater portion of DCA  coming from tetrachloroethylene
24    metabolism does not arise from CYP metabolism, but rather results from further processing of
25    TCVC, whereas the DCA coming from trichloroethylene metabolism results from CYP
26    oxidation.  There are at least three potential routes to DCA in CYP metabolism of
27    trichloroethylene, yet only one likely route—dechlorination of TCA—in the CYP metabolism of
28    tetrachloroethylene. Furthermore, the amount of DCA produced from tetrachloroethylene
29    oxidative metabolism may vary across species.
30
31    3.A.3. CYTOCHROME P450 (CYP) ENZYMES
32          Quantitatively, the liver is by far the predominant site of tetrachloroethylene and
33    trichloroethylene oxidative  metabolism; although most other tissues contain the CYPs that could
34    conceivably metabolize these compounds.  CYP2E1  has been shown to be important in rodent
35    metabolism of trichloroethylene; however, the chemical-specific data are sparse with regard to
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   3-59       DRAFT-DO NOT CITE OR QUOTE

-------
 1    its role in tetrachloroethylene metabolism (Doherty et al., 1996). Still, assuming that CYP2E1 is
 2    important to tetrachloroethylene metabolism is not unreasonable.
 3          CYP3A isoenzymes—and especially CYP2B1/2—may be important for
 4    tetrachloroethylene. Costa and Ivanetich (1980) showed increased hepatic metabolism following
 5    treatment with agents now known to induce these isoenzymes specifically. CYP2B1/2 is
 6    probably the most important CYP isoenzyme involved in oxidative metabolism of
 7    tetrachloroethylene, at least in the rat, although CYPSAs and CYP2E1 are also likely involved.
 8
 9    3.A.4. GLUTATHIONE-DEPENDENT METABOLISM
10          The GSH-dependent pathway for tetrachloroethylene exists in both rodents and humans,
11    and the pathway is also operative for trichloroethylene in these species (Birner et al., 1996;
12    Volkel et al., 1998). The flux through this pathway is thought to be quantitatively less than that
13    through the P450 pathway.  Toxic metabolites can arise from several sources in the pathway;
14    however, for tetrachloroethylene, as well as for trichloroethylene, the GSH pathway is associated
15    with renal toxicity (Anders et al., 1988; Dekant et al., 1989;  U.S. EPA, 1991; IARC, 1995; Lash
16    et al., 2000; Lash and Parker, 2001). For both compounds, recovery of urinary mercapturates,
17    the stable end-products of the GSH pathway, comprise 1% or less of the total dose (Lash and
18    Parker, 2001; Dekant et al 1986a), but this does not reflect the total flux through the GSH
19    pathway. In particular, the TCVC metabolite and the corresponding DCVC and their respective
20    N-acetylated forms derived  from trichloroethylene might also act as substrates for renal beta
21    lyases and  other enzymes: FMO3 and CYP3A (Dekant et al., 1988; reviewed by Anders et al.,
22    1988; Dekant et al., 1989; U.S. EPA, 1991; Lash et al., 2000; Lash and Parker, 2001; see
23    Section 3.2).  It should be noted that a higher cysteine S-conjugate-to-mercapturate ratio exists
24    for tetrachloroethylene when compared to trichloroethylene, which could  influence the relative
25    bioactivation and nephrotoxicity of these two compounds (Lash and Parker, 2001).
26
27    3.A.5. SUMMARY
28          Tetrachloroethylene is closely related structurally to  trichloroethylene and the two
29    chemicals cause similar toxic effects, many of which are attributed to metabolic activation of the
30    parent compounds.  Although tetrachloroethylene is not as extensively metabolized as
31    trichloroethylene, there is little difference in potency between the two chemicals.  TCA, DCA,
32    chloral, and TCOH are reported P450 biotransformation products of both  tetrachloroethylene and
33    trichloroethylene; however, the relative amounts of these metabolites produced, as well as the
34    precursor intermediates in the oxidative pathways, are different for the two compounds.  The fact
35    that the two compounds produce different reactive intermediate P450 metabolites is important to
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  3-60       DRAFT-DO NOT CITE OR QUOTE

-------
 1   consider. Excretion of urinary mercapturates indicates that, relative to P450 oxidation,
 2   tetrachloroethylene is more extensively metabolized via GSH conjugation than is
 3   trichloroethylene.  However, these urinary excretion products do not reflect the total flux through
 4   the GSH pathway since the glutathione and cysteine conjugates of both chemicals have been
 5   shown to undergo further processing to products that are highly reactive.  Thus, regardless of
 6   similarities, both the qualitative and the quantitative differences between tetrachloroethylene and
 7   trichloroethylene in metabolite production could have bearing on toxicity and tumor induction,
 8   and the relative importance of various mechanisms and different modes of action contributing to
 9   their toxic effects,  including tumorgenesis, may vary between the two parent compounds.
10   Recognizing similarities  and differences is important in attempting to understand how each of
11   these two compounds causes its toxic effects.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    3-61       DRAFT-DO NOT CITE OR QUOTE

-------
  1                                    REFERENCES FOR CHAPTER 3
  2
  O

  4     Abbas, R; Fisher, JW. (1997) A physiologically based pharmacokinetic model for trichloroethylene and its
  5     metabolites, chloral hydrate, trichloroacetate, dichloroacetate, trichloroethanol, and trichloroethanol glucuronide in
  6     B6C3F1 mice. Toxicol Appl Pharmacol 147:15-30.
  7
  8     Alberati-Giani, D; Malherbe, P; Kohler, C; et al. (1995) Cloning and characterization of a soluble kynurenine
  9     aminotransferase from rat brain: identity with kidney cysteine conjugate beta-lyase.  J Neurochem 64:1448-1455.
10
11     Altmann, L; Bottger, A; Wiegand, H. (1990) Neurophysiological and psychophysical measurements reveal effects of
12     acute low-level organic solvent exposure in humans.  Int Arch Occup Environ Health 62:493-499.
13
14     Amet, Y; Berthou, F; Fournier, G; et al. (1997) Cytochrome P450 4A and 2E1 expression in human kidney
15     microsomes [published erratum appears in Biochem Pharmacol 1997 Jun 15; 53(12): 1946]. Biochem Pharmacol
16     53:765-771.
17
18     Anders, MW; Lash, L; Dekant, W; et al. (1988) Biosynthesis and biotransformation of glutathione S-conjugates to
19     toxic metabolites. CritRev Toxicol 18:311-341.
20
21     Anna, CH; Maronpot, RR; Pereira, MA; et al. (1994) Ras proto-oncogene activation in dichloroacetic acid-,
22     trichloroethylene- and tetrachloroethylene-induced liver tumors in B6C3F1 mice. Carcinogenesis 15:2255-2261.
23
24     ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
25     (update). Prepared by Sciences, International, under subcontract to Research Triangle Institute, Research Triangle
26     Park, NC.
27
28     Bartels, MJ. (1994) Quantitation of the tetrachloroethylene metabolite N-acetyl-S- (trichlorovinyl)cysteine in rat
29     urine via negative ion chemical ionization gas chromatography/tandem mass spectrometry. Biol Mass Spectrom
30     23:689-694.
31
32     Birner, G; Richling, C; Henschler, D; et al. (1994) Metabolism of tetrachloroethene  in rats: identification of N
33     epsilon-  (dichloroacetyl)-L-lysine and N epsilon-(trichloroacetyl)-L-lysine as protein adducts. Chem Res Toxicol
34     7:724-732.
35
36     Birner, G; Rutkowska, A; Dekant, W.  (1996) N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and 2,2,2-
37     trichloroethanol: two novel metabolites of tetrachloroethene in humans after occupational exposure. Drug Metab
38     Dispos 24:41-48.
39
40     Board, PG; Baker, RT; Chelvanayagam, G; et al. (1997) Zeta, a novel class of glutathione transferases in a range of
41     species from plants to humans. Biochem J 328(Pt 3):929-935.
42
43     Boettner, EA; Muranko, HJ. (1969) Animal breath data for estimating the exposure of humans to chlorinated
44     hydrocarbons. Am Ind Hyg Assoc 130:437-442.
45
46     Bogen, KT; Colston, BW, Jr.; Machicao, LK. (1992) Dermal absorption of dilute aqueous chloroform,
47     trichloroethylene, and tetrachloroethylene in hairless guinea pigs. Fundam Appl Toxicol 18:30-39.
48
49     Bois, FY; Zeise, L; Tozer, TN. (1990) Precision and sensitivity of pharmacokinetic models for cancer risk
50     assessment: tetrachloroethylene in mice, rats, and humans.  Toxicol Appl Pharmacol 102:300-315.
51
52     Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol 70:347-
53     355.
54

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         3-62        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Bolanowska, W; Golacka, J. (1972) Absorption and elimination of tetrachloroethylene in humans under
  2     experimental conditions (English translation). Medycyna Pracy 23:10-119.
  3
  4     Bonse, G; Henschler, D. (1976) Chemical reactivity, biotransformation, and toxicity of polychlorinated aliphatic
  5     compounds.  CRC Crit Rev Toxicol 4:395-409.
  6
  7     Bonse, G; Urban, T; Reichert, D; et al. (1975) Chemical reactivity, metabolic oxirane formation and biological
  8     reactivity of chlorinated ethylenes in the isolated perfused rat liver preparation. Biochem Pharmacol 24:1829-1834.
  9
10     Brown, RP.; Delp, MD; Lindstedt, SL; etal. (1997) Physiological parameter values for physiologically based
11     pharmacokinetic models. Toxicol Industrial Health 13:407-484.
12
13     Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
14     and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
15
16     Bull, RJ. (2000) Mode of action of liver tumor induction by trichloroethylene and its metabolites, trichloroacetate
17     and dichloroacetate. Environ Health Perspect 108(suppl 2):241-259.
18
19     Byczkowski, JZ; Fisher, JW. (1994) Lactational transfer of tetrachloroethylene in rats. Risk Anal 14:339-349.
20
21     Byczkowski, JZ; Fisher, JW. (1995) A computer program linking physiologically based pharmacokinetic model
22     with cancer risk assessment for breast-fed infants.  Comput Methods Programs Biomed 46:155-163.
23
24     Calabrese, E. (1983) Principles of animal extrapolation. New York: John Wiley and Sons.
25
26     Campbell, JA; Corrigall, AV; Guy, A; et al. (1991) Immunohistologic localization of alpha, mu, and pi class
27     glutathione S-transferases in human tissues. Cancer 67:1608-1613.
28
29     Chen, C; Blancato, J. (1987) Role of pharmacokinetic modeling in risk assessment: perchloroethylene as an
30     example. In:  Safe Drinking Water Committee, SoPNRC, eds. Pharmacokinetics in risk assessment. Washington,
31     DC: National Academy Press; pp. 367-388.
32
33     Chien, Y-C. (1997) The influences of exposure pattern and duration on elimination kinetics and exposure
34     assessment of tetrachloroethylene in humans.  Ph.D. Dissertation, Rutgers University, New Brunswick, NJ.
35
36     Chiu, WA; Micallef, S.; Monster, A.C.; and F.Y. Bois. (2007). Toxicokinetics of inhaled trichloroethylene and
37     tetrachloroethylene in humans at 1 ppm: empirical results and comparisons with previous studies. Toxicol Sci  95
38     (l):23-36.
39
40     Clewell, HS, III; Gentry, PR; Covington, TR; et al. (2000) Development of a physiologically based pharmacokinetic
41     model of trichloroethylene and its metabolites for use in risk assessment.  Environ Health Perspect 108 (Suppl)
42     2:283-305.
43
44     Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
45     pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79(2):381-93.
46
47     Clewell, HJ; Gentry, PR; Kester, JE; et al. (2005) Evaluation of physiologically based pharmacokinetic models in
48     risk assessment: an example with perchloroethylene. Crit Rev Toxicol (accepted for publication).
49
50     Cooper, AJ. (1994) Enzymology of cysteine S-conjugate beta-lyases. Adv Pharmacol 27:71-113.
51
52     Costa, AK; Ivanetich, KM. (1980) Tetrachloroethylene metabolism by the hepatic microsomal cytochrome P-  450
53     system. Biochem Pharmacol 29:2863-2869.
54


                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        3-63        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Cummings, BS; Zangar, RC; Novak, RF; et al. (1999) Cellular distribution of cytochromes P-450 in the rat kidney.
  2     Drug Metab Dispos 27:542-548.
  3
  4     Cummings, BS; Lasker, JM; Lash, LH. (2000a) Expression of glutathione-dependent enzymes and cytochrome
  5     P450s in freshly isolated and primary cultures of proximal tubular cells from human kidney.  J Pharmacol Exp Ther
  6     293:677-685.
  7
  8     Cummings, BS; Parker, JC; Lash, LH. (2000b) Role of cytochrome P450 and glutathione S-transferase alpha in the
  9     metabolism and cytotoxicity of trichloroethylene in rat kidney. Biochem Pharmacol 59:531-543.
10
11     Dallas, CE; Chen, XM; Muralidhara, S; et al. (1994a) Use of tissue disposition data from rats and dogs to determine
12     species differences in input parameters for a physiological model for perchloroethylene. Environ Res 67:54-67.
13
14     Dallas, CE; Chen, XM; O'Barr, K; et al. (1994b) Development of a physiologically based pharmacokinetic model
15     for perchloroethylene using tissue concentration-time data. Toxicol Appl Pharmacol 128:50-59.
16
17     Dallas, CE; Chen, XM; Muralidhara, S; et al. (1995) Physiologically based pharmacokinetic  model useful in
18     prediction of the influence of species, dose,  and exposure route on perchloroethylene pharmacokinetics. J Toxicol
19     Environ Health 44:301-317.
20
21     Daniel, J. (1963) The metabolism of 36-Cl-labeled trichloroethylene and tetrachloroethylene in the rat. Biochem
22     Pharmacol 12:795-802.
23
24     Davidson, IW; Beliles, RP. (1991) Consideration of the target organ toxicity of trichloroethylene in terms of
25     metabolite toxicity and pharmacokinetics. Drug Metab Rev 23:493-599.
26
27     Dekant, W; Vamvakas, S; Berthold, K; et al. (1986a) Bacterial beta-lyase mediated cleavage and mutagenicity of
28     cysteine conjugates derived from the nephrocarcinogenic alkenes trichloroethylene, tetrachloroethylene and
29     hexachlorobutadiene. Chem Biol Interact 60:31^4-5.
30
31     Dekant, W; Metzler, M; Henschler, D. (1986b) Identification of S-l,2-dichlorovinyl-N-acetyl-cysteine as a urinary
32     metabolite of trichloroethylene: a possible explanation for its nephrocarcinogenicity in male rats. Biochem
33     Pharmacol 35:2455-2458.
34
35     Dekant, W; Martens, G; Vamvakas, S; et al. (1987) Bioactivation of tetrachloroethylene. Role of glutathione S-
3 6     transferase-catalyzed conjugation versus cytochrome P-450-dependent phospholipid alkylation.  Drug Metab Dispos
37     15:702-709.
38
39     Dekant, W; Berthold, K; Vamvakas, S; et al. (1988) Thioacylating intermediates as metabolites of S-(l,2-
40     dichlorovinyl)-L- cysteine and S-(l,2,2-trichlorovinyl)-L-cysteine formed by cysteine conjugate beta-lyase.  Chem
41     Res Toxicol 1:175-178.
42
43     Dekant, W; Vamvakas, S; Anders, MW. (1989) Bioactivation of nephrotoxic haloalkenes by glutathione
44     conjugation: formation of toxic and mutagenic intermediates by  cysteine  conjugate beta-lyase.  Drug Metab Rev
45     20:43-83.
46
47     Dekant, W; Birner, G; Werner, M; et al. (1998) Glutathione conjugation of perchloroethene in subcellular fractions
48     from rodent and human liver and kidney. ChemBiol Interact  116:31-43.
49
50     Dmitrieva, NV (1967) [On the metabolism  of tetrachloroethylene]. Gig Tr Prof Zabol 11:54-56.
51
52     Dobrev, ID; Andersen, ME; Yang, RSH. (2002) In silico toxicology: simulating interaction thresholds for human
53     exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-trichloroethane. Environ Health Perspect
54     110(10):1031-1039.
55

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         3-64        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Doherty, AT; Ellard, S; Parry, EM; et al. (1996) An investigation into the activation and deactivation of chlorinated
  2     hydrocarbons to genotoxins in metabolically competent human cells. Mutagenesis 11:247-274.
  3
  4     Dohn, DR; Anders, MW. (1982) Assay of cysteine conjugate beta-lyase activity with S-(2-benzothiazolyl)cysteine
  5     as the substrate. Anal Biochem 120:379-386.
  6
  7     Droz, PO;  Guillemin, MP. (1986) Occupational exposure monitoring using breath analysis.  JOccupMed 28:593-
  8     602.
  9
10     Duffel, MW; Jakoby, WB. (1982) Cysteine S-conjugate N-acetyltransferase from rat kidney microsomes. Mol
11     Pharmacol 21:444-448.
12
13     Eger, EL (1963) A mathematical model of uptake and distribution. In: Popper, EM; Ketz, RJ; eds. Uptake and
14     distribution of anesthetic agents. New York: McGraw-Hill; pp. 72-103.
15     Elfarra, A.A.; Krause R.J. (2007) S-(l,2,2-trichlorovinyl)-L-cysteine sulfoxide, a reactive metabolite of S-(l,2,2-
16     Trichlorovinyl)-L-cysteine formed in rat liver and kidney microsomes, is a potent nephrotoxicant. J Pharmacol Exp
17     Ther 321(3): 1095-101.
18
19     Essing, HG; Schacke, G; Schaller, KH; et al.  (1973) [Occupational-medicine studies on the dynamics of
20     perchloroethylene in the organism]. Med Welt 24:242-244.
21
22     Fernandez, J; Guberan, E; Caperos, J. (1976) Experimental human exposures to tetrachloroethylene vapor and
23     elimination in breath after inhalation. Am Ind Hyg Assoc 137:143-150.
24
25     Filser, JG; Bolt, HM. (1979) Pharmacokinetics of halogenated ethylenes in rats. Arch Toxicol 42:123-136.
26
27     Fisher, J; Lumpkin, M; Boyd, J; et al. (2004) PBPK modeling of the metabolic interactions of carbon tetrachloride
28     and tetrachloroethylene in B6C3F1 mice. Environ Toxicol Phar 16:93-105.
29
30
31     Frankel, DM; Johnson, CE; Pitt, HM. (1957) Preparation and properties of tetrachloroethylene oxide.  J Org Chem
32     22:1119-1124.
33
34     Frantz, SW; Watanabe, PG. (1983) Tetrachloroethylene: balance and tissue distribution in male Sprague-Dawley
3 5     rats by drinking-water administration. Toxicol Appl Pharmacol 69:66-72.
36
37     Gamier, R; Bedouin, J; Pepin, G; et al. (1996) Coin-operated dry cleaning machines may be responsible for acute
38     tetrachloroethylene poisoning: report of 26 cases including one death. J Toxicol Clin Toxicol 34:191-197.
39
40     Gearhart, JM; Mahle, DA;  Greene, RJ; et al. (1993) Variability of physiologically based pharmacokinetic (PBPK)
41     model parameters and their effects on PBPK model predictions in a risk assessment for perchloroethylene (PCE).
42     Toxicol Lett 68:131-144.
43
44     Gentry, PR; Covington, TR; Clewell, HJ, III. (2003) Evaluation of the potential impact of pharmacokinetic
45     differences on tissue dosimetry in offspring during pregnancy and lactation. Regul Toxicol Pharmacol 38(1): 1-16.
46
47     Ghantous, H; Danielsson, BR; Dencker, L; et al. (1986) Trichloroacetic acid accumulates in murine amniotic fluid
48     after tri- and tetrachloroethylene inhalation.  Acta Pharmacol Toxicol (Copenh) 58:105-114.
49
50     Goldstein, A; Aronow, L; Kalman, SM. (1969) Principles of drug action. New York: Harper and Row.
51
52     Green, T; Odum, J; Nash, JA; et al. (1990) Perchloroethylene-induced rat kidney tumors: an investigation of the
53     mechanisms involved and their relevance to humans. Toxicol Appl Pharmacol 103:77-89.
54


                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         3-65       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Green, T; Dow, J; Ellis, MK; et al. (1997) The role of glutathione conjugation in the development of kidney tumours
  2     in rats exposed to trichloroethylene.  ChemBiol Interact 105:99-117.
  3
  4     Guberan, E; Fernandez, J. (1974) Control of industrial exposure to tetrachloroethylene by measuring alveolar
  5     concentrations: theoretical approach using a mathematical model. Br J Ind Med 31:159-167.
  6
  7     Guengerich, FP; Macdonald, TL. (1984) Chemical mechanisms of catalysis by cytochrome P450: a unified view.
  8     Ace ChemRes 17:9-16.
  9
10     Hake, CL; Stewart, RD. (1977) Human exposure to tetrachloroethylene: inhalation and skin contact. Environ Health
11     Perspect 21:231-238.
12
13     Hattis, D; White, P; Marmorstein, L; et al. (1990) Uncertainties in pharmacokinetic modeling for perchloroethylene.
14     I. Comparison of model structure, parameters, and predictions for low-dose metabolism rates for models derived by
15     different authors.  Risk Anal 10:449-458.
16
17     Henschler, D; Bonse, G. (1977) Metabolic activation of chlorinated ethylenes: dependence of mutagenic effect on
18     electrophilic reactivity of the metabolically formed epoxides. Arch Toxicol 39:7-12.
19
20     Hinchman, CA; Ballatori, N. (1990) Glutathione-degrading capacities of liver and kidney in different species.
21     BiochemPharmacol40:1131-1135.
22
23     Hu, Y; Hakkola, J; Oscarson, M; et al. (1999) Structural and functional characterization of the 5'-flanking region of
24     the rat and human cytochrome P450 2E1 genes: identification of a polymorphic repeat in the human gene. Biochem
25     Biophys Res Commun 263:286-293.
26
27     IARC (International Agency for Research on Cancer). (1995) Tetrachloroethylene. In: IARC Monographs on the
28     Evaluation of Carcinogenic Risks to Humans, Vol. 63: dry cleaning, some chlorinated solvents and other industrial
29     chemicals.  Lyon, France; pp. 159-221.
30
31     ICRP (International Commission on Radiological Protection). (1975) Report of the Task Group on Reference Man.
32     Publication 23. Annals of the ICRP.
33
34     Ikeda, M. (1977) Metabolism of trichloroethylene and tetrachloroethylene in human subjects.  Environ Health
35     Perspect 21:239-245.
36
37     Ikeda, M; Imanura, T. (1973) Biological half-life of trichloroethylene and tetrachloroethylene in human subjects. Int
3 8     Arch Arbeitsmed  31:209-224.
39
40     Ikeda, M; Otsuji, H. (1972) A comparative study of the excretion of Fujiwara reaction-positive substances in urine
41     of humans and rodents given trichloro- or tetrachloro-derivatives of ethane and ethylene. Br J Ind Med 29:99-104.
42
43     Ikeda, M; Otsuji, H; Imamura, T; et al. (1972) Urinary excretion of total trichloro-compounds, trichloroethanol, and
44     trichloroacetic acid as a measure of exposure to trichloroethylene and tetrachloroethylene. Br J Ind Med 29:328-
45     333.
46
47     Isukapalli, SS; Roy, A; Georgopoulos, PG. (1998) Stochastic response surface methods (SRSMs) for uncertainty
48     propagation: application to environmental and biological systems. Risk Anal 18:351-363.
49
50     Jakobson, I; Wahlberg, JE; Holmberg, B; et al. (1982) Uptake via the blood and elimination of 10 organic solvents
51     following epicutaneous exposure of anesthetized guinea pigs. Toxicol Appl Pharmacol 63:181-187.
52
53     Kaemmerer, K; Fink, J; Keitzmann, M. (1982) Studies on the pharmacodynamics of perchloroethylene. Der
54     Prakitsch Tierarzt 2:171-182.
55

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         3-66        DRAFT-DO NOT  CITE OR QUOTE

-------
  1    Kline, SA; Solomon, JJ; VanDuuren, BL. (1978) Synthesis and reactions of chloroalkene epoxides.  J Org Chem
  2    42:3596-3600.
  3
  4    Koppel, C; Arndt, I; Arendt, U; et al. (1985) Acute tetrachloroethylene poisoning—blood elimination kinetics
  5    during hyperventilation therapy.  JToxicol ClinToxicol23:103-115.
  6
  7    Kundig, S; Hogger, D. (1970) The importance of the determination of tri- and perchloroethylene metabolites in the
  8    urine. Int Arch Arbeitsmed 26:3 06-315.
  9
10    Jakobson, I; Wahlberg, JE; Holmberg, B; et al. (1982) Uptake via the blood and elimination of 10 organic solvents
11    following epicutaneous exposure of anesthetized guinea pigs. Toxicol Appl Pharmacol 63:181-187.
12
13    Larsen, GL. (1985) Distribution of cysteine conjugate beta-lyase in gastrointestinal bacteria and in the environment.
14    Xenobiotica 15:199-209.
15
16    Larsen, GL; Stevens, JL. (1986) Cysteine conjugate beta-lyase in the gastrointestinal bacterium Eubacterium
17    limosum.  Mol Pharmacol 29:97-103.
18
19    Larson, JL; Bull, RJ. (1989) Effect of ethanol on the metabolism of trichloroethylene. J Toxicol Environ Health
20    28:395-406.
21
22    Lash, LH; Parker, JC. (2001) Hepatic and renal toxicities associated with perchloroethylene. Pharmacol Rev
23    53:177-208.
24
25    Lash, LH; Nelson, RM; Van Dyke, RA; et al.  (1990) Purification and characterization of human kidney cytosolic
26    cysteine conjugate beta-lyase activity.  Drug Metab Dispos 18:50-54.
27
28    Lash, LH; Qian, W; Putt, DA; et al. (1998) Glutathione conjugation of perchloroethylene in rats and mice in vitro:
29    sex-, species-, and tissue-dependent differences.  Toxicol Appl Pharmacol 150:49-57.
30
31    Lash, LH; Lipscomb, JC; Putt, DA; et al. (1999) Glutathione conjugation of trichloroethylene in human liver and
32    kidney: kinetics and individual variation. Drug Metab Dispos 27:351-359.
33
34    Lash, LH; Fisher, JW; Lipscomb, JC; et al. (2000) Metabolism of trichloroethylene.  Environ Health Perspect 108
35    Suppl 2:177-200.
36
37    Lash, LH; Qian, W; Putt, DA; et al. (2001) Renal and hepatic toxicity of trichloroethylene and its glutathione-
3 8    derived metabolites in rats and mice: sex-, species-, and tissue- dependent differences. J Pharmacol Exp Ther
39    297:155-164.
40
41    Leibman, KC; Ortiz, E. (1970) Epoxide intermediates in microsomal oxidation of olefins to glycols. J Pharmacol
42    Exp Ther 713:242-246.
43
44    Leibman, KC; Oritz, E. (1977) Metabolism of halogenated ethylenes. Environ Health Perspect 21:91-97.
45
46    Levine, B; Fierro, MF; Goza, SW; et al. (1981) A tetrachloroethylene fatality. J Forensic Sci 26:206-209.
47
48    Lieber, CS. (1997) Cytochrome P-4502E1: its physiological and pathological role. Physiol Rev 77:517-544.
49
50    Lipscomb, JC; Mahle, DA; Brashear, WT; et al. (1995) Dichloroacetic acid: metabolism in cytosol.  Drug Metab
51    Dispos 23:1202-1205.
52
53    Lipscomb, JC; Mahle, DA; Brashear, WT; et al. (1996) A species comparison of chloral hydrate metabolism in
54    blood and liver. Biochem Biophys Res Commun 227:340-350.
55

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        3-67        DRAFT-DO NOT CITE OR QUOTE

-------
  1    Lipscomb, JC; Fisher, JW; Confer, PD; et al. (1998) In vitro to in vivo extrapolation for trichloroethylene
  2    metabolism in humans.  Toxicol Appl Pharmacol 152:376-387.
  3
  4    Loizou, GD. (2001) The application of physiologically based pharmacokinetic modelling in the analysis of
  5    occupational exposure to perchloroethylene. Toxicol Lett 124:59-69.
  6
  7    Lukaszewski, T. (1979) Acute tetrachloroethylene fatality.  Clin Toxicol 15:411-415.
  8
  9    Malherbe, P; Alberati-Giani, D; Kohler, C;  et al. (1995) Identification of a mitochondria! form of kynurenine
10    aminotransferase/glutamine transaminase K from rat brain. FEES Lett 367:141-144.
11
12    Mannervik, B. (1985) The isoenzymes of glutathione transferase.  Adv Enzymol Relat Areas Mol Biol 57:357-417.
13
14    Maronpot, RR; Fox, T; Malarkey, DE; et al. (1995) Mutations in the ras proto-oncogene: clues to etiology and
15    molecular pathogenesis of mouse liver tumors. Toxicology 101:125-156.
16
17    May, R. (1976) Metabolisierung und Ausscheidung von Tetrachlorathylen und Beeinflussung durch Gleichzeitige
18    Aufnahme von Athanol beim Menschen. Thesis, University of Wurzburg, Germany.
19
20    McCarver, DG; Byun, R; Hines, RN; et al.  (1998) A genetic polymorphism in the regulatory sequences of human
21    CYP2E1:  association with increased chlorzoxazone hydroxylation in the presence of obesity and ethanol intake.
22    Toxicol Appl Pharmacol 152:276-281.
23
24    McDougal, JN; Jepson, GW; Clewell, HJ, III; et al. (1990) Dermal absorption of organic chemical vapors in rats and
25    humans.  Fundam Appl Toxicol 14:299-308.
26
27    Miller, RE; Guengerich, FP. (1982) Oxidation of trichloroethylene by liver microsomal cytochrome P-450: evidence
28    for chlorine migration in a transition state not involving trichloroethylene oxide. Biochemistry 21:1090-1097.
29
30    Miller, RE; Guengerich, FP. (1983) Metabolism of trichloroethylene in isolated hepatocytes, microsomes, and
31    reconstituted enzyme systems containing cytochrome P-450. Cancer Res 43:1145-1152.
32
33    Mitchell, AE; Morin, D; Lakritz, J; et al. (1997) Quantitative profiling of tissue- and gender-related expression of
34    glutathione S-transferase isoenzymes in the mouse. Biochem J 325(Pt 1):207-216.
35
36    Moghaddam,  AP; Abbas, R; Fisher, JW; etal. (1996) Formation of dichloroacetic acid by rat and mouse gut
37    microflora, an in vitro study. Biochem Biophys Res Commun 228:639-645.
38
39    Moghaddam,  AP; Abbas, R; Fisher, JW; et  al. (1997) The role of mouse intestinal microflora in the metabolism of
40    trichloroethylene, an in vivo study. Hum Exp Toxicol  16:629-635.
41
42    Monster, AC; Houtkooper, JM.  (1979) Estimation of individual uptake of trichloroethylene, 1,1,1- trichloroethane
43    and tetrachloroethylene from biological parameters.  Int Arch Occup Environ Health 42:319-323.
44
45    Monster, AC; Smolders, JF.  (1984) Tetrachloroethene in exhaled air of persons living near pollution sources. Int
46    Arch Occup Environ Health 53:331-336.
47
48    Monster, AC; Boersma, G; Steenweg, H. (1979) Kinetics of tetrachloroethylene in volunteers: influence of exposure
49    concentration and work load.  Int Arch Occup Environ Health 42:303-309.
50
51    Monster, AC; Regouin-Peeters, W; van Schijndel, A; et al. (1983) Biological monitoring of occupational exposure
52    to tetrachloroethene.  Scand J Work Environ Health 9:273-281.
53
54    Moslen, MT; Reynolds, ES; Szabo, S.  (1977) Enhancement of the metabolism and hepatotoxicity of
55    trichloroethylene and perchloroethylene. Biochem Pharmacol 26:369-375.

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        3-68        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Muller, G; Spassovski, M; Henschler, D. (1974) Metabolism of trichloroethylene in man: II. pharmacokinetics of
  2     metabolites. Arch Toxicol 32:283-95.
  3
  4     Nakai, JS; Stathopulos, PB; Campbell, GL; et al. (1999) Penetration of chloroform, trichloroethylene, and
  5     tetrachloroethylene through human skin. J Toxicol Environ Health A 58:157-170.
  6
  7     Ni, YC; Wong, TY; Lloyd, RV; et al. (1996) Mouse liver microsomal metabolism of chloral hydrate, trichloroacetic
  8     acid, and trichloroethanol leading to induction of lipid peroxidation via a free radical mechanism. Drug Metab
  9     Dispos 24:81-90.
10
11     NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
12     (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3F1 mice. NTP Technical Report 311.
13
14     NYS DOH (New York State Department of Health). (2000) Evaluation of residential exposure to tetrachloroethene
15     using biomarkers of dose and neurological tests (non peer-reviewed draft). Albany, NY.
16
17     Ogata, M; Sugiyama, K; Kuroda, YI. (1962) Investigation of a dry-cleaning shop using tetrachloroethylene, with
18     special reference to Fujiwara's substance in the urine of the employees. Okayama Igakkai Zasshi 74:247-253.
19
20     Ogata, M; Takatsuka, Y; Tomokuni, K. (1971) Excretion of organic chlorine compounds in the urine of persons
21     exposed to vapours of trichloroethylene and tetrachloroethylene. Br J Ind Med 28:386-391.
22
23     Ohtsuki, T; Sato, K; Koizumi, A; et al. (1983) Limited capacity of humans to metabolize tetrachloroethylene.  Int
24     Arch Occup Environ Health 51:381-390.
25
26     Opdam, JJ; Smolders, JF. (1986) Alveolar sampling and fast kinetics of tetrachloroethene in man. I. Alveolar
27     sampling. BrJ Ind Med 43:814-824.
28
29     Otieno, MA; Baggs, RB; Hayes, JD; et al. (1997) Immunolocalization of microsomal glutathione S-transferase in rat
30     tissues. Drug Metab Dispos 25:12-20.
31
32     Overby, LH; Gardlik,  S; Philpot, RM; et al.  (1994) Unique distribution profiles of glutathione S-transferases in
33     regions of kidney, ureter, and bladder of rabbit. Lab Invest 70:468-478.
34
3 5     Pahler, A; Birner, G; Parker, J; et al. (1998) Generation of antibodies to di- and trichloroacetylated proteins and
3 6     immunochemical detection of protein adducts in rats treated with perchloroethene.  Chem Res Toxicol  11:995-1004.
37
3 8     Pahler, A; Volkel, W; Dekant, W. (1999a) Quantitation of N epsilon-(dichloroacetyl)-L-lysine in proteins after
39     perchloroethene exposure by gas chromatography-mass spectrometry using chemical ionization and negative ion
40     detection following immunoaffinity chromatography.  J Chromatogr A 847:25-34.
41
42     Pahler, A; Parker, J; Dekant, W. (1999b) Dose-dependent protein adduct formation in kidney, liver, and blood of
43     rats and in human blood after perchloroethene inhalation. Toxicol Sci 48:5-13.
44
45     Pastino, GM; Yap, WY;  Carroquino, M. (2000) Human variability and susceptibility to trichloroethylene  [In Process
46     Citation].  Environ Health Perspect 108(suppl2):201-214.
47
48     Pegg, DG; Zempel, JA; Braun, WH; et al. (1979) Disposition of tetrachloro(14C)ethylene following oral and
49     inhalation exposure in rats.  Toxicol Appl Pharmacol 51:465-474.
50
51     Pereira, MA. (1996) Carcinogenic activity of dichloroacetic acid and trichloroacetic acid in the liver of female
52     B6C3F1 mice.  Fundam Appl Toxicol 31:192-199.
53
54     Pezzagno, G; Imbriani, M; Ghittori, S; et al. (1988) Urinary concentration, environmental concentration, and
55     respiratory uptake of some solvents: effect of the work load. Am Ind Hyg Assoc J 49:546-552.

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        3-69        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Poet, TS.; Weitz, KK; Gies, RA; et al. (2002) PBPK modeling of the percutaneous absorption of perchloroethylene
  2     from a soil matrix in rats and humans. Toxicol Sci 67: 17-31.
  3
  4     Powell, JF. (1945) Trichloroethylene: absorption, elimination and metabolism. Br J Ind Med 2:142-145.
  5
  6     Quondamatteo, F; Schulz, TG; Bunzel, N; et al. (1998) Immunohistochemical localization of glutathione S-
  7     transferase-Tl in murine kidney, liver, and lung. Histochem Cell Biol 110:417-423.
  8
  9     Ramsey, JC; Andersen ME. (1984). A physiologically based description of the inhalation pharmacokinetics of
10     styrene in rats and humans.  Toxicol Appl Pharmacol 73:159-175.
11
12     Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
13     groundwater for a bathing and showering determination. Risk Anal 13:37^19.
14
15     Raucy, JL. (1995) Risk assessment: toxicity from chemical exposure resulting from enhanced expression of
16     CYP2E1. Toxicology 105:217-224.
17
18     Raunio, H; HusgafVel-Pursiainen, K; Anttila, S; et al. (1995) Diagnosis of polymorphisms in carcinogen-activating
19     and inactivating enzymes and cancer susceptibility-a review. Gene 159:113-121.
20
21     Reitz, RH. (1992) New data on tetrachloroethylene metabolism.  Office of Environmental Health Hazard
22     Assessment Advisory Panel Meeting.  California Environmental Protection Agency. Berkeley.
23
24     Reitz, RH; Nolan, RJ. (1986) Physiological pharmacokinetic modeling for perchloroethylene dose adjustment. Draft
25     Report to EPA's Science Advisory Board, pp. 1-23. Available from: IRIS desk, U.S. Environmental Protection
26     Agency, Washington, DC.
27
28     Reitz, RH; Mendrala, AL; Guengerich, FP. (1989) In vitro metabolism of methylene chloride in human and animal
29     tissues: use in physiologically based pharmacokinetic models. Toxicol Appl Pharmacol 97:230^16.
30
31     Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
32     for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
33     306.
34
3 5     Rhomberg, L. (1992) A cross-species scaling factor for carcinogen risk assessment based on equivalence of
36     mg/kg3/4/day. Federal Register 57:4152-24173.
37
38     Riihimaki, V; Pfaffli, P.  (1978) Percutaneous absorption of solvent vapors in man. Scand J Work Environ Health
39     4:73-85.
40
41     Ripp, SL; Overby, LH; Philpot, RM; et al. (1997) Oxidation of cysteine S-conjugates by rabbit liver microsomes and
42     cDNA- expressed flavin-containing mono-oxygenases: studies with S-(l,2- dichlorovinyl)-L-cysteine, S-(l,2,2-
43     trichlorovinyl)-L-cysteine,  S-allyl- L-cysteine, and S-benzyl-L-cysteine. Mol Pharmacol 51:507-515.
44
45     Rodilla, V; Benzie, AA; Veitch, JM; et al. (1998) Glutathione S-transferases in human renal cortex and neoplastic
46     tissue: enzymatic activity, isoenzyme profile and immunohistochemical localization. Xenobiotica 28:443-456.
47
48     Sakamoto, N. (1976) [Metabolism of tetrachloroethylene in guinea pigs (author'stransl)]. Sangyo Igaku 18:11-16.
49
50     Sata, F; Sapone, A; Elizondo, G; et al. (2000) CYP3A4 allelic variants with amino acid substitutions in exons 7 and
51     12: evidence for an allelic variant with altered catalytic activity.  Clin Pharmacol Ther 67:48-56.
52
53     Sausen, PJ; Elfarra, AA. (1990) Cysteine conjugate S-oxidase. Characterization of a novel enzymatic activity in rat
54     hepatic and renal microsomes.  J Biol Chem 265:6139-6145.
55

                   This document is a draft for review purposes only  and does not constitute Agency policy
       06/06/08                                        3-70        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Savolainen, H; Pfaffli, P; Tengen, M; et al. (1977) Biochemical and behavioural effects of inhalation exposure to
  2     tetrachloroethylene and dichloromethane. J Neuropathol Exp Neurol 36:941-949.
  3
  4     Schreiber, JS. (1993) Predicted infant exposure to tetrachloroethylene in human breast milk.  Risk Anal 13:515-
  5     524.Schreiber, JS; House, S; Prohonic, E; et al. (1993) An investigation of indoor air contamination in residences
  6     above dry cleaners. Risk Anal 13:355-344.
  7
  8     Schreiber, J. (1997) Transport of organic chemicals to breast milk: tetrachloroethene case study. In: Kacew, S;
  9     Lambert, GH; eds.  Environmental toxicology and pharmacology of human development. Washington, DC: Taylor
10     and Francis.
11
12     Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
13     tetrachloroethylene (perc) and deficits in visual contrast sensitivity. Environ Health Perspect  110:655-664.
14
15     Schultz, J; Weiner, H. (1979) Alteration of the enzymology of chloral hydrate reduction in the presence of ethanol.
16     Biochem Pharmacol 28:3379-3384.
17
18     Schumann, AM; Quast, JF; Watanabe, PG. (1980) The pharmacokinetics and macromolecular interactions of
19     perchloroethylene in mice and rats as related to oncogenicity. Toxicol Appl Pharmacol 55:207-219.
20
21     Sellers, EM; Lang, M; Koch-Weser, J; et al. (1972) Interaction of chloral hydrate and ethanol in man. I.
22     Metabolism. Clin Pharmacol Ther 13:37-49.
23
24     Sheldon, L; Handy, R; Hartwell, W; et al. (1985) Human exposure assessment to environmental chemicals: nursing
25     mothers study.  Final report. Research Triangle Institute, Research Triangle Park, NC.
26
27     Skender, LJ; Karacic, V; Prpic-Majic, D.  (1991) A comparative study of human levels of trichloroethylene and
28     tetrachloroethylene after occupational exposure. Arch Environ Health 46:174-178.
29
30     Speerschneider, P; Dekant, W. (1995) Renal tumorigenicity of 1,1-dichloroethene in mice: the role of male- specific
31     expression of cytochrome P450 2E1 in the renal bioactivation of 1,1-dichloroethene. Toxicol Appl Pharmacol
32     130:48-56.
33
34     Stephens, EA; Taylor, JA; Kaplan, N; et al.  (1994) Ethnic variation in the CYP2E1 gene:  polymorphism analysis of
35     695 African-Americans, European-Americans and Taiwanese. Pharmacogenetics 4:185-192.
36
3 7     Stevens, JL. (1985) Isolation and characterization of a rat liver enzyme with both cysteine conjugate beta-lyase and
38     kynureninase activity. J Biol Chem 260:7945-7950.
39
40     Stevens, J; Jakoby, WB. (1983) Cysteine conjugate beta-lyase. Mol Pharmacol 23:761-765.
41
42     Stewart, RD; Dodd, HC. (1964) Absorption of carbon tetrachloride, trichloroethylene, tetrachloroethylene,
43     methylene chloride and 1,1,1-trichloroethane through the human skin. J Ind Hyg 25:429-446.
44
45     Stewart, RD; Erley, DS; Schaeffer, AW; et al. (1961) Accidental vapor exposure to anesthetic concentrations of a
46     solvent containing tetrachloroethylene.  Indus Med Sug 30:327-330.
47
48     Stewart, RD; Baretta, ED; Dodd, HC; et al. (1970) Experimental human exposure to tetrachloroethylene.  Arch
49     Environ Health 20:225-229.
50
51     Stewart, RD; Hake, CL; Forster AJ; et al. (1974) Tetrachloroethylene: development of a biologic standard for the
52     industrial worker by breath analysis. National Institute of Occupational Safety and Health. Cincinnati, OH.
53     NIOSH-MCOW-ENVM-PCE-74-6.
54


                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        3-71        DRAFT-DO NOT CITE OR QUOTE

-------
  1     Stewart, RD; Hake, CL; Wu, A. (1977) Effects of perchloroethylene drag interaction on behavior and neurological
  2     function. DHEW (NIOSH) Publ. No. 77-191. Department of Health, Education, And Welfare, Washington, DC.
  3
  4     Stott, WT; Quast, JF; Watanabe, PG. (1982) The pharmacokinetics and macromolecular interactions of
  5     trichloroethylene in mice and rats. Toxicol Appl Pharmacol 62:137-151.
  6
  7     Tanaka, S; Ikeda, M.  (1968) A method for determination of trichloroethanol and trichloroacetic acid in urine. Br J
  8     IndMed 25:214-219.
  9
10     Tateishi, M; Suzuki, S; Shimizu, H. (1978) Cysteine conjugate beta-lyase in rat liver. A novel enzyme catalyzing
11     formation of thiol-containing metabolites of drags.  J Biol Chem 253:8854-8859.
12
13     Terrier, P; Townsend, AJ; Coindre, JM; et al. (1990) An immunohistochemical study of pi class glutathione S-
14     transferase expression in normal human tissue. Am JPathol 137:845-853.
15
16     Tomisawa, H; Suzuki, S; Ichihara, S; et al. (1984) Purification and characterization of C-S lyase from
17     Fusobacterium varium. A C-S cleavage enzyme of cysteine conjugates and  some S- containing amino acids.  J Biol
18     Chem 259:2588-2593.
19
20     Tomisawa, H; Ichihara, S; Fukazawa, H; et al. (1986) Purification and characterization of human hepatic cysteine-
21     conjugate beta-lyase.  Biochem 1235:569-575.
22
23     Tong, Z; Board, PG; Anders, MW. (1998a) Glutathione transferase zeta-catalyzed biotransformation of
24     dichloroacetic acid and other alpha-haloacids. Chem Res Toxicol 11:1332-1338.
25
26     Tong, Z; Board, PG; Anders, MW. (1998b) Glutathione transferase zeta catalyses the oxygenation of the carcinogen
27     dichloroacetic acid to glyoxylic acid.  Biochem J 33 l(Pt 2):371-374.
28
29     Tsurata, H. (1989) Skin absorption of organic solvent vapors in nude mice in vivo. Ind  Health 27:37-47.
30
31     Tzeng, HF; Blackburn, AC; Board, PG; et al. (2000) Polymorphism- and species-dependent inactivation of
32     glutathione transferase zeta by dichloroacetate.  Chem Res Toxicol 13:231-236.
33
34     U.S. EPA (Environmental Protection Agency). (1985) Health assessment document for tetrachloroethylene
3 5     (perchloroethylene).  Office  of Health and Environmental Assessment, Office of Research and Development,
36     Washington, DC; EPA/600/8-82/005F.  Available from: National Technical Information Service,  Springfield, VA;
37     PB-86-174489/AS.
38
39     U.S. EPA (Environmental Protection Agency). (1986) Addendum to the health assessment document for
40     tetrachloroethylene (perchloroethylene) [review draft]. Office  of Health and Environmental Assessment, Office of
41     Research and Development,  Washington, DC; EPA/600/8-82/005FA. Available from:  National Technical
42     Information Service, Springfield, VA; PB-86-174489/AS.
43
44     U.S. EPA (Environmental Protection Agency). (1989) Risk assessment guidance for Superfund (RAGS), vol. I:
45     human health evaluation manual, part A (1989). Office of Emergency and Remedial Response, Washington, DC;
46     EPA/540/1-89/002. Available online at http://www.epa.gov/superfund/programs/risk/ragsa/index.htm.
47
48     U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
49     carcinogenicity of tetrachloroethylene (perchloroethylene).  Office of Health and Environmental Assessment,
50     Washington, DC; EPA/600/6-91/002F.  Available from: National Technical Information Service,  Springfield, VA.
51
52     U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
53     and application of inhalation dosimetry. Office of Health and Environmental Assessment, Environmental Criteria
54     and Assessment Office, Cincinnati, OH; EPA/600/8-90/066F.  Available from: National Technical Information
55     Service, Springfield, VA, PB2000-500023, and online at http://www.epa.gov/ncea.

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         3-72        DRAFT-DO  NOT CITE OR QUOTE

-------
  1     U.S. EPA (Environmental Protection Agency). (1998) Dichloroacetic acid: carcinogenicity identification
  2     characterization summary.  National Center for Environmental Assessment, Office of Research and Development,
  3     Washington, DC; EPA NCEA-W-0372.
  4
  5     U.S. EPA (Environmental Protection Agency). (2001) Trichloroethylene health risk assessment: synthesis and
  6     characterization. National Center for Environmental Assessment, Office of Research and Development,
  7     Washington, DC; EPA/600/P-01/002A. Available online at
  8     http://oaspub.epa.gov/eims/eimscomm.getfile7p  download id=4580.
  9
10     Uttamsingh, V; Keller, DA; Anders, MW. (1998) Acylase I-catalyzed deacetylation of N-acetyl-L-cysteine and S-
11     alkyl— acetyl-L-cysteines.  Chem Res Toxicol 11:800-809.
12
13     Vainio, H; Parkki, MG; Mamiemi, J. (1976) Effects of aliphatic chlorohydrocarbons on drug-metabolizing enzymes
14     in rat liver in vivo.  Xenobiotica 6:599-604.
15
16     Vamvakas, S; Dekant, W; Berthold, K; et al. (1987) Enzymatic transformation of mercapturic acids derived from
17     halogenated alkenes to reactive and mutagenic intermediates. Biochem Pharmacol 36:2741-2748.
18
19     Vamvakas, S; Dekant, W; Henschler, D. (1989a) Genotoxicity of haloalkene and haloalkane glutathione S-
20     conjugates in porcine kidney cells. Toxicol In Vitro 3:151-156.
21
22     Vamvakas, S; Herkenhoff, M; Dekant, W; et al. (1989b) Mutagenicity of tetrachloroethene in the Ames test-
23     metabolic activation by conjugation with glutathione. J Biochem Toxicol 4:21-27.
24
25     Vamvakas, S; Dekant, W; Henschler, D. (1989c) Assessment of unscheduled DNA synthesis in a cultured line of
26     renal epithelial cells exposed to cysteine S-conjugates of haloalkenes and haloalkanes. Mutat Res 222:329-335.
27
28     Veitch, JM;  Murray, GI; Juronen, E; et al. (1997) Theta-class glutathione S-transferases in human kidney and renal
29     tumours. Biochem Soc Trans 25:8605.
30
31     Vieira, I; Sonnier, M; Cresteil, T.  (1996) Developmental expression of CYP2E1 in the human liver:
32     hypermethylation control of gene  expression during the neonatal period. Eur J Biochem 238(2):476-83.
33
34     Volkel, W; Friedewald, M; Lederer, E; et al. (1998) Biotransformation of perchloroethene:  dose-dependent
3 5     excretion of trichloroacetic acid, dichloroacetic acid, and N-acetyl-S- (trichlorovinyl)-L-cysteine in rats and humans
36     after inhalation.  Toxicol Appl Pharmacol 153:20-27.
37
38     Volkel, W; Pahler, A;  Dekant, W. (1999) Gas chromatography-negative ion chemical ionization mass spectrometry
39     as a powerful tool for the detection of mercapturic acids and DNA and protein adducts as biomarkers of exposure to
40     halogenated olefins. JChromatogr A 847:35-46.
41
42     Ward, RC; Travis, CC; Hetrick, DM; et al.  (1988) Pharmacokinetics of tetrachloroethylene. Toxicol Appl
43     Pharmacol 93:108-117.
44
45     Waters, EM; Gerstner, HB; Huff,  JE. (1977) Trichloroethylene.  I. an overview. J Toxicol Environ Health 2:671-
46     707.
47
48     Weichardt, H; Lindner, J. (1975)  Gesundheitsgefahren durch Perchlorathlen in Chemisch-Reinigungsbetrieben aus
49     arbiet medizinischtoxikologischer Sicht. Staub Reinhalt Luft 35:416^4-20.
50
51     Weiss, G. (1969) [Observation of the course of trichloroacetic acid excretion in occupational tetrachloroethylene
52     poisoning].  Zentralbl  Arbeitsmed 19:143-146.
53
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        3-73        DRAFT-DO NOT CITE OR QUOTE

-------
 1    Werner, M; Birner, G; Dekant, W. (1996) Sulfoxidation of mercapturic acids derived from tri- and tetrachloroethene
 2    by cytochromes P450 3 A: a bioactivation reaction in addition to deacetylation and cysteine conjugate beta-lyase
 3    mediated cleavage. Chem Res Toxicol 9:41-49.
 4
 5    Westlind, A; Lofberg, L; Tindberg, N; et al. (1999) Interindividual differences in hepatic expression of CYP3 A4:
 6    relationship to genetic polymorphism in the 5'-upstream regulatory region. Biochem Biophys Res Commun
 7    259:201-205.
 8
 9    Yllner, S. (1961) Urinary metabolites of 14-C-tetrachloroethylene in mice. Nature 191:820-821.
10
11    Yoo, JS; Guengerich, FP; Yang, CS. (1988) Metabolism of N-nitrosodialkylamines by human liver microsomes.
12    Cancer Res 48:1499-1504.
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        3-74        DRAFT-DO NOT CITE OR QUOTE

-------
 1                               4. HAZARD IDENTIFICATION
 2
 O
 4    4.1.  OVERALL APPROACH
 5          This chapter discusses tetrachloroethylene toxicity on an organ-specific basis, with liver,
 6    kidney, neurotoxicity, and developmental/reproductive effects as the major emphasis in separate
 7    sections. For each of the major organ systems, human effects are presented first, followed by
 8    effects in animals and in in vitro systems. Cancer and noncancer toxicity and mode of action
 9    (MOA) are also included in the discussions. Of note, site concordance of effect between animals
10    and humans is generally not assumed.
11          Evidence  for each organ system is summarized, but an in-depth discussion or data
12    evaluation is not provided for any individual studies, especially those evaluated and discussed in
13    previous EPA documents and other Agency reports. Several existing publications provide more
14    detailed study descriptions: the World Health Organization-International Programme on
15    Chemical Safety (WHO-IPCS, 2006), the ATSDR' s Toxicologic Profile for Tetrachloroethylene
16    (ATSDR, 1997) and the International Agency for Research on Cancer (IARC) review the health
17    effect evidence on tetrachloroethylene, trichloroethylene, and their common metabolites; other
18    studies review this evidence on dry cleaner as an occupational title (IARC,  1995);
19    Tetrachloroethene Ambient Air Criteria Document (NYS DOH, 1997); and Public Health Goal
20   for Tetrachloroethylene in Drinking Water (Cal EPA, 2001). The details for earlier toxicity and
21    carcinogenicity studies may also be found in previous EPA assessments (e.g., U.S. EPA, 1980,
22    1985a, b, 1986a,  1991a).
23
24    4.2.  OVERVIEW OF TETRACHLOROETHYLENE METABOLISM
25          Most tetrachloroethylene toxicity and cancer-causing activity, other than neurotoxicity, is
26    generally attributed to its metabolites.  For example, historically, a direct relationship has been
27    demonstrated between the level of hepatic microsomal cytochrome P450, the extent of
28    metabolism of tetrachloroethylene in vivo, and cellular damage (Bonse et al., 1975; Bonse and
29    Henschler, 1976;  Moslen et al., 1977; Pegg et al., 1979; Schumann et al., 1980; Buben and
30    O'Flaherty, 1985; U.S. EPA, 1985a, 1991a). In addition, several oxidative (P450) and GSH-
31    derived tetrachloroethylene metabolites have been shown to induce toxic and carcinogenic
32    effects in similar targets when they are administered directly (IARC, 1995; Herren-Freund et al.,
33    1987; Bull et al.,  1990; Bull, 2000; Pereira,  1996; DeAngelo et al., 1991,  1999; Daniel,  1963;
34    Carter et al., 2003; Elfarra and Krause, 2007). This metabolism prerequisite for certain toxic
35    effects is true for  related halogenated ethylenes and ethanes as well (IARC, 1995; U.S. EPA,
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-1        DRAFT-DO NOT CITE OR QUOTE

-------
 1    1991a, 2001a; Dekant, 2001).  CNS effects are a notable exception in that they are largely
 2    attributable to the parent compound.
 3          As detailed in Chapter 3, tetrachloroethylene is metabolized through at least two major
 4    pathways. The oxidative P450, or CYP, pathway is quantitatively most important, and it
 5    accounts for the greatest amount of observed metabolite in all species at all doses tested (see
 6    Section 3.3.3.1).  The other is the GSH conjugation pathway, which is associated with renal
 7    toxicity and renal carcinogenicity (see Section 3.3.3.2). A significant portion of absorbed dose in
 8    human studies (as much as 20-40%) cannot be tracked either as parent compound or metabolites
 9    of known pathways, which introduces uncertainty about the identity and the amounts  of the
10    metabolites formed in humans (U.S. EPA, 1991 a; Bogen and McKone, 1988). TCA,  a product
11    of the oxidative pathway, is the major urinary metabolite derived from tetrachloroethylene
12    metabolism.  TCA is a mouse liver carcinogen.  TCA may contribute in part to findings of liver
13    toxicity and cancer observed in tetrachloroethylene-exposed animals; however, according to
14    some investigators (Clewell et al., 2004, 2005), the amount of TCA produced from
15    tetrachloroethylene in rodent bioassays is insufficient to account in total for observed
16    hepatocarcinogenicity (see Appendix 4A). DC A is another known tetrachloroethylene urinary
17    metabolite that is formed in both the oxidative pathway by dechlorination of TCA and, in organs
18    other than the liver, in the GSH pathway. DC A is known to cause liver cancer in both rats and
19    mice.  Whether DC A contributes to tetrachloroethylene-induced toxicity or carcinogenicity in the
20    liver is not known.
21          Tetrachloroethylene oxide, trichloroacetyl chloride, and chloral/chloral hydrate are
22    proposed reactive intermediates in tetrachloroethylene P450 oxidation.  Tetrachloroethylene
23    oxide and trichloroacetyl chloride have the potential to contribute to tetrachloroethylene toxicity
24    and carcinogenicity, particularly in the liver. Detection of TCOH in the urine of
25    tetrachloroethylene-exposed humans and animals would provide evidence for the existence of
26    the chloral hydrate intermediate.  However, TCOH—and therefore the evidence that its
27    chloral/chloral hydrate precursor is formed from tetrachloroethylene—is not consistently
28    detected and it might be an artifact of the methodology used in some but not all studies. Chloral
29    hydrate is a liver carcinogen in mice.
30          The glutathione pathway entails initial glutathione-S-transferase catalyzed conjugation of
31    tetrachloroethylene with GSH to form TCVG.  The cellular damage and genotoxic effects of
32    these conjugation products are thought to be from their further metabolism via beta-lyase, FMO3,
33    and/or P450 metabolism to highly reactive toxic products.
34          All of these metabolites have effects that may contribute to the toxicity and
35    carcinogenicity of tetrachloroethylene, although the role of specific intermediates has not been
36    elucidated.  Genotoxicity of the oxidative metabolites TCA, DC A, chloral hydrate,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-2        DRAFT-DO NOT CITE OR QUOTE

-------
 1    tetrachloroethylene oxide and the GSH-derived intermediates TCVC, TCVG, and NAcTCVC is
 2    discussed in Section 4.3; for TCA and DCA, see Section 4.4.4.3 for peroxisome proliferator-
 3    activated receptor alpha (PPAR-a) form activation and Section 4.4.4.4 for hypomethylation.
 4    Section 4.10.3 provides a summary of the cancer MOA conclusions for tetrachloroethylene.
 5
 6    4.3.  GENOTOXICITY
 7          Tetrachloroethylene has been extensively studied for genotoxic activity in a variety of in
 8    vitro assay systems such as bacteria, yeast, and mammalian cells (see reviews by U.S. EPA,
 9    1985c, 1991a; IARC, 1995; ATSDR, 1997). Also, a review of the mutagenicity of
10    trichloroethylene (Moore and Harrington-Brock, 2000) contains a discussion of several of known
11    (TCA, DCA) and proposed (chloral hydrate) tetrachloroethylene metabolites.
12          The application of mutagenicity data to the question of potential carcinogenicity is based
13    on the premise that genetic alterations are found in all cancers.  Mutagenesis is the ability of
14    chemicals to alter the genetic material in a manner that permits changes to be transmitted during
15    cell division. Although most tests for mutagenicity detect changes in DNA or chromosomes,
16    modifications of the epigenome, including proteins associated with DNA or RNA, can also cause
17    transmissible changes. Genetic alterations can occur via a variety of mechanisms including gene
18    mutations, deletions, translocations or amplification; evidence of mutagenesis provides
19    mechanistic support for the inference of potential for carcinogenicity in humans.
20          The following discussion focuses on the conclusions of the earlier studies and includes
21    details of recent studies that may provide some insight into the potential genotoxicity of
22    tetrachloroethylene. Positive findings were reported in some experiments using technical-grade
23    tetrachloroethylene that contained impurities or used epichlorohydrin or epoxybutane as
24    stabilizers, both of which are clearly mutagenic in a number of biological systems.  Purified
25    tetrachloroethylene was negative in the same systems tested without or with mixed-function
26    oxidation activity provided by either rat or hamster liver  S9 (Haworth et al., 1983).  The results
27    of a large number of in vitro genotoxicity tests in which tetrachloroethylene was the test agent do
28    not clearly support the conclusion that tetrachloroethylene exhibits direct mutagenic activity,
29    although the few studies of conditions that would generate the GSH conjugate were positive (U. S.
30    EPA, 1991a; IARC, 1995; ATSDR, 1997).
31          An increased level of DNA single-strand breaks (SSB) was seen in liver and kidney
32    tissues but not in the lung tissue of mice 1 hr after single intraperitoneal (i.p.) injections of 4-8
33    mmol/kg (663-1326 mg/kg) of tetrachloroethylene (Walles, 1986).  Potter et al. (1996) found no
34    increases in DNA strand breaks in kidneys of male F344 rats after a single gavage treatment with
35    1,000 mg/kg tetrachloroethylene. However, differences in species and/or route of exposure
36    preclude direct comparisons of these studies. Muzzullo (1987) found DNA binding of

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-3         DRAFT-DO NOT CITE OR QUOTE

-------
 1    tetrachloroethylene in mouse liver and rat kidney. Cytosols from several organs were more
 2    effective than liver microsomes in enhancing in vitro DNA or protein binding of
 3    tetrachloroethylene, and enrichment with GSH enhanced the activity of liver microsomes
 4    (Muzzullo, 1987).
 5          Toraason et al. (1999) found no increase in 8-hydroxydeoxyguanosine (8-OHdG) in the
 6    urine, the liver, or the kidneys of male F344 rats after a single i.p. injection of
 7    tetrachloroethylene at 100, 500, or 1,000 mg/kg (8-OHdG in peripheral lymphocytes was
 8    measured only in the 500 mg/kg group). In a subsequent paper, Toraason et al. (2003) reported
 9    no increase in 8-OHdG in urine of 18 dry cleaner workers sampled pre- and post-shift work
10    (time-weighted average [TWA] concentration of tetrachloroethylene was 3.8 +  5.3 ppm).
11          Tetrachloroethylene induced damage was observed in the sister chromatid exchange
12    (SCE) assay and in the single-cell gel test in human blood culture treated with up to 5 mM (-830
13    rng/L) tetrachloroethylene, at which viability was reduced by 40% (Hartmann and  Speit, 1995).
14    Tetrachloroethylene exposure increased the frequency of micronuclei in peripheral blood
15    reticulocytes or hepatocytes of ddY mice given single i.p. injections at 1,000 or 2,000 mg/kg
16    tetrachloroethylene given after, but not prior to partial hepatectomy (Murakami and Horikawa,
17    1995). Tetrachloroethylene-induced micronuclei have also been reported in cultured Chinese
18    hamster kidney cells (Wang et  al., 2001) and in human cells (Doherty et al., 1996; White et al.,
19    2001). Micronucleus induction was enhanced by tetrachloroethylene exposure  in human
20    lymphoblastoid cells by stable  expression of cDNAs encoding either CYP2E1 (hEl cells) or
21    human CYP1A2, 2A6, 3A4, 2E1 and microsomal epoxide hydrolase (Doherty et al., 1996). In
22    contrast to these findings, neither chromosome aberrations nor SCE were induced in Chinese
23    hamster ovary cells following in vitro exposure to tetrachloroethylene (Galloway et al., 1987).
24          Tetrachloroethylene when incubated with rat liver GST, GSH, and a rat kidney fraction,
25    exhibited a clear dose-response in the Ames test (Vamvakas et al., 1989b). In addition, it was
26    demonstrated that TCVG was produced from tetrachloroethylene in isolated perfused rat liver
27    and excreted into bile; in the presence of a rat kidney fraction, the collected bile was mutagenic
28    in Salmonella, as was purified TCVG (Vamvakas et al., 1989b).  Dreesen (2003) also
29    demonstrated, for TCVG, an unequivocal dose-dependent mutagenic response in the TA 100
30    strain in the presence of the rat kidney S9-protein fraction; TCVC was mutagenic without
31    metabolic activation in this strain.  In a separate study, the tetrachloroethylene metabolite TCVC
32    was also positive in Salmonella (strains TA 98 and TA 100) and inhibition of beta lyase activity
33    blocked the effect (Dekant et al., 1986). A subsequent study indicated that Salmonella also were
34    capable of deacetylating the urinary metabolite NAcTCVC when TA 100 showed a clear positive
35    response without exogenous activation (Vamvakas et al., 1987).  Vamvakas et al. (1989a) also
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-4        DRAFT-DO NOT CITE OR QUOTE

-------
 1    reported concentration-related increases in unscheduled DNA synthesis (UDS) in LLC-PK1 (a
 2    porcine kidney cell line) exposed to TCVC, with the effect abolished by a beta lyase inhibitor.
 3          Several identified or putative P450 metabolites of tetrachloroethylene are mutagenic.
 4    Tetrachloroethylene-epoxide, a hypothesized intermediate in tetrachloroethylene P450 oxidative
 5    metabolism (Henschler et al., 1977a, b), is mutagenic in bacteria (Kline et al., 1982).  As
 6    reviewed by Moore and Harrington-Brock (2000), the oxidative metabolite TCA, the major
 7    urinary excretion product, exhibits little, if any, genotoxic activity. However, in vitro
 8    experiments with TCA should be interpreted with caution if steps have not been taken to
 9    neutralize pH changes caused by the compound. TCA was positive in genotoxicity studies
10    conducted by Bhunya and Behera (1987), Bhunya and Jena (1996), and Birner et al. (1994) in in
11    vivo mouse and chick test systems. TCA has also been reported to induce DNA SSB in hepatic
12    DNA of mice.  A single dose of TCA was administered to Sprague-Dawley rats and B6C3F1
13    mice by gavage (Nelson and Bull, 1988).  The animals were sacrificed 4 hrs later and SSB in
14    liver DNA were analyzed by alkaline unwinding assay. SSB were observed in a dose-dependent
15    manner. The lowest dose of TCA that produced significant SSB in the rats was 0.6 mmol/kg (98
16    mg/kg).  For mice, the lowest dose of TCA that produced significant increases was 0.006
17    mmol/kg (0.98 mg/kg). Further, in another study by the same authors (Nelson et al., 1989), the
18    incidence of SSB was elevated at 1 hr after a single i.p. dose TCA exposure of 500 mg/kg; the
19    level returned to control levels by 8 hrs. In a second experiment, no increase in SSB in hepatic
20    DNA was observed 24 hrs after 10 days of daily gavage of 500 mg/kg TCA.  A later study by
21    Styles et al. (1991), using essentially the same procedures, failed to detect any increase in SSB.
22          Chang et al. (1992) observed a marginally significant increase in SSB in hepatocyte DNA
23    of mice but not rats at 4 hrs after a single TCA dose of 10 mmol/kg (1,633.9 mg/kg)
24    administered orally. However, the authors considered this finding to be not biologically
25    significant, because SSB  were not increased at 1 hr and there were no detectable SSB in isolated
26    hepatocytes exposed to concentrations of TCA as high as 10 mM (-1,650 mg/L).  Storer et al.
27    (1996), after evaluating 81 chemicals (carcinogens, noncarcinogens, mutagens, and
28    nonmutagens) for SSB using the alkaline unwinding assay, demonstrated that increased DNA
29    SSB at high doses can be the result of cytotoxicity involving endonucleocytic degradation of
30    DNA.
31          As reviewed elsewhere (see Salmon et al., 1995; Moore and Harrington-Brock, 2000),
32    chloral hydrate is mutagenic in the standard battery of screening assays.  Effects include positive
33    results in bacterial mutation tests for point mutations and in the mouse lymphoma  assay for
34    mutagenicity at the Tk locus (e.g., Haworth et al., 1983). In vitro tests showed that chloral
35    hydrate also induced micronuclei and aneuploidy in human peripheral blood lymphocytes or
36    Chinese hamster pulmonary cell lines. Micronuclei were induced in Chinese hamster embryonic

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-5         DRAFT-DO NOT CITE OR QUOTE

-------
 1    fibroblasts.  Several studies demonstrate that chloral hydrate induces aneuploidy (loss or gain of
 2    whole chormosomes) in both mitotic and meiotic cells, including yeast (Singh and Sinha, 1976,
 3    1979; Kafer, 1985; Gualandi, 1987; Sora and Agostini-Carbone, 1987), cultured mammalian
 4    somatic cells (Degrassi and Tanzarella, 1988), and spermatocytes of mice (Russo et al., 1984;
 5    Liang and Pacchierotti, 1988). Chloral hydrate has also been shown to block spindle elongation
 6    in insect spermatocytes (Ris, 1949). Chloral hydrate was negative for sex-linked recessive lethal
 7    mutations in drosophila (Yoon et al., 1985). It induces SSB in hepatic DNA of mice and rats
 8    (Nelson and Bull,  1988) and mitotic gene conversion in yeast (Bronzetti et al., 1984). Schatten
 9    and Chakrabarti (1998) showed that chloral hydrate affects centrosome structure, which results
10    in the inability to reform normal microtubule formations and causes abnormal fertilization and
11    mitosis of sea urchin embryos.
12           The chloroacid metabolite, DCA, is also mutagenic in the standard battery of screening
13    tests (reviewed by Moore and Harrington-Brock,  2000). DCA was positive in bacterial mutation
14    tests, in the in vitro mouse lymphoma assay, the micronucleus induction test, the Big Blue mouse
15    system and  other tests (DeMarini et al., 1994; Fuscoe et al., 1996; Nelson and Bull, 1988;
16    Harrington-Brock  et al., 1998; Leavitt et al., 1997; Chang et al., 1989; Bignami et al., 1980).
17    Anna et al. (1994) compared mutations in the ras gene in liver tumors in mice treated orally with
18    tetrachloroethylene, DCA and trichloroethylene with those in untreated mice.  The frequency of
19    mutations at codon 61 ofH-ras was significantly  lower in liver tumors of tetrachloroethylene-
20    exposed mice but not in DCA or trichloroethylene tumors.  Thus, the phenotype of
21    tetrachloroethylene-induced mouse tumors appeared to differ from trichloroethylene, DCA or
22    spontaneous occurring tumors.  While not sufficient to indicate the MO A, tumor phenotype data
23    regarding H-ras codon 61 suggests that tetrachloroethylene-induced liver tumors differ from
24    those induced by DCA, TCA, or trichloroethylene and those arising spontaneously in the mouse.
25           In summary, tetrachloroethylene has been shown to induce some genotoxic effects
26    (micronuclei and SCEs following in vitro exposure, DNA binding and SSBs in tumor tissue).
27    Results of in vitro  mutagenicity (Ames) or DNA binding assays of tetrachloroethylene have
28    largely been negative except in the few tests of conditions where metabolites of the GSH
29    pathway are generated. The GSH metabolites are clearly mutagenic. TCVC is the most potent
30    bacterial mutagen  of the tetrachloroethylene metabolites and induces UDS in a porcine kidney
31    cell line; TCVG and NAcTCVC are also mutagenic in bacteria. The known (DCA) or putative
32    (tetrachloroethylene oxide, chloral hydrate) P450 metabolites also exhibit mutagenicity.
33           Uncertainties with regard to the genotoxicity characterization include that not all
34    tetrachloroethylene metabolites have been identified, nor have all the known or postulated
35    metabolites been sufficiently tested in the standard genotoxicity screening battery. Of note,
36    bacterial mutation testing protocols typically specify the inclusion of cytotoxic concentrations of

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-6        DRAFT-DO NOT CITE OR QUOTE

-------
 1    the test article, and the relative potency of the metabolites in vitro may not necessarily inform
 2    their relative contribution to the overall mechanistic effects of the parent chemical. This may be
 3    especially relevant when evaluating in vitro testing results for tetrachloroethylene, which can
 4    undergo inter-organ metabolic processing involving multiple enzyme systems to yield highly
 5    reactive species.  In addition, such tests do not provide data for all effects that are relevant for
 6    carcinogenesis. Thus, other data gaps include incomplete characterization of the metabolites in
 7    tests beyond the standard battery of genotoxicity tests including on important genetic and
 8    epigenetic endpoints.
 9           Section 4.10.3 addresses the contribution of mutagenicity of tetrachloroethylene and its
10    oxidative and GSH-derived metabolites to the MO A of carcinogeniticy for tetrachloroethylene.
11    Overall, the finding is that the MOA for tetrachloroethylene-induced carcinogenesis is not yet
12    fully characterized, completely tested, or understood. The database for hepatocarcinogenesis is
13    especially limited with regard to chemical-specific studies. It is concluded that the role of
14    genotoxicity in hepatocarcinogenicity, an effect that is thought to be related to products of CYP
15    metabolism, is uncertain (see Section 4.4.4.5).  While the complete mechanisms are not yet
16    understood, the weight of evidence, including the known mutagenicity of GSH-derived
17    metabolites produced in the kidney, suggests a mutagenic MOA cannot be ruled out for
18    tetrachloroethylene-induced renal carcinogenesis (see Section 4.5.4.3.3).
19
20    4.4. LIVER TOXICITY
21    4.4.1.  Human Effects
22           A number of hepatotoxic effects,  including hepatomegaly, hepatocellular damage, and
23    elevations of several hepatic enzymes and bilirubin degradation byproducts, have been observed
24    after acute high-level exposure to tetrachloroethylene (levels not identified; Meckler and Phelps,
25    1966; Coler and Rossmiller, 1953; Hake  and Stewart, 1977; Saland, 1967; Stewart et al., 1961,
26    as reported in ATSDR, 1997). One case  report noted obstructive jaundice and hepatomegaly in
27    an infant exposed orally to tetrachloroethylene (1 mg/dL; Bagnell and Ellenberger, 1977, as
28    reported in ATSDR, 1997).
29
30    4.4.1.1. Liver Damage
31           Four cross-sectional studies were  available that evaluated the prevalence of liver damage
32    among dry cleaner populations (Lauwerys et al., 1983; Cai et al.,  1991; Gennari et al.; 1992;
33    Brodkin et al., 1995). These studies assessed serum concentration of a number of hepatic
34    enzymes in  dry cleaner and control populations.  Additionally, sonographic changes to hepatic
35    parenchymal tissue were examined in one study (Brodkin et al., 1995). An elevated
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-7        DRAFT-DO NOT CITE OR QUOTE

-------
 1    concentration of the serum enzyme GOT and mild hepatic changes were notable observations in
 2    two studies (Gennari et al., 1992; Brodkin et al., 1995).
 3          Gennari et al. (1992) measured the electrophoretic fractionation patterns of serum GGT
 4    isozymes among  141 tetrachloroethylene-exposed dry cleaners and 130 nonexposed controls
 5    selected from staff and students from the academic institution of the principal investigators.
 6    Both the exposed subjects and the controls had similar lifestyle (smoking, alcohol consumption)
 7    and clinical medical histories. The TWA tetrachloroethylene concentration in the dry cleaning
 8    facilities was 11.3 ppm.  Total GGT was higher in exposed workers (exposed: mean of 12.4
 9    international units per liter [U/L; standard deviation, 6.9 U/L]; controls:  8.8 U/L  [4.9 U/L],/? <
10    0.01).  The GGT-2 isoenzyme component was higher in exposed workers (6.8 U/L [5.7 U/L] in
11    exposed vs. 3.5 U/L [3.3 U/L] in controls,/* < 0.01) and the GGT-4 component was detectable in
12    exposed workers but not measurable in controls. The authors regarded a GGT-2/GGT-3 ratio of
13    greater than 1 as a sensitive index of the reciprocal behavior of the two isoenzymes.  GGT-2 is
14    generally associated with activation of liver microsomal enzymes.  GGT-4 is common in liver
15    diseases and indicates hepato-biliary impairment.
16          This study excluded individuals who presented values for GGT, or other liver enzymes
17    above a normal range, and individuals who had past or current liver disease. None of the
18    workers showed any clinical symptoms of liver disease, and their enzymatic profiles, including
19    GGT,  aspartase amino transaminase (AST), alanine amino transaminase, 5'-nucleotidase, and
20    alkaline phosphotase, were within the clinically normal reference limits. Given the study's
21    exclusion criteria, it is not surprising that liver enzyme concentrations were within a normal
22    range. The authors stated that more research is required to develop this GGT fractionation assay
23    into a  clinically useful method of measuring liver function.  Nevertheless, the study showed that
24    these dry cleaners had markers of tetrachloroethylene oxidative metabolism (GGT-2) and liver
25    impairment (GGT-4).
26          The study by Brodkin et al. (1995) examined liver function and carried out sonography
27    measurements in a population of 27 dry cleaners and 26 nonexposed laundry workers.  Dry
28    cleaners were older and had a longer duration of employment than did laundry workers. The
29    noninvasive imaged penetration of ultrasound into liver tissue can reveal the presence of fat
30    accumulation and fibrous structures. The mean TWA exposure (8 hrs) among all dry cleaners
31    was 15.8 ppm (range: 0.4-83 ppm). The investigators found a higher prevalence of abnormal
32    hepatic sonograms among the dry cleaners (67%) than among laundry workers (38%;p < 0.05),
33    the control group. Hepatic parenchymal changes, as assessed by sonography, were graded as
34    mild, moderate, or severe. The prevalence of hepatic parenchymal changes increased both with
35    increasing current concentration and with cumulative exposure (p < 0.05).  Subjects with
36    serological evidence of active hepatitis infection were excluded from these analyses.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-8        DRAFT-DO NOT CITE OR QUOTE

-------
 1          Brodkin et al. (1995) fit logistic regression models to examine possible associations
 2    between mild or greater parenchymal changes and tetrachloroethylene exposure. These analyses
 3    included adjustment for the effects of ethanol consumption within the past six months, sex, body
 4    mass index, age, and serological evidence of active and past hepatitis infection. Subjects with
 5    serological evidence of active hepatitis infection were included in the  logistic regression analysis
 6    due to the ability of the statistical method to account for the effects associated with this factor.
 7    These analyses showed subjects exposed during older wet or dry-to-dry transfer processes
 8    (average concentration: 19.8 ppm; range: 1.8-83 ppm) was strongly—but imprecisely—
 9    associated with mild or greater sonographic  changes (odds ratio [OR]  = 4.2, 95% confidence
10    interval [CI] = 0.9-20.4) as compared with controls. No association was shown with subacute
11    exposure in new dry-to-dry operations (OR = 0.7, 95%  CI = 0.1-5.9).  An inverse dose-response
12    association was found with cumulative exposure after adjustment for age due to a strong but
13    imprecise association between tetrachloroethylene exposure and hepatic sonographic changes in
14    younger workers (workers less than 35 years of age, OR =15; 95% CI =  1.33-170).
15          Only 21% of the exposed study subjects who had changes graded as mild or greater had
16    increases in any hepatic enzyme (Brodkin et al., 1995).  Mean concentrations of GGT, AST, and
17    alanine transferase (ALT) tended to be higher among the dry cleaners  as compared with laundry
18    workers; however, the differences were not statistically significant and all mean values were
19    within the normal range of reference values. However, all of the  subjects who had elevated ALT
20    concentrations had moderate or severe sonographic changes. Hence, sonographic imaging of the
21    liver appeared to be a more sensitive indicator of toxicity than was measurement of serum
22    hepatic enzymes.
23          Lauwerys et al. (1983) performed behavioral, renal, hepatic, and pulmonary tests on 22
24    subjects  exposed to tetrachloroethylene in six dry  cleaning shops  and compared the results with
25    those obtained for 33 subjects nonoccupationally exposed to organic solvents.  The mean TWA
26    concentration was 21 ppm. The investigators found no statistically  significant differences in
27    mean serum hepatic enzyme concentration between exposed subjects and controls, but they did
28    not describe the statistical  methods used to test  for differences between the exposed and control
29    groups.
30          Cai et al. (1991) investigated subjective symptoms, hematology, serum biochemistry, and
31    other clinical signs in 56 dry cleaners exposed to tetrachloroethylene at 20 ppm (as a geometric
32    mean of 8 hr TWA) and compared the results with findings for 69 nonexposed controls from the
33    same factories. Exposure-related increases were observed in the prevalence of subjective
34    symptoms during the workday as well as in the  past 3-month period, whereas no significant
35    changes  in hematology were seen. There was no effect on liver and kidney function,  as
36    measured by enzyme activities, blood urea nitrogen (BUN), and creatinine in the serum.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-9       DRAFT-DO NOT  CITE OR QUOTE

-------
 2
 O
 4
 5
 6
 7
 8
 9
10
11
       Table 4-1 presents a summary of the human liver toxicity studies in dry cleaners. Two of
the four studies (Brodkin et al., 1995; Gennari et al., 1992) showed clinical signs of liver toxicity,
namely, sonographic changes in the liver and higher serum concentrations of liver enzymes
indicative of liver injury in the absence of frank toxicity. Subjects in these two studies were
exposed to tetrachloroethylene for a longer duration than were subjects in Cai et al. (1991) or
Lauwerys et al. (1983), and for this reason these two studies carry greater weight in this analysis.
Moreover, the studies by Brodkin et al. (1995) and Gennari et al. (1992) assessed potential liver
damage using a different set of markers than those of Cai et al. (1991) or Lauwerys et al. (1983).

       Table 4-1.  Summary of studies of human liver toxicity
Subjects
27 tetrachloroethylene -
exposed dry cleaners
26 nonexposed laundry
workers
141 tetrachloroethylene-
exposed dry cleaners
130 controls
24 tetrachloroethylene -
exposed dry cleaners
33 controls non-
occupationally exposed to
organic solvents
56 tetrachloroethylene -
exposed dry cleaners
69 nonexposed factory
controls
Effects
Sonographic scattering of
fat in liver (in vivo)
Severity greater with
higher cumulative exposure
No liver toxicity
Elevation of total GGT due
to GGT-2
GGT-4 detected in exposed
but not in control workers
No effect on serum hepatic
enzymes
Increased subjective
symptoms
No effects on serum
indicators of liver and
kidney toxicity
Exposure
Group mean TWA =
15.8 ppm
Mean duration of exposure =
12 years
Mean TWA =11.3 ppm
Mean duration of
exposure = 20 years
Mean TWA = 21 ppm
Mean duration of
exposure = 6 years
Geometric mean TWA
= 20 ppm
Mean duration of
exposure = 3 years
Author
Brodkin et
al. (1995)
Gennari et
al. (1992)
Lauwerys
etal.
(1983)
Cai et al.
(1991)
12
13
14
15
16
17
       Biological markers of liver effects permit the early identification of adverse effects of
xenobiotic exposure. They are an important link between biological markers of exposure and
frank liver toxicity, and they offer the most potential for clinical intervention before irreversible
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-10       DRAFT-DO NOT CITE OR QUOTE

-------
 1    effects have occurred (NRC, 1995). The observations of Brodkin et al. (1995) and Gennari et al.
 2    (1992) support the indication that tetrachloroethylene exposure affects liver function; hence, the
 3    lowest-observed-adverse-effect level (LOAEL) for liver effects in humans can be established as
 4    a range from 12 to 16 ppm (TWA).
 5
 6    4.4.1.2. Liver Cancer
 1          Cohort and case-control studies assessing possible association between liver cancer and
 8    dry cleaner and laundry workers, or tetrachloroethylene specifically, are identified in Tables
 9    4B-la, 4B-lb, and 4B-3 (Appendix 4B). An incidence study by Andersen et al. (1999) of dry
10    cleaning and laundry workers in Denmark, Finland, Norway, and Sweden reported the following
11    liver cancer risks:  males, standardized indicence ratio (SIR) =1.3 (95% CI = 0.6-2.3); females,
12    SIR= 1.3 (95% CI = 0.9-1.9); combined, SIR= 1.3 (95% CI = 0.9-1.8).  This study included
13    some of the same subjects as the studies by Lynge and Thygesen, who also reported an elevation
14    in liver cancer incidence among Danish female dry cleaners and laundry workers (Lynge and
15    Thygesen, 1990; Lynge, 1994), and the study by Travier et  al. (2002) of Swedish dry cleaners,
16    launderers, and pressers.
17          Risk for primary liver cancer in these analyses was larger than the risk for the liver and
18    biliary tract cancer, which indicates the potential for bias due to disease misclassification in
19    studies that examine liver cancer as a broad category. A nested case-control  study (Lynge et al.,
20    1995) suggests that the excess primary liver cancer risk among females observed in Lynge
21    (1994) is attributable to laundry workers rather than to dry cleaners (Table 4B-3  in Appendix 4B).
22    This type of information is not available for primary liver cancer cases in Andersen et al. (1999).
23    Mortality studies are biased due to misclassification of liver cancer on death certificates; and
24    these studies do not report statistically significant elevated risks for liver and biliary tract cancer.
25    Primary liver cancer mortality was not elevated, and observations from case-control studies that
26    assessed generic organic solvents or dry cleaning fluid mixtures did not show a consistent liver
27    response (Wartenberg et al., 2000).
28
29    4.4.2.  Animal Studies
30    4.4.2.1. Liver Toxicity
31          Hepatic ffects observed after subchronic or chronic inhalation exposure to
32    tetrachloroethylene include increased liver weight  (Kjellstrand et al., 1984; Kyrklund et al.,
33    1990); hypertrophy (Odum et al., 1988); fatty changes (Kylin et al., 1965; Odum et al.,  1988);
34    peroxisome proliferation, an increase in the size and numbers of peroxisome  organelles (Odum et
35    al., 1988; Goldsworthy and Popp, 1987; Bergamaschi et al., 1992); other histological lesions
36    (Kjellstrand et al., 1984; NTP, 1986a); and necrosis and tumors (NTP, 1986a; JISA, 1993). Liver
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-11       DRAFT-DO NOT CITE OR QUOTE

-------
 1    toxicity observed in animal studies has been reviewed (see U.S. EPA, 1980, 1985a, b, 1986a,
 2    1991a; IARC, 1995; ATSDR, 1997; NYS DOH, 1997; Cal EPA, 2001).
 3          Species differ in their susceptibility to tetrachloroethylene-induced hepatic toxicity. For
 4    example, mice appear to be more sensitve than rats to the adverse liver effects caused by
 5    tetrachloroethylene exposure (U.S. EPA, 1985a; NTP, 1986a; Lash and Parker, 2001). In Rowe
 6    et al. (1952), guinea pigs exposed to 100 to 2,500 ppm proved to be more susceptible than rabbits,
 7    monkeys, and rats to liver toxicity. The lowest reported level for liver effects in laboratory
 8    animals is in tetrachloroethylene-exposed NMRI mice at 9 ppm (61 mg/m3; Kjellstrand et al.,
 9    1984). These investigators exposed male and female mice to 9 ppm and higher concentrations of
10    tetrachloroethylene for 30 days and observed changes indicative of adverse health effects
11    including statistically significant increases in liver weight as well as changes in liver morphology.
12    Increases in levels of blood plasma enzyme butyrylcholinesterase (BuChE) were reported at all
13    tetrachloroethylene concentration levels at or above 9 ppm. A recovery period reversed the
14    effects on BuChE, although liver weight was still slightly elevated at 120 days after cessation of
15    tetrachloroethylene exposure for 30 days at 150 ppm.
16          Chronic lifetime inhalation bioassays of tetrachloroethylene in mice have been conducted
17    by the National Toxicology Program (NTP,  1986a), the Japan Industrial Safety Association
18    (JISA, 1993), and Nagano et al. (1998).  In the NTP study, B6C3F1 mice were exposed to 0, 100,
19    and 200 ppm tetrachloroethylene for 104 weeks. In addition to liver tumors in mice of both
20    sexes, the authors reported liver degeneration in 2/49, 8/49, and 14/50 males and in 1/49, 2/50,
21    and 13/50 females. Liver necrosis was seen in some of the mice (1/49, 6/49, and 15/50 males;
22    3/48, 5/50 and 9/50 females). The authors also observed nuclear inclusions in male mice (2/49,
23    5/49, and 9/50).  No dose-related liver effects were reported in the rats.
24          In the Japan Industrial Safety Association (JISA, 1993) study (some results reported in
25    Nagano et al., 1998), male and female Crj/BDFl mice were exposed to 0, 10, 50, and 250 ppm
26    tetrachloroethylene for 104 weeks and sacrificed at  110 weeks. In addition to hepatocellular
27    carcinomas and adenomas in the mice, telangiectasis (vascular lesions formed by dilation of a
28    group of small blood vessels) and focal necrosis occurred in males at 50 ppm and above. Liver
29    degeneration was observed at 250 ppm in both sexes.  Liver hemangiosarcomas were also
30    reported in the male mice. The authors described effects in F344/DuCrj rats exposed to 0, 50,
31    200, 600 ppm for 104 weeks and sacrificed at 110 weeks. Male, but not female, rats had excess
32    incidence of spongiosis hepatitis at 200 ppm and above and hyperplasia at 600 ppm. Liver
33    tumors were not observed in either male  or female rats.
34          Tetrachloroethylene was found to cause liver toxicity in laboratory animals by the oral
35    route in several studies (e.g., Buben and O'Flaherty, 1985; Story et al., 1986; Hayes et al., 1986).
36    The observed effects included increased liver weights, biochemical changes, histological lesions,

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-12       DRAFT-DO NOT CITE OR QUOTE

-------
 1    necrosis, and polyploidy. Buben and O'Flaherty treated male Swiss-Cox mice with
 2    tetrachloroethylene doses of 0, 20, 100, 200, 500, 1,000, 1,500, or 2,000 mg/kg-day for 5
 3    days/week for 6 weeks. These investigators demonstrated that indices of tetrachloroethylene
 4    hepatotoxicity (increased liver weight, liver triglyceride accumulation, glucose-6-phosphotase
 5    activity, and serum glutamic pyruvic transaminase activity) were highly correlated with the
 6    amount of tetrachloroethylene metabolized by the mice. The degree of liver response, as
 7    measured by each toxicity parameter when plotted against total urinary metabolites, was linear in
 8    all cases.  The several dose-related liver effects were reported at doses above the lowest dose of
 9    20 mg/kg-day.  Increased liver triglycerides and increased liver-to-body weight ratios were seen
10    in mice receiving 100 mg/kg-day and higher doses. At doses of 500 mg/kg-day and higher,
11    effects in the treated mice also included reduction of DNA content, increased serum levels of
12    liver enzymes, liver degeneration, necrosis, and polyploidy.  The LOAEL was 100 mg/kg-day.
13          Ebrahim et al. (1996) administered 3 g/kg/day tetrachloroethylene in sesame oil to mice
14    for 15 days and observed a significant increase in liver weight  and degeneration and necrosis of
15    hepatocytes. These changes occurred simultaneously with a decrease in blood glucose; elevated
16    activities of enzymes hexokinase, aldolase, and phosphoglucoisomerase; and decreased activities
17    of gluconeogenic enzymes.  Table 4-2 presents a summary of liver toxicity studies in animals.
18
19    4.4.2.2. Liver Cancer
20          In carcinogenicity bioassays, tetrachloroethylene has been shown to cause a statistically
21    significant increase in the incidence of hepatocellular carcinomas in both sexes of B6C3F1 mice
22    following either oral gavage administration or inhalation exposure (NCI, 1977; NTP, 1986a).
23    Both sexes of Crj :BDF 1 mice have also been shown to develop an increased incidence of
24    hepatocellular carcinomas when exposed to tetrachloroethylene by inhalation (Nagano et al.,
25    1998; JISA, 1993). The National  Cancer Institute (NCI) and NTP bioassays were reviewed
26    previously by EPA (U.S. EPA, 1985a, 1986a,  1991a) and are briefly summarized here.
27    Observations regarding liver cancer from the more recent study (Nagano et al., 1998; JISA,
28    1993), which confirms the earlier findings of liver tumors in B6C3F1 mice, is also briefly
29    summarized. The tumor incidence data from these studies, with the accompanying tables and
30    figures, are presented in Section 5.3.2.
31          Several metabolites of tetrachloroethylene have been found to be carcinogenic in mice,
32    and it is thought that the hepatocarcinogenicity of the parent compound is mediated through the
33    action of one or more of its metabolites. Metabolites of tetrachloroethylene, including TCA,
34    DCA, and the putative metabolite chloral hydrate, have been observed to cause liver cancer in
35    mice (Daniel et al., 1992; Rijhsinghani et al., 1986; Herren-Freund et al., 1987; Bull et al., 1990;
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
       Table 4-2. Summary of rodent liver toxicity studies
Subjects
NMRI mice (both
sexes), inhalation
Swiss-Cox mice
(male), gavage
B6C3F1 mice
Crj/BDFl mice
(both sexes)
F344/DuCrj rats
(both sexes)
Effects
Increase in liver weight
Morphological changes
Increased plasma butylcholinesterase
Increased liver/body weight ratio at 100
mg/kg-day
Increased triglycerides at 100 mg/kg-day
No change at 20 mg/kg-day
Liver degeneration and necrosis at 100
ppm and higher in males and at 200 in
females
Focal necrosis in males at 50 ppm and
higher
Liver degeneration in males and females
at 250 ppm
Spongiosis hepatitis in males at 200 ppm
and higher
Hyperplasia in males at 600 ppm
Exposure
9 ppm and above for
4 weeks, inhalation
0, 20, 100, 200, 500,
1,000, 1,500, 2,000
mg/kg-day for 6
weeks, gavage
0, 100, 200 ppm for
104 weeks
0, 10, 50, 250 ppm
for 110 weeks
0, 50, 200, 600 ppm
for 110 weeks
Authors
Kjellstrand et
al. (1984)
Buben and
O' Flaherty
(1985)
NTP, (1986a)
JISA(1993)
JISA(1993)
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
Richmond et al., 1995; Pereira, 1996; DeAngelo et al., 1991, 1996, 1999; NTP, 2000a, b).  In
addition, DCA causes liver cancer in rats (DeAngelo et al., 1996; Richmond et al., 1995).
       In the mouse gavage study (NCI, 1977), groups of 50 male mice received TWA doses of
536 or 1,072 mg/kg tetrachloroethylene in corn oil by intragastric gavage for 78 weeks (450 or
900 mg/kg for  11 weeks, then 550 or  1,100 mg/kg for 67 weeks). Groups of 50 female mice
received TWA doses of 386 or 772 mg/kg of tetrachloroethylene in corn oil by gavage (300 or
600 mg/kg for  11 weeks, then 400 or  800 mg/kg for 67 weeks). Mice were dosed 5 days/week.
The tetrachloroethylene used in the study was greater than 99% pure, but impurities were not
identified (NCI, 1977; U.S. EPA, 1985a).  The test sample was estimated to contain
epichlorohydrin concentrations of less than 500 ppm (U.S. EPA, 1985a). It was considered
unlikely, however, that the tumor response resulted from this low concentration of
epichlorohydrin.  Tetrachloroethylene caused statistically significant increases (p < 0.001) in the
incidences of hepatocellular carcinoma in both sexes of mice in both treatment groups when
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  4-14       DRAFT-DO NOT CITE OR QUOTE

-------
 1    compared with untreated controls or vehicle controls.  The time to tumor was significantly
 2    decreased in treated mice.
 3          The inhalation study (NTP, 1986a) confirmed the finding of hepatocellular carcinoma in
 4    B6C3F1 mice.  Groups of 50 mice of each sex were exposed to (epichlorohydrin free)
 5    tetrachloroethylene concentrations of 0, 100, or 200 ppm, 6 hrs/day, 5 days/week, for 103 weeks.
 6    Tetrachloroethylene caused statistically significant dose-related increases in the incidences of
 7    hepatocellular carcinoma and in combined hepatocellular adenoma and carcinoma in both sexes.
 8          More recent tetrachloroethylene inhalation studies conducted in Japan using Crj :BDF1
 9    mice resulted in the observation of hepatocellular carcinomas in both sexes (JISA, 1993 [results
10    reported in Nagano et al., 1998]).  Groups of 50 male and 50 female mice were exposed to 0, 10,
11    50, and 250 ppm tetrachloroethylene, 6 hrs/day, 5 days/week, for 104 weeks, and the terminal
12    sacrifice was performed at 110 weeks. Both males and females showed dose-related increased
13    incidences of liver carcinomas and combined liver adenomas and carcinomas. Malignant liver
14    hemangioendotheliomas were also increased in males.  Both malignant and combined benign and
15    malignant hemangioendotheliomas in the spleen were increased in males.  The investigators also
16    observed Harderian gland adenomas and enlargement of the nucleus in the kidney proximal
17    tubular cells in male mice at the highest dose.
18
19    4.4.3.  Summary of Liver Effects in Humans and Animals
20          Two of four studies of occupationally exposed dry cleaners showed indications of liver
21    toxicity, namely sonographic changes of the liver and altered serum concentrations of liver
22    enzymes indicative of liver injury.  Frank liver disease was not seen among these workers for a
23    number of possible reasons: individuals with frank liver disease may not have been included in
24    cross-sectional studies because they had left the workforce due to their conditions, the healthy
25    worker effect, and other selection biases. LOAELs in these human studies were between  12 and
26    16 ppm (TWA).
27          Primary liver cancer incidence was not consistently elevated across incidence studies and
28    appeared to be  associated with laundry work (Andersen et al., 1999; Travier et al., 2002; Lynge
29    and Thygesen,  1990; Lynge et al., 1995).  Additionally, elevated risks were also seen for
30    incidence in the combined category of primary liver cancer and cancer of the biliary passages.
31    Primary liver cancer is often misclassified on death certificates; hence, mortality studies that
32    examine mortality from liver and biliary passage cancer are less informative than are studies of
33    incidence.  For this reason, greatest weight is placed on the observations of incident studies of
34    primary liver disease. Observations from case-control studies that assess organic solvent
35    generically or dry cleaning fluid mixtures do not show a consistent carcinogenic effect on the
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-15       DRAFT-DO NOT CITE OR QUOTE

-------
 1    liver (Wartenberg et al., 2000). There are no human studies of drinking water or other oral
 2    exposure.
 3          In animals, liver toxicity, manifested by fatty changes, liver enlargement, and enzyme
 4    changes in blood, has been observed in rats and mice in several studies. The LOAEL for the
 5    inhalation studies, 9 ppm, is from a 30-day-exposure mouse study. A chronic mouse inhalation
 6    bioassay showed liver necrotic foci at 50 ppm and higher. In two lifetime inhalation cancer
 7    bioassays, increases in liver cancer occurred at 100 ppm and above, and there was a significant
 8    dose-response trend in both studies. With oral administration, liver effects have been observed at
 9    100 mg/kg-day, although these were not considered to be irreversible effects.  The lowest dose at
10    which liver tumors have appeared is 386 mg/kg-day, administered long term.
11
12    4.4.4.  Mode of Action for Liver Toxicity
13          This section  summarizes scientific data regarding the MOA for tetrachloroethyene-
14    induced hepatic toxicity and carcinogenicity in mice and its relevance to humans. The MOA for
15    tetrachloroethylene-induced mouse liver cancer is not well understood, and it is highly likely that
16    more than one MOA is operative.  The following topics are relevant to the MOA for liver
17    toxicity.
18          (1) Tetrachloroethylene metabolites and liver toxicity.  Metabolic activation of
19    tetrachloroethylene is required to produce adverse effects in the liver. TCA is the major urinary
20    excretion product, and it is also a hepatocarcinogen in  mice; however, insufficient amounts of
21    TCA are produced from tetrachloroethylene metabolism to quantitatively account for the mouse
22    liver tumor incidences observed in cancer bioassays. In addition, the liver tumor phenotypes
23    with regard to H-ras codon 61 mutation do not appear to be similar between TCA, DC A, and
24    tetrachloroethylene.  Therefore, it is likely that other tetrachloroethylene metabolites, such as the
25    potentially reactive trichloroacetyl chloride, are contributing to the production of liver tumors.
26    The potential role of GST conjugates of tetrachloroethylene in liver toxicity, although unknown,
27    is presumed to be less than that in the kidney.
28          (2) Role of receptor activation. Data exist to advance the hypothesis that peroxisome
29    proliferators can contribute to liver tumorigenesis in rodents; however, the causal role of PPAR-
30    mediated events in tumorigenesis, and human sensitivity to these effects needs further scientific
31    examination and analysis. Data suggest that tetrachloroethylene is a very weak peroxisome
32    proliferator. The strongest evidence supporting the PPAR MOA for tetrachloroethylene is the
33    data for TCA; however, TCA also has other MO As, TCA alone does not account quantitatively
34    for tetrachloroethylene induced tumors, and tetrachloroethylene- and TCA-induced tumors are
35    phenotypically distinct (Bull et al., 2002).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-16       DRAFT-DO NOT CITE OR QUOTE

-------
 1          (3) Genotoxic effects.  Tetrachoroethylene has been shown to induce some genotoxic
 2    effects (micronuclei and SCEs following in vitro exposure, DNA binding and SSBs in liver).
 3    Results of in vitro mutagenicity (Ames) or DNA binding assays of tetrachloroethylene have
 4    largely been negative except in the few studies of conditions where metabolites of the GSH
 5    pathway are generated. The GSH metabolites are clearly mutagenic. In addition, several known
 6    (DCA) and putative (tetrachloroethylene oxide) P450 metabolites exhibit mutagenicity.  The
 7    mutagenic potential of reactive metabolites of tetrachloroethylene has not been adequately
 8    studied.  Moreover, the identity of all metabolites is not known.
 9          (4) Nongenotoxic effects. Existing data suggest the involvement of events related to
10    tumor induction that are nonspecific to activation of PPAR-a, being common to other
11    nongenotoxic MO As. Hypomethylation is a common early molecular event in most tumors, and
12    alterations in DNA methylation following exposure to chemicals, both hypomethylation and
13    hypermethylation, may be factors in tetrachloroethylene-induced  tumorigenesis.  Although
14    tetrachloroethylene-specific data are lacking, its metabolites DCA and TCA are known to induce
15    hypomethylation  of DNA and protooncogenes in mouse liver.
16
17    4.4.4.1. Background
18          Although  hepatocellular tumors are common endpoints in mouse carcinogenicity studies,
19    their biological significance with respect to identifying human hazard has long been a subject of
20    intense controversy and debate (Tomatis et al., 1989; Ward et al., 1979; Nutrition Foundation,
21    1983; U.S. EPA,  1985c; U.S. EPA, 1986b, 1991a; Popp, 1984; Stevenson et al., 1990).  The
22    current controversy in the case of tetrachloroethylene-induced hepatocellular carcinoma in  mice
23    involves identifying the operative MO As and their relevance to human situations.
24          Hemangiosarcomas, unlike hepatocellular carcinomas, are not a common finding in
25    mouse bioassays  (U.S. EPA, 2002, 2000b); in fact, they are considered relatively rare, and their
26    relevance to human health hazard, therefore, is generally accepted. Findings of a positive trend
27    for liver and spleen hemangiosarcomas in the most recent mouse  carcinogenicity bioassay of
28    tetrachloroethylene (JISA, 1993; Nagano et al.,  1998) in a strain of mice not known to have any
29    type of high background tumor incidence constitute important information about risk of exposure
30    to humans.
31          The focus of the human relevance of the hepatocellular carcinomas observed in mice has
32    turned to the emerging information on modes of carcinogenic action. Peroxisome proliferation,
33    which is associated with certain rodent liver carcinogens, has gained increasing attention due to
34    its possible relationship to a hypothesized MOA.  Studies of tetrachloroethylene and its
35    chloroacid metabolites suggest that the compound is a peroxisome proliferator chemical, albeit a
36    very weak one.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-17       DRAFT-DO NOT CITE OR QUOTE

-------
 1          A lack of human relevance for the MOA associated with peroxisome proliferator
 2    carcinogens has been proposed (Klaunig et al., 2003; Meek et al., 2003).  However, agreement is
 3    lacking on the extent to which the MOA hypothesis has been validated or whether the MOA or
 4    quantitative differences among species are sufficiently understood to rule out a potential risk of
 5    carcinogenicity to humans (Melnick, 2001; U.S. EPA, 2005a). EPA's Federal Insecticide,
 6    Fungicide, and Rodenticide Act Scientific Advisory Panel (SAP) reviewed a draft EPA proposed
 7    science policy report (U.S. EPA, 2003d) and considered the role of PPAR-a agonists in
 8    activation of PPAR-a leading to an increase in cell proliferation, a decrease in apoptosis, and
 9    eventual clonal expansion of preneoplastic cells leading to liver cancer, and the human relevance
10    of PPAR-a agonist-induced hepatocarcinogenesis. While the majority  of the panel concluded
11    that there was sufficient evidence in support of the proposed MOA for PPAR-a agonist-induced
12    rodent hepatocarcinogenesis, some panel members complete disagreed. The majority of the
13    panel agreed that there are relevant data indicating that humans are less sensitive than rodents to
14    the hepatic effects of PPAR-a agonists. However, it was  noted that humans are not refractory to
15    the effects of PPAR agonism and many questions remain  regarding the specific events PPAR
16    activation entails.
17          This assessment attempts to evaluate scientific information and review the current issues
18    on MOA hypotheses pertinent to tetrachloroethylene.
19
20    4.4.4.2.  Relationship of Metabolism to Potential Mode of Action and Organ Toxicity
21          Metabolic activation of tetrachloroethylene is required for adverse effects to occur in the
22    liver. The cancer-causing activity of tetrachloroethylene and other chlorinated ethylenes is
23    generally considered to reside in metabolites rather than in the parent compounds (U.S. EPA,
24    1991a; Davidson and Bellies, 1991; Lash et al., 2000a; Lash and Parker, 2001; see Section 3.3).
25    Certain metabolites of tetrachloroethylene—specifically, TCA, DC A, and chloral hydrate—have
26    been shown to cause liver tumors in mice (IARC, 1995; Daniel et al., 1992; Rijhsinghani et al.,
27    1986; Herren-Freund et al., 1987; Bull et al., 1990; Pereira,  1996; Odum  et al., 1988; DeAngelo
28    et al., 1991, 1999; DeAngelo, 2000; NTP, 2000a,  b; Carter et al.,  2003), and they may be
29    involved in hepatocarcinogenicity following exposure to the parent compound.
30          TCA is the metabolite that has received the most attention as being potentially associated
31    with tetrachloroethylene-induced liver tumorigenesis. Although it may play a role in
32    tetrachloroethylene hepatocarcinogenicity, not enough TCA is produced from metabolism to
33    account for all tetrachloroethylene-induced mouse liver tumors (see Appendix 4A) observed in
34    bioassays. Although TCA can be further metabolized to DCA, most DCA originating from
35    tetrachloroethylene is proposed by some investigators to be derived predominantly from the renal
36    beta lyase-mediated cleavage of the TCVC conjugate, where DCA production ultimately results

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-18       DRAFT-DO NOT CITE OR QUOTE

-------
 1    from dichlorothioketene in the kidney (Volkel et al., 1998). If this is the case, most
 2    tetrachloroethylene-derived DCA would not occur in the liver target organ, implying that DCA
 3    would not likely be critically involved in tetrachloroethylene-induced liver tumorigenesis.
 4          The potential significance of precursor metabolites—e.g., trichloroacetyl chloride, a
 5    major and potentially reactive P450 intermediate—should not be underestimated.  The possible
 6    roles of key reactive precursor intermediates in causing hepatotoxicity need to be better
 7    elucidated, as they may be important to understanding MOA for tetrachloroethylene.
 8
 9    4.4.4.3. Description of a Hypothesized Mode of Action (MOA): Peroxisome Proliferator-
10           Activated Receptor (PPAR) Mediated Hepatocarcinogenesis
11    4.4.4.3.1. Background summary. Some peroxisome proliferators cause increased incidence of
12    rodent liver tumors.  Several investigators (Reddy et al., 1980; Reddy and Lalwai, 1983; Moody
13    et al., 1991; Ashby et al., 1994) have hypothesized a causal relationship between proliferation of
14    peroxisomes and hepatocellular carcinogenicity because peroxisome proliferation in rodent
15    hepatocytes frequently occurs alongside hepatocyte hypertrophy and liver hyperplasia and
16    disproportionate transcriptional increases of peroxisomal enzymes involved in B-oxidation of
17    fatty acids (reviewed by Cattley et al., 1998).
18          Insight into the possible MOA by which chemicals induce peroxisome proliferation—and
19    possibly cancer—was revealed by the discovery of the PPAR receptor, which was shown to be
20    activated by peroxisome proliferators (Issemann and Green, 1990). Activation of this receptor
21    regulates transcription of the genes that encode the enzymes responsible for biochemical changes,
22    including peroxisomal enzymes responsible for beta-oxidation, liver fatty acid-binding protein
23    (Issemann et al., 1993), certain microsomal P450 (CYP4A, CYP2B, and CYP2C) family
24    enzymes (Heuvel, 1999; Gorton et al., 1998; Fan et al., 2003; Simpson et al., 1995, 1996) and
25    other enzymes (Barbier et al., 2003).
26          The evidence regarding whether peroxisome proliferation induced by tetrachloroethylene
27    or its metabolite TCA is the primary or sole mode of action for carcinogenesis is equivocal at
28    best (Ashby et al., 1994; IARC, 1995a; Goldsworthy and Popp, 1987; Odum et al., 1988;
29    Elcombe, 1985; Elcombe et al., 1985; Goldsworthy and Popp, 1987; DeAngelo et al., 1989;
30    Laughter et al., 2004).
31
32    4.4.4.3.2. Summary description  of postulated mode of action (MOA)—peroxisome
33   proliferation via modification of cell signal pathways through the peroxisome proliferator-
34    activated receptor (PPAR) receptor. A recent, in-depth review by Klaunig et al. (2003)
35    summarized the PPAR MOA and supporting data; see also the OPP draft science policy paper
36    and the SAP review (U.S. EPA, 2003d). Klaunig et al. (2003) proposed three events to be

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-19       DRAFT-DO NOT CITE OR QUOTE

-------
 1    causally related to tumorigenesis: activation of PPAR-a, perturbation of cell proliferation and
 2    apoptosis, and selective clonal expansion. The causal role is largely based on evidence that the
 3    induction of these events is attenuated in PPAR-a -null mice (or in hepatocytes isolated from
 4    such mice) in response to the prototypical agonist WY 14,643 (Lee et al., 1995; Peters et al.,
 5    1997).  A number of intermediary events are considered associative including: expression of
 6    peroxisomal and nonperoxisome genes, peroxisome proliferation, inhibition of gap junction
 7    intracellular communication, hepatocyte oxidative stress, as well as Kupffer cell-mediated events.
 8           Historically, the increase in peroxisomal organelles, peroxisomal fatty acid beta-
 9    oxidation, and alteration in ratios and production of marker enzymes such as increased  acyl-CoA
10    oxidase, observed in rodents treated with the peroxisome proliferator chemicals, was thought to
11    induce  oxidative stress in hepatocytes and potentially result in oxidative damage to proteins and
12    DNA, leading to carcinogenesis. Two key factors—oxidative injury and enhanced cell
13    proliferation—were implicated in the rodent hepatocarcinogenicity of peroxisome proliferating
14    agents (Cattley et al., 1998; Klaunig et al., 2003).  PPAR-a has been identified as the specific
15    PPAR receptor associated with cell proliferation and hepatocarcinogenesis in mouse liver (Lee et
16    al., 1995; Peters et al., 1997a; Corton et al., 2000).  PPAR-a activation has been shown to trigger
17    multiple events, and events other than the proliferation of peroxisomes—or at least
18    manifestations not limited to this phenomenon only—are thought relevant to tumorigenesis.
19           Peroxisome proliferators have been shown to alter hepatocyte growth and survival by
20    induction of DNA synthesis and suppression followed by enhancement and then depression of
21    apoptosis (cell death; Cattley and Popp, 1989; Roberts et al., 1995; Bursch et al., 1984; Marsman
22    et al., 1992).  Cell proliferation is thought to play an important role through specifically
23    enhanced proliferation of normal hepatocytes, resulting in an increase in the frequency  of
24    initiated cells or in the selective growth of pretransformed hepatocytes, with subsequent
25    tumorigenesis (Cattley and Popp, 1989; Kraupp-Grasl et al., 1990, 1991; Grasl-Kraupp et al.,
26    1993; Cattley et al., 1991; Marsman et al., 1988; Eacho et al.,  1991; Marsman, 1991; Marsman
27    and Popp, 1994).  The currently hypothesized MOA for liver carcinogenesis assumes that events
28    such as the increased cell proliferation, inhibition of apoptosis, and clonal expansion of
29    preneoplastic lesions  are linked directly to PPAR-a activation.
30
31    4.4.4.3.3. Identification of potential key events in mode of action (MOA) for liver. Certain
32    biochemical and cellular events have been associated with hepatocarcinogenic effects of
33    peroxisome proliferating chemicals.  Whether these key events are causally related to liver
34    tumorigenesis remains to be determined.  Potential key events include (a) peroxisome
35    proliferation—an increase in the number of peroxisomes and also an increase in their volume
36    density (Meijer and Afzelius, 1989; Ganning et al., 1983; Thangada et al., 1989); (b) certain

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-20      DRAFT-DO NOT CITE OR QUOTE

-------
 1    disproportionate alterations in levels of peroxisomal enzymes, especially increases in fatty acyl-
 2    CoA oxidase levels of 20- to 30-fold, whereas catalase and urate oxidase are increased only 2- to
 3    3-fold, with resulting excess production of hydrogen peroxide, which may affect hepatocytes by
 4    oxidative injury; (c) increases in members of the cytochrome P450 CYP4A subfamily; (d)
 5    increases in rates of cell proliferation due to increased DNA synthesis and suppression of
 6    apoptosis, particularly in basophilic preneoplastic lesions; (e) hepatomegaly; and (f) expression
 7    and activation of the a subtype of the PPAR (PPAR-a; Hardwick et al., 1987; Dreyer et al, 1992;
 8    Marcus et al., 1993; Tugwood et al., 1992; Zhang et al., 1992; Cattley et al., 1998; Chevalier and
 9    Roberts, 1998).
10
11    4.4.4.3.3.1. Peroxisome proliferator-activated receptor alpha (PPAR-a) activation. The
12    hypothesized causal event most associated—and best supported by existing data—with the
13    proposed peroxisome proliferation MOA is the one having true specificity for PPAR-a MOA.
14    This is the activation of the PPAR-a receptor. PPAR-a expression and activation, inducing
15    transcription of selected genes, possibly in concert with altered cell  signal transduction initiated
16    by release of cytokines by hepatic macrophages (Kupffer cells), is probably the important
17    process to evaluate as a potential key event to the development of liver hyperplasia and
18    hepatocyte carcinogenesis. The strongest support for a causal relationship between PPAR-a
19    activation and hepatocellular tumorigenesis is found in studies in null mice, i.e., mice lacking
20    PPAR-a, particularly the eleven-month study of WY-14643 (Peters et al., 1997a, b).  Such
21    "knockout" mice do not respond to this prototype peroxisome proliferator with increased cell
22    proliferation and decreased apoptosis or with development of other events potentially associated
23    with PPAR-a activation leading to liver cancer. Although a short-term study in null mice has
24    been performed for TCA (Laughter et al., 2004), such a study has not been conducted for
25    tetrachloroethylene. This short-term study of TCA provides data consistent with a relationship
26    between PPAR-a activation and peroxisome proliferation, but provides minimal, if any, support
27    for PPAR-a activation and liver cancer.  Although null mouse studies have flaws, some due to
28    deficiencies in the altered mice such as physiological and biological differences in response to
29    stress when compared to wild mice, a well-designed, lifetime tetrachloroethylene carcinogenicity
30    study in null mice could provide valuable information.
31
32    4.4.4.3.3.2. Alterations in cell replication and death rate.  Other events could be claimed to be
33    causal for rodent liver tumor induction. One such event is selective clonal expansion. Stauber et
34    al. (1998) and Bull et al.  (2004) reported data indicating that TCA acts to induce liver tumors by
35    increasing clonal expansion of a specific group of initiated cells in mouse liver. TCA stimulates
36    growth of colonies of hepatocytes expressing c-Jun phenotype, which is representative of tumors

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-21        DRAFT-DO NOT CITE OR QUOTE

-------
 1    caused by TC A. TCA also lowers replication rates of normal liver cells. TCA treatment,
 2    therefore, results in negative selection, or growth advantage being given to specific characteristic
 3    tumor cells over normal cells.  Although selective clonal expansion has been associated with
 4    PPAR-a MO A, it is not clear that this is the case with the tetrachloroethylene metabolite. In fact,
 5    these investigators recently stated that they believe TCA causes cancer independently of
 6    peroxisome proliferation (Bull, 2004). The DCA metabolite also selectively stimulates the
 7    growth of clones of cells.  Clonal expansion is thought to occur with all cancer-causing agents,
 8    so it is not limited to PPAR-a MOA. The available data indicate that the MO As for the two
 9    chloroacid metabolites are different.
10          Another possible causal key event is cell proliferation and apoptosis, although it is not
11    unique to peroxisome proliferator chemicals.
12
13    4.4.4.3.3.3.  Potential key events specific for peroxisome proliferator-activated receptor alpha
14    (PPAR-a).  Several other events have potential for being key events in the PPAR-a MOA for
15    liver tumors observed in rodents exposed to peroxisome proliferator chemicals. Two events
16    considered specific to PPAR-a activation are the actual proliferation of peroxisome organelles
17    and the expression of peroxisomal genes. These events  can be considered biomarkers for
18    peroxisome proliferator chemicals, but a cause-and-effect relationship to liver tumor induction
19    cannot be made. These events are historically linked to  the PPAR-a MOA.
20
21    4.4.4.3.3.4.  Other potential key events not limited to prottferator-activated receptor alpha
22    (PPAR-a) mode of action (MOA). Yet another key event to consider is an alteration in the
23    expression and activities of nonperoxisomal lipid-metabolizing enzymes that mediate
24    hypolipidemia. PPAR-a agonists shown to cause liver tumors in rodents also induce genes that
25    encode lipid metabolizing enzymes, although these same genes can be altered by other agents.
26          Alterations in DNA methylation, both hypomethylation and hypermethylation, may be
27    factors in tumorigenesis and occur following exposure to chemicals that also cause peroxisome
28    proliferation (Pereira et al., 2004a, b). Hypomethylation is a common early molecular event in
29    most tumors (Pereira et al., 2004b).  DCA and TCA are  known to induce hypomethylation of
30    DNA and protooncogenes in mouse liver (Tao et al., 1998, 2000; Pereira et al., 2004a).
31          Inhibition of gap junction cellular communication has been attributed to certain
32    peroxisome proliferator chemicals (Klaunig et al., 1988; Dybing et al., 1995).  This event is thus
33    correlated with the rodent tumorigenesis caused by such chemicals, although similar inhibition of
34    the gap junction cellular communication process also occurs with other nongenotoxic liver
35    carcinogens (Klaunig et al., 2003) and, therefore, cannot be considered specific to peroxisome
36    proliferators and PPAR-a MOA.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-22       DRAFT-DO NOT CITE OR QUOTE

-------
 1          Oxidative stress and resulting DNA damage to hepatocytes secondary to that stress have
 2    been attributed to peroxisome proliferator chemicals causing liver cancer in rodents, but the
 3    exact role is not clear and remains a controversial issue. As discussed above, alterations in the
 4    ratios of peroxisomal enzymes by induction of beta-oxidation enzymes results in an imbalance
 5    that causes overall increases in hydrogen peroxide leading to oxidative damage.
 6           The nonparenchymal Kupffer cell macrophages may be involved in peroxisome
 7    proliferator chemical tumor induction (Klaunig et al., 2003; Rusyn et al., 2000a, 2001). Kupffer
 8    cells do not express PPAR-a (Peters et al.,  2000).  Peroxisome proliferator chemicals activate
 9    Kupffer cells directly (Rose et al., 1999; Peters et al., 2000) and independently of PPAR-a
10    activation (Peters et al., 2000). Klaunig et al. (2003) suggest that Kupffer cell mediated events
11    are associated with (i.e., are not causally related to) hepatic tumors induced by the PPAR-a
12    MO A; it is noted that responses of these cells are independent of PPAR-a and are not restricted
13    only to peroxisome proliferator chemicals.
14
15    4.4.4.3.4. Correlation between proliferator-activated receptor alpha (PPAR-a)
16    activation/peroxisomeproliferation and tumor induction.  Historically, chemicals have been
17    characterized as peroxisome proliferators on the basis of either observations of increases in
18    volume density of peroxisomes or increases in peroxisomal fatty acid beta-oxidation enzyme
19    activity, with characterization by both of these parameters being preferable. Demonstration of
20    induction of the cyanide-insensitive palmitoyl CoA enzyme is viewed as a key biochemical
21    marker acceptable for the detection and quantitation of peroxisome proliferation. Usually,
22    palmitoyl CoA oxidation (PCO) is measured, although palmitoyl CoA oxidase activity can be
23    determined directly where hydrogen peroxide production is measured.
24          The potential key events described  above in Section 4.4.4.3.3 have been correlated with
25    tumorigenesis, although some of these events are not restricted to PPAR-a MO A, and a cause-
26    and-effect relationship is questionable for others.  Certain key events are associated with
27    PPAR-a activation and can also be associated with tumorigenesis; however, evidence  supporting
28    the link between the receptor activation and tumorigenesis through these key events lacks
29    compelling persuasiveness.
30          The strongest case currently available for a cause-and-effect link is the results of cancer
31    studies using PPAR-a null mice (see Peters et al., 1997, and Ito et al, 2007). When exposed to
32    the peroxisome proliferator WY-14,643, the null mice do not show increased cell proliferation or
33    decreased apoptosis or evidence of developing hepatocellular carcinogenesis (Peters et al., 1997).
34    The occurrence of events known to be associated with tumorigenesis following exposure to
35    peroxisome proliferator chemicals allows an association to be made between PPAR-a activation
36    and these other events. The events clearly  specific to PPAR-a receptor activation—actual

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-23       DRAFT-DO NOT CITE OR QUOTE

-------
 1    peroxisome proliferation and expression of peroxisomal genes and enzymes—are likely markers
 2    for the receptor activation and not cause-and-effect events for carcinogenesis.  The other events
 3    may be related to tumorigenesis but are not restricted to PPAR-a activation MOA. The TCA
 4    metabolite of tetrachloroethylene has been studied in null mice,  and although no tumors were
 5    observed, this was only a short-term study and findings are only minimally supportive of
 6    PPAR-a being related to liver tumorigenesis. Section 4.4.4.3.5  describes the study in greater
 7    detail. Tetrachloroethylene has not been studied in such mice. Although some investigators
 8    attributed tetrachloroethylene-induced hepatocarcinogenesis to TCA, analysis of the amount of
 9    tetrachloroethylene metabolism at the doses administered in the mouse carcinogenicity bioassays
10    demonstrates that not enough TCA is produced to account for the tumor response (see Appendix
11    4A).
12          The evidence for peroxisome proliferation by tetrachloroethylene and the chloroacid
13    metabolites is published in Zhou and Waxman (1998), Zanelli et al. (1996), Bruschi and Bull
14    (1993), Nelson et al. (1989), Elcombe (1985), Elcombe et al. (1985), Goldsworthy and Popp
15    (1987), Odum et al. (1988), Channel et al. (1998), DeAngelo et  al. (1989), and Daniel  et al.
16    (1993).
17          Studies by Goldsworthy and Popp (1987) indicate that tetrachloroethylene and its
18    metabolite TCA elevate cyanide-insensitive PCO activity in mouse liver, yet only TCA caused
19    increased PCO activity in rat liver. The elevation in PCO  activity in mouse liver caused by
20    tetrachloroethylene was not great.  In a study conducted by Zanelli et al. (1996), TCA  was shown
21    to increase PCO activity  in the liver of treated rats of a different strain, so the metabolite does
22    cause the effect, and it may be responsible for the response of the parent compound in mice.
23    Different rat strains were used in the two studies (F344 and Wistar, respectively),  and  the
24    increase reported by Goldsworthy and Popp was a relatively weak response.
25    Tetrachloroethylene increased both PCO enzyme activity and peroxisome volume density in
26    exposed mice in the study by Goldsworthy and Popp.  In the tetrachloroethylene study conducted
27    by Odum et al. (1988), peroxisome proliferation was increased in the livers of mice but not rats.
28    Elcombe (1985) found TCA to cause  peroxisome proliferation—as measured by an increase in
29    peroxisomal enzyme activity—in hepatocytes of both rats and mice,  both in vivo and in vitro,
30    after  short-term exposure. Interestingly, Elcombe reported that  the Wistar rat showed  a greater
31    peroxisome proliferation response than did mice, as measured by increases in cyanide-insensitive
32    acyl-CoA oxidase activity induction.  Clearly, strain and species differences exist.
33          DeAngelo et al. (1989) demonstrated peroxisome proliferation induction by TCA and by
34    DCA exposures in mice and rat livers, as indicated by increased PCO activity and peroxisomal
35    volume and possibly the  observed increased carnitine acetyl transferase activity as well. The
36    investigators examined peroxisome proliferation activity in three strains of rats (Sprague-Dawley,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-24       DRAFT-DO NOT CITE OR QUOTE

-------
 1    F344, and Osborne-Mendel) and in four strains of mice (Swiss-Webster, C57BL/6, C3H, and
 2    B6C3F1).  The conclusion from the DeAngelo et al. (1989) study is that mice are more sensitive
 3    than rats with respect to the enhancement of liver peroxisome proliferation by TCA.  More recent
 4    studies conducted by Waxman and colleagues (Maloney and Waxman, 1999; Zhou and Waxman,
 5    1998) showed induction of peroxisome proliferation in rodents by the tetrachloroethylene
 6    metabolites TCA and DCA.  DCA and TCA activated the PPAR-a receptor in both mouse and
 7    human cells in these studies. Walgren and colleagues (Walgren et al., 2000a, b, 2004) reported
 8    expression of PPAR-a in human hepatocytes, activation by TCA and DCA, and peroxisome
 9    proliferation by a series of acetates. High concentrations of TCA were used in the studies.
10
11    4.4.4.3.5. Strength, consistency, specificity of association of the hepatocellular tumor response
12    with key events. Whether or not any cause-and-effect relationship exists between peroxisome
13    proliferation per se  and cancer-causing activity leading to rodent liver cancer is not clear
14    (Capone, 1994; Cattley et al., 1998; Cattley and Roberts, 2000; Youssef and Badr, 1999). The
15    current majority opinion regards activation of PPAR-a as the important causal key event—the
16    obligatory  step—for the MOA of rodent liver carcinogenesis.  Even so, not all scientists agree,
17    and not all data support this hypothesis (see Section 4.4.4.1).
18          The strongest support for a causal relationship between the PPAR-a activation MOA and
19    hepatocellular tumorigenesis using the compound WY-14643 is from studies in null mice by
20    Peters et al. (1997a, b). The existing data show null mice to be refractive to other possible key
21    events—suppression of apoptosis and cell proliferation—and also refractive to tumor formation
22    following 11 months of exposure to this prototype peroxisome proliferator. The null mouse,
23    when challenged by 11 months of exposure to WY-14643, did not respond to a dose that causes
24    100% tumor response in wild-type mice.
25          Results from null  mouse studies would be convincing data in support of the PPAR-a
26    MOA hypothesis, except for serious shortcomings. For example, the studies are less-than-
27    lifetime studies, or they are in vitro studies conducted in tissues from animals exposed to test
28    compounds.  Such studies clearly cannot be considered equal to the standard rodent lifetime
29    bioassays conducted to detect carcinogenic activity, and because of this deficiency, they are not
30    considered adequate for assessing lifetime cancer risk. Also, they cannot be used to demonstrate
31    conclusively that PPAR-a activation is an obligatory  step in rodent hepatocellular tumorigenesis
32    simply because some of the key events that could be associated with tumorigenesis are not
33    observed.  The complexity of the multitude of effects and the lack of understanding about which
34    of the myriad downstream events result at particular dose levels—mechanisms and steps linking
35    those events to PPAR-a activation—also render the null mouse data somewhat less than
36    adequate for understanding the PPAR-a activation relationship to tumorigenesis. Additionally,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-25       DRAFT-DO NOT CITE OR QUOTE

-------
 1    the null mice are innately dissimilar from wild mice—for example, they respond to stress
 2    differently (Watanabe et al., 2000; Huss and Kelly, 2005). Their differences include both
 3    physiological and biochemical aspects.  Some are likely related to PPAR-a-dependent changes in
 4    gene expression, but others are not PPAR-a dependent (Valles et al., 2003; Jalouli et al., 2003;
 5    Hasmall et al., 2002; Meyer et al, 2003). Because of such inherent differences between null mice
 6    and wild mice prior to exposure to test chemicals, the data from studies using these mice should
 7    be interpreted with caution.
 8          The most relevant data pertinent to tetrachloroethylene comes from studying TCA.
 9    Laughter et al. (2004) attempted to determine whether effects of TCA in the liver associated with
10    carcinogenesis were mediated by PPAR-a.  Male wild-type and PPAR-a-null mice were given
11    TCA at 0.25, 0.5, 1, or 2g/L in the drinking water for 7 days.  TCA increased liver-to-body
12    weight ratios, but the increases were not significant. The livers from wild-type but not PPAR-a-
13    null mice exposed to 2 g/L TCA exhibited centrilobular hepatocyte hypertrophy. Further, global
14    gene expression was assessed using the mouse Atlas cancer 1.2 array.  The induction of CYP4a
15    and acyl-CoA oxidase was examined in the livers of mice after exposure to TCA. In wild-type
16    mice, but not PPAR-a-null mice, CYP4a was induced at Ig TCA/L and above, Aacyl-CoA was
17    induced by TCA at 2g/L, and palmitoyl-CoA oxidase activity was induced at 2 g/L.  These data
18    suggest that peroxisome proliferation induced by such compounds could be potentially mediated
19    by PPAR-a (Nakajima et al., 2000). They do not indicate a cause-and-effect relationship
20    between PPAR-a and liver tumorigenesis, however. These studies of TCA were not designed to
21    examine tumor development or show any evidence for a cause-and-effect relationship  between
22    receptor activation and tumor development. No comparable study exists for tetrachloroethylene.
23          If peroxisome proliferation is causally related to the induction of liver cancer, then a
24    detectable quantitative relationship between the two events could be expected. That is, potent
25    peroxisome proliferators should also be potent hepatocarcinogens.  However, this does not
26    appear to be the case (Elcombe and Mitchell, 1986; Marsman et al., 1988, 1992;  Eacho et al.,
27    1991; U.S. EPA, 1991a). A comparison of the tetrachloroethylene chloroacid metabolites
28    indicates that DCA is a more potent hepatocarcinogen than TCA.  For example, in a chronic
29    65-week study of DCA and TCA in male B6 mice, Herren-Freund et al. (1987) found that equal
30    concentrations in drinking water resulted in a nearly threefold higher incidence of liver cancer in
31    DCA-dosed animals than in TCA-dosed animals. DeAngelo et al. (1989), however, reported that
32    TCA was more potent than DCA as a peroxisome proliferator in male B6 mice. Nelson et al.
33    (1989) also reported that TCA produced greater peroxisome proliferation than did DCA in B6
34    mice dosed for only 10 days.  Additionally, evaluation of TCA and DCA indicates that these two
35    metabolites act through different MO As because they exhibit clearly unparallel dose-response
36    curves.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                 4-26      DRAFT-DO NOT CITE OR QUOTE

-------
 1          Compared with oxidative damage, hepatomegaly caused by cell proliferation appears to
 2    be better correlated with hepatocarcinogenesis in rodents (Marsman et al., 1988, 1992; Barrass et
 3    al., 1993; Cattley et al., 1998). A reversible increase is seen in cell death along with the
 4    increased cell proliferation (Bursch et al., 1984; Roberts et al., 1995; Marsman et al., 1992;
 5    Cattley et al., 1998). Both TCA and DCA can produce hepatomegaly at carcinogenic doses,
 6    although, at least in the case of DCA, the increase is more likely due to cytomegaly (i.e., an
 7    increase in cell size), whereas in the case of TCA it is more likely due to an increase in the
 8    number of liver cells (Bull, 2000).
 9          Tetrachloroethylene and its major oxidative metabolite, TCA, cause liver tumors in mice,
10    yet do not induce liver tumors in rats (NCI, 1977; NTP, 1986a; DeAngelo et al., 1997). TCA has
11    been demonstrated, however, to produce hepatic peroxisome proliferation in rats as well as in
12    mice (Elcombe, 1985; Zanelli et al., 1996; DeAngelo et al., 1989).
13          Tetrachloroethylene is not as potent a peroxisome proliferator as are its metabolites. This
14    information is meaningful because some investigators have postulated  that although
15    tetrachloroethylene may possess an intrinsic ability to induce peroxisomes, it may be less
16    effective as a peroxisome proliferator and carcinogen in the rat due to a metabolic inability of
17    that species to form sufficient amounts of peroxisome proliferator metabolites, whereas TCA is
18    formed in mice in sufficient amounts from the parent compound bioassay  doses to result in a
19    sustained level of peroxisome proliferation.  Carcinogenicity bioassay  studies in rats disprove
20    that theory because TCA does not cause liver tumors in rats.
21          The hepatocellular cancer-causing activity of tetrachloroethylene has not been heavily
22    associated with DCA. DCA is thought not to be produced in the liver in sufficient quantity
23    because its only source from tetrachloroethylene oxidative metabolism is the  further
24    biotransformation of TCA. TCA, on the other hand, is generally considered to contribute to the
25    mode of carcinogenic action for tetrachloroethylene. Because DCA may be somewhat rapidly
26    further metabolized to other compounds, such a conclusion may be in error.
27          It is important to note that the peroxisome proliferation effects  observed in rodents
28    exposed to tetrachloroethylene are equivocal.  In a summary plot of tumor incidence versus
29    peroxisome proliferation from the two studies reporting tetrachloroethylene data, the effects are
30    significant in male mice but not in females (NCI, 1977; NTP, 1986a; Ashby et al., 1994;
31    Goldsworthy and Popp, 1987; Odum et al., 1988).  Compared with the effects of potent
32    peroxisome proliferator chemicals, the effects caused by tetrachloroethylene are relatively weak
33    (Ashby etal., 1994).
34          PPAR-a isolated from mouse liver can be activated by certain tetrachloroethylene
35    chloroacid metabolites.  Both the TCA and the DCA metabolites have  been shown to activate
36    PPAR-a (Maloney and Waxman, 1999).

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-27       DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.4.4.3.6. Dose-response relationship. The data for a clear dose-response for increased
 2    peroxisome proliferation in mouse liver resulting from tetrachloroethylene exposure are
 3    equivocal, especially in female mice (Section 4.4.4.3.5).  Some evidence exists for increased
 4    tumor response with increased tetrachloroethylene dose in both gavage and inhalation
 5    carcinogenicity bioassays, although the dose/concentration levels in those studies were relatively
 6    high.  Approaching saturation of metabolism blurs the dose-response at these levels, most
 7    noticeably in the oral gavage study.  There is also some evidence, from studies conducted
 8    separately from the cancer bioassays, to support increase in tetrachloroethylene peroxisome
 9    proliferation with increase in dose, although these data are not particularly convincing (Ashby et
10    al., 1994).
11          Likewise, positive dose-responses for cancer-causing activity in lifetime carcinogenicity
12    bioassays and for peroxisome proliferation in short-term studies have been observed in mice
13    following exposure to relatively high doses of TCA (Ashby et al., 1994; Daniel et al., 1993;
14    Goldsworthy and Popp, 1987; Elcombe, 1985; Herren-Freund et al., 1987).
15          Other potentially  adverse effects associated with DCA exposure (e.g., changes in
16    carbohydrate metabolism, as well as other alterations in cell signaling) are expected to occur, in
17    some cases, at lower doses than are required for peroxisome proliferation (Bull, 2000), indicating
18    occurrence of events possibly associated with tumorigenesis at doses below those causing the
19    peroxisome proliferation response.
20
21    4.4.4.3.7. Temporal association.  Increases in peroxisome volume density as well as in marker
22    cyanide-insensitive (PCO) oxidation indicative of peroxisome proliferation have been shown to
23    occur  following a few days of treatment with tetrachloroethylene, DCA, or TCA (Goldsworthy
24    and Popp, 1987; Elcombe, 1985; Odum et al.,  1988; Daniel et al., 1993).   On the other hand,
25    following DCA treatment, DNA SSB has  been observed prior to peroxisome proliferation
26    (Nelson and Bull, 1988; Nelson et al., 1989), although investigators using a different
27    methodology did not observe the DCA-induced  SSB (Chang et al.,  1992) after DCA treatment.
28
29    4.4.4.3.8. Species similarities and differences:  human evidence.  The relevance to humans of
30    rodent hepatocellular carcinomas thought to be induced specifically by peroxisome proliferator
31    chemicals has been questioned.  Humans do have a functional PPAR receptor (Sher et al., 1993),
32    which is comparable to PPAR receptors of mice and rats in its affinity  for PPAR-a ligands
33    (Klaunig et  al., 2003) and it is capable of activating many of the genes regulated in the mouse by
34    PPAR-a (Yu et al.,  2001).
35          PPAR-a, the PPAR subtype considered to be the causal factor for peroxisome
36    proliferation in rodent hepatocytes, has been found in tissue from several species, including mice,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-28       DRAFT-DO NOT CITE OR QUOTE

-------
 1    rats, and humans as well as dogs, guinea pigs, hamsters, and nonhuman primates (Yousef and
 2    Badr, 1999; Schultz et al., 1999; Lake et al., 1993; Roberts et al., 2000; Reddy et al., 1984;
 3    Graham et al., 1994; Kurata et al., 1998). Humans express PPAR-a in liver (Auboeuf et al.,
 4    1997), although reportedly to a lesser extent than do rats and mice (Klaunig et al., 2003).
 5    PPAR-a mRNA in human liver samples have been reported by some investigators to be one
 6    order of magnitude lower than those observed in mice (Palmer et al., 1998; Tugwood et al.,
 7    1996).
 8          Only a few human liver samples have been examined for quantification of PPAR-a
 9    transcription factors (Klaunig et al., 2003), and one study by Walgren et al. (2000a) reported that
10    one of six human samples was equivalent to mice in expression of PPAR-a protein. Although
11    the number of human liver samples examined is limited, evidence exists for mutations in
12    PPAR-a (Flavell et al., 2000; Sapone et al., 2000; Vohl et al., 2000; Yamakawa-Kobayashi et al.,
13    2002), which could contribute to the large variations in PPAR-a levels (Walgren et al., 2000a).
14    Intel-individual variability  (Tugwood et al.,  1996), along with the inducibility of PPAR-a
15    expression by chemicals and other factors (Sterchele et al., 1996), indicate the likelihood  of a
16    susceptible subpopulation (Heuvel, 1999).
17          Although some studies have shown no increase in DNA synthesis in primary human
18    hepatocytes following treatment with several peroxisome proliferator chemicals, other evidence
19    indicates that humans may indeed be responsive to adverse effects of peroxisome proliferators.
20    For example, investigations of human hepatocytes following treatment with certain fibrate
21    chemotherapeutic agents found dose-dependent induction of acyl-CoA oxidase activity and, in
22    one case, increased peroxisome density (Cimini et al., 2000; Perrone et al., 1998). Increases of
23    liver peroxisomes have been reported in human patients taking the hypolipidemic therapeutic
24    agents clofibrate and ciprofibrate (Hanefeld et al., 1983; Bentley et al.,  1993; Hinton et al., 1986).
25    The increases in volume density of peroxisomes (23-30%) are comparable to or greater than
26    those observed in rodents exposed to tetrachloroethylene.  Also, both humans and rodents
27    respond to peroxisome proliferators with reduction of serum lipids, indicating similar capabilities
28    for modification of gene expression. Epidemiologic evidence of cancer from exposure to
29    peroxisome proliferator chemicals is limited to only a few studies in patients taking fibrate drugs,
30    and is inconclusive (Newman and Hulley, 1996; Melnick, 2001).
31          Taking into account kinetic and dynamic factors, the proposed animal MOA is plausible
32    in humans.  If peroxisome proliferation is involved in a mode of carcinogenic action for
33    tetrachloroethylene, the cancer-causing activity cannot be dismissed for humans, especially since
34    PPAR-a has been identified in human liver,  and both TCA and DCA metabolites similarly
35    activate human as well as mouse PPAR-a, even if to different  degrees.  Chemical-specific data
36    regarding the ability of the major tetrachloroethylene metabolite TCA to activate PPAR-a in

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-29        DRAFT-DO NOT CITE OR QUOTE

-------
 1    humans as well as in mice indicate cross-species relevance (Maloney and Waxman, 1999),
 2    although there exists a quantitative difference between these two species in hepatic PPAR-a
 3    activation (Walgren et al., 2000a, b).  Site concordance is not a requirement for extrapolation of
 4    tumorigenesis in animal models to the human situation, however, and PPAR-a is found in
 5    several different organs.  It is highly expressed in cells having active fatty acid capacity, such as
 6    hepatocytes in liver, and also in renal proximal tubule cells, cardiomyocytes, and enterocytes.
 7    Some PPAR-a agonists  cause tumors in rodents at sites other than the liver.
 8          PPAR-a is proposed to be involved in causing Leydig cell tumors. The human relevance
 9    of PPAR-a induction of Leydig cell tumors is not so controversial.  The understanding of the
10    science is not good enough to explain why humans have a functional PPAR-a capable of gene
11    expression modulation in a manner similar to that of rodents but does not respond similarly.
12    PPAR-a in humans may be capable of modulating lipid homeostasis through alteration of
13    expression of other, different, genes in the liver and genes in other target organs that express
14    higher levels of PPAR-a.  Similar events occurring in various tissues and organs could lead to
15    carcinogenesis in those tissues and organs; therefore, target organs could differ among species.
16    The liver carcinogenesis in mice could be a red flag for tumorigenesis at some other site in
17    humans.  This is the case for other chemicals, such as arsenic.
18          A different PPAR receptor also involved in lipid homeostasis may be involved in human
19    liver cancer, because cross-talk is known to occur among these receptors. Glinghammar et al.
20    (2003) found that PPAR delta (PPAR-5) receptor activation in human hepatocellular carcinoma
21    cells induced COX2 expression, a factor associated with carcinogenesis, and increased cellular
22    proliferation.  Such results suggest a potential role for PPAR-5 in human hepatocellular
23    carcinoma induction.  PPAR-a also can mediate mRNA alterations associated with prostaglandin
24    synthesis in rodents, i.e., COX2 expression (Peters and Vanden Heuvel, 2002).  PPAR gamma
25    has been shown to induce COX2 expression as well.  Because cross-talk is known to occur
26    among the PPAR receptors, the potential exists for cross-talk to be involved in receptor
27    activation related to carcinogenicity.
28
29    4.4.4.3.9. Biological plausibility and coherence of the database. Tetrachloroethylene induces
30    liver cancer in treated mice, but it has not been shown to cause liver cancer in treated rats; thus,
31    an inconsistency exists between rodent species. Epidemiologic evidence is insufficient for
32    determining whether tetrachloroethylene causes liver cancer in humans (see Section 4.3.1.2).
33          Inadequate amounts of TCA are produced from tetrachloroethylene metabolism to
34    account totally for the liver tumor induction observed in bioassays (Appendix 4A). DCA is less
35    likely to contribute to tetrachloroethylene hepatocarcinogenesis because relatively small amounts
36    would be produced in the liver target organ from metabolism of tetrachloroethylene. Limited

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-30       DRAFT-DO NOT CITE OR QUOTE

-------
 1    data on tumor phenotype indicate that tetrachloroethylene tumors differ phenotypically from
 2    TCA tumors or DCA tumors (Maronpot et al., 1995; Anna et al., 1994).
 3          An important distinction needs to be drawn between biomarkers that are causally related
 4    to carcinogenic activity and those that are merely correlative. There is much confusion and
 5    debate over the reliability of hepatic peroxisome proliferation as a marker for
 6    hepatocarcinogenesis. As reviewed in Section 4.4.4.1, some scientists maintain that there is a
 7    causal relationship between peroxisome proliferation and hepatocarcinogenesis, whereas others
 8    have questioned the validity of this relationship and suggest that the degree of peroxisome
 9    proliferation correlates poorly with relative hepatocarcinogenic effectiveness and potency.
10    Although many of the responses generally observed in the overall peroxisome proliferation
11    phenomena are often manifested in tumorigenesis, causality is uncertain,  and in the specific case
12    of tetrachloroethylene, the data are especially limited.
13          Several recent studies have expanded the scientific understanding of the PPAR-a mode of
14    action (see Caldwell et al., 2008).  First, Yang et al. (2007) demonstrated that PPAR-a activation
15    in hepatocytes induces peroxisome proliferation but not liver tumors.  The approach entailed
16    targeting expression of PPAR-a to hepatocytes by placing the VP16 PPAR-a transgene gene
17    under control  of the liver enriched activator protein (LAP) promoter. LAP-VP16 PPAR-a
18    transgenic mice showed a number of PPAR-a-mediated effects:  decreased serum triglycerides
19    and free fatty  acids, peroxisome proliferation, enhanced hepatocyte proliferation, and induction
20    of cell-cycle and PPAR-a target genes. However, compared with wild-type mice exposed to
21    Wy-14,643, the extent of hepatomegaly was reduced  and no hypertrophy or eosinophilic
22    cytoplasms was seen in LAP-VP16 PPAR-a mice.  Also in contrast with wild-type mice exposed
23    to Wy-14,643, no evidence of non-parenchymal cell proliferation was observed in the LAP-
24    VP16 PPAR-a transgenic mice. Moreover, at one year of age no evidence of preneoplastic
25    hepatic lesions or hepatocellular neoplasia was observed in LAP-VP16 PPAR-a transgenic mice.
26    As noted by the authors, PPAR-a activation only in mouse hepatocytes is sufficient to induce
27    peroxisome proliferation and hepatocyte proliferation but"... is not sufficient to induce liver
28    tumors."
29          Secondly, Ito et al. (2007) found that di (2-ethylhexyl)phthalate (DEHP), a proposed
30    robust example of PPAR-a agonism-induced hepatocarcinogenesis, yields liver tumors in a
31    2-year study in PPAR-a knock-out mice.  This study  demonstrates the limitations, cited by the
32    FIFRA SAP, of drawing conclusions from the one-year bioassay of high doses of WY-14,643
33    referenced above (e.g., Peters, 1997). It supports the  view that knock-out mouse bioassays
34    should be carefully characterized and conducted for 2 years to assess whether PPAR-a activation
35    is indeed necessary for induction of liver cancer.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-31       DRAFT-DO NOT CITE OR QUOTE

-------
 1           In summary, limited evidence supports the hypothesis that tetrachloroethylene tumor
 2    induction could be related to PPAR-a activation, but critical review of the scientific literature
 3    reveals significant data gaps regarding the relationship between the PPAR-a activation and
 4    neoplasia induced by tetrachloroethylene. If PPAR-a does play a role in tetrachloroethylene-
 5    induced tumorigenesis, available information suggests relevance  to humans cannot be ruled out.
 6
 7    4.4.4.4. Effects That Could Be Related to Other Potential Modes of Action
 8    4.4.4.4.1. Mutagenicity and genotoxic effects. The available evidence is inconclusive regarding
 9    mutagenicity of tetrachloroethylene or its metabolites and hepatocarcinogenesis (see Section 4.3).
10    Tetrachloroethylene induces SSB and DNA binding in liver tissue, and existing data implicate a
11    potential role for genotoxic effects of certain metabolites, such as DCA and the proposed
12    intermediate chloral hydrate; the epoxide tetrachloroethylene oxide is a bacterial mutagen.
13    Interestingly, and the phenotype and frequency of tumors produced by DCA and
14    tetrachloroethylene tumors differ (Bull, 2000; Anna et al., 1994; Maronpot et al., 1995; Moore
15    and Harrington-Brock, 2000; also see Section 4.4.4.2).  The mutagenic potential of several
16    metabolites has not been studied. Not all of the P450 metabolites, including the unstable,
17    potentially reactive intermediates, such as trichloroacetyl chloride, for example, have been
18    sufficiently tested in the standard genotoxicity screening battery.
19
20    4.4.4.4.2. Immunosuppressive effects. Although tetrachloroethylene-specific data are lacking,
21    it is possible that inhibition of the natural immune surveillance could be related to
22    hepatocarcinogenic properties of tetrachloroethylene (see also Section 4.8.3). Immune
23    suppression could play a role in the induction of cancer as many immunosuppressive agents are
24    human carcinogens (Tomatis et al., 1989). Exposure to organic solvents has been generally
25    associated with autoimmune diseases such as scleroderma (Nietert et al., 1998). A strong
26    association has been reported between exposure to solvents structurally similar to
27    tetrachloroethylene and systemic sclerosis in patients who have autoantibodies (Nietert et al.,
28    1998).
29           Binding of reactive compounds to cellular macromolecules has been proposed as an
30    important step  in the pathogenesis of several diseases, both for cancer (Hinson and Roberts,
31    1992) and for chemically induced autoimmune disease (Uetrecht et al., 1988). Reactive
32    metabolites of  tetrachloroethylene have been shown to bind irreversibly to cellular
33    macromolecules in vitro (e.g., Costa and Ivanetich, 1980) and in vivo (Pegg et al., 1979;
34    Schumann et al.,  1980). Binding occurs proportionally to the amount metabolized, and
35    metabolism is proportional to toxicity (e.g., Buben and O'Flaherty, 1985).  Several published
36    studies have demonstrated formation of trichloroacylated protein adducts, for example, in liver

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-32       DRAFT-DO NOT CITE OR QUOTE

-------
 1    and kidney of rats (Birner et al., 1994) and in plasma of rats and humans (Pahler et al., 1999)
 2    following exposures to tetrachloroethylene. Another example is the detection of
 3    trichloroacetylated protein adducts formed in mice treated with tetrachloroethylene (Green et al.,
 4    2001).  Further studies designed to identify the adducted proteins may help to elucidate an MOA
 5    for tetrachloroethylene-induced autoimmune response, which, in turn, may be related to cancer-
 6    causing activity.
 7
 8    4.4.4.4.3. Effects on the insulin receptor/glucose metabolism,  Tumorigenesis is associated
 9    with changes in enzymes involved in carbohydrate metabolism (Ahn et al., 1992) and tumor cells
10    depend uniquely on glucose as  an energy source.  Transformation of many cell types is
11    associated with an increase in glucose metabolism including increases in important glycolytic
12    enzyme activities (Baggetto,  1992), and often these enzyme activities correlate with malignancy
13    (Weber and Lea,  1966; Harap, 1975).  Ebrahim et al. (1996) observed tetrachloroethylene-
14    induced alterations in glycolytic and gluconeogenic enzymes in liver and kidney of mice treated
15    with the chemical. Administration of 2-deoxy-D-glucose and vitamin E controlled the changes
16    in glycolytic and gluconeogenic enzymes induced by tetrachloroethylene.
17
18    4.4.4.4.4. Alteration in DNA methylation.  No tetrachloroethylene-specific data are available
19    regarding a role of alteration in DNA methylation in tumorigenesis. Such  changes are reported
20    to be a common early molecular event in most tumors (Zingg and Jones, 1997; Baylin et al.,
21    1998).  Alterations in DNA methylation following exposure to chemicals,  both hypomethylation
22    and hypermethylation, may be factors in tetrachloroethylene-induced tumorigenesis. Although
23    tetrachloroethylene-specific data are lacking, its metabolites DCA and TCA are known to induce
24    hypomethylation of DNA and protooncogenes in mouse liver.
25
26    4.4.4.4.5. Alterations in cell replication and death rate. Although modification of cell
27    replication and death rates may be important to tetrachloroethylene liver tumorigenesis, no
28    tetrachloroethylene-specific data are available.
29
30    4.4.4.4.6.  Cytotoxicity and compensatory hyperplasia. Cytotoxicity and reparative hyperplasia
31    are not marked findings resulting from tetrachloroethylene exposures capable of causing liver
32    cancer in mice.
33
34    4.4.4.4.7. Hepatomegaly/cytomegaly.  Increase in liver size is highly correlated with liver
35    tumorigenesis in mice.  Treatment with tetrachloroethylene can lead to increased liver weight. In
36    carcinogenicity studies, hepatomegaly occurred following exposures in the dose ranges that

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-33        DRAFT-DO NOT CITE OR QUOTE

-------
 1    cause liver tumors and at experimental exposures well below the carcinogenicity bioassay dose
 2    levels.  It is not clear exactly how the phenomenon is related to tumorigenesis in the case of
 3    tetrachloroethylene.
 4
 5    4.4.4.5. Summary and Conclusions
 6          At the present time, the specific mechanisms or MOA for tetrachloroethylene-induced
 7    hepatocarcinogenesis in mice are not known.  The cancer-causing activity of tetrachloroethylene
 8    is thought to be metabolism-dependent, however the specific contribution of particular
 9    metabolites has not been elucidated. TCA induces liver tumors with different phenotype and is
10    not produced in sufficient amounts to account quantitatively for the liver tumor response
11    observed with tetrachloroethylene. Not all metabolites have been identified or characterized, but
12    several known metabolites including those derived from P450 as well as GSH pathways are
13    clearly mutagenic in the standard battery of tests.  Tetrachloroethylene is mutagenic in bacterial
14    assays in the presence of GST and GSH whereas the standard S9 fraction has typically yielded
15    negative results.  Tetrachloroethylene at higher concentrations also induces SSBs and DNA
16    binding in liver tissue. The metabolite DC A is the most potent mutagen of the P450-derived
17    metabolites, exhibiting mutagenic activity in a number of assays.  A putative P450 derived
18    metabolite,  1,1,2,2-tetrachloroethylene oxide,  is also mutagenic;  the mutagenicity of this epoxide
19    would be predicted from structure-activity relationships.  Given the demonstrated mutagenicity
20    of several tetrachloroethylene metabolites, it is expected that mutagenicity contributes to the
21    MOA for tetrachloroethylene carcinogenesis, although the specific metabolic species or
22    mechanistic effects are not known.
23          Chemical-specific data for PPAR-a activation are limited but suggest that this is not the
24    primary MOA for hepatocarcinogenesis.  The  existing data for tetrachloroethylene show minimal
25    peroxisome proliferator activity, and no chemical-specific data correlate peroxisome
26    proliferation with tumor induction for tetrachloroethylene.  As noted above, TCA produces
27    tumors of a different phenotype than tetrachloroethylene and TCA is not produced in sufficient
28    amounts to  account quantitatively for the tetrachloroethylene liver tumor response.  Moreover,
29    several recent studies have expanded the scientific understanding of the PPAR-a mode of action
30    (see Caldwell et al., 2008) including the demonstration that PPAR-a activation in hepatocytes
31    induces peroxisome proliferation but not liver tumors (Yang et al., 2007). In particular,
32    peroxisome proliferation and hepatocyte proliferation, but not liver tumors, were observed with
33    PPAR-a activation in mouse hepatocytes. Furthermore, Ito et al. (2007) found that DEHP, a
34    proposed robust example of PPAR-a agonism-induced hepatocarcinogenesis, yields liver tumors
35    in a 2-year study in PPAR-a knock-out mice.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-34       DRAFT-DO NOT CITE OR QUOTE

-------
 1          In summary, the MOA for tetrachloroethylene-induced liver toxicity and tumorigenesis is
 2    not understood.  Data are lacking particularly for tetrachloroethylene P450 intermediates that
 3    could be involved in mutagenicity and carcinogenicity of the parent compound.  Among the data
 4    gaps is the incomplete characterization of the metabolites in tests beyond the standard battery of
 5    genotoxicity tests, including on important genetic and epigenetic endpoints.
 6
 7    4.5.  KIDNEY TOXICITY
 8    4.5.1. Human Studies
 9    4.5.1.1. Kidney Toxicity in Humans
10          High concentrations of inhaled tetrachloroethylene given acutely as an anaesthetic are
11    associated with symptoms of renal dysfunction, including proteinuria and hematuria (Hake and
12    Stewart, 1977, AT SDR, 1997). Controlled inhalation exposure to tetrachloroethylene at levels of
13    0, 20, 100, or  150 ppm for up to 11 weeks did not affect a number of urine parameters or BUN (a
14    measure of kidney function) in 12 healthy individuals (Stewart et al., 1977, as reported in
15    ATSDR, 1997).  Whether renal effects would occur from these acute exposure levels in a larger,
16    more diverse population than the one studied by Stewart et al. (1977) is not known.
17          The evidence for kidney effects from chronic inhalation of tetrachloroethylene is limited
18    because many of the available reports do not include information on even a minimal core battery
19    of tests for kidney function. The ATSDR (Amler et al., 1998; Lybarger et al., 1999)
20    recommends a core battery of kidney function tests that includes serum creatinine, urinalysis
21    with microscopic examination of urine sediment, albumin, retinol binding protein (RBP),
22    N-acetyl-p-D-glucosaminidase (NAG), alanine aminopeptidase (AAP), osmolality, and urine
23    creatinine (Lybarger et al.,  1999). These indicators evaluate a range of toxicity,  from effects on
24    general kidney function to effects on specific segments of the nephron. For example, the overall
25    integrity of the nephron can be evaluated from the urinalysis, and albumin is an indicator of the
26    integrity of the glomerulus; three indicators—RBP, NAG, and AAP—assess damage to the
27    proximal tubules (Lybarger et al., 1999). The proximal tubules house  beta lyase enzymes and
28    are hypothesized to be a target of tetrachloroethylene toxicity due to the bioactivation of reactive
29    metabolites produced from the further metabolism of TCVC (see Section 4.2). For this reason,
30    this analysis places greater weight on urinary indicators of proximal tubule function.
31          The epidemiologic studies are suggestive of subtle damage to the renal tubules. Five
32    studies (Trevisan et al., 2000; Verplanke et al., 1999; Mutti et al., 1992; Solet and Robins, 1991;
33    Lauwerys et al.,  1983) have examined the three core  indicators of tubule function—RBP, NAG,
34    or AAP—in urine of dry cleaners. Three studies measured RBP, with two of the studies
35    reporting a statistically significant elevated prevalence of abnormal values among study
36    participants (Mutti et al., 1992) or a statistically significant elevated geometric mean
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-35       DRAFT-DO NOT CITE OR QUOTE

-------
 1    concentration of RBP (Verplanke et al., 1999) for tetrachloroethylene-exposed workers as
 2    compared with controls.  The mean concentration of RBP for exposed subjects (75.4 ug/g
 3    creatinine) in the Verplanke et al. (1999) study is within a normal range,1 indicating the absence
 4    of concurrent tubul e toxi city.
 5          As a comparison, Nomiyama et al. (1992) suggest a critical level of RBG of 200 ug/g
 6    creatinine as indicative of cadmium-induced kidney toxicity. Exposure levels were to a median
 7    of 15 ppm (range: limit of detection to 85 ppm) in Mutti et al. (1992) and 1.2 ppm (range:
 8    0.3-6.5 ppm) in Verplanke et al. (1999). Lauwerys et al. (1983), the only other study to assess
 9    RBP, did not observe any differences in the geometric mean concentration of RBP between dry
10    cleaners with mean tetrachloroethylene exposure of 21 ppm and their controls; however, this
11    study contained fewer exposed subjects with a shorter duration of exposure than did that of Mutti
12    etal. (1992).
13          The four studies that measured urinary excretion of NAG (Solet and Robins, 1991;  Mutti
14    et al., 1992; Verplanke et al., 1999; Trevisan et al., 2000) and the one study that measured AAP
15    (Verplanke et al., 1999) did not observe any differences between exposed subjects and controls.
16    These findings are not surprising; NAG is not a sensitive and specific marker of tubular
17    dysfunction (Lybarger et al., 1999). Mean exposures were 14 ppm in Solet and Robins (1991)
18    and 9 ppm in Trevisan et al. (2000); both studies assessed exposure from personal monitoring of
19    exhaled breath.
20          The above findings are further supported by the observation of elevated urinary excretion
21    of other proteins that are also indicators of damage to the proximal tubules: beta2-microglobulin,
22    intestinal alkaline phosphatase (LAP), tissue non-specific alkaline phosphatase  (TNAP),
23    lysozyme, beta2-glucuronidase, and glutamine synthetase. Both IAP and TNAP are indicators of
24    proximal tubule brush border integrity (Price et al.,  1996), whereas lysozyme and
25    beta2-microglobulin indicate a failure of the tubule to reabsorb protein (Lybarger et al.,  1999;
26    Bernard and Lauwerys, 1995; Kok et al., 1998).  Glutamine synthetase is a mitochondrial
27    enzyme located in the proximal tubules and has been recently suggested as a marker of tubular
28    damage in rats exposed to 1,3-hexachlorobutadiene (Trevisan et al., 1999).
29          Mutti et al. (1992) observed an elevated prevalence of abnormal values for
30    beta2-microglobulin and brush border antigens, a higher geometric mean concentration of brush
31    border antigens in urine, and a higher concentration of TNAP in urine among 50  exposed dry
32    cleaners as compared with 50 blood donors matched by sex and age with the exposed subjects.
33    Furthermore, markers of renal damage were highly  predictive of exposure status in discriminant
             1 Lapsley et al. (1998) found a median and an upper 98% confidence limit of 67 and 143 ug/g creatinine,
      respectively, in a survey of 70 adults, and this range closely matches the findings of Topping et al. (1986), who
      observed a mean and a 98% upper limit of 64 and 185 ug/g creatine in 118 subjects.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-36        DRAFT-DO NOT CITE OR QUOTE

-------
 1    analysis.  Beta2-microglobulin, however, was not elevated among exposed subjects as compared
 2    with controls in the other two studies that examined this protein (Lauwerys et al., 1983; Vyskocil
 3    et al., 1990), although the mean concentration of beta2-microglobulin appeared higher in subjects
 4    studied by Vyskocil et al.  than the mean concentration in controls. Both these studies contained
 5    fewer numbers of exposed subjects than did the study by Mutti et al. (1992), and reduced power
 6    as a consequence of fewer subjects may be a reason for the null observations. Further,
 7    tetrachloroethylene exposure appears to affect reabsorption in the renal tubules. Two studies that
 8    assessed lysozyme or beta-glucuronidase observed a statistically significant elevated mean
 9    concentration of these proteins among dry cleaners as compared with controls (Franchini et al.,
10    1983; Vyskocil et al., 1990).
11           It is not clear whether tetrachloroethylene exposure engenders an effect on other parts of
12    the kidney. The study by Mutti et al. (1992) is suggestive of damage to the glomerulus; however,
13    the lack of an elevated excretion of albumin, an indicator of glomerular function (Lybarger et al.,
14    1999), in the study by Verplanke et al. (1999) argues for further assessment of possible
15    glomerular effects.
16           Taken together, the epidemiologic studies support an inference of subtle effects on the
17    renal proximal tubules.  Effects are seen in populations of both males and females, and potential
18    differences in susceptibility due to sex-related differences in rates of metabolism (see Section
19    4.2) cannot be determined from the available evidence. Median exposure levels in the studies
20    that observed  alterations in renal enzymes were 9 ppm (Trevisan et al.,  2000), 10 ppm  (Franchini
21    et al., 1983), and 15 ppm (Mutti et al., 1992), representing LOAELs for these studies.  Only the
22    study by Trevisan et al. (2000) observed an exposure-response relationship, a correlation
23    between urinary tetrachloroethylene and the concentration of glutatmine synthatase (p  < 0.001).
24    None of the other studies reported exposure-response relationships, which is a limitation on the
25    inference of an association between tetrachloroethylene and renal damage. However, as pointed
26    out by Mutti et al. (1992), this is a common finding among solvent-exposed populations, and
27    inadequate definition of the dose metric most likely contributes to the null finding.  Table 4-3
28    summarizes the human kidney toxicity studies.
29
30    4.5.1.2. Kidney Cancer
31           The evidence  supporting a hypothesis of an association between tetrachloroethylene
32    exposure and kidney cancer consists of the observation of elevated risks in the larger case-
33    control studies (Asal et al., 1988; Aschengrau et al., 1993; Dosemeci et al., 1999; Mandel et al.,
34    1995; McCredie and Stewart, 1993; Mellemgaard et al., 1994; Schlehofer et al., 1995;  Partanen
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-37      DRAFT-DO NOT CITE OR QUOTE

-------
        Table 4-3.  Summary of human kidney toxicity marker studies in dry cleaners
Author
Mutti et
al. (1992)
Verplanke
etal.
(1999)
Solet and
Robins
(1991)
Lauwreys
etal.
(1983)
Vyskocil
etal.
(1990)
Franchini
etal.
(1983)
Trevisan
etal.
(1999)
Number
of subjects
50 dry cleaners and 50
matched controls
82 exposed and 19
nonexposed dry cleaners
192 exposed dry cleaners,
no control group
24 exposed dry cleaners
and 33 unexposed controls
16 exposed dry cleaners
and 13 nonoccupationally
unexposed controls
57 exposed dry cleaners
and (a)
81 unexposed and (b) 80
unexposed factory workers
40 dry cleaners
and 45 laundry workers
Estimated mean
exposure
concentration"
(ppm)/duration
(yrs)
15/10
PM
1.2 (TWA)/
5.1
PM
14/11.6
PM
21/6.4
PM
23 (TWA)/9
U
10 (TWA)2/14
PM
8.8/15
PM, B, U
RBP
D
D

NS



NAG
NS
NS
NS




AAP

NS





Beta2-
micro-
globulin
D


NS
NS


Albumin
D

NS

NS
NS

Total
protein
NS
NS
NS


NS

Lyso-
zyme




D
D

TNAP
NS






Beta-
glucuro-
nidase





D

Gluta-
mine
synthe-tase






Dose-
response
observed
a TWA tetrachloroethylene concentration developed using the relationship of between excretion of TCA in urine and breathing-zone, TWA concentration of
  Ikeda and Otsuji (1972).
B       = Biological monitoring of blood
D       = Detected with statistically significant elevation with respect to controls
IA      = Indoor air monitoring
NS     = Not statistically significant
PM     = Personal monitoring of breath
U       = Biological monitoring of urine for trichloroacetic acid (U-TCA)

-------
 1    et al., 1991; Pesch et al., 2000a).  The studies by Aschengrau et al. (1993), Partanen et al. (1991),
 2    and Pesch et al. (2000a) are of high quality because they had good exposure information,
 3    controlled adequately for confounding, and used histologic confirmation of outcomes. For these
 4    reasons, observations in these two case-control studies carry greater weight than observations in
 5    the other case-control studies identified in Table 4B-4 (Appendix 4B). The remaining studies
 6    included large numbers of cases self-reported to determine exposure.  These types of reports are
 7    more subject to misclassification errors.
 8           In many of the case-control studies there are concerns about selection bias, blinding of
 9    investigators or interviewers, and, particularly, exposure characterization (Wartenberg et al.,
10    2000).  Three studies (Pesch et al., 2000a; Dosemeci et al., 1999; Schlehofer et al.,  1995) present
11    risks for tetrachloroethylene exposure explicitly. The studies by Pesch et al. (2000a) and
12    Dosemeci et al. (1999) both suggest that there may be gender differences in  renal cell carcinoma
13    risk with occupational exposure to tetrachloroethylene; in both studies the risks were higher in
14    males than in females. Exposure-response gradients were not observed in any of the three
15    studies.
16           Cohort studies of kidney cancer incidence among dry cleaners and laundry workers in
17    Sweden and Denmark (Andersen et al., 1999; Travier et  al., 2002; Lynge and Thygesen, 1990;
18    McLaughlin et al., 1987) did  not observe excess risks of kidney cancer (see Table 4B-la and
19    Appendix 4B)—an inconsistency with the case-control studies. Few kidney cancer deaths were
20    observed in cohort studies assessing mortality among dry cleaners (Ruder et al., 2001; Blair et al.,
21    2003; see Table 4B-lb and Appendix 4B). The highest risks (not statistically significant) were
22    reported for tetrachloroethylene-exposed subjects (Ruder et al., 2001) and for subjects identified
23    with higher levels of exposure as compared with subjects with little or no exposure (Blair et al.,
24    2003).  There are too few cases of kidney cancer in the tetrachloroethylene subcohorts (degreaser
25    studies) to assess any relationship with tetrachloroethylene (see Table 4B-2 and Appendix 4B).
26
27    4.5.2.  Animal Studies
28    4.5.2.1. Kidney Toxicity in Animals
29           Tetrachloroethylene causes renal toxicity across several species, including rats, mice,
30    rabbits, dogs, guinea pigs, and humans (for reviews, see  U.S. EPA,  1985a, ATSDR, 1997; NYS
31    DOH, 1997; Cal EPA, 2001).
32           Adverse effects  on the kidney have been observed in studies of animals exposed to high
33    concentrations of tetrachloroethylene by inhalation, oral  intake, and i.p. injection.  These effects
34    include hyperplasia and increased kidney-to-body weight ratios, hyaline droplet formation,
35    glomerular "nephrosis," karyomegaly (enlarged nuclei),  cast formation, and other lesions or
36    indicators of renal toxicity. The effects occurred following very high doses  or chronic, relatively
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-39       DRAFT-DO NOT CITE OR QUOTE

-------
 1    high-doses of tetrachloroethylene exposures.  The LOAEL for renal toxicity reported in the
 2    scientific literature is 100 ppm (678 mg/m3) for inhalation exposure in mice (NTP, 1986a).
 3          Oral administration of tetrachloroethylene in sesame oil (3 g/kg/day for 15 days) to mice
 4    caused an increase in kidney weight as well as increases in glomerular nephrosis and
 5    degeneration (Ebrahim et al., 1996). A lifetime animal carcinogenicity study in which
 6    tetrachloroethylene was administered to rats and mice by oral gavage in corn oil for 78 weeks
 7    resulted in clear evidence of kidney toxicity in both species (NCI, 1977).  The TWA doses
 8    (mg/kg-day) used in the bioassay were 471 and 941 for male rats, 474 and 949 for female rats,
 9    536 and  1,072 for male mice, and 386 and 772 for female mice.  Nephropathy was observed in
10    almost all of the test animals.
11          Hayes et al. (1986) reported renal effects in rats exposed to 400 mg/kg-day
12    tetrachloroethylene in drinking water for 90 days. In a study by Jonker et al. (1996),
13    tetrachloroethylene nephrotoxicity was observed in female Wistar rats administered the chemical
14    in corn oil by oral gavage for 32 days.  Nephrotoxic effects were noted  at 2,400 mg/kg.
15
16    4.5.2.2. Kidney Cancer in Animals
17          In the studies conducted by NTP (1986a), groups of 50 male and 50 female F344/N rats
18    were exposed for 6 hrs/day, 5 days/week, for 103 weeks by inhalation to atmospheres containing
19    0, 200, or 400 ppm tetrachloroethylene. Tubule cell hyperplasia was observed in male rats
20    (control, 0/49; low dose, 3/49; high dose, 5/50) and in one high-dose female rat. Renal tubule
21    adenomas and adenocarcinomas were observed in male rats (control, 1/49; low dose, 3/49; high
22    dose, 4/50).
23          The spontaneous incidence rate for renal tubule tumors in F344/N rats, the strain used in
24    the NTP bioassay, as well  as for other rat strains reported by NTP was less than 1%,  making the
25    appearance of tubule neoplasms in 8% of the treated animals in the NTP study (low-dose and
26    high-dose groups combined) convincing evidence of a treatment-related effect (Goodman et al.,
27    1979; Solleveld et al., 1984;  U.S. EPA, 1986a, 1991a).  Also notable is the fact that no malignant
28    renal tubule neoplasms had ever been observed in any control rats examined by NTP—including
29    chamber controls from the performing laboratory and the untreated controls and vehicle controls
30    from gavage studies—whereas two of the tumors observed in high-dose animals in the NTP
31    study were carcinomas. The probability of two rare carcinomas appearing by chance in a group
32    of 50 animals has been calculated to be less than 0.001 (U.S. EPA, 1987a, 1991a; NTP, 1986a).
33          In addition, when statistically compared with historical control incidences of renal tubule
34    tumors, a significant dose-related positive trend exists, and tumor incidences in both low-dose
35    and high-dose groups are significantly elevated.  Standard statistical analyses of tumor incidence
36    data did not reveal a significant increase in kidney tumors,  and the tumor incidence is not

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-40       DRAFT-DO NOT CITE OR  QUOTE

-------
 1    statistically significant when compared with concurrent controls; however, when the incidences
 2    of tubule cell hyperplasia and neoplasms and tumor severity are all considered, a dose-response
 3    relationship is apparent.
 4          A slight increase in renal tumors was observed in other studies in male Sprague-Dawley
 5    (SD) rats receiving tetrachloroethylene by gavage or by inhalation (Maltoni and Cotti, 1986;
 6    Rampy et al., 1978; cited in U.S. EPA, 1991a), consistent with the findings reported in the NTP
 7    studies.  However, in the rat chronic bioassay reported by JISA (1993), there was no increase in
 8    the incidence of kidney tubular cell adenoma or carcinoma in excess of that in the concurrent or
 9    historical control animals (see Tables 5-7 and 5-8) at administered concentrations of 50, 200, and
10    600 ppm.
11          The findings of rare kidney tumors in some cancer bioassays constitute suggestive
12    evidence that tetrachloroethylene can induce kidney cancer in humans. The findings of rare
13    kidney tumors in tetrachloroethylene bioassays in laboratory animals have been reviewed in EPA
14    assessment documents on tetrachloroethylene (U.S. EPA, 1985a, 1986a, 1991a) and by an EPA
15    risk assessment forum technical panel (U.S. EPA, 1991b) reporting on the association of alpha-
16    2u-globulin with renal lesions in the male rat. The closely related tetrachloroethylene congener
17    trichloroethylene also induces low increased incidences of rare renal tumors in rats (U.S. EPA,
18    200la) and NTP has found low incidences of tubule neoplasms in rats dosed with other
19    chlorinated ethanes and ethylenes (NTP,  1983, 1988, 1986a, b, 1987, 1989, 1990b).
20
21    4.5.3. Summary of Kidney Effects in Humans and Animals
22          Taken together, the epidemiologic studies support an inference of subtle effects on the
23    renal proximal tubules from inhalation exposure in tetrachloroethylene.  The elevated urinary
24    RBP levels seen in two studies (Mutti et al., 1992; Verplanke et al., 1999) provide some evidence
25    for effects  to the proximal tubules from tetrachloroethylene exposure. Exposures in the two
26    studies that observed renal toxicity were  1.2 ppm and 15 ppm (means), representing an
27    observational LOAEL for human kidney effects. None of the reviewed studies reported
28    exposure-response relationships, and this is an important limitation of the available data.
29    However, as pointed out by Mutti et al. (1992), this is a common finding among solvent-exposed
30    populations, and inadequate definition of the dose metric most likely contributes to the absence
31    of exposure-response relationships. No human studies investigating drinking water or other oral
32    exposures on kidney toxicity have been published.
33          Positive associations between kidney cancer (renal cell carcinoma) and exposure to dry
34    cleaning and laundry workers or to tetrachloroethylene specifically were observed in several
35    well-conducted studies (Mandel et al., 1995; McCredie and Stewart, 1993; Pesch et al., 2000a;
36    Schlehofer et al., 1995).

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-41       DRAFT-DO NOT CITE OR QUOTE

-------
 1           Adverse effects on the kidney have been observed in studies of animals exposed to high
 2    concentrations of tetrachloroethylene by inhalation, oral gavage, and i.p. injection. These effects
 3    include hyperplasia and increased kidney-to-body weight ratios, hyaline droplet formation,
 4    glomerular "nephrosis," karyomegaly, enlarged nuclei, cast formation, and other lesions or
 5    indicators of renal toxicity.  Increased incidences of relatively rare renal tumors have been
 6    observed in multiple studies of male rats.  The renal effects occurred following very high (or
 7    chronic, relatively high) doses of tetrachloroethylene exposures. The LOAEL for renal toxicity
 8    reported in the scientific literature is 100 ppm (678 mg/m3) for inhalation  exposure in mice (NTP,
 9    1986a).
10
11    4.5.4. Mode of Action for Kidney Toxicity and Carcinogenicity
12    4.5.4.1. Background
13           The data support the conclusion that the chronic administration of tetrachloroethylene
14    produces nephrotoxicity in both sexes of mice and rats and an increased incidence of
15    proliferative lesions of the kidney tubules in male rats.  The renal tumors observed in male rats
16    exposed to tetrachloroethylene are of a rare type and include carcinomas.  However, the use of
17    these data to infer risk of carcinogenesis to humans has been a focus of scientific debate. Of
18    particular consequence in this debate are two issues:  the possibility that quantitative species
19    differences in conjugative metabolism of tetrachloroethylene may greatly reduce the potential
20    risk of human hazard and the possibility that the induction of renal tubule tumors by
21    tetrachloroethylene may be unique to male rats and, therefore, is inappropriate for deducing
22    potential human health hazard.
23           There are multiple hypothesized MO As for kidney  toxicity induced with
24    tetrachloroethylene exposure, including mutagenicity, alpha-2u-globulin accumulation, and
25    cytotoxicity unrelated to alpha-2u-globulin.  When clearly demonstrated to develop from the
26    sequence of events induced by alpha-2u-globulin accumulation, kidney tumors in male rats
27    caused by exposure to a test chemical, are generally considered to be species and sex specific
28    and not relevant for assessing human hazard. Limited data from studies of tetrachloroethylene
29    indicating hyaline droplet formation provide some evidence for the alpha-2u-globulin MOA.
30    The phenomenon occurs only at very high doses of tetrachloroethylene, however, above the
31    doses used in cancer bioassays in which tumors were observed. There is also data supporting
32    other MO As for tetrachloroethylene-induced renal tumors, particularly findings of mutagenicity,
33    and also cytotoxicity not associated with alpha-2u-globulin accumulation.  This mutagenicity
34    and cytotoxicity are attributed to the further biotransformation of glutathione and cysteine
35    conjugates of tetrachloroethylene to reactive chemical intermediates.  Humans are known to
36    conjugate tetrachloroethylene with glutathione and excrete the mercapturate end product,
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-42       DRAFT-DO NOT CITE OR QUOTE

-------
 1    therefore, exhibiting evidence for operation of the same metabolic pathway.  Quantitative
 2    measurements of urinary excretionmetabolites from this pathway do not provide data for
 3    estimations of the amount of chemical coverted to toxic intermediates; therefore, relative
 4    amounts of tetrachloroethylene processed by enzymes activating the conjugates to toxic products
 5    in rats versus humans are not known.
 6
 7    4.5.4.2. Summary Description of a Postulated Mode of Action—alpha-2u-globulin
 8           Accumulation
 9          A variety of organic compounds have been shown to cause sex- and species-specific
10    lesions in the renal tubules of male  rats in the form of what is known as "hyaline droplet
11    nephropathy" (NTP, 1983, 1986a, b, 1987,  1990a; U.S. EPA 1991b; Alden, 1985; MacNaughton
12    and Uddin, 1984; Alden and Repta, 1984; Phillips et al., 1987). These chemicals have been
13    associated  with interference of normal renal proximal tubule reabsorption of protein from the
14    glomerular filtrate, resulting in accumulation of alpha-2u-globulin in phagolysosomes of renal
15    proximal tubule cells (U.S. EPA, 1991a, b). This accumulation is believed to be the reason for
16    an excessive number of hyaline droplets (Stonard et al., 1986;  Olson et al., 1987) and associated
17    nephropathy. The sequence of functional changes in the epithelial cells of proximal tubules, with
18    subsequent tubule necrosis and compensatory cell proliferation, is hypothesized to culminate in
19    the renal tubule tumors observed in the male rats exposed to these compounds in bioassays
20    (UAREP, 1983; Alden et al., 1984;  Haider  et al., 1984; Swenberg et al.,  1989; U.S. EPA, 1991b).
21          Alpha-2u-globulin is considered unique to the male rat and is the major component of its
22    urinary protein load. Alpha-2u-globulin is  synthesized in the liver under hormonal control, but it
23    has not been detected in the liver of female rats or in other species (Roy  et al., 1975; U.S. EPA,
24    1991b), although homologous proteins do exist in other species, including humans (Flower et al.,
25    1993).
26          The renal tubule tumors associated with alpha-2[j, nephropathy appear to be the end
27    product in  the following histopathological sequence of functional changes in the epithelial cells
28    of proximal tubules:
QQ
m
31       1.  Excessive accumulation due to increased number and size of hyaline droplets in the P2
32          segment of renal proximal tubule cells, representing lysosomal overload, leads to tubule
33          cell degeneration, cell loss, and regenerative cellular proliferation.  The excessive hyaline
34          droplet accumulation occurs in male rats only,  and the  accumulating protein is identified
35          as alpha-2u-globulin.
36
37       2.  Cell debris in the form of granular casts accumulates at the corticomedullary junction,
38          with associated dilation of the affected tubule segment and, more distally, mineralization
39          of tubules within the renal medulla.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-43       DRAFT-DO NOT CITE OR QUOTE

-------
 1       3.  The chronic progressive nephropathy characteristically found in aging rats is exacerbated
 2          as a consequence of the induced nephrotoxicity.
 O
 4       4.  Renal tubule hyperplasia and neoplasia develop subsequently. The increased cellular
 5          proliferation is thought to cause development of renal cell tumors due to increases in
 §          DNA damage  in replicating cells.
 8
 9          This proposed MOA for kidney tumorigenesis seems plausible to many scientists and
10    may provide an adequate explanation of the specific susceptibility of the male rat to renal tubule
11    tumor induction by certain chemicals.  However, data gaps still exist,  and the mechanism of
12    cellular damage in alpha-2u-globulin nephropathy is not known (Melnick et al., 1997).  Several
13    chemical compounds have been shown to cause the acute renal nephropathy associated with
14    alpha-2u-globulin accumulation, but not to cause tumors (Melnick et al.,  1997; Swenberg and
15    Lehman-McKeeman,  1999) even though a critical level of regenerative cellular proliferation may
16    have to be attained for renal tumorigenesis to occur (Swenberg and Lehman-McKeeman,  1999).
17          EPA has developed specific criteria for use in evaluating the likelihood of a chemical's
18    inducing renal tumors through the hypothesized alpha-2u-globulin MOA (U.S. EPA, 1991b).
19    Although EPA downgrades the finding of kidney cancer in male rats as being unimportant to the
20    human situation if it can be shown that the criteria for alpha-2u-globulin are clearly met, the
21    proposed MOA, although reasonable, is still hypothetical, and other reasonable alternative
22    hypotheses have been proposed. As described and discussed below, in the case of
23    tetrachloroethylene, evidence for alpha-2u-globulin accumulation exists only at doses above
24    cancer-causing doses, and other alternative MO As are well supported.
25
26    4.5.4.2.1. Human relevance of alpha-2u-globulin nephropathy.  The U.S. EPA has specific
27    guidance (U.S. EPA, 1991b) for evaluating chemically induced male rat renal tumors to
28    determine the use of the data for human risk assessment. It is interesting to note, however, that
29    controversy still exists within the scientific community regarding this mode of carcinogenic
30    action and its relevance  to human health risk assessment (Lash et al., 2000b; Ashby, 1996; de la
31    Iglesia et al., 1997; Dietrich, 1997; Huff, 1995, 1996; Melnick et al., 1997; Melnick, 2001, 2002).
32
33    4.5.4.2.2. Identification of tetrachloroethylene-specific key events in support of the alpha-2u
34    hypothesis. Goldsworthy et al. (1988) observed increases in alpha-2u-hyaline droplets in
35    exposed male but not  female F344 rats following 10 days of gavage with 1,000 mg/kg
36    tetrachloroethylene. This finding was correlated with both protein droplet nephropathy
37    (crystalloid accumulation) and increases in cellular proliferation.  The cell replication was
38    enhanced in the male rats specifically in damaged P2 segments, suggesting a link between the

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-44       DRAFT-DO NOT CITE OR QUOTE

-------
 1    alpha-2u-globulin accumulation and kidney tumors. These investigators reported similar
 2    findings for pentachloroethane in the same study, but at a dose of 150 mg/kg for 10 days.
 3    Trichloroethylene has a similar structure but did not cause any alpha-2[j, accumulation or increase
 4    in protein droplets, nor did it stimulate cellular proliferation in either male or female rats in this
 5    study when a dose of 1,000 mg/kg was administered for 10 days. Bergamaschi et al.  (1992) also
 6    demonstrated alpha-2u-accumulation in P2 segments of rat proximal tubule cells resulting from  a
 7    daily exposure of rats to 500 mg/kg tetrachloroethylene in corn oil for 4 weeks.
 8           In short-term, high-dose studies, Green et al. (1990) found that the oral administration of
 9    from 1,000 to 1,500 mg/kg of tetrachloroethylene daily for up to 42 days caused an accumulation
10    of alpha-2u-globulin in the proximal tubules of male rats.  The animals were sacrificed within
11    24 hrs of the last dose of tetrachloroethylene. The effect was accompanied by  evidence of
12    nephrotoxicity, with the formation of granular tubular casts and  evidence of tubular cell
13    regeneration. These effects were not observed in female rats or  in mice. Inhalation exposure to
14    1,000 ppm tetrachloroethylene for 10 days resulted in the formation of hyaline droplets in the
15    kidneys of male rats, but granular casts and tubule cell regeneration were not observed, although
16    the time period may have been too short to allow progression to this stage.  These results show
17    that very high doses of tetrachloroethylene are capable of precipitating hyaline droplet
18    nephropathy in male rats.  The results also show that male rats are more sensitive to the effect
19    than are female rats or mice of either sex. It is possible, therefore, that alpha-2u-globulin
20    accumulation may indeed play a role in  the tumorigenesis observed in male rats exposed to
21    tetrachloroethylene. EPA has listed tetrachloroethylene as an alpha-2u-accumulator (U.S.  EPA,
22    1991b) and in the same report specifically identified trichloroethylene as not being an alpha-2u-
23    accumulator.
24           It is interesting to note that tetrachloroethylene-induced alpha-2u-globulin accumulation
25    is probably more likely  to be caused by  the parent molecule rather than by its metabolites,
26    because its occurrence is related more to the charge and lipid solubility of the inducing agent
27    than to specific interactions with reactive chemical species (Lash and Parker, 2001).
28
29    4.5.4.2.3. Points relevant to biological plausibility and coherence of this mode of action
30    (MOA) for tetrachloroethylene-induced kidney tumors.  The following points show that factors
31    other than the specific protein droplet nephropathy may have as  much—or more—of a
32    significant role in explaining renal tumor formation resulting from tetrachloroethylene exposure,
33    although some contributions of alpha-2u-globulin accumulation cannot be entirely ruled out.
34           The alpha-2u-globulin response  reported following exposure to tetrachloroethylene is
35    relatively mild, and the  fact that renal  tumors have been observed at doses lower than the ones
36    shown to cause the alpha-2u-globulin response is inconsistent with this phenomenon  being

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-45       DRAFT-DO NOT  CITE OR QUOTE

-------
 1    responsible for tumorigenesis. Although the alpha-2u-globulin response occurs in male rats
 2    exposed to tetrachloroethylene, it has been observed following only high doses. Green et al.
 3    (1990) tested lower inhaled tetrachloroethylene doses in rats—up to 400 ppm for 6 hrs/day for 28
 4    days, with the animals being sacrificed within 18 hrs of termination of the final exposure—but
 5    found no evidence of hyaline droplet formation; however, there may have been time for recovery
 6    prior to sacrifice.
 7          It is noteworthy that the 400 ppm concentration was the same exposure level used for the
 8    high-dose rats in the NTP inhalation carcinogenicity bioassay (NTP, 1986a). In the NTP study,
 9    the 400 ppm concentration caused a high incidence of nontumor nephropathy and resulted in the
10    formation of kidney tubule adenomas and adenocarcinomas. The renal pathology of rats in the
11    NTP study was reported to be different from the specific alpha-2u-globulin nephropathy, but the
12    age of the rats as well as the length of time that elapsed between final exposure and sacrifice may
13    explain some of the differences.  However, mineralization in the inner medulla and papilla of the
14    kidney—a characteristic trait of alpha-2u-globulin nephropathy—was not seen.
15          Green et al. (1990) proposed the  possibility that longer-term exposure to the 400 ppm
16    concentration of tetrachloroethylene is required for the hyaline droplet accumulation in the
17    kidney of rats.  Alpha-2u-globulin accumulation can be demonstrated, however, after only short-
18    term exposures (even a single administration) to several agents, such as d-limonene, decalin,
19    unleaded gasoline, and trimethylpentane (Charbonneau et al., 1987; NTP, 1988). Lack of
20    hyaline droplet formation, increase in alpha-2u-globulin, or signs of the characteristic renal
21    nephropathy at the high dose  level of the NTP inhalation study (NTP, 1986a) may  indicate a
22    threshold effect and thus diminish the likelihood that the renal tumors associated with exposure
23    to tetrachloroethylene are induced through this mechanism (Green et al., 1990).
24    Pharmacokinetic differences between oral and inhalation exposure may contribute to the
25    observed discrepancies in some of the results.
26          NTP did not report the presence of hyaline droplets in rats that had been exposed to either
27    200 or 400 ppm tetrachloroethylene for up to 2 years.  These doses were associated with the
28    production of renal tubule neoplasms in male rats. However, the fact that NTP did not report the
29    presence of hyaline droplets in the 14-day, 90-day, or 2-year studies is not definitive, because the
30    NTP protocol at that time was not designed specifically to detect hyaline droplets or alpha-2u-
31    globulin accumulation in the kidney (NTP, 1990a). Thus, the procedures followed at the time of
32    the study were not necessarily conducive to detecting  hyaline droplets.  For example, in the
33    chronic study of tetrachloroethylene, at least 1 week elapsed between the final
34    tetrachloroethylene exposure  and the scheduled sacrifice of the surviving animals.  It is possible
35    that had hyaline droplets been present, they could have regressed. Also, the nephropathy
36    observed at the end of a 2-year bioassay  could be difficult to distinguish from the old-age

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-46       DRAFT-DO NOT CITE OR QUOTE

-------
 1    nephropathy that occurs in these rats. Other investigators (Goldsworthy et al., 1988; Green et al.,
 2    1990) observed hyaline droplets containing alpha-2u-globulin following very high doses of
 3    tetrachloroethylene administered to male rats.
 4          On the other hand, the renal pathology reported in the NTP bioassay is not entirely
 5    consistent with the results generally found for chemicals where there is alpha-2u-globulin
 6    accumulation (NTP, 1986a; letter from Scot Eustis, National Toxicology Program, to William
 7    Farland, Director, Office of Health and Environmental Assessment, EPA, 1988).  For example,
 8    there was no mineralization in the inner medulla and papilla of the kidney, a frequent finding in
 9    bioassays of chemicals that induce alpha-2u-globulin accumulation (e.g., for pentachloroethane,
10    the incidence of renal papillar mineralization was 8% in controls, 59% in the low-dose group,
11    and 58% in the high-dose group).  In addition, it is important to note that some aspects of toxic
12    tubular nephropathy were also observed in female rats and male mice exposed to
13    tetrachloroethylene, clearly contrary to sex and species specificity.
14          Thus, chronically induced tetrachloroethylene nonneoplastic kidney lesions exhibit
15    neither species nor sex specificity.  Unlike with other chemicals that induce alpha-2u-globulin
16    accumulation and have been tested by NTP in chronic carcinogenicity bioassays, renal lesions
17    occurring in animals exposed to tetrachloroethylene were not limited to the male rat. Although
18    the female rat did not develop any renal tubule tumors, the incidence of karyomegaly was
19    significantly elevated in the female rat as well as in the male rat; 1 of 50 female rats exposed at
20    the high dose developed tubule  cell hyperplasia.
21          In the mouse, "nephrosis" was observed at increased incidences in dosed females, casts
22    were observed  at increased incidences in dosed males and high-dose females, and karyomegaly
23    of the tubular cells was observed at increased incidences  in both sexes of treated mice.  The
24    severity  of the renal lesions was dose related, and one low-dose male had a renal tubular cell
25    adenocarcinoma.
26          In the NCI gavage study of tetrachloroethylene (NCI,  1977), toxic nephropathy, which
27    was not  detected in the control animals, occurred in both male and female Osborne-Mendel rats
28    administered tetrachloroethylene.  Tetrachloroethylene also clearly caused nephropathy in both
29    sexes of mice in the study. Unfortunately, animal survival in the rat study was not adequate to
30    support any conclusions about tetrachloroethylene carcinogenicity.
31          Other chlorinated ethanes and ethylenes also produce nephrotoxicity and renal tubule
32    tumors in laboratory animals. Hexachloroethane causes accumulation of hyaline droplets and
33    renal tubule tumors in male rats (NTP, 1989). On the other hand, trichloroethylene, which was
34    also tested by NTP, induces kidney tumors in male rats and also possibly in female rats, but it
35    does not cause  an accumulation of hyaline droplets or an increase in levels of alpha-2u-globulin
36    (Goldsworthy et al., 1988). Consequently, kidney tumors induced by this compound are not

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-47       DRAFT-DO NOT CITE OR QUOTE

-------
 1    considered to be associated with alpha-2u-globulin accumulation (U.S. EPA, 1991b, 2001a).
 2    Tetrachloroethylene is related in structure to trichloroethylene, and both chemicals have been
 3    shown to be metabolized in the kidney to cytosolic and mutagenic compounds.
 4
 5    4.5.4.3.  Other Modes of Action for Tetrachloroethylene-Induced Renal Tumors in Rats
 6    4.5.4.3.1. Genotoxicity. The glutathione conjugation of tetrachloroethylene in the kidney,
 7    discussed in Chapter 3, leads sequentially to S(l,2,2-trichlorovinyl)glutathione and
 8    S(l,2,2-trichlorovinyl)cysteine—TCVG and TCVC. TCVC can be further processed by beta-
 9    lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to a highly reactive
10    thioketene, a chemical species that can form covalent adducts with cellular nucleophiles
11    including DNA. Additionally, sulfoxidation of both TCVC and its N-acetylated product via
12    FMO3 or P450s occurs, resulting in reactive metabolites (Ripp et al,  1997, 1999; Werner et al.,
13    1996). While most of these intermediates have not been characterized for mutagenic potential,
14    TCVG, TCVC, and NAcTCVC are clearly mutagenic in Salmonella tests. In addition,
15    tetrachloroethylene exhibited mutagenicity in Salmonella in the few studies of conditions that
16    could generate GSH-derived metabolites and, following in vivo exposures, induces SSB and
17    DNA binding in kidney. See Section 4.3 for more details of genotoxicity.
18
19    4.5.4.3.2. Peroxisomeproliferation.  Peroxisome proliferation has been linked to tumorigenesis
20    in rodents; however, the mechanisms involved have not been clearly  elucidated (see Section 4.3).
21    The PPARs, a class of nuclear receptors, are believed to be transcriptionally activated to mediate
22    the effects of peroxisome proliferators (Issemann et al., 1993; Desvergne and Wahli, 1995).
23    Although most of the focus on peroxisome proliferation and PPAR receptor activation, and their
24    relationship to tumor development, has been on the liver (see Section 4.3.4 for a more detailed
25    discussion), the phenomenon can also occur in other tissues.  In fact,  peroxisomes were first
26    noted in rodent renal tubule epithelial cells, and peroxisome proliferation in these cells in
27    response to peroxisome proliferating agents is not unusual  (reviewed by Stott, 1988; Klaunig et
28    al., 2003). Data exist to support increased peroxisome proliferation in rodent kidney following
29    exposure to tetrachloroethylene (Goldsworthy and Popp, 1987; Zanelli et al., 1996).
30          Goldsworthy and Popp (1987) investigated the  ability of tetrachloroethylene to induce
31    peroxisome proliferation in both liver and kidney of rats and mice using increases in cyanide-
32    insensitive PCO activity as a marker enzyme.  Tetrachloroethylene caused elevations in enzyme
33    activity in mouse kidney as well as liver but not in rat kidney. It seems somewhat unlikely that
34    any peroxisome proliferation observed following tetrachloroethylene exposure would be
35    associated with renal tumors if one considers the magnitude of the measurable peroxisome
36    proliferation effect across species.  Renal tumors occur in rats, but greater peroxisome

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-48       DRAFT-DO NOT CITE OR QUOTE

-------
 1    proliferation is observed in the kidneys of mice (Goldsworthy and Popp, 1987). Modification of
 2    cell signaling pathways that control rates of cell division and apoptosis, for example, may occur
 3    through activation of PPAR receptors.  The occurrence of peroxisome proliferation per se may
 4    be only a marker for PPAR receptor activation, the key event having the most support for being
 5    causally related to other tumor types. The dissimilarity in the peroxisome response observed in
 6    different tissues may be related to variability in the levels of PPAR receptors in these tissues.
 7          It is important to note that, relatively speaking, chlorinated ethylenes, particularly
 8    tetrachloroethylene, and their chloroacid metabolites are not very potent peroxisome proliferators
 9    compared to many other compounds that are known to cause the phenomenon. PPAR receptors
10    belong to the superfamily of proteins that control almost every metabolic and developmental
11    event in mammals, and PPAR receptor activation is known to result in numerous biochemical,
12    physiological, and molecular events. Therefore, the possibility of PPAR receptor activation
13    being related to tetrachloroethylene-induced kidney tumorigenesis in rats is plausible,  especially
14    since peroxisome proliferation, the phenomenon that could be considered a biomarker for PPAR
15    activation, has been shown to occur in the kidneys of mice following exposure to
16    tetrachloroethylene. Moreover, a causal link between PPAR activation and kidney tumorigenesis
17    has yet to be established for tetrachloroethylene or other PPAR agonists.
18
19    4.5.4.3.3. Cytotoxicity/sustained chronic nephrotoxicity not associated with alpha-2fi-globulin
20    nephropathy. The kidney is a major target organ for tetrachloroethylene-induced toxicity
21    through the reactive metabolites of TCVC.  Tetrachloroethylene has been reported to produce
22    nephrotoxicity  across species, although its relative potency is not extremely high.  Other
23    chlorinated ethanes and ethylenes also induce nephrotoxicity, although the toxicity manifests
24    itself differently with specific chemicals. The observed effects vary across species and between
25    sexes and may  include tubular cell cytomegaly, karyomegaly and pleomorphism, tubular cell
26    dilation, or the  formation of granular casts.  There may be a link between renal toxicity and
27    tumorigenesis,  and sustained kidney damage may be a risk factor for tumorigenesis. It is
28    reasonable, therefore, to suspect that renal tubule neoplasia observed in tetrachloroethylene-
29    exposed male rats may be influenced by cytotoxicity and subsequent cellular regeneration.
30          It has been suggested that renal neoplasms induced by tetrachloroethylene may be
31    secondary to renal cytotoxicity and subsequent cellular proliferation without regard to alpha-2u-
32    accumulation.  Thus, sustained chronic nephrotoxicity, independent of alpha-2u-globulin
33    accumulation and its resulting neuropathic cascade of events, may be a possible MO A for
34    tetrachloroethylene carcinogenesis.  If this is the case, renal  tubule neoplasia observed to occur in
35    male rats would not be expected to be a species- or sex-specific response because the nontumor
36    lesions appear in both sexes of both rodent species tested. In support of this expectation, the

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-49      DRAFT-DO NOT CITE OR QUOTE

-------
 1    renal lesions occurring in animals exposed to tetrachloroethylene are not limited to the male rat.
 2    Signs of tetrachloroethylene-induced kidney damage appeared in both rats and mice during the
 3    early phases of the NTP inhalation study, for example, indicating that animals of both species
 4    surviving to the scheduled termination of the study had long-standing nephrotoxicity. Although
 5    the female rats did not develop any renal tubule tumors, the incidence of karyomegaly was
 6    significantly elevated in females as well as in males, and 1 of 50 female rats exposed at the high
 7    dose developed tubule cell hyperplasia.
 8          In the NTP study of the mouse, "nephrosis" was observed at increased incidences in
 9    dosed females, casts were observed at increased incidences in dosed males and high-dose
10    females, and karyomegaly of the tubule cells was observed at increased incidences in both sexes
11    of treated mice. The severity of the renal lesions was dose related, and one low-dose male had a
12    renal tubule cell adenocarcinoma.  In the NCI gavage study of B6C3F1 mice and Osborne-
13    Mendel rats exposed to tetrachloroethylene, toxic nephropathy was not detected in control
14    animals but did occur in both male and female rats as well as in mice. On the other hand,
15    findings using in vitro models studied by Lash et al. (2002) suggest a marked sex difference
16    between male and female rats in the severity of acute renal toxicity caused by both
17    tetrachloroethylene and its TCVG metabolite. Tetrachloroethylene and TCVG also produced
18    signs of toxicity in mitochondria; i.e., mitochondrial dysfunction, such  as inhibition of state 3
19    respiration by specific inhibition of several sulfhydryl-containing enzymes in both sexes of mice
20    (Lash et al., 2000, 2001,2002).
21          Mechanistic  studies of tetrachloroethylene nephrotoxicity are relatively sparse.  More and
22    better data are available for trichloroethylene.  Most studies performed  to elucidate information
23    related to understanding tetrachloroethylene renal  toxicity have concentrated on the GSH
24    pathway metabolites rather than on the parent chemical; this is because much available data for
25    both tetrachloroethylene  and trichloroethylene suggest that it is flux through this pathway  that
26    generates reactive chemical species responsible for nephrotoxicity. Vamvakas et al. (1989c, d)
27    have shown the tetrachloroethylene conjugate metabolites TCVG and TCVC to cause dose-
28    related cytotoxicity in renal cell preparations and prevention of this toxicity by beta lyase
29    enzyme inhibitor.  Renal beta lyases are known to cleave TCVC to yield an unstable thiol,
30    1,2,2-trichlorovinylthiol, that may give rise to a highly reactive thioketene, a chemical species
31    that can form covalent adducts with cellular nucleophiles. Additionally, sulfoxidation of both
32    TCVC and its N-acetylated product occurs, resulting in toxic metabolites (Ripp et al, 1997, 1999;
33    Werner et al., 1996). Contribution to overall toxicity is unknown, however, and it is interesting
34    to note that human CYP3A4 catalyzed sulfoxidation of the N-acetyl metabolite of a structurally
35    related chemical, HCBD, at rates comparable to those of rat CYP3A1 (Werner et al., 1995),
36    indicating a relative  value of the human-to-rat rate constant of 1 (Lash and Parker, 2001).

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-50       DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.5.4.3.4. Immunotoxicity/immunosuppression. Although specific data about
 2    tetrachloroethylene are lacking, immune suppression could contribute to the induction of kidney
 3    tumors caused by tetrachloroethylene exposure. Many immunosuppressive therapeutic agents
 4    are human carcinogens (see Tomatis et al., 1989), although they are usually associated almost
 5    exclusively with lymphoma.  Organic solvent exposure in general is associated with autoimmune
 6    disease such as schleroderma and other autoimmune responses (Nietert et al.,  1998).  Several
 7    published studies have demonstrated formation of trichloroacylated protein adducts, for example,
 8    in liver and kidney of rats (Birner et al., 1994) and in plasma of rats and humans following
 9    exposures to tetrachloroethylene (Pahler et al., 1999). Another recent example is the detection of
10    trichloroacetylated protein adducts formed in mice treated with tetrachloroethylene (Green et al.,
11    2001).
12
13    4.5.4.4.  Summary
14           The kidney is a target organ in mammalian species for tetrachloroethylene and other
15    related chlorinated ethanes and ethylenes, and tetrachloroethylene causes kidney cancer in male
16    rats.  It is likely that several mechanisms contribute to tetrachloroethylene-induced kidney
17    toxicity, including cancer, and that the relative importance of specific MO As varies from high-
18    dose to low-dose exposure. Peroxisome proliferation, alpha-2u-globulin nephropathy,
19    mutagenicity, and cytotoxicity not associated with alpha-2u-globulin accumulation are MO As
20    that have been investigated. The pathogenesis of immunosuppression is another potential MOA
21    that may also be related to tumorigenesis.
22           Tetrachloroethylene-induced renal toxicity is likely associated with its metabolites rather
23    than with the parent compound, except for toxicity associated with alpha-2u-globulin
24    accumulation, which is more likely due to tetrachloroethylene itself (Lash and Parker, 2001).
25           The GSH conjugate and reactive metabolites generated from further processing of TCVC
26    or its acetylated metabolite NAcTCVC by beta lyase, FMO3/P450 and/or CYP3 A, are the most
27    likely tetrachloroethylene metabolites to induce renal toxicity and tumorigenicity, as opposed to
28    the metabolites resulting from oxidative CYP processing. These conjugate metabolites are
29    known to be mutagenic, and they are known to occur in rats, mice, and humans.
30          Due to tetrachloroethylene's nephrotoxic  effects, it has been suggested that the low-level
31    renal tumor production observed in exposed rats is secondary to sustained cytotoxicity and
32    necrosis leading to activation of repair processes  and cellular regeneration. However,
33    "nephrotoxicity" occurs in both sexes of rats and mice whereas cell replication and
34    tumorigenesis occurs in male but not in female rats In addition, tetrachloroethylene induces
35    kidney tumors at lower doses than those required to cause alpha-2u-globulin accumulation,
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-51       DRAFT-DO NOT CITE OR QUOTE

-------
 1    raising serious doubt that alpha-2u-globulin plays a key role—especially any major role—in the
 2    rat kidney tumor formation.
 3          Because tetrachloroethylene has been shown to induce peroxisome proliferation, an
 4    indicator of PPAR activation, the possibility exists that certain responses resulting from
 5    activation of PPAR receptors might be involved in cancer-causing activity leading to
 6    tetrachloroethylene-induced renal tumors.  However, there is no evidence causally linking
 7    PPAR-a activation to kidney tumorigenesis for tetrachloroethylene or other compounds.
 8          The weight of evidence suggests that for tetrachloroethylene the further processing of
 9    conjugative metabolites by beta lyase, FMO3 and/or CYP3 A leads to reactive and mutagenic
10    metabolites responsible for nephrotoxicity and carcinogenicity.  The glutathione conjugation of
11    tetrachloroethylene in the kidney, discussed in Chapter 3, leads sequentially to
12    S(l,2,2-trichlorovinyl)glutathione and S(l,2,2-trichlorovinyl)cysteine—TCVG and TCVC.
13    TCVC can be further processed by beta-lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol,
14    that may give rise to a highly reactive thioketene, a chemical species that can form covalent
15    adducts with cellular nucleophiles including DNA.  TCVC can also undergo FMO3 or P450
16    oxidation to reactive intermediates; additionally, sulfoxidation of both TCVC  and its N-
17    acetylated product occurs, resulting in reactive metabolites (Ripp et al, 1997, 1999; Werner et al.,
18    1996). While most of these intermediates have not been characterized for mutagenic potential,
19    TCVG, TCVC, and NAcTCVC are clearly mutagenic in Salmonella tests.  In addition,
20    tetrachloroethylene exhibited mutagenicity in Salmonella in the few studies of conditions that
21    could generate GSH-derived metabolites. Tetrachloroethylene, following in vivo exposures, also
22    binds to kidney DNA and induces SSB in kidney.
23          In summary, the complete mechanisms of tetrachloroethylene-induced renal
24    carcinogenesis are not yet understood.  Given the known mutagenicity of the GSH-derived
25    tetrachloroethylene metabolites that are formed in the kidney, and the observed in vitro
26    mutagenicity of tetrachloroethylene under conditions that would generate these metabolites, a
27    mutagenic MOA contributing to the development of the kidney tumors clearly cannot be ruled
28    out.
29
30    4.6. NEUROTOXICITY
31    4.6.1. Human Studies
32          A wide range of effects on neurologic function are well-documented for both acute and
33    chronic exposure to tetrachloroethylene. Acute controlled inhalation exposures of 100 ppm and
34    higher have induced symptoms consistent with depression of the CNS and included dizziness
35    and drowsiness (ATSDR, 1997). Changes in electroencephalograms (EEGs) have also been
36    noted with controlled inhalation  exposures at this level (Stewart et al., 1977).  Acute exposure to
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-52       DRAFT-DO NOT CITE OR QUOTE

-------
 1    lower levels of tetrachloroethylene has induced alterations in neurobehavioral function.  For
 2    example, Altmann et al. (1990, 1992) reported increases in the latency of visually evoked
 3    potentials (VEPs) and significant performance deficit for vigilance and eye-hand coordination in
 4    subjects with inhalation exposure to 50 ppm for 4 hrs/day for 4 days as compared with that seen
 5    among subjects exposed to 10 ppm (the control group in these  reports). These observations
 6    indicate visual system dysfunction, including delayed neuronal processing time, is related to
 7    tetrachloroethylene exposure. ATSDR (1997) developed an acute exposure minimal risk level
 8    from this study and considered 10 ppm as the no-observed-adverse-effect level  (NOAEL).
 9          Studies by Stewart et al. (1997) and Hake and Stewart (1977), funded primarily by the
10    National Institutes of Occupational Safely and Health (NIOSH), are third-party studies and are
11    considered by EPA to be protective of human  subjects. EPA and other federal agencies
12    subscribe fully to principles articulated in EPA's Protection of Human Subjects Rule ("the
13    Common Rule"), 40 CFR Part 26.  EPA recently issued an interim policy on accepting human
14    test data stating that the Agency will continue to accept third-party test data on a case-by-case
15    basis (U.S. EPA, 2005d). A description of the studies by Altmann et al. (1990,  1992) is also
16    included because ATSDR used these studies to develop an acute minimal risk level (MRL;
17    ATSDR, 1977).  EPA also considers these studies to be third-party. No information is provided
18    in the published papers regarding the procedures the study investigators adopted for informed
19    consent or protection of human subjects, and staff of the National Center for Environmental
20    Assessment (NCEA) contacted study investigators requesting this information (e-mail
21    communication from Robert McGaughy, U.S. EPA, to L. Altmann, Heinrich-Heine University,
22    Dusseldorf, Germany, October 8, 2003).
23          Epidemiologic studies of workers or residents with chronic exposure to
24    tetrachloroethylene show that the nervous system is a target, with decrements reported in several
25    nervous system domains (Altmann et al., 1995; Schreiber et al., 2002;  Seeber, 1989; Ferroni et
26    al., 1992; Cavalleri et al., 1994; Gobba et al., 1998; Spinatonda et al., 1997; Echeverria et al.
27    1994, 1995). Table 4-4 presents select details of available chronic studies evaluating
28    neurological function.  Furthermore, neurotoxic effects from chronic-duration exposure to
29    tetrachloroethylene are similar to neurotoxic effects reported for other solvents (Arlien-Sorborg,
30    1982).  Case reports and case studies also show the nervous system as a target of organic solvent
31    exposures such as tetrachloroethylene  (Seppalainen and Antti-Poika, 1983; Antti-Poika,
32    1982a, b; Onofirj et al., 1998). Electrophysiological abnormalities in tetrachloroethylene- and
33    other organic solvent-exposed patients, with diagnosed chronic solvents intoxication, persisted
34    post-exposure (Seppalainen and Antti-Poika, 1983; Antti-Poika, 1982b).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-53       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Table 4-4.  Summary of human neurotoxicology studies
Subjects
24 dry cleaners exposed to
tetrachloroethylene, 33
subjects nonoccupationally
exposed to other solvents
101 dry cleaners (both
sexes), 84 nonexposed
controls
56 tetrachloroethylene-
exposed dry cleaning
workers, 69 non-exposed
factory controls
64 dry cleaners exposed to
tetrachloroethylene,
120 controls
Dry cleaners, 60 females
and 30 nonsolvent-exposed
controls
22 dry cleaners, 13 ironers,
35 controls
Exposure
Mean TWA = 21 ppm, PM
Mean duration of exposure
= 6 years
Low-exposure group: 12
ppm TWA, IA, PM, 11.8
years
High-exposure group: 53
ppm TWA IA, PM, 10.6
years
Geometric mean TWA =
20 ppm, PM
Mean duration of exposure
= 3 years
Mean duration of
employment not reported
Geometric mean =15 ppm
(males), 1 1 ppm; (females),
PM
15 ppm IA, 10.1 years
Mean TWA = 6 ppm
(7 ppm, dry cleaning
workers; 5 ppm, ironers),
PM, 8.8 years (exposed
subjects)
Effects
Statistically significant
differences for simple
reaction time and critical
flicker fusion
Decrease in information
processing speed
(perceptual threshold,
choice reaction time),
visual scanning
(cancellation dZ test),
visuo-spatial function
No fine motor function
deficits
Neurological signs
Statistically significant
increase in the prevalence
of several CNS symptoms
such as nasal irritation and
dizziness
No statistically significant
differences between
exposed and controls in
color vision loss
Impaired performance on
three tests (simple reaction
time, vigilance, stress)
No fine motor function
deficit
No effects on digit symbol
test
CCI statistically
significantly elevated
among dry cleaners with a
statistically significant
exposure (TWA) -response
relationship.
No effect on CCI in ironers
Author(s)
Lauwerys et al.
(1983)
Seeber(1989)
Caietal. (1991)
Nakatsuka et al.
(1992)
Ferroni et al.
(1992)
Cavalleri et al.
(1994)
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   4-54       DRAFT-DO NOT CITE OR QUOTE

-------
       Table 4-4. Summary of human neurotoxicology studies (continuted)
Subjects
45 dry cleaners matched to
69 laundry workers, 59
pressers or counter clerks
65 dry cleaners, no
unexposed group
Residents near dry cleaning
facilities, 14 exposed and 23
age- and gender-matched
nonexposed controls
24 solvent-exposed workers
and 24 controls
35 dry cleaners, 39 age- and
education-matched controls
Exposure
<0.2 ppm, 3 ppm, 9 ppm,
PM, 2. 6 to 11 years
1 1 ppm, 23 ppm, 41 ppm,
PM, 1.2 to 14.6 years
0.7 ppm (mean), IA (7-day
monitoring period), B, 10.6
years
Past use of solvent mixture,
including tetrachloro-
ethylene (10%) for at least
3 years, solvent exposure
for 6 years average
8 ppm TWA, IA, grab
sample
Mean duration of
employment not reported
Effects
Statistically significant
association between chronic
lifetime tetrachloroethylene
exposure and reduced test
performance on three
cognitive tests: switching,
pattern memory, and
pattern recognition.
Statistically significant
differences between "high"
and "low" lifetime
exposure groups in three
tests of visuo-spatial
function
No effect on digit span test
Increase in simple reaction
time
Decrements in continuous
performance and visuo-
spatial function
No fine motor function
deficits
Color vision impairment
(p < 0.05) among exposed
subjects as compared with
controls
Increase in vocal reaction
time to visual stimuli
Concentration-response
relationship
Author(s)
Echeverria et al.
(1994)
Echeverria et al.
(1995)
Altmann, et al.
(1995)
Muttray et al.
(1997)
Spinatonda et al.
(1997)
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   4-55       DRAFT-DO NOT CITE OR QUOTE

-------
        Table 4-4. Summary of human neurotoxicology studies (continuted)
         Subjects
        Exposure
         Effects
   Author(s)
33 dry cleaners and ironers,
self control (follow-up of
Cavalleri et al, 1994,
subjects)
Group A, 4 ppm, PM

Group B, 0.7 ppm, PM
Worse CCI in subjects who
experienced a higher mean
TWA exposure than that
reported in Cavalleri et al.

No impairment in CCI in
subjects who experienced a
lower mean TWA exposure
than that reported in
Cavalleri et al.
Gobba et al.
(1998)
30 patients with solvent-
induced encephalopathy
Several solvents, including
tetrachloroethylene, 20
years
No peripheral neuropathy
Albers et al.
(1999)
Apartment residents living
above dry cleaning
facilities:  17 exposed and
17 age- and gender-matched
unexposed controls
0.4 ppm (mean,
monitoring taken before
closure of dry cleaners) IA,
PM, B, U

Duration of residence, 6
years (mean)
No effects on visual acuity
or color discrimination
(comparison of group
means)

Lower (worse) scores on
tests of visual contrast
sensitivity
Schreiberetal.
(2002)
Employees of a day care
facility located in the same
building as a dry-cleaning
business, 9 exposed and 9
age- and gender-matched
unexposed controls
0.32 ppm (mean,
monitoring taken before
closure of dry cleaners) IA

No information on duration
of employment
No effects on visual acuity
color discrimination
(comparison of group
means)

Lower (worse) scores on
tests of visual contrast
sensitivity
Schreiberetal.
(2002)
 14 dry cleaners exposed to
tetrachloroethylene, 27 non-
exposed (Control 1) and 27
support staff of Universiti
Kebangsaan Malaysia
(Control 2)
No exposure information
presented in paper other
than tetrachloroethylene
was used for dry cleaning
No effect on color vision
using Ishihara plates. 43%
and 93% of dry cleaners
compared to no controls
had errors on the color
vision D-15 test and FM
100 Hue test, respectively.
Dry cleaners had more
errors on FM 100 Hue than
control group 2 (p < 0.05)
but not control group 1.
Sharanj eet-Kaur
et al. (2004)
            This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                      4-56       DRAFT-DO NOT CITE OR QUOTE

-------
             Table 4-4.  Summary of human neurotoxicology studies (continuted)
              Subjects
        Exposure
          Effects
   Author(s)
     65 households (67 adults
     and 68 children) in
     residential buildings with
     co-located dry cleaners, 61
     households (61 adults and
     71 children) in residential
     buildings without dry
     cleaners
Geometric mean = 5 ppb
(0.005 ppm), IA

Duration of residence, 10
years (mean)
Association (p < 0.05)
between tetrachloroethylene
concentrations in indoor air
and in blood and
performance on test of
visual contract sensitivity in
children. No association
observed in adults.

Color vision impairment
(p < 0.05) among children
but not adult exposed
subjects as compared with
controls
NYSDOH
(2005a),
McDermott et al.
(2005)
     4-year follow-up of 13
     former students, 13 children
     matched to exposed
     children on age, gender, and
     daycare experience
Exposure had ceased 4
years earlier.
No difference in visual
function (VCS, color
vision) or neurobehavioral
function between exposed
children and controls.
NYSDOH
(2005c)
     88,820 births from
     1964-1976 identified from
     the Jerusalem Perinatal
     Study and linked to Israel's
     national Psychiatric
     Registry for hospitalization
     with a schizophrnia-related
     diagnosis as of 1-1-98.
Occupation of mother and
father on birth certificate as
dry cleaner.
Four cases were identified
in 144 offspring of dry
cleaners. Unadjusted
relative risk (RR) of 3.4
(95%CI=1.3-9.2)for
schizophrenia in the
offspring of dry cleaners
using proportional hazard
modeling. Control for a
number of potentially
confounding variables did
not lead to appreciable
different RR estimates.
Perrin et al.
(2007)
2    B      = Biological monitoring of blood
3    IA     = Indoor air
4    PM    = Personal monitoring of breath
5    U      = Biological monitoring of urine for trichloroacetic acid
6    VCS   = visual contrast sensitivity
                 This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                      4-57        DRAFT-DO NOT CITE OR QUOTE

-------
 1          Additionally, other reports (Laslo-Baker et al., 2004; Till et al. 200la, b, 2005) suggest a
 2    vulnerability of the fetus to organic solvent exposures, including tetrachloroethylene exposure.
 3    Deficits in neurobehavioral parameters and in visual system functioning in young children of
 4    mothers exposed during pregnancy were observed in these reports. These reports are not fully
 5    discussed in this section. Case series help identify target organ toxicity and can support
 6    generating hypotheses; however, the lack of information in an unexposed population in these
 7    types of studies limits the ability to infer observations reported among cases to other populations.
 8          Most human data on tetrachloroethylene exposure and nervous system effects—from
 9    cross-sectional or prevalence studies—is of dry cleaner and laundry workers and assessment of
10    neurobehavioral effects; one report on neurobehavioral  effects is of residents living in close
11    proximity to a dry cleaning establishment.  Three studies assessed the visual system; two reports
12    of the same population are of dry cleaner and laundry workers, and one report is of two
13    populations living or working in a building co-located with a dry cleaning establishment.  Few
14    studies are available on neurologic diseases such as Parkinson's disease, amyotrophic lateral
15    sclerosis, and Alzheimer's disease and organic solvents (IOM, 2002), and none of these reports
16    uniquely assess tetrachloroethylene.
17          Tetrachloroethylene  concentrations reported in the dry cleaning and laundry worker
18    studies ranged from an 8-hr  TWA mean of 6 ppm for dry cleaner and laundry workers in
19    Cavalleri et al.  (1994) to an  8-hr TWA of 41 ppm for operators of wet-transfer dry cleaning
20    machine in Echeverria et al.  (1995). Tetrachloroethylene concentrations reported for exposed
21    residents were 0.4 ppm (mean) to residents living in an  apartment building containing an
22    operating dry cleaning business (Schreiber et al., 2002) and 0.7 ppm  (mean) in indoor air of
23    residents in close proximity  to a dry cleaning business (Altmann et al., 1995).
24
25    4.6.1.1. Environmental Chamber Studies
26    4.6.1.1.1. Stewart, R.O., E.D. Baretta, H.C. Doddand T.R. Torkelson. 1970.  Experimental
27    human exposure to tetrachloroethylene. Arch. Environ. Health. 20:225—229.  In a study by
28    Stewart et al. (1970), 12 healthy adults were exposed to 100 ppm for 7 hrs; eye and nose
29    irritation was reported by 60% of the subjects,  a slight frontal headache by 26%, slight light-
30    headedness by 26%, feeling  slightly sleepy by  40%, and difficulty in speaking  by 25%.  Some of
31    these complaints were made during the first 2 hrs.  Of five healthy men exposed to 100 ppm for
32    7 hrs/day on 5 consecutive days, one reported a mild frontal headache during each exposure and
33    two consistently reported mild eye and throat irritation.  Other symptoms were not reported, and
34    individual responses during  exposures to 0 ppm were not assessed. Subjects reported that their
35    ability to detect the odor of tetrachloroethylene decreased during the  course of daily and weekly
36    exposure.

                 This document is a  draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-58      DRAFT-DO NOT  CITE OR QUOTE

-------
 1          Three tests of equilibrium (a modified Romberg test,2 where an individual stands on one
 2    foot with eyes closed and arms at side; a heel-to-toe test; and a fmger-to-nose test) were
 3    performed every 60 minutes during each day of exposure. After 6 hrs, neurobehavioral tests of
 4    motor function (the Crawford manual dexterity and Flanagan coordination tests), cognitive
 5    function (arithmetic test), and motor/cognitive function (inspection test) were also performed.
 6    Three of the subjects had increased difficulty in maintaining their equilibrium when tested within
 7    the first 3 hrs of exposure (i.e., performance on the Romberg equilibrium test was impaired).
 8    The three subjects were able to perform the test normally when given a second chance.
 9    Performance on the other tests was not impaired. An additional subject, exposed during the third
10    day of testing, showed a slight deterioration in his Romberg test and complained of slight
11    dizziness and slight impairment of his intellectual faculties after 1 hr of exposure.  No
12    improvement in his Romberg test occurred during the next hour, and he was removed from the
13    test chamber.  The subject performed the test normally when retested 30 minutes later.
14          The investigators concluded that there were CNS effects in some subjects exposed to 100
15    ppm and that there exists a large range of individual susceptibility to tetrachloroethylene.  The
16    latter conclusion was based on the observations that only 3 of 17 subjects reported light-
17    headedness and 4 of 17 subjects had an abnormal Romberg test.
18
19    4.6.1.1.2. Hake, C.L., andR.D. Stewart. 1977. Human exposure to tetrachloroethylene:
20    inhalation and skin contact Environ. Health Perspect. 21:231-238.  As part of a 6-week study,
21    four healthy men were exposed 7.5 hrs/day to 0 ppm (2 days in week one,  1 day in week three,
22    and 2 days in week six), 21  ppm (4 consecutive days in week three), 100 ppm (5 consecutive
23    days in week two), and a TWA of 100 ppm (5  consecutive days in week four) when exposure
24    levels were more than 53, 100, or 155 ppm (5 consecutive days during week five). In addition,
25    four healthy women were exposed to 100 ppm for 7.5 hrs/day on 5 consecutive days and to 0
26    ppm on 2 days. A fifth woman became sick with influenza during the study and was exposed to
27    100 ppm on only 2 days.
28          All subjects were cautioned to either abstain from the use of alcohol and drugs (or to limit
29    their use to very moderate amounts) and they were asked not to drink coffee until 1 hr after the
30    end of each exposure period.  The subjects were told that they would be exposed to various
31    concentration of tetrachloroethylene, but they were not told their sequence of exposures (a
32    single-blind protocol). All subjects were sedentary during exposure except that men exercised
33    on a bicycle ergometer for 6 minutes at 1 and 5 hrs of exposure.
            2 The Romberg test measures CNS depression.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-59       DRAFT-DO NOT CITE OR QUOTE

-------
 1          Reports of symptoms (e.g., headache) varied among individuals but, overall, complaints
 2    during exposures were similar to those during exposures to 0 ppm of tetrachloroethylene. All
 3    subjects were able to detect the odor of tetrachloroethylene at all levels of exposure immediately
 4    upon entering the chamber; thereafter, they varied in their ability to detect odors.  Some subjects
 5    retained the ability to detect odors during the entire experimental period, particularly at 155 ppm.
 6    A few other subjects detected no odor after the first hour of the first day.
 7          The evaluation of EEG recordings made during exposure suggested altered patterns
 8    indicative of cortical depression in three of four men and four of five women exposed to 100
 9    ppm (constant or TWA). In five subjects, altered EEG recordings occurred during hours 4 to 7
10    of exposure;  another subject had altered recordings within 10 minutes of exposure, which
11    gradually returned to normal during continued exposure, and the seventh subject showed changes
12    between 30 minutes and 6-7 hrs  of exposure.  Recordings of VEPs in response to bright flashes
13    of light (i.e.,  neurophysiological  measurements of the  electrical  signals generated by the visual
14    system in response to visual stimuli) and equilibrium tests (Romberg and heel-to-toe) were
15    normal in men and women.
16          The performance of men  on neurobehavioral tests of cognitive function (arithmetic),
17    motor function (alertness), motor/cognitive function (inspection), and time estimation were not
18    significantly affected by any exposure.  The performances of men on a second test of motor
19    function (Flanagan coordination) were significantly decreased (p < 0.05) on 1 of 3 days during
20    each of 2 weeks of exposure to 100 ppm and on 2 of 3 days during the week of exposure to 155
21    ppm, but the investigators concluded that only the results at 155 ppm were related to
22    tetrachloroethylene.  In women, alertness (the only neurobehavioral endpoint evaluated) was not
23    affected by exposure to tetrachloroethylene.
24          The study authors concluded that (1) there is considerable interindividual variation in
25    response to tetrachloroethylene vapors, (2) EEG analysis indicates preliminary signs  at narcosis
26    in most subjects exposed to  100 ppm for 7.5 hrs, (3) impairment of coordination may occur in
27    subjects exposed to 155 ppm for 7.5 hrs, and (4) the effects are likely due to tetrachloroethylene
28    itself, given its slow metabolism  in humans. They also reported that their data suggested that a
29    threshold limit value of 100  ppm contains no margin of safety for susceptible subjects—both
30    subjectively and neurologically—to the vapors of tetrachloroethylene,  a surprising finding,  given
31    the study's sample size.
32
33    4.6.1.1.3. Hake, C.L., R.D.  Stewart, A. Wu, J. Kalbfleisch, RE. Newton, S. K. Marlow andM.
34    Vucicevic-Salama. 1977. Effects of perchloroethylene/dmg interaction on behavior and
35    neurological function. DHEW (NIOSH) Publ No. 77-191.  Stewart et al. (1977) conducted a
36    complex study with 12 healthy adults (6 men and 6 women) on the behavioral effects of inhaled

                This document is a draft for review purposes only  and does not constitute Agency policy
      06/06/08                                   4-60       DRAFT-DO NOT CITE OR QUOTE

-------
 1    tetrachloroethylene and oral doses of alcohol or Valium. Individuals were typically exposed for
 2    6.5 hrs to 0 ppm on Monday or Tuesday, 100 ppm on Wednesday and Friday, and 25 ppm on
 3    Thursday during each of the 11 weeks of exposure and were given a placebo capsule, alcohol,
 4    Valium, or nothing during each period.  Numerous neurological tests were performed throughout
 5    each exposure, and all testing was done in a double-blind mode (neither the testers nor the
 6    subjects were told the exposure level).
 7          Exposure to 25 or 100 ppm of tetrachloroethylene for 6.5 hrs did not increase the overall
 8    prevalence of reported symptoms (e.g., headache) or alter the subjects' mood. There were
 9    exposure-related increases in the strength and persistence of the tetrachloroethylene odor
10    perceived by the subjects.  Exposure did not significantly reduce performance on two
11    equilibrium tests (Romberg and heel-to-toe) and two neurobehavioral tests of motor function
12    (Michigan eye-hand coordination test and rotary pursuit test).  At 100 ppm (but not 25  ppm)
13    there was a significant decrease (p < 0.05) in scores  on a third test of motor function (Flanagan
14    coordination test) on some days of exposure. Statistical analyses performed by the investigators
15    revealed no effect of tetrachloroethylene exposure alone on EEGs and no significant interactive
16    effects between tetrachloroethylene and either alcohol or Valium.
17          The study authors reported that exposure to 100 ppm tetrachloroethylene did not have a
18    significant consistent effect on performance, although it did have a significant but inconsistent
19    detrimental effect on the performance of the Flanagan coordination test (given during the 3rd and
20    4th hrs of exposure).
21
22    4.6.1.1.4. Altmann, L., A. Bottgor andH.  Wiegand. 1990. Neurophysiological and
23   psychophysical measurements reveal effects of acute low-level organic solvent exposure in
24    humans. Int. Arch. Occup. Environ. 3:493-499.  Altmann, L., H. Wiegand, A. Bottger, F.
25    Eistwmelor and G. Winneke.  1992. Neurobehavioral and neurophysiological outcomes of
26    acute repeatedperchloroethylene exposure. Appl. Psych. 41:269-279. This study, conducted in
27    Germany, reports intentional inhalation exposure of human subjects to  tetrachloroethylene for
28    the purpose of measuring potentially adverse health  outcomes.  No information is provided about
29    ethical principles espoused by the U.S. government for exposure to human subjects.  Therefore,
30    the principal  author was contacted by EPA staff (e-mail communication from R. McGaughy,
31    EPA to L. Altmann, October 8, 2003). Information was requested regarding procedures that
32    were used to select the subjects and inform them about the  nature of the exposure, institutional
33    procedures that were taken to  review the design of the study, and ethical standards and guidelines
34    that the institution was operating under at the time of the study. No response had been received
35    by EPA staff as of October 19, 2004. Although the report is not of crucial importance  in the
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-61       DRAFT-DO NOT CITE OR QUOTE

-------
 1    evaluation of chronic neurotoxic effects of tetrachloroethylene, the scientific and ethical issues
 2    associated with intentional dosing of human subjects is of importance to EPA.
 3          Altmann et al. (1990, 1992) used neurophysiological and neurobehavioral techniques to
 4    evaluate the neurological effects of tetrachloroethylene on healthy adults exposed to 10 ppm or
 5    50 ppm for 4 hrs on 4 consecutive days. All subjects denied prior occupational exposure to
 6    solvents and drug use at the time of the study. Visual acuity of all subjects was normal or
 7    corrected to normal.  The study was a single-blind study (subjects were not told their level of
 8    exposure), and subjects were randomly assigned to either group.  Sixteen subjects were exposed
 9    to 10 ppm and 12 subjects were exposed to 50 ppm, but neurophysiological measurements were
10    made on only 22 subjects (12 at the low level and 10 at the high level).  There was no unexposed
11    control group.
12          Three neurophysiological measurements were taken on the day before exposure started
13    and on each of the four exposure days:  (1) VEPs in response to black-and-white checkerboard
14    patterns were measured; the VEPs of some subjects (exact number not reported) were also
15    measured on the day after exposure ceased; (2) a visual contrast sensitivity (VCS) test (a test of
16    the central spatial vision that determines the minimum contrast necessary for an individual to see
17    patterns of various sizes) was given to five subjects (three from the low-exposure group and two
18    from the high-exposure group); (3) recordings of brainstem auditory-evoked potentials
19    (neurophysiological measurements of the electrical signals generated by the hearing system in
20    response to auditory  stimuli) were made to evaluate peripheral hearing loss. All measurements
21    were started 2 hrs after a subject entered the chamber and were completed within  1 hr.
22          A German version of the Neurobehavioral Evaluation System was used to assess motor,
23    motor/cognitive, and cognitive function of subjects. The battery included nine tests (finger
24    tapping,  eye-hand coordination, simple reaction time, continuous performance, symbol digit,
25    visual retention, pattern recognition, digit span, and paired associates).  A vocabulary test and a
26    test of emotional state (moods) were also given. Each subject was assessed with a complete
27    battery of tests during the pre-exposure baseline assessment and at the end of the study.  Subsets
28    of the battery covering motor function and mood were given repeatedly at the beginning and end
29    of each 4-hr exposure period.
30          Tetrachloroethylene was not detected in blood samples collected before the start of the
31    first exposure period. The detection limit was less than 0.0005 mg/L. Mean tetrachloroethylene
32    blood levels increased slightly over the 4-day period.  Among subjects exposed to 10 ppm, mean
33    blood levels were 0.33, 0.36, 0.4, and 0.38 mg/L at the end of days one, two, three, and four of
34    exposure, respectively. Among subjects exposed to 50 ppm, mean blood levels were 1.1, 1.2, 1.4,
35    and 1.5 mg/L at the end of days one, two, three, and four of exposure, respectively.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-62       DRAFT-DO NOT CITE OR QUOTE

-------
 1          On the first day of testing, a faint solvent odor was reported by 33% and 29% of the
 2    subjects exposed to 10 ppm and 50 ppm, respectively. On the fourth day, these incidences
 3    changed to 17% and 36%, respectively. The VEP latencies of subjects during the 3rd hr of
 4    exposure to 50 ppm on days one, two, three, and four of exposure were significantly longer
 5    (p< 0.05) from those measured on the control day, and the differences became progressively
 6    longer on successive exposure days. One set of VEP latencies on the day after the end of the
 7    exposure period remained longer than the control day values (statistical significance not
 8    reported).  VEP  latencies in subjects with exposure to 10 ppm were not statistically significantly
 9    longer than those recorded on the control day. There were significant differences (p < 0.05)
10    between the VEP latencies of subjects exposed to 10 ppm and those exposed to 50 ppm.
11          Data on contrast sensitivity indicated greater effects at 50 ppm than at 10 ppm; effects
12    were most pronounced on the last day of exposure.  However, statistical analysis was not
13    reported, and the data are limited by the small number of subjects. There were no indications of
14    peripheral hearing loss at either exposure level.
15          Neurobehavioral tests results were reported for only those tests given repeatedly on 4
16    consecutive days (finger tapping, eye-hand coordination test, simple reaction time, continuous
17    performance, and moods). There were significant post-exposure performance deficits (p 0.05)
18    among subjects exposed to 50 ppm when compared with the group exposed to 10 ppm in tests of
19    motor/cognitive function (continuous performance test for vigilance) and motor function (eye-
20    hand coordination), and a near-significant difference (p = 0.09) on a test of motor function
21    (simple reaction time). In all  cases, the degree of improvement shown by the subjects exposed to
22    50 ppm was less than that shown by the subjects exposed to 10 ppm. There were no exposure-
23    related effects on the finger-tapping or moods test.
24          The study authors concluded that visual function in healthy, young, adult males is mildly
25    affected by tetrachloroethylene exposures to 50 ppm maintained for 4 hrs on each of 4 days
26    (Altmann et al.,  1990), and they stated that the impaired performance on tests of motor/cognitive
27    and motor function suggests that 50 ppm cannot be considered a NOAEL for neurobehavioral
28    end-points indicative of CNS  depression (Altmann et al., 1992).
29
30    4.6.1.2. Chronic Exposure Studies
31    4.6.1.2.1. Lauwerys, R., J. Herbrand, J.P. Bucket, A. Bernard and J. Gaussin. 1983. Health
32    surveillance of workers exposed to tetrachloroethylene in dry-cleaning shops. Int. Arch. Occup.
33    Environ. Health. 52:69-77. Lauwerys et al. (1983) studied 263  workers (24 women and 2 men)
34    occupationally exposed to tetrachloroethylene in six dry cleaning shops in Belgium for a mean of
            3
             Abstract of paper reports 22 subjects were exposed to perchloroethylene.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-63       DRAFT-DO NOT CITE OR QUOTE

-------
 1    6.4 years.  The authors evaluated potential effects on the CNS, the kidney, the liver, and the
 2    lungs in these workers and in controls (31 women and 2 men) working in a chocolate factory
 3    (20) or an occupational health service (13) who did not report occupational exposure to organic
 4    solvents. No information is provided in the paper on the methods used to identify subjects or
 5    their reasons for participating in the study. Several characteristics of the two groups were similar
 6    (sex ratio, mean age [32.9 vs. 34.5 years], and level of education).  However, 13 of the 26 dry
 7    cleaning workers—but only 9 of the 33 controls—were smokers.  Neurobehavioral tests of motor
 8    function (simple and choice reaction time), sensory function (critical flicker fusion), and
 9    cognitive function (sustained attention test) were given twice to each worker, once before work
10    and once after work. Both groups were tested in the middle of the work week.  Individuals also
11    were questioned about chronic symptoms related to nervous system disturbances.  Blood samples
12    were collected both before and after work. Urine samples for kidney function tests were
13    collected after work.
14           The mean tetrachloroethylene air concentration (8-hr TWA) was 21 ppm and the range of
15    TWA values was 9 to 38 ppm, using results from active sampling of personal air.  The mean
16    tetrachloroethylene blood level (30 minutes after the end of work) was 1.2 mg/L (range of means
17    from the shops was 0.6 to 2.4 mg/L).  There was no significant connection between air
18    concentrations and blood levels. Trichloroacetic acid, a metabolite of tetrachloroethylene, was
19    not detected (LOD not identified in published paper) in urine specimens from exposed subjects.
20           Seventeen of 22 symptoms related to nervous  system disturbances were reported by study
21    investigators as being more prevalent among the workers than among unexposed controls.
22    Although statistical testing was conducted, the study investigators did not describe statistical
23    methods or tests in the published paper. The investigators reported no statistically significant
24    differences in prevalences for most symptoms and no relationships with duration of exposure.
25    Lauwerys et al. (1983) reported more complaints in the exposed group than in control workers,
26    particularly memory loss (7/26 exposed vs. 3/33controls) and difficulty falling asleep (11/26
27    exposed vs. 6/33 controls). EPA analysis of the data found the latter complaint to be statistically
28    significant using Fisher's exact test (p = 0.04). The mean cycles/second, the score of the critical
29    flicker fusion test (a test  of sensory function) was significantly increased (better performance) in
30    the exposed group than in controls when given both before work and after work.  The dry
31    cleaning workers did not differ from controls on the other three neurobehavioral tests.  The
32    prevalence of abnormal scores (those beyond the 5th or 95th percentile  of the control group) did
33    not vary significantly between the two groups.
34
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-64      DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.6.1.2.2. Seeber, A. 1989. Neurobehavioral toxicity of long-term exposure to
 2    tetrachloroethylene. Neurotoxicol. Teratol. 11:579—583.4  Seeber (1989) evaluated the
 3    neurobehavioral effects of tetrachloroethylene on 101 German dry cleaning workers (machine
 4    operators, ironers, touch-up workers, counter attendants, and other employees) who were
 5    employed in coin-operated or while-you-wait shops, all affiliated with one organization.  The
 6    workers were separated into a low-exposure group (50 women, 7 men) and a high-exposure
 7    group (39 women, 5 men). A third group of 84 sales personnel (64 women, 20 men) from
 8    several department stores and receptionists from large hotels served as unexposed controls. No
 9    information was provided on the methods used to identify subjects or their reasons for
10    participating in the study, although the authors reported that 29 service technicians were
11    excluded from the study because of either discontinuous exposure conditions with peak
12    concentrations or long periods of no exposure. Predominant characteristics of both groups
13    included primarily standing work, contact with customers, and moderate physical exercise. The
14    mean ages of the low-exposure, high-exposure, and control groups were 38.2, 38.4, and 31.8
15    years, respectively.
16           Details on air monitoring methods were sparsely reported, but mean tetrachloroethylene
17    concentrations (8-hr TWA) for the low- and high-exposure groups were 12 (+8) ppm and 53
18    (±17) ppm, respectively, using results from active sampling of room air and passive sampling of
19    personal air. The mean duration of occupational exposure for the low- and high-exposure groups
20    was 11.8 and 10.6 years, respectively.
21           A number of tests of neuropsychological functioning were administered, including
22    standardized tests of symptoms and personality; tests of sensorimotor function, including finger
23    tapping and aiming; and the Mira and Santa Ana dexterity tests, which are published
24    standardized tests.  Threshold of perceptual speed was assessed by recognition of stimuli flashed
25    briefly on a screen; whether this procedure used a standardized instrument was not noted.
26    Choice reaction time was also determined using "nine light and tone stimuli." It is not clear
27    whether the auditory and visual stimuli occurred together or whether some trials consisted of an
28    auditory stimulus and others a visual stimulus. Details of the timing of the stimulus presentation
29    were not provided. One of the response variables, "delayed reactions," was not defined.  The
30    typical dependent variable measured in this task—response reaction time—apparently was not
31    measured;  only the number of correct reactions was reported. Subtests of the Wechsler
32    Intelligence Test (digit span, digit symbol, and cancellations) were used, as was recognition of
33    words, faces, and digits. The instrument used and the scoring of the last three tests were not
            4 Dr. Seeber provided additional information on this study in written correspondence to the New York State
      Department of Health dated January 19 and May 20, 1996. This information appears in NYS DOH (1997).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-65       DRAFT-DO NOT CITE OR QUOTE

-------
 1    described. Intelligence was assessed using the logical thinking subtest of the German
 2    Performance Test System.
 3          Each subject was examined during a 1.5-day stay at a clinic located at a large institute for
 4    occupational medicine. Each subject came to the clinic in the evening hours, stayed overnight,
 5    and started the examination and testing process the next morning. The clinic examined
 6    numerous people daily, and the dry cleaners and the control group took up only a small part of
 7    the daily routine of the clinic staff. Neurobehavioral tests were given by two specialized clinic
 8    staff who did not question the subjects about exposure status. However, clinic psychologists (six
 9    at the time of the study) did inquire about the exposure and living conditions of the subjects.
10    Because the dry cleaner groups and the control group differed in gender ratios, age, and scores
11    on the intelligence test, stratified analysis was used to statistically control the influence of these
12    confounding factors on test scores. As discussed in the section that describes Altmann et al.
13    (1995), the use of dichotomous or categorical variables may not fully control for confounding
14    effects of these factors on the endpoint.  The groups  also differed in alcohol consumption, so a
15    separate analysis was used to examine the role of alcohol on effects associated with
16    tetrachl oroethy 1 ene.
17          Performance of both the low-exposure and high-exposure groups differed significantly
18    (p< 0.01) from that  of the unexposed control group on the threshold  of perceptual speed and
19    "delayed responses" on a choice reaction time task, both of which are measures of information
20    processing speed (p = 0.08 and 0.03 for low exposure and high exposure, respectively). Both
21    exposed groups also had worse scores (p < 0.01) on two tests of attention (digit reproduction and
22    digit symbol) and on visual scanning (cancellations). Group mean scores for digit reproduction
23    and digit symbol did not appear to increase from the low-exposure to the high-exposure group.
24    The low-exposure group also showed significantly higher scores than did the control group on
25    questionnaires, on neurological signs (p < 0.01), and emotional liability (p < 0.05).  Scores of the
26    high-exposure group for these measures appeared higher than those for the control group;
27    however, the scores  did not show a statistically significant difference. There were no differences
28    between groups on the other tests. Controlling for group differences  in alcohol consumption did
29    not alter any test results.
30
31    4.6.1.2.3. Cm, S.X., M. Y. Huang, Z. Chen,  Y. T.  Liu, C. Jin, T.  Watanabe, H. Nakatsuka, K.
32    Seiji, O. Inoue andM. Ikeda. 1991. Subjective symptom increase among dry-cleaning workers
33    exposed to tetrachloroethylene vapor. Ind. Health. 29:111-121.  Cai et al. (1991) evaluated the
34    CNS effects of tetrachloroethylene exposure among  56 dry cleaning workers (27 women and 29
35    men) from three shops in China.  The control group (37 women and 32 men) were of similar
36    mean age (34 years vs. 35 years for dry cleaning workers), but the male dry cleaning workers

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-66       DRAFT-DO NOT CITE OR QUOTE

-------
 1    were 4 years younger than the male controls and the women were 4.9 years older than the female
 2    controls. The controls were recruited from the same factories as the dry cleaning workers but
 3    from workshops without known solvent exposures. No information is provided in the paper on
 4    the methods used to identify subjects or their reasons for participating in the study. Further, no
 5    information was provided on test procedures or the questionnaire used to assess subjective
 6    symptoms.  The geometric mean tetrachloroethylene air concentration (8-hr TWA) was 20 ppm
 7    and the range of TWA values was 4 to 97 ppm, using results from passive sampling of personal
 8    air.  The mean duration of occupational (tetrachloroethylene) exposure was 3 years.
 9          The prevalence of symptoms of tetrachloroethylene exposure was significantly higher
10    among the dry cleaning workers (men, women, and men and women combined; p < 0.001) than
11    among the unexposed controls. Five symptoms (dizziness, drunken feeling, floating sensation, a
12    heavy feeling in the head, and facial flushes) in men and women (combined) were significantly
13    more prevalent in the dry cleaning workers than in the controls (p < 0.001). Nasal irritation and
14    unusual smell were also reported significantly more often by the dry cleaning workers than by
15    controls (p < 0.05). Similar findings were reported when the workers were asked about the
16    symptoms they had noticed during the 3 months before the study. The investigators found
17    exposure-related increases in the prevalence of subjective symptoms among dry cleaning
18    workers exposed to 21 ppm (8-hr TWA).
19
20    4.6.1.2.4. Nakatsuka, H.,  T. Watanabe, Y. Takeuchi, N. Hisanaga, E. Shibata, H. Suzuki, M. Y.
21    Huang, Z. Chen, Q.S. Qu andM. Ikeda. 1992. Absence of blue-yellow color vision loss among
22    workers exposed to toluene or tetrachloroethylene, mostly at levels below occupational
23    exposure limits. Int.  Arch. Occup. Environ. Health. 64:113-117.  Nakatsuka et al. (1992)
24    evaluated the effects of tetrachloroethylene exposure on the color vision of 64 dry cleaning
25    workers (34 women and 30 men) in China. The workers were from the same shops studied by
26    Cai et al. (1991). Control workers (72 women and 48 men) were recruited from the clerical
27    sections of dry cleaning shops and from other factories (paint production plants or plants
28    producing tetrachloroethylene from trichloroethylene).  No information is provided in the paper
29    on the methods used to identify subjects or their reasons for participating in the study. The mean
30    ages of the dry cleaning workers (34 years for men, 35 years for women) were lower than those
31    of the controls (34 years for men, 33 years for women). A screening color test (Lanthony's new
32    color test) and a test used for confirmation of red-green vision  loss were carried out by
33    ophthalmologists or occupational health doctors in charge of the factories under one of two
34    lighting conditions (natural sunlight or a daylight fluorescent light). The published report does
35    not identify what procedure was used on which test; illumination is a critical component in
36    administering color vision tests to subjects (Geller and Hudnell,  1997).

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-67      DRAFT-DO NOT CITE OR QUOTE

-------
 1           The geometric mean air concentrations of tetrachloroethylene (averaging time not
 2    reported) were 16 and 11 ppm for the men and women, respectively, using results from passive
 3    sampling of personal  air. The overall geometric mean was 13 ppm.
 4           There was no  significant difference in the performance of the dry cleaning workers and
 5    unexposed controls on Lanthony's new color test. The study authors reported that the
 6    percentages of dry cleaning workers who correctly separated colored caps from monochromatic
 7    caps were not significantly different from the percentages in the corresponding control group.  A
 8    statistical analysis of these data reported in public comments of the Halogenated Solvents
 9    Industry Alliance to EPA (HSIA, 2004) on the Neurotoxicity of Tetrachloroethylene Discussion
10    Paper (U.S. EPA, 2003b) showed—using a chi-square test for differences in proportions—that
11    tetrachloroethylene-exposed women were more likely to have normal color vision as compared
12    with unexposed women. An EPA analysis of male workers did not show any differences, either
13    better color vision or worse color vision, in exposed males compared with unexposed male
14    controls.  Nakatsuka et al. concluded, overall, that they found no distinct case of color vision loss
15    among the dry cleaning workers.
16
17    4.6.1.2.5. Ferroni, C., L. Selis, A. Mutti, D. Folli, E. Bergamaschi and I. Franchini. 1992.
18    Neurobehavioral and neuroendocrine effects of occupational exposure to perchloroethylene.
19    Neurotoxicol. 13:243—247.5 Ferroni et al. (1992) evaluated neuroendocrine and neurobehavioral
20    effects of tetrachloroethylene exposure among 60 female dry cleaners and 30 unexposed female
21    controls who were comparable in age (mean ages 39.7 and 37.6 years, respectively) and
22    vocabulary level. Each dry cleaning  shop in a small town outside of Parma, Italy was visited.
23    The workers were invited to participate in the study, which was part of a preventive health
24    program implemented by the local health office and professional associations of small businesses.
25    There were no refusals.  Controls were selected from the workers at a hospital who cleaned
26    clothes using a water-based process.  Their jobs were essentially the same as those of the dry
27    cleaners, but they were not exposed to any organic solvents.  Both  groups filled out a
28    questionnaire on their health status, medication (including oral contraceptives), lifestyle, and
29    current and past jobs.  Both groups met the following criteria: no history of metabolic disorders,
30    no history of psychiatric disorders, and low level of daily alcohol intake. The two groups were
31    similar in height, weight, body mass index, smoking habits, and use of medication, but alcohol
32    intake was about 5% higher (p < 0.03) in the control group than in  the dry cleaner group.
            5 Dr. Mutti provided details on the selection process of exposed and control subjects and also clarified
      reported results to Dr. Ken Bodgen, New York State Department of Health, in written correspondence dated July 29
      and September 5, 1995 (see NYS DOH, 1997).
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-68       DRAFT-DO NOT CITE OR QUOTE

-------
 1          Workers and controls were given five neurobehavioral tests (part of the Swedish
 2    Performance Evaluation System, "adapted" Italian version: finger tapping with both dominant
 3    hand and nondominant hand, simple reaction time, digit symbol test, shape comparison-vigilance,
 4    and shape comparison-response to stress).  All subjects were examined in  the morning before
 5    their work shift in the same room by the same examiners (NYS DOH, 1997). The tests were part
 6    of a computer-based battery, and the same machines and software were used to administer the
 7    tests and score the results. The same sequence of tests and protocols were used for all subjects.
 8    Although the examiners were not blind to the status of the subjects (dry cleaner or control), they
 9    were blind to the worker's exposure level (NYS DOH, 1997). Serum prolactin levels were
10    measured in all subjects; blood samples were collected during the working day during summer
11    and winter.  Prolactin secretion by the pituitary is under control by hypothalamic dopamine;
12    dopamine is also important to neurotransmitter systems.
13          One proposed alternative for assessment of nervous system toxicity is the study of
14    biochemical signals in peripheral tissues as biomarkers of nervous system function (Manzo et al.,
15    1996). Samples from dry cleaners and controls were alternated and analyzed in the same
16    experimental runs. For women, only those samples obtained during the proliferative phase of the
17    menstrual cycle were used for comparison between groups (41 dry cleaners and 23 controls).
18          Workplace air samples were randomly collected throughout the work week during
19    summer and winter to account for variability related to either the work cycle or seasonal
20    environmental fluctuations.  The median tetrachloroethylene air concentration (4-hr TWA) was
21    15 ppm and the range of TWA values was 1 to 67 ppm.  The subjects' range of
22    tetrachloroethylene blood levels was 0.012 to 0.864 mg/L (median = 0.145 mg/L; incorrectly
23    expressed in Ferroni et al., 1992, as 12,864 and 145 mg/L [NYS DOH, 1997]).  The mean
24    duration of occupational exposure was 10 years.
25          The dry cleaners showed significantly reduced performance when  compared with the
26    unexposed matched controls in three tests (simple reaction time, p < 0.0001; vigilance,
27    p< 0.005; and stress,/? < 0.005), as reported by Ferroni et al. (1992).  Performance on the fmger-
28    tapping test (both hands) and digit symbol test was not affected (NYS DOH, 1997). Additionally,
29    the mean serum level of prolactin was significantly higher in the workers than in the matched
30    controls (p < 0.001). None of the three measures of exposure (duration of exposure and air or
31    blood concentration of tetrachloroethylene) was significantly associated with decreased test
32    scores or increased serum prolactin levels among the dry cleaners.
33          The study authors concluded that tetrachloroethylene exposure in dry cleaning shops may
34    impair performance and affect pituitary function but that the cross-sectional  design prevented
35    distinguishing acute effects from chronic effects.  Ferroni et al. (1992) also reported that the most
36    likely bias of cross-sectional studies is a spontaneous selection of the sample (i.e., workers who

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-69       DRAFT-DO NOT CITE OR QUOTE

-------
 1    believe exposure is making them sick or workers who actually become sick may quit work
 2    prematurely and not be included in the study) and, as a result, the actual risk was likely to be
 3    underestimated rather than overestimated, although no data are presented in the paper with which
 4    to evaluate this statement. On the other hand, the exposed and unexposed study population of
 5    women was tested during the proliferation phase of menstruation, which may better capture
 6    changes in prolactin secretion, but also may potentially confound findings if there are individual
 7    differences in severity of menstruation and in the timing of test session relative to  the day of
 8    menstruation (U.S. EPA, 2004).
 9
10    4.6.1.2.6. Cavalleri, A., F. Gobba, M. Paltrinieri, G. Fantuzzi, E. Righi and C.L. Aggazzoti.
11    1994. Perchloroethylene exposure can induce colour vision loss. Neuroscience Lett. 179:162-
12    166.6  Gobba, F., E. Righi, G. Fantuzzi,  G. Predieri, L. Cavazzuti and G. Aggozzotti. 1998.
13    Two-year evaluation of perchloroethylene-induced color-vision loss. Arch. Environ. Health
14    53:196-198. Cavalleri et al. (1994) reported on a control-matched, cross-sectional,
15    observational, occupational study that evaluated the effects of tetrachloroethylene  exposure on
16    the color vision of dry cleaners. The investigators compiled a list of all the dry cleaning shops in
17    the municipality of Modena, Italy, (110 shops employing  189 workers) and randomly selected 60
18    dry cleaners from 28 premises for recruitment into the study (Aggazzotti et al., 1994a). Only
19    full-time workers (n = 52) were asked to participate, and two declined. No information is
20    provided in the paper on  a subject's motivation for participating or not participating in the study.
21    All 50 workers provided, via questionnaires, information on work history, health status,
22    occupational and hobby use of solvents, drinking and smoking habits, and drug use. Thirty-five
23    of the 50 dry cleaners (33 women, 2 men) met the inclusion criteria; others were excluded for
24    hypertension, smoking more than 30 cigarettes a day, alcohol consumption exceeding 50 g of
25    alcohol a day, oculo-visual pathology, or working less than 1 year. Another worker was
26    excluded because a matched control could not be found.
27          The controls were factory workers who were not occupationally exposed to solvents or
28    other neurotoxic chemicals; they were selected and recruited into the study using the same
29    methods that were used for dry cleaners.  The controls (n = 35) were from factories in the
30    Modena area and met the same inclusion criteria as the dry cleaners.  They were matched to dry
31    cleaners by gender, age (+3 years), alcohol consumption (+10 g/day),  and cigarette use (+5
32    cigarettes a day). The mean age of both groups (35 years) and the percentages of each group that
33    were smokers (43%) or alcohol drinkers (71%) were comparable.
            6 Dr. Cavalleri provided additional information on this study in written correspondence to the New York
      State Department of Health dated October 8, 1996 (see NYS DOH, 1997).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-70      DRAFT-DO NOT CITE OR QUOTE

-------
 1          All subjects appeared healthy and met minimal status of visual acuity. None of the
 2    subjects reported hobby exposure to solvents or other substances toxic to the eye. There were no
 3    known systematic differences between exposed and control groups or between machine
 4    operators and ironers.
 5          Color vision was assessed using Lanthony's D-15 desaturated panel test, in which
 6    subjects are asked to put a series of small round "caps" in order by color. The types of errors
 7    made can distinguish specific types of color vision deficiency; e.g., red-green color confusion
 8    errors (blindness) is a common condition in males, mostly but not entirely of congenital origin,
 9    whereas blue-yellow color confusion errors are very rarely due to congenital conditions and
10    therefore are considered as a hallmark of an acquired condition. Impairments in color vision,
11    beginning as blue-yellow confusion errors, have been reported in numerous populations exposed
12    to organic solvents (Mergler and Blain, 1987; Mergler et al., 1987, 1988a, b, 1991; Campagna et
13    al., 1995, 1996).  Test scores are based on the ability of each subject to arrange a set of 15 caps
14    colored with desaturated colors according to a definite chromatic sequence, with each mistake
15    increasing the score above a perfect score of 1.00. A formula (the Color Confusion Index [CCI])
16    is used to calculate total errors.
17          Exposed and control subjects  were tested in a random order (NYS DOH, 1997).  All
18    subjects were tested at the same time of day (in the morning, before work) under the same
19    lighting  conditions by the same investigator. With respect to exposed subjects, the investigator
20    was unaware of both the exposure levels and the job (operator or ironer) of each dry cleaner.
21          For all dry cleaners, the mean tetrachloroethylene air concentration (8-hr TWA) was 6
22    ppm and the range of TWA values was 0.4-31 ppm, using results from passive sampling of
23    personal air. For operators (n = 22), the mean air concentration 8-hr TWA was 7 ppm and the
24    range of TWA values was 0.4-31 ppm. For ironers (n = 13), mean air concentration (8-hr TWA)
25    was 5  ppm and the range of TWA values was 0.5-11 ppm.  The mean duration of occupational
26    exposure was 8.8 years. Tetrachloroethylene  concentrations were also measured in alveolar air
27    for a subset of these dry cleaners, with a high  correlation observed between tetrachloroethylene
28    concentration in alveolar air and the 8-hr TWA levels in ambient air (r = 0.8, p < 0.001;
29    Aggazzotti et al., 1994a).
30          Only three dry cleaning workers, as opposed to 13  controls, scored a perfect test score
31    (p < 0.01). Mistakes were made mainly in the blue-yellow range. Overall, the workers  showed
32    poorer performance on the test as compared to controls, and they had a significantly higher mean
33    CCI using a Student's t-test  (p = 0.03). The effect was statistically significant among operators
34    but not among ironers. Study investigators also evaluated whether CCI values were normally
35    distributed, which is important if using a Student's t-test, but they did not present any
36    information about the result of the Kolmogorov-Smirnov test.  The observation for ironers may

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-71       DRAFT-DO NOT CITE OR QUOTE

-------
 1    reflect a lower statistical power in this group due to fewer subjects (13 ironers vs. 22 operators).
 2    There also was a statistically significant positive correlation (p < 0.01) between TWA air
 3    concentrations and the CCI (r = 0.52), which remained after multivariate analysis considered
 4    previous tetrachloroethylene exposure duration, age, number of cigarettes a day, and daily intake
 5    of alcohol as covariates.
 6          The effect  on color vision may not be rapidly reversible; preliminary data showed that the
 7    scores of some workers did not improve when retested after 4 weeks of vacation (NYS DOH,
 8    1997). Moreover, some of these workers showed poorer performance on this test in the follow-
 9    up study by Gobba et al. (1998), described below, suggesting color vision impairment as a
10    chronic effect. The CCI values were not associated with two other measures of
11    tetrachloroethylene exposure (mean duration and an integrated index of exposure, yearly TWA
12    level). The study authors  suggested that this may reflect the difficulty in controlling for the
13    interactive effects  of age and exposure and accurately evaluating exposure.
14          Gobba et al. (1998) reexamined color vision after a period of 2 years in 33 of the 35 dry
15    cleaners and ironers examined by Cavalleri et al. (1994). Two subjects had retired during the
16    2-year period between examinations. These investigators used the Lanthony D-15 test, the test
17    used by  Cavalleri et al. (1994) to assess color vision, and performance was compared with the
18    subject's score from the initial survey (self-control). Tetrachloroethylene concentration in the
19    occupational setting was determined in the breathing zone using personal passive samplers.
20    Monitoring was carried out during the afternoon shift, as Cavalleri et al. (1994) did not show any
21    differences between morning and afternoon samples.  Gobba et al. (1998) found that
22    tetrachloroethylene concentration had increased during the 2-year period for 19 subjects,
23    identified as Group A, (geometric mean, from 1.67 ppm at the first survey to 4.35 ppm at the
24    second survey) and had decreased for 14 subjects, identified as Group B (geometric mean, from
25    2.95 ppm to 0.66 ppm). For the 33 workers overall, tetrachloroethylene concentration did not
26    change over the 2-year period (geometric mean, from 2.4 ppm to 1.94 ppm at the second survey,
27    /?>0.05).
28          Color vision deteriorated between the two surveys for the entire group,  a reflection of the
29    color vision loss among Group A subjects, whose exposure had increased by the second survey.
30    As found in the  first survey, color vision was impaired primarily in the blue-yellow range of
31    color, with few subjects presenting a red-green deficit. Color vision performance for the entire
32    group was related  significantly to age (r = 0.45) and tetrachloroethylene concentration (r = 0.39).
33    The mean CCI score for Group A subjects showed a statistically significant difference between
34    the two surveys (p < 0.05). Analysis of variance methods that controlled for an effect of age
35    further supported the finding of color vision deterioration among these subjects. For Group B
36    subjects, who experienced lower exposure concentrations by the second survey, the CCI score

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-72        DRAFT-DO NOT CITE OR QUOTE

-------
 1    did not change from that of the initial survey. The findings in Groups A and B were also
 2    supported using analysis of variance methods that adjusted for age, alcohol consumption, or
 3    cigarette smoking between the subgroups.
 4
 5    4.6.1.2.7. Echeverria, D., N. Hyer, J.A. Bitner, H. Checkoway, G. Toutonghi andN.
 6    Ronhovde. 1994. Behavioral effects of low level exposure to perchloroethylene (PCE) among
 1    dry cleaners. In:  Battelle Centers for Public Health Research and Evaluation. A behavioral
 8    investigation of occupational exposure to solvents: perchloroethylene among dry cleaners and
 9    styrene among reinforced fiberglass laminations. Final Report SSRC-10OM4/040. Seattle,
10    WA: pp. 6.1—6.37. Echeverria et al. (1994) reported a study designed to evaluate a hypothesis in
11    a previous study (Echeverria et al., 1995)7 of frontal/limbic system effects8 as the site underlying
12    tetrachloroethylene pathology, where degradation in behavior may be the earliest indicator of
13    acute, subchronic, or chronic neurotoxicity.  The study was conducted in the Seattle/Tacoma,
14    Washington area from 1989 through 1993, when the area's dry cleaning industry was switching
15    from wet-transfer to dry-to-dry machines. Initially, 320 dry cleaning shops and laundries were
16    sent introductory letters requesting permission to allow their employees to participate in the
17    study. Of the 181 owners who responded, 39 agreed to participate. Of the owners who did not
18    agree to participate, 22% expressed no  specific reason, 19% cited time constraints, 17% feared
19    legislative reprisal from federal agencies, 17% did not speak English, 15% were unavailable or
20    never contacted, and 10% cited various other reasons.
21          Recruitment ended when a total of 45 operators were enrolled in the study (total n = 173).
22    Each operator was matched with a less-exposed person from the same shop.  The subjects
23    included laundry workers  (n = 69), pressers or counter clerks (n = 59), and operators or former
24    operators (n = 45). The mean ages of the groups were 42.5, 34.2, and 46.2 years, respectively.
25    Women comprised 63% of the study population (109/173). The subjects, who were paid
26    volunteers, were eligible if they spoke English, had no history of diabetes or CNS disorders, and
27    had worked for more than one year in the trade.  The final sample excluded three subjects for
28    limited knowledge of English and reading skills and six subjects for not wearing glasses or
29    missing covariate information such as vocabulary performance on the test.
             7 Although published a year after this study, the study by Echeverria et al. (1995), discussed in Section
      4.6.1.2.8, was conducted in 1986, 3 years before this study.
             8 Echeverria et al. (1994) hypothesized that exposure to solvents particularly affected attention, executive
      function, visuospatial skill, short-term memory, and mood, leaving motor, language-based skills, and long-term
      memory intact.  The frontal system has been associated with executive function, such as that measured by Switching
      Attention, Trailmaking A and B, and the Wisconsin Card Sort.  Tests associated with the limbic system include
      mood, as measured by the Profile of Mood States, short-term memory as measured by digit span, visual
      reproductions, and pattern memory.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-73       DRAFT-DO NOT CITE OR QUOTE

-------
 1          An index of chronic exposure and measures of subchronic and acute exposure were
 2    developed for each subject. The chronic exposure index was based on a detailed work history,
 3    including consideration of the type of dry cleaning machine, job title, percentage of time at each
 4    job title, estimated air levels associated with each job title, and employment duration. The
 5    measures of subchronic and acute current exposure were based on mean 8-hr TWA air
 6    concentrations measured on the day of neurobehavioral  testing.  Mean chronic indices were zero
 7    for the never-exposed group of laundry workers, 68 for the dry cleaning workers with low
 8    exposure (pressers/clerks), and 1,150 for the dry cleaning workers with high exposure (operators).
 9    Mean exposures (8-hr TWA, using results from passive sampling of personal air) for workers
10    placed in these chronic exposure categories were <0.2 ppm (laundry workers), 3 ppm
11    (pressers/clerks), and 9 ppm (operators). Dry cleaning workers placed in the chronic exposure
12    categories of low and high had been employed in their current job for 2.6 and 11 years,
13    respectively.
14          The subjects also were placed in acute and subchronic exposure categories of <1 ppm
15    (laundry workers and some dry cleaning workers, e.g., clerks), low (mainly pressers), and high
16    (operators), with corresponding current tetrachloroethylene 8-hr mean concentrations of 0.5, 3,
17    and 20 ppm. Dry cleaning workers placed in the low and high categories had been employed in
18    their current job for 5 and 9 years, respectively. Because of the changes in dry  cleaning practices
19    over the course of the study, many subjects who were placed in the high chronic-exposure
20    category, which was based on detailed work history, were frequently found in the low acute- or
21    subchronic-exposure group, which was based on air concentrations on the day of testing.
22          The test battery included tests of cognitive function, including  visuospatial ability, motor
23    skills, mood, CNS symptoms, and basic verbal and arithmetic skills. The chronic and subchronic
24    assessment was based on tests given during the morning of each subject's day off and on pre-
25    shift  scores. Additional tests considered to have an acute component were given 1 hr before
26    work on the first day of the work week and again at the  end of the day, allowing acute effects to
27    be examined by using pre-shift performance to predict post-shift performance and then
28    comparing predicted with observed performances. On their day off, the subjects were tested at
29    home. At the work site, subjects were tested in a minivan.  Each subject signed a consent form,
30    provided a breath sample at each test session,  and completed a questionnaire covering transient
31    factors that could affect performance (e.g., headache). This was followed by questionnaires on
32    medical history, medication, drug and alcohol use, occupational and nonoccupational exposure to
33    chemicals, symptoms, and mood. The subject was then administered the neurobehavioral tests.
34          Multivariate analysis was used to evaluate the relationship between exposure indices and
35    levels and performance  on neurobehavioral tests after adjusting for the variables of age, gender,
36    race, vocabulary level (surrogate for education and test-taking), and alcohol consumption. After

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-74        DRAFT-DO NOT CITE OR QUOTE

-------
 1    adjustment for those variables that were significant confounders, associations between increased
 2    indices of chronic (lifetime) exposure and reduced test performance were found in three tests of
 3    cognitive function:  switching (p = 0.1), pattern memory (p = 0.03), and pattern recognition
 4    (p = 0.09). The magnitude of change attributable to tetrachloroethylene was a 3% loss in
 5    function for the latency of pattern memory and an 11% loss in function for the correct number in
 6    visual reproductions; losses in function that are well within pre-clinical values. Subjective
 7    measures of mood and symptoms were not significantly associated with exposure. Dry cleaning
 8    workers scored lower (but not significantly) on all but one of the remaining tests (the digit span
 9    test).
10           Analysis of the association between test scores and measures of subchronic exposure
11    (8-hr TWA tetrachloroethylene concentrations on the day of testing) confirmed the findings of
12    the chronic analysis: reduced scores on tests of switching (p = 0.1) and pattern recognition
13    (p = 0.04) as exposure increased.  Analysis of effects  of acute exposures showed no relationship
14    between workday exposure at any level and post-work performance on nine neurobehavioral
15    tests.
16           Echeverria et al. (1994)  detected deficits in visuospatial function (reduced performance in
17    tests of pattern memory and pattern  recognition) in the dry cleaning workers categorized as
18    having high lifetime chronic exposure and whose current exposure level was 9 ppm, 8-hr TWA.
19    However, the exposure level of 9 ppm was not considered representative of past chronic
20    exposure levels because the industry in the study area was switching from wet-transfer to dry-to-
21    dry machines during the study.  The investigators attributed the reduced performance to prior
22    exposures that were about two to four times higher 3 to 5 years previously, and they
23    hypothesized that a few years of reduced exposure may not be long enough to eliminate the
24    residual effects on visuospatial  skills caused by the exposures associated with wet-transfer
25    machines.
26
27    4.6.1.2.8. Echeverria, D., R.F. White and C. Sampaio. 1995. A behavioral evaluation ofPCE
28    exposure in patients and dry cleaners: a possible relationship between clinical andpreclinical
29    effects. J. Occup. Environ. Med.  37:667—680. Echeverria  et al. (1995) assessed neurobehavioral
30    effects and mood disturbances in four patients diagnosed with tetrachloroethylene
31    encephalopathy.  Subject 1 was exposed chronically over a 1-year period when the interior
32    woodwork of her home was mistakenly treated with tetrachloroethylene.  The three other cases
33    were occupationally  exposed. Subject 2 was exposed during two separate periods:  first, for 3
34    years in a dry cleaning establishment and, second, for 7 years cleaning parts. Subject 3 was
35    exposed for 16 years as a dry cleaning worker. Subject 4 was also exposed as a dry cleaning
36    worker, but her duration of employment was not reported.  Subjects 2, 3, and 4 were working

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-75       DRAFT-DO NOT CITE OR QUOTE

-------
 1    with tetrachloroethylene when first tested.  Air monitoring data were not available; however,
 2    occupational health physicians diagnosed each case with tetrachloroethylene encephalopathy on
 3    the basis of symptoms, neurophysiological assessment, and their own examinations.
 4          A large battery of standard neurobehavioral tests was given to each subject.  For most
 5    tests, impairment was inferred clinically when a subject's score was greater than one standard
 6    error of measurement below expectation, which  is less restrictive than the criterion (more than
 7    two standard deviations below mean) commonly used in neurobehavioral testing to separate
 8    normal from abnormal scores (Lezak, 1995). Test results for the four subjects most consistently
 9    indicated complaints of fatigue and confusion, accompanied by cognitive deficits on tests
10    assessing memory and motor, visuospatial, and executive function. Repeated testing of subjects
11    3 and 4 indicated post-exposure improvement on neurobehavioral tests of all the affected
12    functional domains, although performance on some of the more difficult tests in each domain
13    remained impaired. These results suggest an association between CNS effects and
14    tetrachloroethylene exposure, but a conclusion of a causal relationship is precluded by the lack of
15    data on the duration and severity of the tetrachloroethylene exposure.
16          The investigators also assessed the performance of 66 dry cleaning workers on
17    neurobehavioral tests designed to detect the same impairments noted in the clinical cases. The
18    testing was  conducted in 1986.  The owners of 125 shops in Detroit, Michigan were contacted,
19    and 23 agreed to allow their workers to participate in the study. Within each shop, operators
20    were matched on education and age (+5 years) with a lower-exposure subject.
21          The subjects (35 men and 30 women) were grouped into three categories of chronic
22    tetrachloroethylene exposure (low, moderate, and high), based on type of shop (wet-transfer or
23    dry-to-dry), job title (counter clerk, presser, or operator), and years of employment. All the
24    operators were placed in the high-exposure category.  There was no unexposed control group.
25    Dry cleaning workers placed in the chronic exposure categories of low, moderate, and high had
26    been employed at their main job  for 2.1, 3.9, and 14.6 years, respectively. Their mean age was
27    40.9, 40.6, and 43 years. The three groups were also characterized by estimates of current
28    exposure (low, medium, and high), which corresponded to mean tetrachloroethylene air
29    concentrations (8-hr TWA) of 11, 23, and 41 ppm, respectively, for counter clerks, pressers, and
30    operators in the more common wet-transfer shops (17 of 23 shops).  Estimated air concentrations
31    for counter clerks, pressers, and operators in the dry-to-dry shops were 0.5, 10,  and 11 ppm. The
32    estimates were based  on a relationship between breath and air concentrations derived from a
33    larger independent study (Solet et al., 1990). The study authors noted that the estimates were
34    comparable to those found in other surveys of dry cleaning facilities in the United States.
35          All subjects were tested in groups of two in the afternoon after work on the first or
36    second day  of their work week. The tests were conducted in a minivan.  Each subject provided a

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-76       DRAFT-DO NOT CITE OR QUOTE

-------
 1    breath sample and completed a medical, symptom, work history, and hobby questionnaire. The
 2    subjects were administered six neurobehavioral tests, a test of verbal skills, and questionnaires
 3    on emotional states (moods) and CNS symptoms. The neurobehavioral test battery consisted of
 4    one test of motor/cognitive function (symbol digit) and five tests of cognitive function (digit span,
 5    trailmaking A and B, visual reproduction, pattern memory, and pattern recognition), including
 6    three tests of an individual's ability to process and remember visuospatial stimuli (the latter three
 7    tests).
 8           Multivariate analysis was used to evaluate the relationship between a chronic index of
 9    lifetime exposure and performance on neurobehavioral tests, after adjusting for the confounding
10    variables  of current exposure, a 3-year index of exposure, age, education, verbal skill, alcohol
11    consumption, hours of sleep, fatigue, mood, symptoms, medication, and secondary exposures to
12    neurotoxicants. After adjustment for factors affecting performance, the scores of the dry
13    cleaning workers with high chronic exposure were statistically significantly lower (p < 0.01) than
14    those of the workers with low chronic exposure in three tests of visual function:  visual
15    reproduction, pattern memory, and pattern recognition.  Adjusted scores were reduced from 6 to
16    15%; the  two most sensitive tests were those that measured short-term memory of visual designs.
17    These impairments of visually mediated function were consistent with the impairment of
18    visuospatial functions observed in the four patients previously studied by Echeverria et al. who
19    were diagnosed with tetrachloroethylene encephalopathy.  Other effects seen in the patients
20    (mood changes and decreased cognitive function in nonvisual tests) were not found in the dry
21    cleaning workers with high lifetime exposures. Among complaints by the dry cleaning workers,
22    only the number of complaints of dizziness from standing up rapidly and "solvent-induced
23    dizziness" over the previous 3 months was significantly elevated (p<0.04) in the high-exposure
24    group.
25           The study authors concluded that effects on visuospatial function were consistently found
26    in subjects employed as operators for an average of 14.6 years and exposed to an estimated
27    tetrachloroethylene 8-hr TWA air concentration of 41 ppm, suggesting a vulnerability of visually
28    mediated  functions with tetrachloroethylene exposure.  This conclusion was based  on the
29    impaired  performance of the high-exposure group when compared with a group of dry cleaning
30    workers with low lifetime exposure, including 16/22 workers who were probably clerks  in wet-
31    transfer shops where the mean current exposure  level was  12 ppm. This exposure level is
32    substantially above background ambient levels, and whether the performance of the low-
33    exposure  group was impaired when compared with that of a group without occupational
34    exposure  (i.e., an unexposed control group) is not known.  The lack of an unexposed control
35    group limits the ability of the study to fully characterize the magnitude of the effects on
36    visuospatial ability and to detect exposure-related symptoms or effects on tests of nonvisual

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-77       DRAFT-DO NOT CITE OR QUOTE

-------
 1    cognitive ability. It also limits the extrapolation of the results to other populations exposed to
 2    tetrachloroethylene.
 O
 4    4.6.1.2.9. Altmann, L., H.F. Neuhann, U. Kramer, J. Witten andE. Jermann. 1995.
 5    Neurobehavioral and neurophysiological outcomes of chronic low-level tetrachloroethylene
 6    exposure measured in neighborhoods of dry cleaning shops. Environ. Res. 69:83-89.
 1    Altmann et al. (1995) used neurophysiological and neurobehavioral techniques to assess the
 8    effects of long-term exposures to tetrachloroethylene. A total of 19 tetrachloroethylene-exposed
 9    subjects (residents of Mulheim, Germany) were chosen from a population of 92 subjects living in
10    neighborhoods close to dry cleaning facilities. Three criteria were used to select subjects: a
11    tetrachloroethylene blood level above 0.002 mg/L, a period of living above or next to a dry
12    cleaning facility for at least 1 year, and no occupational exposure to organic solvents. The mean
13    age of the exposed subjects was 39.2 years (range: 27-58 years) and the mean duration of living
14    near a dry cleaning facility was 10.6 years (range: 1-30 years). The daily activity pattern of the
15    exposed subjects was not reported. A total of 30  controls were  selected from volunteers;  their
16    mean age was 37.2 years (range: 24-63 years).  One or two controls, matched for age (+1 year,
17    but 3 years in one case and 6 years in another case) and gender, were chosen for each exposed
18    subject. The control subjects were recruited mainly from the staff of a public health office or an
19    institute for environmental hygiene, and none reported a history of solvent exposure.  No
20    information is provided in the paper on the motivation for exposed and control subject to
21    participate in the study. Voluntary consent was obtained from all  subjects prior to the initiation
22    of testing.
23          All subjects were given medical examinations.  Five exposed (26%) and seven control
24    subjects (23%) were excluded for various medical reasons, including impaired vision, diseases
25    with potential neuropathy, hypertension, and joint impairment.  The reasons for exclusion were
26    similar in both groups. All subjects met standards for visual acuity and vibration perception.
27    The final exposed group was composed of 5 men and 9 women and the control group was
28    composed of 9 men and 14 women. The two groups did not differ with regard to consumption of
29    alcoholic beverages, regular medication, smoking, or body mass index, but they did differ in
30    degree of education, which was used as an indicator of social status.
31          The effect of tetrachloroethylene exposure on the neurophysiological and
32    neurobehavioral measurements was evaluated using both univariate and covariate linear
33    regression.  The multivariate regression analysis accounted for age, gender, and education as
34    covariates. Degree of education was defined as "low," "medium," or "high" level.
35          VEPs in response to black-and-white checkerboard patterns were recorded for all
36    individuals. Vibration perception using a tuning fork—a crude measure of peripheral

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-78       DRAFT-DO NOT CITE OR QUOTE

-------
 1    neuropathy—was assessed at the ankle. Five tests included in the Neurobehavioral Evaluation
 2    System developed in the United States and adapted for testing on a German population were
 3    used:  (1) finger-tapping speed with the index finger of both the dominant and the nondominant
 4    hand; (2) hand-eye coordination using a joy stick to follow a sine wave on a computer screen; (3)
 5    a continuous performance test for assessment of vigilance, which requires a response to a
 6    specific stimulus appearing on the computer screen and failure to respond to other stimuli; (4)
 7    simple reaction time, which requires the fastest possible response to a simple visual stimulus
 8    (measured twice); and (5) visual memory on the Benton visual retention test, which requires a
 9    match of a previously displayed stimulus out of several choices after a short delay interval. All
10    of these tests are commonly used to assess occupationally exposed adults, and the software for
11    testing and analysis is available for purchase. All testing was completed in a single 3-hr session;
12    testing times were selected randomly for both exposed or control subjects.
13          Blood samples were taken once in the examination room immediately before testing (all
14    subjects) and, if possible, once when the exposed subjects were at home. The mean blood level
15    for exposed subjects, based on samples collected in the examination room, was 0.0178 mg/L
16    (standard deviation, 0.469 mg/L). For seven of the nine exposed subjects, blood concentrations
17    in samples collected at home were higher than those in samples collected in the  examination
18    room.  None of the blood concentrations in the control group exceeded the detection limit of
19    0.0005 mg/L. For the exposed subjects (data from 13  apartments), indoor air sampling indicated
20    that the mean (7-day TWA) air concentration was 0.7  ppm (standard deviation,  1 ppm) and the
21    median was 0.2 ppm. For the control group, the mean and median values were 0.0005 ppm
22    (standard deviation, 0.0005 ppm) and 0.0003 ppm. There was a good correlation between home
23    indoor air concentrations and blood levels of tetrachloroethylene in the exposed subjects
24    (r = 0.81). The correlation was much lower when the  examination room blood samples were
25    used (r = 0.24).
26          After adjusting for covariates and possible confounders of age, gender, and education,
27    there were statistically significant group differences between the adjusted mean  scores of
28    exposed and control subjects on three neurobehavioral tests (simple reaction time,/? < 0.05 for
29    the first test andp < 0.01  for the second test; continuous performance, p < 0.05;  and visual
30    memory,/' < 0.05).  In all cases, the exposed subjects  had slower response times or more errors
31    than did the unexposed controls.  No statistically significant differences  were observed between
32    the performance of the exposed and control groups on the finger-tapping or hand-eye
33    coordination tests, which are measures of fine motor function; on VEP, which may be less
34    sensitive than direct measurement of visual function; or on vibration perception at the ankle
35    using a tuning fork.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-79       DRAFT-DO NOT CITE OR QUOTE

-------
 1          The relationship between indoor tetrachloroethylene concentration and individual
 2    performance was not reported, so it was not possible to evaluate concentration-response
 3    relationships. The small numbers of study subjects in this study compared to the occupational
 4    studies is a limitation; however, this study appeared to have sufficient power to detect
 5    associations with tetrachloroethylene exposure.  Additionally, exposed and control groups did
 6    not differ with regard to consumption of alcoholic beverages, regular medication, smoking, or
 7    body mass index, but they did differ in degree of education. Statistical analysis of the data took
 8    into account  effects from important covariates such as age, gender, and education.  Univariate
 9    analyses showed that age, gender, and education were not predictors of deficits in tests of
10    continuous performance, visual memory,  and the second reaction time, although education and
11    gender were  predictors of deficits in the first reaction time.
12          The use of three categories for education in the multivariate regression analyses may not
13    fully account for all effects from these covariates (U.S. EPA, 2004), although it is not possible to
14    evaluate whether residual  confounding from these covariates may explain observations on the
15    neurobehavioral tests. Furthermore, because the responses in the exposed group for the tests
16    highlighted above (simple reaction time, continuous performance, visual retention) were
17    statistically significantly different from those of the control group, whether or not the covariates
18    were considered, an approximate estimate of the impact of the tetrachloroethylene exposures can
19    be derived by comparing the reported response levels for the two groups.  The degree of change
20    from control  was approximately  15-20%  for this subset of tests.
21
22    4.6.1.2.10. Spinatonda, G., R. Colombo,  E.M. Capodaglio, M. Imbriani, C. Pasetti, G. Minuco
23    and P. Pinelli. 1997. [Study on speech production processes:  application for a group of
24    subjects chronically exposed to organic solvents (part II).J Med. Lav. 19:85—88. Spinatonda et
25    al. (1997) assessed the effect of tetrachloroethylene exposure on vocal reaction times among 35
26    dry cleaners  and 39 unexposed controls. Controls were matched to exposed individuals for age
27    (mean age of 35 years for  both groups) and education.  The published paper did not identify the
28    population from which exposed and controls were drawn, the inclusion criteria for exposed
29    subjects and  controls—and hence, whether potential  study subjects may have been excluded—
30    and duration of exposure in a tetrachloroethylene-exposed j ob.
31          Exposure was assessed by a "grab sample" and not as a weighted average (as often
32    reported in other occupational studies reviewed in this section).  Exposure monitoring indicated a
33    median concentration of tetrachloroethylene of 8 ppm (range: 2-136 ppm).  An index of
34    cumulative exposure to tetrachloroethylene was  also developed for each exposed subject by
35    multiplying the tetrachloroethylene concentration by the number of years worked.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-80       DRAFT-DO NOT CITE OR QUOTE

-------
 1           Latency to and duration of vocal response to the stimulus (reading) were measured in
 2    each subject after the presentation of a sequence of words on a computer screen. For each
 3    condition, subjects were asked to say the word immediately or following delays of 0.1 or 0.5
 4    seconds. The test was performed using a random sequence of concrete or meaningless disyllabic
 5    words.  These tests were carried out at the place of employment for dry cleaners and in a clinical
 6    setting for controls, indicating that the investigators were not blinded as to a subject's exposure
 7    status. Testing conditions may  have differed between exposed group and controls.
 8           Compared with the control group, the exposed group had statistically significant longer
 9    mean reaction times and/or vocalization durations under all response conditions (immediate or
10    delayed response) with either real or meaningless words. Furthermore, statistically significant
11    positive correlations were observed between cumulative tetrachloroethylene exposure and
12    immediate reading and delayed reading tasks r = 0.69 and r = 0.73, respectively). No
13    information on alcohol consumption or other potential differences between exposed subjects and
14    controls was reported, precluding an analysis of how these factors may have affected the
15    observed association between tetrachloroethylene and reaction time.
16
17    4.6.1.2.11.  Schreiber, J.S., H.K. Hudnell, A.M. Getter, D.E. House, KM. Aldous, M.E. Force,
18    K. W. Langguth, E.J. Prohonic and J. C. Parker. 2002. Apartment residents' and day care
19    workers' exposure to tetrachloroethylene (perc) and deficits in visual contrast sensitivity.
20    Environ, Health Perspect. 110:655-664.  Schreiber et al. (2002) reported the findings from
21    investigations using visual tests to assess neurologic function in two populations:  apartment
22    residents and day care employees who had potential environmental tetrachloroethylene exposure
23    due to close proximity to dry cleaning facilities.9  Residential exposure to tetrachloroethylene
24    can result in nearly  continuous exposure (NYS OAG, 2004a) and is distinct from the pattern of
25    tetrachloroethylene exposure experienced by the occupational populations described in the
26    preceding paragraphs.  Objectives of the residential and day care investigations were to
27    characterize tetrachloroethylene exposure and to screen  for subclinical neurological effects using
28    a battery of visual function tests.  All participants—or their guardians in the case of the
29    residential study—signed voluntary consent forms prior to study commencement.
30           For the residential  study, the exposed group consisted of 17 tetrachloroethylene-exposed
31    subjects (11 adults between the ages of 20 and 50, 2 adults over the age of 60, and 4 children)
             9 The apartment residents lived in two separate buildings in New York City that each contained a dry
      cleaning business. The residential study served as a pilot for a larger study that is investigating visual effects among
      tetrachloroethylene-exposed residents. The day care study was part of an investigation of staff and children carried
      out by the NYS DOH and the Centers for Disease Control and Prevention. The day care facility, located in Albany,
      New York, was in a building that also housed a business that did dry cleaning. Visual testing for both studies was
      carried out by the same investigator using the same testing apparatus.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-81       DRAFT-DO NOT CITE OR QUOTE

-------
 1    from six families residing for an average of 5.8 years (6 years median) in two apartment
 2    buildings in New York City.  Preliminary monitoring of these buildings indicated
 3    tetrachloroethylene concentrations were elevated compared to eight other buildings also
 4    monitored by the NYS DOH. These eight buildings were identified by NYS DOH from
 5    discussions with the New York City Department of Health and from a review of files on dry
 6    cleaning facilities (NYS OAG, 2004b).
 7           Study subjects were identified through several methods: (1) both families in the first
 8    building (Buildling A) had been referred to the NYS DOH for information about participating in
 9    the study by Consumer Union/Hunter College researchers, (2) one family in the second building
10    (Buildling B) had previously contacted NYS DOH about exposure concerns and desired to
11    participate in a study, and (3) three other families in Building B were recruited by a participating
12    family (NYS OAG, 2004b).  Exposed residents were an affluent, English-speaking, Caucasan
13    population living near New York City's Central Park (telephone communication from K.
14    Hudnell, EPA, to D. Rice, EPA, February 2003). Exposed participants were generally unaware
15    of the tetrachloroethylene exposure, although some study participants did observe
16    tetrachloroethylene-like odors prior to the study period.
17           Control subjects were recruited from among NYS DOH Albany, New York employees
18    and their families. They were considered representative of the general population not living near
19    dry cleaning facilities.  All controls were Caucasan, except for one Asian individual, and were
20    age- and sex-matched to exposed apartment residents.  In some cases, more than one control
21    participant was matched to an exposed subject, and an average of the multiple control  visual
22    function test scores was used for comparison to that of an exposed subject. Mean age was 34.5
23    years for exposed apartment residents and 33.2 years for control subjects.
24           Nine adult staff (all females) of a day care facility agreed to participate in the day care
25    study.  Controls were age- and gender-matched acquaintances of the exposed participants, local
26    retail shop employees, NYS DOH employees, or staff from other local day care centers with no
27    known  tetrachloroethylene exposure. All subjects in the exposed and control groups were
28    Caucasan (telephone communication from K. Hudnell, EPA, to D. Rice, EPA, February 2003).
29    Mean age was 27.7 years for control participants and 27.2 years for day care staff; mean duration
30    of employment for exposed subjects was 4 years at the center.
31           Information on sociodemographics; lifestyle factors such as exposure to direct or passive
32    smoke, alcohol consumption, and exercise; medical history; and neurotoxicant exposure in
33    addition to the visual tests was obtained by questionnaire from both study populations and their
34    controls.  Exposed participants had no known exposure to other neurotoxicants, ongoing illness,
35    current use of neuroactive drugs, or a medical history indicative of neurologic dysfunction, and
36    both exposed participants and controls reported low or moderate alcohol consumption that did

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-82       DRAFT-DO NOT CITE OR QUOTE

-------
 1    not differ between either exposed group and their controls. Moreover, the profile of moods test
 2    scores of all residential exposed subjects were within normal limits.  The investigators also
 3    administered visual tests of acuity, contrast sensitivity, and color discrimination to exposed
 4    subjects and their referents. The investigators were not blinded as to a subject's status as either
 5    exposed or nonexposed.
 6           The assessment of tetrachloroethylene exposure of residents consisted of
 7    tetrachloroethylene concentrations in indoor air and personal air samples, exhaled breath, and
 8    blood, which were collected at the time of visual testing.  Testing was performed  during a period
 9    of active dry cleaning for four of the families and one month after closure of the facility for the
10    remaining two families in the residential study.  Additionally, adult residents provided urine
11    samples, which were analyzed for tetrachloroethylene as well as for three products of its
12    metabolism: TCA, trichloroethanol, and the urinary acetyl metabolite. Breast milk samples
13    were provided from two exposed breastfeeding mothers.
14           Ambient concentrations of tetrachloroethylene were available for all study participants
15    for an earlier time frame (from 1 to 3 months before the date of visual testing), when active dry
16    cleaning was occurring in both apartment buildings.  These measurements were used by NYS
17    DOH to identify study sites. Concentrations of airborne tetrachloroethylene levels in apartment
18    rooms were higher in these samples than in the monitoring data obtained at the time of the visual
19    testing.  Median concentrations in these samples, which were taken during the day during active
20    periods of dry cleaning, were 0.21 ppm (mean = 0.36 ppm; range: 0.1-0.9 ppm).  Airborne
21    tetrachloroethylene concentrations had decreased in samples collected at the time of visual
22    testing; median tetrachloroethylene concentration was 0.09 ppm (mean = 0.18 ppm; range:
23    0.01-0.78  ppm).  Tetrachloroethylene levels in blood correlated well with levels in room air,
24    personal air, and breath.
25           Atmospheric monitoring of the day care facility before closure of the dry cleaning
26    business showed airborne concentrations of tetrachloroethylene ranging from 0.27 to 0.35 ppm,
27    with median and mean concentrations of 0.32 ppm. Samples obtained at the time of visual
28    testing, five weeks after removal of the dry cleaning machines, approached background
29    concentrations (range:  0.0012-0.0081 ppm).
30           Visual function testing consisted of near visual acuity, near VCS, and color vision. All
31    study participants who wore corrective lenses for reading wore their lenses when  taking the
32    vision tests. The visual acuity test measured the ability to discriminate high- frequency (i.e.,
33    small) images at high contrast; e.g., reading successively  smaller black-on-white letters as part of
34    an examination for corrective lenses.  This measure typically is dependent on the  optics  of the
35    eye (and corrective lenses when needed) and is insensitive to subclinical deficits in neurologic
36    function. In neither assessment did the groups differ in visual acuity.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-83       DRAFT-DO NOT CITE OR QUOTE

-------
 1           The contrast sensitivity test is sensitive to subclinical deficits in neurologic function in
 2    the visual pathways. The test measured the least amount of luminance difference between dark
 3    and light bars that was needed to detect the bar pattern. Luminance varied between the bars in
 4    sine-wave fashion, and each test pattern represented one size of bars or spatial frequency.  The
 5    bar patterns were presented at five different spatial frequencies, thereby breaking spatial visual
 6    function into its essential components.  The least amount of luminance contrast needed to detect
 7    each bar size was measured.  The contrast sensitivity data are presented in Figure 4-1.  A
 8    strength of this study is that the test of contrast sensitivity employed a forced-choice procedure,
 9    providing better reliability  and consistency than other approaches.
10           Multivariate analysis of variance was used to analyze the VCS  data.  Group mean scores
11    for VCS across spatial frequencies were statistically significantly lower in exposed residents than
12    in controls and in day care  employees as compared with controls, indicating poorer visual
13    function in the exposed groups.  An exposure-response analysis did not show an association
14    between poorer performance and increasing tetrachloroethylene concentration. Among
15    apartment residents, mean  scores of VCS in all four children and in both older adults (60 years of
16    age) were lower than the 12th percentile score of all control subjects. (The 12th percentile
17    represents the two control subjects with the poorest performance out of the 17 total data points.)
18    In contrast,  5 of the 11 adults aged less than 60 years scored below the 12th percentile.  It is
19    unknown whether the difference between groups would have been statistically significant on the
20    basis of the adults under 60 years alone. However, there was a statistically significant lower
21    group mean VCS score across all spatial frequencies when day care employees were compared
22    with the control group (data not shown).
23           In the residential study, exposed subjects were  retested twice after the initial assessment,
24    6 to 10 months and 17 to 21 months after closure of the dry cleaning facility. Performance
25    appeared to worsen over successive evaluations, although statistical comparisons were not
26    performed (NYS DOH, 2000).  Control subjects from the initial testing were not retested,
27    preventing a comparison with observations from exposed subjects.
28           Color vision was also assessed in both the residential and the day care groups.  Subjects
29    were asked to put a series of small round "caps" in order by color. The types of errors made
30    could distinguish specific types of color vision deficiency:  e.g., red-green color blindness, which
31    is common in males, or blue-yellow color blindness, which is associated with solvent exposure
32    (Mergler and Blain, 1987; Mergler et al.,  1987,  1988a, b, 1991; Campagna et al., 1995, 1996).
33    Group differences in the CCI were assessed using two-tailed Student's t-tests for matched-pair
34    analyses. CCI scores of exposed groups did not show  statistically significant impairment as
35    compared with referents, although the performance of the exposed groups, particularly the
36    residential group, appeared worse than that of control.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-84       DRAFT-DO NOT CITE OR QUOTE

-------
                                 140
                               >, 120
                               a 100
                               W
                               ra
                               o
                               15
                               3
                               Ifl
                                 140
                               >, 120
                               S 100
                               P  80
                               Ifl
                               ra
                               o
                               O
                                  60 -
                               O
                               T5
                              Control, <18 years old
                              Exposed, <18 years old
                                        1.5      3       6       12   18
                                         Spatial Frequency (Cycles/Degree)
                                        Spatial Frequency (Cycles / Degree)
                                          Control, >60 years old
                                          Exposed, >60 years old
                                         Spatial Frequency (Cycles/Degree)
 4
 5
 6
 7
 8
 9
10
11
Figure 4-1. Visual contrast sensitivity functions for control and exposed
children (top), adults that were identified as having impaired function (i.e.,
5 of the total 11) and their matched controls (middle), and the control and
exposed individuals over 60 years of age.  The x-axis represents the frequency
of the stimulus bars, with finer bars toward the right.  The y-axis represents the
inverse of the contrast at which the subject could no longer distinguish the
orientation of the bars (threshold).  For any frequency, a higher contrast
sensitivity threshold represents better visual function. It is apparent that the group
of children is relatively more impaired than the impaired group adults.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                     4-85        DRAFT-DO NOT CITE OR QUOTE

-------
 1          Observations in the study paper have been questioned, particularly on selection bias in
 2    the residential investigation as an explanation of observed VCS deficits (HSIA, 2004). Although
 3    motivation for study participation is not known, the New York State Office of Attorney General
 4    (NYS OAG, 2004b) noted that test results were provided to individual study subjects, which
 5    probably encouraged participation; however, the principal investigator does not believe selection
 6    bias was a factor for study participation. Some information in the study paper may also be used
 7    to judge the potential for selection bias. The study authors noted that the profile of moods test
 8    scores of all exposed residential subjects were within normal limits, with no cases of clinical
 9    depression or other neuropsychiatric conditions. Hence, it does not appear that exposed residents
10    had major psychological impairments.  Additionally, bias may be introduced through the use of
11    controls living in Albany for comparison with exposed residents living in New York City.
12    Information on covariates is lacking, and the impact of these covariates on VCS function cannot
13    be adequately assessed.
14          Some general information is available to evaluate potential  confounding  due to education,
15    occupation, and residential location. Factors such as education, socioeconomic status, and
16    smoking do not affect the VCS test (NYS OAG, 2004b; Hudnell et al., 2001; Mergler et al.,
17    1991; Frenette et al., 1991; U.S. EPA, 2004). Occupation is highly correlated with
18    socioeconomic status (Deonandan et al., 2000) and is not likely to confound the  VCS test.
19    Moreover, urban-rural  differences between exposed and control subjects are not thought to
20    strongly bias findings.  For example, Kaufman et al. (1988) did not show that urban or rural
21    residence was related to performance on specific subtests of the Wechsler Adult Intelligence
22    Scale, although  associations were seen with other variables such as sex, age, and education,
23    variables that are similar or matched for exposed and referent subjects in the current study.
24          Public comments to EPA (NYS DOH, 2004) discuss that two of the four children in the
25    residential study had medically verified diagnoses of learning disabilities or developmental
26    delays; however, no information was provided in these public comments about these conditions
27    in referent children. Without comparable information on control children, it is difficult to draw
28    any conclusions about whether these conditions may or may not have also contributed to the
29    VCS deficits observed in residential subjects.
30          Finally, as with all other studies discussed in this section, unmeasured differences or
31    residual confounding between exposed and referent groups may possibly explain observations;
32    however, in the  absence of information, it is not possible to evaluate the unmeasured variables.
33
34    4.6.1.2.12. Sharanjeet-Kaur, MursyidA, Kamamddin A, Ariffln A. 2004. Effect of petroleum
35    derivatives and solvents on colour perception. Clin Exp Optom 87:339-343. Fourteen healthy
36    subjects of ages 24-53 years working in 3 dry cleaning facilities using tetrachloroethylene are

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-86       DRAFT-DO NOT CITE OR QUOTE

-------
 1    included in a study assessing color vision. This study was part of a larger study assessing color
 2    vision in two other occupationally-exposed populations, 39 workers in a factory producing
 3    polyethylene resins plastic storage containers and 40 workers manufacturing polystyrene plastic
 4    bags.  The published study is poorly reported, lacks many details, and adopts post-hoc statistical
 5    testing. The paper reports neither how facilities were identified nor recruitment methods for
 6    study subjects. Furthermore, the paper does  not present any information on tetrachloroethylene
 7    concentrations or on tetrachloroethylene biomarkers, making it difficult to judge the degree of
 8    exposure to tetrachloroethylene. Control selection criteria are not identified in the published
 9    paper other than 27 healthy subjects (mean age 27 + 4 years) composed Control Group 1 and 2
10    healthy subjects (mean age 33 + 4 years) who were support staff of Universiti Kebangsaan
11    Malaysia.  Dry cleaning workers differed from controls on several variables: work duration,
12    hours worker per day, cigarette smoking, mean age (compared to Control Group 1), and race.
13    Also, no information is presented on possible difference between dry cleaners and controls on
14    socio-economic status (SES).  Voluntary consent was provided by all subjects.
15          Visual testing was carried out at the factory or dry cleaner, for exposed subjects, and at
16    the Optometry Clinic in the Universiti Kebangsaan Malaysia for control subjects.  Given these
17    testing conditions, a subject's exposure status was known, i.e., no blinding.  Visual acuity was
18    measured at distance using the Snellen chart and at near using a reading chart.  Subjects were
19    excluded with poor visual acuity or with systemic, ocular, or neurological diseases; the number
20    of excluded subjects is not identified in the published paper. Color vision was assessed
21    binocularly using Ishihara plate, D-15 test, and Farnsworth Munsell 100 Hue test under a light
22    box at illumination of 1,000 lux. Subjects wore the best corrective lens and testing was carried
23    out at a distance of 35 to 40  cm.
24          The number of subjects with abnormal scores, using criteria of Vingrys and King-Smith
25    (1988), is presented but not group-mean color confusion index scores. None of the controls or
26    dry cleaners had color vision errors with the  Ishihara plates. In contrast, 6 dry cleaners (43%)
27    and 13 subjects (93%) compared to no controls were identified with errors on the D-15 test and
28    FM 100 Hue test, respectively. Statistical testing of differences is lacking.  Total error scores for
29    the FM 100 Hue test differed between dry cleaners and control group 2 (p < 0.05) but not with
30    control group  1. It is difficult to interpret these findings due to the lack of exposure information
31    on potential tetrachloroethylene exposure other than job title, and differences between dry
32    cleaners and controls regarding test conditions, SES, and smoking.
33
34    4.6.1.2.13. New York Department of Health (NYS DOH). 2005a. Improving human risk
35    assessment for tetrachloroethylene by using biomarkers  and neurobehavioral testing. Final
3 6    Technical Report to US EPA Star Grant #R82 7446.  Accessed 5 December 2006,

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-87       DRAFT-DO NOT CITE OR QUOTE

-------
 1    http://cfpub.epa.gov/ncer_abstracts/index.cfm/fuseaction/display.abstractDetail/abstract/97
 2    7/report/O. NYS DOH (2005a) examines the effect of tetrachloroethylene exposure on visual
 3    function in two populations, residents living in a building co-located with a dry cleaning
 4    establishment and among employees and former students of a day care establishment exposed 4
 5    years previously during a period when the day care was co-located near a dry cleaner. The first
 6    study, the New York City Perc Project, did not include subjects studied by Schreiber et al. (2002)
 7    and employed different methods for testing visual contrast sensitivity and color vision.  NYS
 8    DOH (2005a) enlisted 65 households in 24 residential buildings with dry cleaners using
 9    tetrachloroethylene on-site and 61 households in 36 buildings without dry cleaners located in the
10    study area in Manhattan, New York City.  Health outcome and tetrachloroethylene
11    concentrations as measured from indoor air monitoring and in exposed subject's breath and
12    blood were obtained over the period from 2001 to 2003. The full report of the residential study
13    has not received public peer review nor has it been published as a literature paper although
14    McDermott et al. (2005) presents exposure monitoring findings from the dry cleaner households.
15           The second project, the Pumpkin Patch Day Care Center (PPDCC) Follow-up Evaluation,
16    is a 5-year follow-up of visual function among some employees and neurobehavioral function
17    among children in a day care center that had been previously co-located in a building with a dry
18    cleaning establishment. The PPDCC Follow-up Study also included first-time visual function
19    tests of former students. NYS DOH together with the U.S. Centers for Disease Control and
20    Prevention carried out the initial evaluation of PPDCC staff and students in 1998 (NYS DOH,
21    2005b). Funding to NYS DOH for the residential study and the PPDCC Follow-up study was
22    provided though U.S.  EPA  STAR Grant #827446010 (NYS DOH, 2005a, 2005c).
23
24    4.6.1.2.13.1. New York City Perc Project. The objectives of the New York City Perc Project are
25    as follows: 1) to document tetrachloroethylene exposures in buildings where dry cleaners were
26    present; 2) to evaluate whether living in a building with a dry cleaner was associated with CNS
27    effects; 3) to evaluate the relationship(s) between measures of tetrachloroethylene exposure and
28    CNS effects; and, 4) to assess whether children were disproportionately exposed to and/or
29    affected by tetrachloroethylene compared to adults. Study design and protocols were approved
30    by Institutional Review Boards at the NYS DOH and other collaborating institutes (Mr. Sinai
31    Medical Center and U.S. CDC).
32          Subjects were identified in buildings from eleven zip code areas surrounding Central Park,
33    New York City, contiguous with one another, but different in demographic and socioeconomic
34    characteristics. Eligible households for participating in this study include at least one adult
35    (20-55 years old) and one child (5-14 years old), so as to assess whether residential
36    tetrachloroethylene exposure would disproportionately affect children. Initial monitoring

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-88       DRAFT-DO NOT CITE OR QUOTE

-------
 1    indicated few residences in dry cleaner buildings with elevated indoor air concentration of
 2    tetrachloroethylene above the current NYS DOH residential air guideline of 0.015 ppm
 3    (0.1 mg/m3).  The study area was broadened to include buildings subject of a resident complaint
 4    and to include buildings in additional zip codes, primarily characterized by lower SES or higher
 5    percentage of minority residents. This decision was made after the finding of elevated
 6    tetrachloroethylene levels in households in dry cleaner buildings in one zip code area, a low
 7    income, minority area, compared to other zip codes area. Mail and telephone contact were the
 8    primary methods of recruiting subjects, with door-to-door recruitment by a not-for-profit child
 9    advocacy organization (Northern Manhattan Perinatal Partnership, Inc.) providing assistance
10    with recruiting bilingual  subjects primarily living in 3 of the 11 zip code areas.  Of the 1,261 dry
11    cleaner and 1,252 reference  households contacted, 132 dry cleaner households and 175 reference
12    households included age-eligible adult-child pairs and a total of 65 dry cleaner (67 adults, 68
13    children) and 61 reference households (61 adults, 71 children) participated in the study. All
14    participants or their guardians signed voluntary consent forms prior to study commencement.
15           Tetrachloroethylene  indoor air concentrations in dry cleaner buildings had decreased
16    since 1997, the period of the pilot study (Schreiber et al., 2002), and ranged up to around 0.77
17    ppm (5 mg/m3) with a geometric mean of 0.005 ppm (0.035 mg/m3).  Monitoring was carried out
18    using passive monitoring badges. In comparison, tetrachloroethylene concentrations in building
19    without dry cleaners ranged up to 0.014 ppm (0.09 mg/m3) with a geometric mean of 0.0004
20    ppm (0.003 mg/m3). Both breath and blood tetrachloroethylene levels were significantly
21    (p< 0.05) correlated with indoor air concentration for adult and for child subjects of dry cleaner
22    buildings. Levels of detection (LODS) were 5 |ig/m3 air and 0.048 mg/ml blood. Air, breath,
23    and blood tetrachloroethylene concentrations were inversely correlated with income and were
24    higher among minority compared to non-minority subjects.
25          NYS DOH staff visited participants in their residences to collect  24-hr indoor air samples,
26    breath samples, and to give adult participants a questionnaire seeking information on residential,
27    occupational, and medical history for themselves and their children.  Opthalmologic
28    examinations were scheduled at the same time for the Mt Sinai School of Medicine Department
29    of Opthalmology research clinic. Participants received financial compensation after completing
30    the home visit ($50.00) and  ophthalmology clinic visit ($50.00).
31           No differences between exposure groups were observed for participants recruited using
32    the mail and telephone method, although this was not so for participants recruited using door-to-
33    door methods. For these individuals, language and adult age differed significantly between
34    exposed and non-exposed groups with more English speaking households participating in the
35    non-exposed group and non-exposed adults were slightly older than exposed adults.  Overall,
36    differences between adult residents  of reference buildings or buildings with dry cleaners in SES

                 This document is a  draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-89      DRAFT-DO NOT CITE OR QUOTE

-------
 1    characteristics, residence duration, education level, age smoking or alcohol use are not apparent.
 2    Differences between child residents in gender or residence duration are not apparent, but the
 3    highest exposure group is about a year younger and has about one less year of education than
 4    children in the other exposure groups.
 5          Ophthalmologic examinations and visual function tests were given to study participants
 6    at the Mt. Sinai Medical School of Medicine.  The final report does not describe whether
 7    examiners were or were not blinded as to a subject's exposure status (NYS DOH, 2005a).  The
 8    examination included determination of past ocular and medical history; measurement of visual
 9    acuity, pupil size, extrocular motility, and intraocular pressure; and anterior and posterior
10    segment exams. Subjects with abnormalities or taking medications that could influence VCS
11    and/or color vision were excluded from further testing. Furthermore, visual functional tests for
12    some children were excluded from the statistical analysis because of their young age or because
13    they were identified by their parents as learning disabled or having attention deficit hyperactivity
14    disorder.  VCS was determined using the Functional Acuity Contrast Test (FACT) distance chart
15    placed 10 feet from the participant under light conditions of 68-240 cd/m2.  These testing
16    conditions differ from those employed by Schreiber et al. (2002) in their residential study where
17    visual test was carried out assessing near contrast sensitivity.
18          Adults and children demonstrated a ceiling effect with VCS performance, i.e., a
19    maximum score at 1.5, 3, 6, 12, and 18 cycles per degree (cpd) is achieved by some study
20    participants. VCS scores among adults were not correlated with any SES factor or personal
21    characteristics (smoking, alcohol use, education level, duration of residence).  Among all
22    children, poorer VCS at 1.5, 3, and 6  cpd were significantly correlated with speaking primarily
23    Spanish at home.
24          NYS DOH examined possible association between VCS and tetrachloroethylene
25    exposure by, (1) comparing the percent of exposed subjects with maximum VCS score (no
26    errors) to referents, (2) comparing mean  differences in VCS scores between adult and child
27    subjects living in the same residence  across exposure categories, and (3) using logistic regression
28    to assess the effect of tetrachloroethylene in indoor air, blood or breathe on the achievment of
29    maximum VCS score at 6 and  12  cpd. Analyses examining relationships between
30    tetrachloroethylene and visual function were conducted with the referent exposure group
31    (background exposure, living in a building without a dry cleaner), <0.015 ppm (<100 |ig/m3),
32    and >0.015  ppm (>100 |ig/m3).
33          Several analyses suggest a susceptibility of exposed subjects, particularly among children,
34    to tetrachloroethylene on VCS performance at higher spatial frequencies.  A decreasing trend
35    (p < 0.05) was observed between increasing residential tetrachloroethylene exposure and the
36    proportion of adults achieving the maximum contrast sensitivity score at 6 cpd and in the

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-90       DRAFT-DO NOT CITE  OR QUOTE

-------
 1    proportion of children achieving the maximum contrast sensitivity score at 6 and 12 cpd; i.e., a
 2    lower proportion of participants with a maximum VCS score in the highest exposure category
 3    compared to referents living in a building without a dry cleaner.  Stratified analyses suggested a
 4    lower percentage of low income and minority children with maximum VCS scores at a given cpd
 5    than higher income and non-minority children, but sample sizes in the highest exposure group,
 6    especially in higher income, non-minority groups, limit reliability of this observation. Race did
 7    not appear to confound the association in adults between VCS at 6 cpd and tetrachloroethylene.
 8          VCS scores in children at a given cpd were generally higher (better contrast vision) than
 9    the VCS score of an adult living in the same apartment.  Using differences between adult-child
10    pairs in each exposure grouping to assess possible tetrachloroethylene effects, the advantage of
11    children over adults appeared much smaller in the >0.015 ppm (>100 |ig/m3) category at 12 cpd
12    (mean difference of 10.9) compared to referents (mean difference of 21.6) but was not
13    statistically significant from the mean difference in the referent population (p = 0.16).
14          Results from logistic regression analyses further support susceptibility of children but not
15    adults. Whereas adult VCS at 6 or 12 cpd was not significantly influenced by any measure of
16    tetrachloroethylene exposure, VCS performance at 12 cpd among children was significantly
17    influenced (p < 0.05) by tetrachloroethylene concentrations in either indoor air or in blood; that
18    is, a lower percentage of children achieved a maximum VCS score with higher
19    tetrachloroethylene exposure. Analyses of tetrachloroethylene breath concentrations and VCS
20    performance at 12 cpd in children appeared to support the findings with indoor air and blood, but
21    were of borderline statistical significance. Logistic regression models examining VCS findings
22    in either children or adults are not adjusted for potentially confounding factors such as SES,
23    education, smoking, alcohol use, age (for children) and gender (for children); these variables
24    were correlated with one another as well as with tetrachloroethylene, but not with VCS
25    performance.
26          Color vision was assessed biocularly using both the Farnsworth D15 and Lanthony's
27    Desaturated 15 Hue Test. Both tests were administered under light conditions specified by the
28    manufacturer. The number of errors for each eye was recorded by noting instances of inversions
29    involving a single cap (minor error) and instances of inversions involving two or more caps
30    (major errors).  Total Color Distance Scores (TCDS) were determined and a CCI was calculated
31    for each participant according to Geller (2001) and Bowman (1982).
32          Analyses were carried out using the proportion of subjects with no errors, comparing
33    quantitative differences in CCI, and logistic regression modeling to assess associations between
34    tetrachloroethylene exposure measures and occurrence of any major errors.  A comparison of
35    differences in CCI and major error between children and adults residing in the same household
36    was used to assess the possible vulnerability of children.  A high proportion of adult and child

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-91       DRAFT-DO NOT CITE OR QUOTE

-------
 1    participants scored perfectly on both the Farnsworth and Lanthony color vision tests. Lower
 2    annual household income, being a member of a minotiry group, speaking primarily Spanish at
 3    home, and fewer years of education were all significantly associated with increased CCI on both
 4    color vision tests.
 5          Tetrachloroethylene measures of exposure were unrelated to color vision performance
 6    among adults; however, similar to VCS performance, children appear as a susceptible population.
 7    There were no differences between exposure groups for either adults or children in the percent of
 8    subject with major effort on both color vision tests. A comparison of mean CCI between
 9    exposure groups showed that children in the highest exposure category performed worse (mean
10    CCI of 1.3, range 1.0-1.9) than children in the low exposure category (mean CCI of 1.1, range
11    1.0-1.7) and to referent children (mean CCI of 1.2, range 1.0-2.0) on the Lanthony test; the test
12    for trend for the three exposure groups was statistically significant (p < 0.05). Performance
13    (mean CCI) on the less sensitive Farnsworth test was not associated with tetrachloroethylene
14    exposure in either adults or children. Moreover, for children, tetrachloroethylene in breath was
15    significantly associated (p < 0.05) with making one or more major errors on the Lanthony color
16    vision test in logistic regression analyses that adjusted for the effects of age and gender.  Logistic
17    regression analyses examining color vision and other tetrachloroethylene measures such as
18    indoor tetrachloroethylene concentration or breath concentration were not discussed in NYS
19    DOH (2005a.).  Last, the higher mean difference in CCI between children and adults in the
20    highest exposure category, >0.015 ppm (>100  |ig/m3), and referents was statistically significant.
21    Children in the high exposure group were a year younger than in other exposure groups; age was
22    correlated  with CCI and with tetrachloroethylene exposure in this study. The highly correlated
23    variables and the few numbers of children in the high exposure group limits analysis of age
24    effects on the association between breath tetrachloroethylene concentration and CCI.
25          In summary, this  study adopts a different approach than Schreiber et al. (2002) to assess
26    vision, using far vision methods as opposed to the near vision methods of Schreiber et al. (2002).
27    For both contrast vision and color vision, a number of analyses in NYS DOH (2005a) are
28    suggestive of vulnerability among children. The association with vision effects in children and
29    exposure to >0.015 ppm  (>100 |ig/m3) tetrachloroethylene support findings from the earlier pilot
30    study (Schreiber et al., 2002). Exposure to >0.015 ppm (>100 |ig/m3) tetrachloroethylene was
31    highly correlated with race and children's age, and the sample sizes in the highest  exposure
32    group, especially in higher income, non-minority groups, makes it difficult to fully examine
33    possible effects of income, race, and age on vision. However, association of tetrachloroethylene
34    exposure >0.015 ppm (>100 |ig/m3) with visual deficits suggests a susceptibility of the
35    population studied.
36

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-92        DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.6.1.2.13.2. Pumpkin Patch Day Care Center follow-up evaluation.  The objective of the
 2    PPDCC Follow-up Evaluation was to assess neurobehavioral function in former students of
 3    PPDCC after a 5 year post exposure period and to carry out first-time testing of visual function
 4    of the former students. Additionally, visual function testing was carried out on five staff exposed
 5    to tetrachloroethylene 5 years previously.  The NYS DOH final report to EPA (NYS DOH,
 6    2005c) provides a full description of testing in children but not adults.  The discussion of visual
 7    tests on former PPDCC is contained in NYS DOH (2005b). The initial testing in 1998 of vision
 8    in PPDCC staff and of neurobehavior in children is contained in NYS DOH (2005b).
 9           Children eligible for testing in the current evaluation were enrolled in the New York State
10    Volatile Organic Chemical (VOC) Registry and had attended PPDCC.  There were 115 children
11    who met this criteria.  Of this group, 27 children with the highest number of hours spent at
12    PPDCC were asked through letters or by phone to participate; 17 children completed vision
13    testing  and 13 children completed some or all of the neurobehavioral assessment.  Referents
14    were children who attended other day care centers and who were about the same age as PPDCC
15    participants. No information is provided on methods employed for referent participation.
16    Exposed and referent subjects were matched on day care experience, age,  and gender. Overall,
17    17 PPDCC and 13 comparison children completed vision testing and 13 PPDCC and 13
18    comparison children completed neuropsychological testing. Of these subjects, 13  matched pairs
19    completed vision  test; but only 8  matched pairs completed the neurobehavioral test.  No
20    information is provided in the NYS DOH final report on how many of the 13 former PPDCC
21    students were part of the PPDCC student group who underwent neurological testing  5 years
22    previously one month after the close of the dry cleaner facility.
23           Neurobehavioral evaluations consisted of a battery of tests that assess general intellectual
24    function, attention/information processing speed, visuospatial ability, reasoning and logical
25    analysis, memory, motor functions, and sensory-perceptual functions. Tests were  administered
26    in fixed order on two different days. All children completed the same tests with the exception of
27    the Halstead-Reitan Neuropsychological Batteries. Children age eight or younger were
28    administered the Reitan-Indiana Neuropsychological Test Battery and the Halstead-Reitan
29    Neuropsychological Test Battery for Old Children was administered  to children age  nine or older.
30    Children also performed portions of the computerized Neurobehavioral Evaluation System-2
31    (NES-2) which assessed perceptual-motor skills, attention, visual memory, and mood.  A parent
32    or guardian completed the Child Behavioral Checklist and a background history questionnaire.
33    All neurobehavioral evaluations were conducted at the office of Albany Psychological
34    Associated, P.C.,  in Albany, NY.
35           Independent samples t-tests were performed on the scores from the Wechsler Intelligence
36    Scale for Children, Children's Memory Scale, the Halstead-Reitan Neuropsychological Test

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-93       DRAFT-DO NOT CITE OR QUOTE

-------
 1    Battery for Old Children, and Reitan-Indiana Neuropsychological Test Battery. Age was
 2    significantly correlated with performance on the Purdue Pegboard and many subtests in the NES-
 3    2 and analysis of covariance was completed on subtests from the NES-2 and Purdue Pegboard
 4    with age as a covariate.  Each child's performance level on the neurobehavioral tests was
 5    determined by comparing his/her test score to normative information for the specific test or
 6    battery. For NES-2, performance of the referent children, children who attended other day care
 7    centers in Albany and who were about the same age as PPDCC participants, was used as the
 8    normative basis, with  scores 2 S.D. below the mean of the same age and gender from the
 9    normative data being classified as impaired.
10          Neurobehavioral function of the 13 PPDCC children evaluated in this follow-up study did
11    not differ from that of the 13 referent children. PPDCC children performed better than referent
12    children on several tests but performance was within normative ranges. These results are not
13    surprising. Neuropsychological or behavioral testing was conducted in October 1998 by the
14    auspices of the U.S. Centers for Disease Control (U.S. CDC) on children then of ages four and
15    five. No consistent differences in neurological function were found between 18 children who
16    then attended the day  care center and 18 age- and gender-matched control children who did not
17    attend the day care center, although a statistically significant association was found between
18    duration of attendance at PPDCC and poorer performance on the Purdue pegboard test with the
19    dominant hand (NYS  DOH, 2005b).
20          Visual function testing consisted of visual acuity, far VCS,  and color vision.  Visual
21    contrast sensitivity was determined monocularly using the Functional Acuity Contrast Test
22    distance chart placed 10 feet from the participant under light conditions specified by the
23    manufacturer. Scores for each eye were recorded on a graph showing a normal range (90%
24    confidence interval) of VCS at each spatial frequency. Color vision was assessed using both the
25    Farnsworth D15 and Lanthony's Desaturated 15 Hue Test under light conditions specified by the
26    manufacturer. For both tests, for each eye, participants were shown a rectangular box containing
27    16 color caps arranged in chromatic order.  The test administrator removed  15 caps, leaving the
28    first as a stand, and randomized them in front of the participant. Participants were asked to place
29    the cap which most closely  matched the stand in hue in the box next to the stand, and to continue
30    the process until all colored caps were in the box.  When the participant was done, the order  of
31    cap placement was recorded and diagramed on templates accompanying the tests.  Both color
32    vision and contrast sensitivity tests were performed monocularly.  Ophthalmologic examinations
33    and visual function testing was performed by Cornea Consultants of Albany, NY.  Examiners
34    were not blinded but were not told whether participants were associated with the PPDCC.
35          VCS results for all 13 matched pairs of children were analyzed using the Wilcoxon
36    matched-pairs signed-ranks test. Two pairs of PPDCC and comparison children were six years

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-94       DRAFT-DO NOT CITE OR QUOTE

-------
 1    old during vision testing; all other children were aged seven or more years. So as to examine the
 2    effect of age on visual function, analyses were conducted using the 13 child pairs and excluding
 3    two pairs who were <6 years old. PPDCC children performed better on the VCS test compared
 4    to referent children.
 5           Color vision results were evaluated in several ways.  Statistical analyses were performed
 6    for matched pairs (n= 13) only.  Proportions of pairs of children with discordant clinical
 7    judgements and with discordant numbers of major errors were assessed using McNemar's Exact
 8    Test for Correlated Proportions.  Furthermore, difference in CCI between matched PPDCC and
 9    referent children were assessed using Wilcoxon Matched-Pairs signed-rank test.  As for VCS
10    results, analyses were conducted using the all child pairs and excluding pairs who were <6 years
11    old. No significant difference in proportions of children with abnormal color vision or with
12    children making major errors between PPDCC and comparison children for either of the color
13    vision tests were  found. Similarly, PPDCC and referent children were not significantly different
14    on CCI for either color vision test.
15
16    4.6.1.2.14.  Perrin, MC; Opler, MG; Harlap, S; Harkavy-Friedman, J; Kleinhau,s K; Nation,
17    D; Fennig, S; Susser, ES; Malaspina, D.  2007. Tetrachloroethylene exposure and risk of
18    schizophrenia: Offspring of dry cleaners in apopulaton birth cohort, preliminary findings.
19    Schizophr Res. 90(1-3):251-254. Perrin et al. (2007) evaluates the time to a diagnosis of
20    schizophrenia among a cohort of 88,829 births born between 1964-1976 in the Jerusalum
21    Perintal Project, a population-based cohort. Births in this cohort were linked to the database of
22    Israel's Psychiatric Case Registry (PCR), with cases identified using a broad definition of
23    schizophrenia-related disorders as recorded as hospital discharge codes. Diagnoses for
24    individuals with psychosis were validated  and the date of onset was identified as the date of first
25    psychiatric admission. Of the 88,829 births, 136 offspring were born to parents identified as
26    having a job title of dry cleaner on the birth certificate;  120 offspring whose fathers but not
27    mothers were dry cleaners, 20 whose mothers but not fathers were dry cleaners; and 4 with both
28    parents as dry cleaners;  4 of the 136 births had a later diagnosis of schizophrenia. The relative
29    risk (crude) between schizophrenia and parental employment in dry cleaning was 3.9 (95% CI =
30    1.3-9.2) using proportional hazard methods. The investigators noted risk estimates did not
31    greatly change when fitting proportional hazard models that adjusted for a number of potentially
32    confounding variables; although adjusted relative risk (RR) estimates are not reported in the
33    paper.  Variables considered as possible confounders were parents' age, father's social class,
34    duration of marriage,  rural residence, religion, ethnic origin, parental immigration status,
35    offspring's birth order, sex, birth weight and month of birth.  Family history of mental illness
36    was not included as a covariate; rates of schizophrenia are higher among relatives of patients

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-95       DRAFT-DO NOT CITE OR QUOTE

-------
 1    than in the general population (Mueser and McGurk, 2004).  The magnitude of this possible bias
 2    on the association between parental occupational employment as a dry cleaner and schizophrenia
 3    in offspring can not be judged given the information provided in the paper.
 4
 5    4.6.1.3. Summary  of Neuropsychological Effects in Low- and Moderate-Exposure Studies
 6           It is important to compare outcomes across studies in order to determine whether it is
 7    possible to identify a pattern of neuropsychological  deficits produced by tetrachloroethylene.
 8    Table 4-5 is presented as an aid for this comparison. Primarily these studies have assessed
 9    neurobehavioral and, to a limited extent, neurophysiological  effects of tetrachloroethylene
10    exposure using a number of statistical methods of varying sensitivity, from simple methods that
11    are more susceptible to multiple comparison errors to regression analyses that control for
12    potentially confounding effects.  A clinical neurological examination that includes the Romberg
13    test, tests of body balance, and neuroradiological examination has not been widely incorporated
14    into the tetrachloroethylene epidemiologic studies. Neurophysiological tests such as EEGs,
15    nerve conduction tests, and evoked potentials (EPs)  have seen limited use for assessing
16    neurotoxicologic effects in tetrachloroethylene-exposed populations. Although statistically
17    significant alterations in VEPs were reported by Altmann et al. (1990, 1992)  with 4-hr acute
18    exposure at 10 ppm, they were not altered in residents exposed chronically to a median of around
19    1 ppm tetrachloroethylene (Altmann et al., 1995).
20           Acute and chronic exposures are of different patterns, short-term peak exposure versus
21    longer duration exposure, and, therefore, may result in a different pattern of toxicity.
22    Furthermore,  studies  assessing peripheral neuropathy and tetrachloroethylene uniquely were not
23    found, and studies reporting tetrachloroethylene exposure as  one of a number of solvent
24    exposures (Albers et al., 1999; Antti-Poika, 1982a, b) are not informative, as discussed in
25    Section 4.6.1.
26           Several occupational studies of dry cleaner and laundry workers and the residential study
27    by Altmann et al. (1995) share a common  set of tests from a neurobehavioral battery. Tests in
28    this battery have been widely administered to occupational populations in different settings with
29    a reasonably high degree of reliability (Anger et al.,  2000). Moreover, these tests have been used
30    in clinical or experimental research to assess normal nervous system functioning, and they
31    measure a range of sensory and cognitive function.  Studies that used a test battery include
32    Ferroni et al. (1992),  Seeber (1989), Echeverria et al. (1994,  1995), and Altmann et al. (1995).
33    Both the Seeber and the Echeverria et al. studies involved more subjects than did the studies by
34    Ferroni et al. and Altmann et al.  and statistical analyses, such as in Altmann et al., controlled for
35    a number of potentially confounding factors.  The Ferroni et  al.  study was not well-reported and
36    was methodologically poorer than the other studies.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-96        DRAFT-DO NOT CITE OR QUOTE

-------
Table 4-5. Summary of neuropsychological effects of tetrachloroethylene in humans
Neurological
effects
Contrast
sensitivity
(spatial vision)
CCI
(color vision)
VEP
(vision)
Fine motor
function
Simple RT
(attention,
motor)
Continuous
performance
(vigilance)
Visuo spatial
function
Study, number of subjects, exposure level (ppm)
Cavalleri
etal.
(1994)
70
7

+/-b





Echeverria
etal.
(1994, 1995)a
173
<0.2
3
9






+,+,+
pattern
recognition,
reproduc-tion,
memory
65
11
23
41






+,+,+
pattern
recogni-tion,
reproduc-tion,
memory
Ferroni et
al. (1992)
90
15



-
+
+

Seeber
(1989)
185
12
53



-


+
digit
reproduc-
tion
Spinatonda
et al. (1997)
74
8







Altmann et
al. (1995)
37
0.7


-
-
+
+
+
Benton
Schreiber et al. (2002)
Residents
34
0.4
+
trend +





Daycare
18
0.3
+






Lauwerys
et al. (1983)
59
21




—


Nakatsuka
et al. (1992)
184
15

C






-------
        Table 4-5.  Summary of neuropsychological effects of tetrachloroethylene in humans (continued)



Neurological





Information
processing
speed




Digit span,
digit symbol
Cancellation
(visual
scanning)
Trailmaking
(executive
function)
Study, number of subjects, exposure level (ppm)
Cavalleri
et al.
(1994)

70
7

















Echeverria
et al.
(1994, 1995)a

173
<0.2
3
9
















65
11
23
41
_






-, - (both)








Ferroni et
al. (1992)

90
15


_






-








Seeber
(1989)

185
12
53

+
perceptual
threshold
"delayed
reaction" on
choice
reaction time
+/-,+

+






Spinatonda
et al. (1997)

74
8


+
vocal
reproduction













Altmann et
al. (1995)

37
0.7

















Schreiber et al. (2002)


Residents

34
0.4


















Daycare

18
0.3


















Lauwerys
et al. (1983)

59
21


















Nakatsuka
et al. (1992)

184
15

















a Field study, no unexposed controls.
b Positive in dry cleaners, negative in ironing workers with lower exposure.
0 Using less sensitive test instrument and data analysis procedure than the other studies.

RT     = Reaction time
+       = Impaired performance in exposed group
-       = No effect of tetrachloroethylene

-------
 1           Cognitive domains affected by tetrachloroethylene include visuospatial function,
 2    attention, vigilance, and speed of information processing (choice reaction time; Table 4-5).
 3    Effects on visuospatial function are of particular interest, given the finding in the four studies
 4    that examined this domain and  similar reports for other solvents (Morrow et al.,  1990; Daniel et
 5    al., 1999). Echeverria et al. (1995) found effects on tests of pattern memory, visual reproduction,
 6    and pattern recognition  in the absence of effects on attention (digit symbol  and digit span) or
 7    executive function (Trailmaking A and B).  Further, Echeverria and colleagues (1994) confirmed
 8    these findings in an independent sample of dry cleaners in their follow-up study  (U.S. EPA,
 9    2004).
10           Seeber (1989) also reported impaired visuospatial recognition in both exposure groups,
11    and Altmann et al. (1995) observed deficits on a test of visuospatial function in residents with
12    much lower exposure concentrations than those of the two occupational studies.  These studies
13    are considered to provide strong weight, given the numbers of subjects and their use of
14    appropriate statistical methods, including adjustment for potentially confounding factors.
15    Additionally, they considered potential bias and confounding more carefully than did other
16    studies in this review.
17           Altmann et  al. (1995) and Ferroni et al. (1992) assessed vigilance using a continuous
18    performance procedure  in which the subject faces a screen that presents one of several different
19    stimuli at random intervals. The subject  must make a response to a specified stimulus and not to
20    the others. This test measures sustained attention and is correlated with performance on tests of
21    executive function. Both studies found deficits as a result of tetrachloroethylene exposure on
22    this task. Seeber (1989) found  effects on two tests of attention (cancellation d2 and digit
23    symbol) that are subsets of the Weschler  IQ tests and were designed to be sensitive to
24    performance within the  normal range. These investigators also found positive effects on a visual
25    scanning test that is usually used to assess laterality of brain  damage but has also proved
26    sensitive to toxicant (lead) exposure (Bellinger et al., 1994).  In contrast, Echeverria et al. (1995)
27    and Ferroni et al. (1992, as described in NYS DOH, 1997) did not find effects on digit span,
28    which is given as a test  of attention and memory, or digit symbol, despite higher levels of
29    exposure than in Seeber (1989).
30           Two of these studies—an occupational study with relatively higher exposure (Ferroni et
31    al., 1992) and the Altmann et al. (1995) residential study—also assessed simple reaction time, a
32    task that uses a motor response and demands a relatively modest amount of attention; results
33    were positive in both studies. Speed of information processing was assessed in two studies,
34    Seeber (1989) and Spinatonda et al. (1997).  Seeber used two tasks: recognition and choice
35    reaction time. Effects were observed  in both groups on a task requiring recognition of briefly
36    presented stimuli. In a choice reaction time task, effects were borderline in the lower-exposure

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-99       DRAFT-DO NOT CITE OR QUOTE

-------
 1    group and negative in the higher-dose group, with no exposure-response relationship.
 2    Spinatonda et al. (1997) found effects on response to vocal and visual stimuli. A third study,
 3    Lauwerys et al. (1983), reported better performance on simple and choice reaction times.
 4           Of the occupational studies, greatest weight is placed on the Seeber (1989) observations
 5    due to the larger number of study subjects and to their consideration in the statistical analysis of
 6    potentially confounding factors. Ferroni et al. (1992), Spinatonda et al. (1997), and Lauwerys et
 7    al. (1983) all reported limited information in their published papers, particularly regarding
 8    potential confounding and bias, and because of this, they have greater inherent uncertainties than
 9    does Seeber (1989).
10           Tetrachloroethylene exposure has not been reported to affect fine motor tests. Seeber
11    (1989), Ferroni et al. (1992),  and Altmann et al. (1995) each assessed fine motor control using
12    various instruments and all three found no significant decrements in fine motor performance.
13           Deficits in blue-yellow color vision, a well established effect of solvents, were observed
14    in the high-exposure group (mean tetrachloroethylene concentration of 7 ppm) but not the low-
15    exposure group (mean tetrachloroethylene concentration of 5  ppm) in Cavalleri et al. (1994) and
16    in Muttray et al. (1997)—a study carrying lesser weight than that of Cavalleri et al.—of workers
17    previously exposed to a mixture of solvents that contained tetrachloroethylene. Overall, the
18    findings of the Cavalleri et al. study and its follow-up study (Gobba et al., 1998) are in
19    agreement with previous reports on other solvents (Geller and Hudnell, 1997; Mergler et al.,
20    1996; Mergler and Blain, 1987): the blue-yellow range of color vision was primarily affected in
21    the dry cleaners, with only a few workers showing an effect on red-green perception.
22           The absence of a color vision effect in Nakatsuka et al. (1992), who used confirmatory
23    methods to augment their screening method of Lanthony's new color test, may not be
24    inconsistent with the findings of Cavalleri et al. (1994) and Gobba et al. (1998).  There are
25    uncertainties regarding testing lighting conditions in Nakatsuka et al. (1992)—an important
26    determinant of a subject's response (Geller and Hudnell, 1997)—and the fewer subjects in this
27    study than in Cavalleri et al. (1994). A pilot study of residents living above dry  cleaners with
28    mean tetrachloroethylene exposure during active dry cleaning of 0.4 ppm (Schreiber et al., 2002)
29    also reported a trend of decreasing color vision, although this  finding was not statistically
30    significant.  The follow-up study of NYS DOH (2005a), reported to U.S. EPA as a final grant
31    report, is further suggestive of tetrachloroethylene effects on color vision, particularly in children
32    compared to their parents. Tetrachloroethylene exposure concentrations had  decreased since
33    Schreiber et al. (2002),  making it it difficult to find higher-exposed subjects.  Higher
34    tetrachloroethylene exposure, that is, exposure at or over 0.1 ppm, was highly correlated with
35    SES and belonging to a minor population. This study is not able to adjust for these possible
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-100      DRAFT-DO NOT CITE OR QUOTE

-------
 1    confounders given the sample size. Studies of a larger number of residents with similar exposure
 2    concentrations are needed to draw more definitive conclusions.
 3           Only Schreiber et al. (2002) and NYS DOH (2005a) assessed spatial vision, an effect
 4    reported for exposure to other solvents (Bowler et al., 1991; Broadwell et al., 1995; Campagna et
 5    al., 1995; Donoghue et al., 1995; Frenette et al., 1991; Hudnell et al., 1996a, b; Mergler et al.,
 6    1991).  VCS deficits in subjects with normal visual acuity were observed at low exposure
 7    concentrations in residential populations that were subject to very different dose-rates than those
 8    incurred by occupational workers (U.S. EPA, 2004).  This finding is based on few subjects in
 9    this study and is noteworthy for this reason, even  in light of questions regarding potential biases.
10    Potential bias and confounding could be introduced, in part, from a lack of blinding of testers,
11    differences in motivation between exposed and referent subjects for participating in the  study,
12    and individual differences in exposed and control populations (U.S. EPA, 2004). As discussed in
13    Section 4.6.1.2.11, occupation has not been found to  strongly affect contrast sensitivity (Hudnell
14    et al., 2001), nor does motivation (U.S. EPA, 2004).
15           Peer consultation comments on EPA's earlier draft "Neurotoxicity of Tetrachloroethylene
16    (Perchloroethylene) Discussion Paper" (U.S. EPA, 2003b) noted that the deficit in contrast
17    sensitivity could reflect a sensitivity of the visual  system to tetrachloroethylene, or it may be that
18    this test was simply carried out by a superior test method (U.S. EPA, 2004). Furthermore, the
19    peer consultants also suggested that contrast sensitivity loss may reflect impaired function
20    throughout the brain, because  contrast sensitivity  is affected by retinal, optic nerve, or central
21    brain dysfunction  (U.S. EPA, 2004). Nonetheless, drawing strong conclusions from a single
22    study is difficult, particularly in light of the paucity of data on this test in occupational
23    populations with higher exposure concentrations and in animal studies. The finding of poorer
24    performance among children with exposure to >0.1 ppm tetrachloroethylene compared to adults
25    in NYS DOH,  (2005a) adds some support for observations in Schreiber et al. (2002).
26           Mental disease of neurologic origin has not been well studied with respect to
27    environmental factors.  Perrin et al. (2007), who reports an association between schizophrenia
28    and parental exposure  in dry cleaning, is the only  such study. Other studies are needed to
29    understand the role of parental tetrachloroethylene exposure  in the development of mental
30    disease in children.
31           The epidemiologic studies all have limitations.  The body of evidence is characterized
32    generally by a  lack of studies that adopt sensitive  tests of important functions affected by
33    solvents.  The exceptions are the studies employing vision assessment  and the Echeverria et al.
34    (1994,  1995) studies.  In most studies, investigators were not blinded to subject status, a potential
35    source of bias, particularly in situations in which the investigator interacted directly with the
36    subject during  testing.  All studies used a cross-sectional design, which is weaker than a

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-101      DRAFT-DO NOT CITE OR QUOTE

-------
 1    longitudinal design for a number of reasons, including a greater potential for selection bias and
 2    exposure misclassification (the latter of which would bias the results toward the null).  One
 3    possible source of selection bias is motivational differences between exposed and control
 4    populations. The designs of the Schreiber et al. (2002) and Ferroni et al. (1992) studies may
 5    have introduced such unwanted bias, although motivation has been found to more strongly
 6    influence performance on color vision but not contrast sensitivity tests (U.S. EPA, 2004).
 7    Several studies provided insufficient details on the population from which controls were selected
 8    (Seeber, 1989; Spinatonda et al., 1997; Ferroni et al., 1992),  or the details provided raise
 9    concerns regarding the appropriateness of the control group (Seeber, 1989; Spinatonda et al.,
10    1997; Schreiber et al., 2002 [residents only]).
11          For some of the occupational studies, the descriptions of behavioral testing procedures or
12    results were insufficient or ambiguous (Ferroni et al.,  1992; Seeber,  1989; Nakatsuka et al.,
13    1992; Spinatonda et al., 1997). A number of studies had either insufficient control for possible
14    influences of education (Cavalleri et al., 1994; Gobba et al.,  1998; Schreiber et al. (2002) [day
15    care study]) or provided insufficient detail on the study populations (Schreiber et al., 2002
16    [residents]; Nakatsuka et al., 1992; Ferroni et al., 1992; Spinatonda et al., 1997).  Further,
17    adjustment for education in the statistical analysis may not have been fully adequate due to the
18    use of categorical variables (Seeber, 1989; Altmann et al., 1995).  Additionally, Spinatonda et al.
19    (1997) did not provide detailed information on variables other than age of subjects, precluding a
20    determination of whether subjects may or may not have been comparable.
21          Alcohol by itself cannot explain the observed deficits in neurobehavioral functions,
22    because either study designs excluded subjects who were moderate to heavy drinkers, or
23    statistical analyses of the epidemiologic observations controlled for this covariate. However,
24    effects from the interaction between tetrachloroethylene exposure and alcohol consumption were
25    not well investigated in these studies. Valic et al. (1997) showed greater decrements in color
26    vision among subjects with exposures to both tetrachloroethylene and ethanol when compared
27    with individuals with solvent exposure only to the solvents or to neither substance.
28          Many studies did not include exposure monitoring of individual subjects, and the
29    statistical analyses compare groups using t-tests or chi-square tests, with the result of a greater
30    dependency on the performance in the control group.  Dose response analyses are statistically
31    more powerful. However, despite the use of t-tests or chi-square tests, deficits in
32    neurobehavioral or neurophysiological functions were reported in these studies. A number of
33    statistical comparisons were made in these studies, increasing the possibility of a type I, or false
34    positive, error. The issue of multiple comparisons is always present in risk assessment and
35    evaluation activities; however, unfortunately, reducing the type I error increases the type II, or
36    false negative, error for those associations that are not null. As mentioned above, inferences

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-102      DRAFT-DO NOT CITE OR QUOTE

-------
 1    about associations between exposure and effects are best drawn from a body of evidence where
 2    consistency across studies of different designs, populations, and statistical methods may be
 3    obtained; value and richness can be found when consistency emerges from the diversity and
 4    despite the flaws.
 5           These studies do have important strengths.  They describe susceptibility to
 6    tetrachloroethylene toxicity in humans, providing evidence to augment findings from animal
 7    toxicity testing. Further, the majority of studies, with the exception of Schreiber et al. (2002),
 8    exceeded a smallest cell size of 40 exposed subjects that is generally considered sufficient to
 9    detect preclinical effects in a group that range from 3 to 18, or 20%, from normal function.
10    Several studies (Altmann  et al., 1995; Schreiber et al., 2002; Echeverria et al., 1995) employed
11    multiple measures of exposure (indoor air monitoring, personal monitoring, and in some cases,
12    biological monitoring), with a high degree of correlation between tetrachloroethylene
13    concentration as assessed from indoor air monitoring or personal monitoring and biological
14    metrics such as blood tetrachloroethylene  concentration, suggesting indoor air concentration as a
15    reasonable exposure metric.
16           Several independent lines of evidence can be found in the occupational and residential
17    studies to support an inference of a broad range of cognitive and behavioral deficits following
18    tetrachloroethylene exposure (U.S. EPA, 2004). First, adverse effects on visuospatial function
19    are reported in three studies (Seeber, 1989; Altmann et al., 1995; Echeverria et al., 1995), with
20    Echeverria et al. (1994) as a confirmatory  study of Echeverria et al. (1995).  The results across
21    these three studies appear reasonably consistent, despite substantial differences in study design.
22           A second line of evidence can be found  in both the occupational (Nakatsuka et al., 1992;
23    Cavalleri et al., 1994) and residential studies (Schreiber et al., 2002), both of which evaluated
24    performance on the Lanthony color vision test.  Cavalleri et al. (1994) reported a decrement in
25    color vision in the high exposure group, but not the low exposure group, and a significant dose-
26    response relationship between CCI value and tetrachloroethylene concentration. The lack of an
27    association between color vision and tetrachloroethylene exposure in Nakatsuka et al. (1992)
28    may not be inconsistent, given significant  weaknesses in this study (U.S. EPA, 2004).
29    Performance of the residents and day care workers who worked in buildings with a co-located
30    dry cleaner appeared worse (particularly that of residents) than the performance of controls,
31    although CCI scores were not statistically  significantly different from referents (Schreiber et al.,
32    2002).  Last, VCS deficits were observed in these residents and day care workers.  These
33    subjects received exposures of lower dose rates, but a different and, for residents, a more
34    prolonged daily exposure duration than typical occupational exposures occurred.  Overall, the
35    evidence reveals a high degree of consistency in visually mediated function.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-103      DRAFT-DO NOT CITE OR QUOTE

-------
 1          Effects on spatial vision are well-known consequences of solvent exposure in industrial
 2    workers (Bowler et al., 1991; Broadwell et al.,  1995; Campagna et al., 1995; Donoghue et al.,
 3    1995; Frenette et al., 1991; Hudnell et al., 1996a; Mergler et al., 1991). Other organic solvents,
 4    as well as alcohol, induce effects on memory and color vision (Altmann et al., 1995; Mergler et
 5    al., 1991; Hudnell et al., 1996a, b). By analogy, the observations on other solvents also support
 6    an inference of neurobehavioral deficits following exposure to tetrachloroethylene.
 7          In conclusion, the weight of evidence across the available studies of humans exposed to
 8    tetrachloroethylene—and by analogy to other organic solvents—indicates that chronic exposure
 9    to tetrachloroethylene may be associated with adverse decrements in nervous system function.
10
11    4.6.2. Animal Studies
12    4.6.2.1.  Inhalation Studies
13          Mattsson et al. (1998) studied the effects of acute exposure to tetrachloroethylene for 13
14    weeks observing flash-evoked potentials (FEPs), somatosensory-evoked potentials (SEPs), EEGs,
15    and rectal temperature in F344 rats.  During the acute (pilot) study, male rats were exposed to 0
16    or 800 ppm tetrachloroethylene for 6 hrs/day for 4 days and tested before and after exposure on
17    the 4th day.  Changes in FEP, SEP, and EEG components were observed after acute exposure. In
18    the subchronic study, the above evoked potentials and caudal nerve conduction velocity were
19    determined in male and female rats exposed to 0, 50,  200, or 800 ppm for 6 hrs/day for 13 weeks.
20    Testing was performed during the week following cessation of exposure.  Changes in FEP were
21    observed at the highest dose (800 ppm). Several measures of the evoked potential were affected,
22    at 50 ppm but not at higher doses. Other measures were not affected, and no dose response was
23    observed. The finding of an overall greater effect following short-term (4-day) exposure as
24    compared with longer-term exposure is similar to the  findings of Moser et al. (1995) on a
25    number of measures of a neurotoxicity battery.
26          The effects of exposure to 90-3,600 ppm tetrachloroethylene for 1 hr on motor activity
27    were examined in male MRI mice (Kjellstrand et al.,  1985). A strong odor (cologne) was used
28    as the control  condition.  Total activity was monitored during the dark period during exposure
29    and for several hours thereafter.  All doses produced increased activity during exposure; activity
30    decreased over several hours after cessation of exposure. Although apparently no statistical
31    analyses were performed,  it is clear from the figures that the lowest dose produced an average
32    performance that was well outside the boundary of the 95% CIs of the cologne-treated controls
33    and was dose-dependent.  Tetrachloroethylene  induced motor activity at concentrations lower
34    than those of any of the other organic solvents tested (methylene chloride, toluene,
35    trichloroethylene, 1,1,1-trichloromethane).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-104      DRAFT-DO NOT CITE OR QUOTE

-------
 1          De Ceaurriz et al. (1983) exposed male Swiss OF1 mice to 596, 649, 684, or 820 ppm
 2    tetrachloroethylene for 4 hrs.  Immediately following exposure, subjects were immersed in a
 3    cylinder filled with water and the duration of immobility was observed for 3 minutes.  The term
 4    "behavioral despair" has been coined for this initial immobility, and the length of immobility is
 5    shortened by antidepressant administration.  Tetrachloroethylene exposure also shortened the
 6    period of immobility, with a no-observed-effect level (NOEL) of 596 ppm.
 7          Nelson et al. (1980) of NIOSH, investigated developmental neurotoxicity in SD rats by
 8    exposing pregnant dams to tetrachloroethylene at concentrations of 100 ppm or 900 ppm during
 9    both early pregnancy (gestation days 7 to 13) or late pregnancy (gestation days  14 to 20). The
10    investigators made morphological examinations of the fetuses and performed behavioral testing
11    and neurochemical analysis of the offspring. There were no alterations in any of the measured
12    parameters in the 100 ppm groups. At 900 ppm there were no skeletal abnormalities, but the
13    weight gain of the offspring as compared with controls was depressed about 20% at weeks 3-5.
14    Developmental delay was observed in both the early and late pregnancy groups. Offspring of the
15    early pregnancy-exposed group performed poorly on an ascent test and on a rotorod test, whereas
16    those in the late pregnancy group underperformed on the  ascent test only at postnatal day 14.
17    However, later in development (days 21 and 25), their performance was higher than that of the
18    controls  on the rotorod test. These pups were markedly more active in the open field test at days
19    3 land 32.
20          There were no effects  on running in an activity wheel on days 32 or 33  or avoidance
21    conditioning on day 34 and operant conditioning on days 40 to 46.  Neurochemical analyses of
22    whole brain (minus cerebellum) tissue in 21-day-old offspring revealed significant reductions in
23    acetylcholine levels at both exposure periods, whereas dopamine levels were reduced among
24    those exposed on gestation days 7-13. Unfortunately, none of the statistics for the 100 ppm
25    treatments was presented. The authors observed that more behavioral changes occurred in
26    offspring exposed during late pregnancy than in those exposed during early pregnancy.
27          Szakmary et al. (1997) exposed CFY rats to tetrachloroethylene via inhalation throughout
28    gestation (i.e., gestation days  1—20) for 8 hrs/day at concentrations of 0,  1,500, or 4,500 mg/m3
29    tetrachloroethylene.  The primary focus of the study was prenatal developmental evaluations (see
30    Section 4.7.2).  However a cohort of rats (15 litters/group) was allowed to deliver, and the
31    offspring (standardized to 8 pups/litter) were maintained on study until postnatal day 100 and
32    evaluated for growth, development and neurotoxic effects. The report did not specify whether
33    the animals were exposed to tetrachlorotehylene after birth. Pre-weaning observations included
34    weekly body weights, developmental landmarks (pinna detachment, incisor eruption, and eye
35    opening), and functional assessments (forward movement, surface righting reflex, grasping
36    ability, swimming ontogeny, rotating activity,  auditory startle reflex, and examination of

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-105      DRAFT-DO NOT CITE OR QUOTE

-------
 1    stereoscopic vision). After weaning, exploratory activity in an open field, motor activity in an
 2    activity wheel, and development of muscle strength were assessed. The study authors reported
 3    that adverse findings included a decreased survival index (details were not provided), a minimal
 4    decrease of exploratory activity and muscular strength in treated offspring (presumably at both
 5    exposure levels) which normalized by postnatal day 51, and significantly increased motor
 6    activity on postnatal day 100 of females exposed to 4,500 mg/m3. Litter was evaluated as the
 7    statistical unit of measure for all outcomes. There is no clear indication of group means for
 8    postnatal measures reported.  The lack of experimental detail in the postnatal evaluation part of
 9    this study reduces the overall confidence in the findings. There was no evaluation of postnatal
10    histopathology of the nervous system reported or cognitive testing during the post weaning
11    period or during adulthood.
12          Wang et al. (1993) exposed male SD rats to 300 ppm tetrachloroethylene continuously
13    for 4 weeks or 600 ppm for 4 or 12 weeks. Exposure to 600 ppm at either duration resulted in
14    reduced brain weight gain, decreased regional brain weight, and decreased DNA in frontal cortex
15    and brain stem but not hippocampus. Four specific proteins (S-100 [an astroglial protein], glial
16    fibrallary acidic protein, neurone specific enolase, and neurofilament 68 kD polypeptide) were
17    decreased at 4 and/or 12 weeks exposure to 600 ppm; 300 ppm had no effect on any endpoint.
18          The effects of exposure to 200 ppm tetrachloroethylene 6 hrs/day for 4 days in male SD
19    rats were examined on a number of endpoints (Savolainen et al., 1977a,  b).  Rats were killed on
20    the 5th day following a further 0-6 hrs of exposure. Tetrachloroethylene levels were highest in
21    fat, followed by liver, cerebrum, cerebellum, lung, and blood.  Tissue levels increased in all
22    tissues over the 6 hrs of exposure. Brain RNA content decreased, and brain nonspecific
23    cholinesterase was increased on the 5th day, although no statistical comparisons were performed.
24    Locomotion in an open field was increased immediately following the end of exposure on the 4th
25    day, with no difference 17 hrs after exposure, although no statistical comparisons were made.
26    Brain protein, GSH, and acid proteinase were unaffected.
27          A series of experiments were performed on the effects of tetrachloroethylene on brain
28    lipid patterns. Exposure to 320 ppm for 90 days (Kyrklund et al., 1990) or 30 days (Kyrklund et
29    al., 1988) in male SD rats resulted in changes in the fatty acid composition of cerebral cortex,
30    which persisted after a 30-day recovery period (Kyrklund et al.,  1990).  Similar results were
31    observed in cerebral cortex and hippocampus after exposure to 320 ppm in the Mongolian gerbil
32    (sex unspecified) in the presence of reduced brain weight (Kyrklund et al., 1987). Exposure of
33    male Mongolian gerbils to 120 ppm for  12 months also resulted in decreases in long-chain,
34    linolenic acid-derived fatty acids in cerebral cortex and hippocampus  (Kyrklund et  al., 1984).
35          The effect of tetrachloroethylene on neurotransmitter levels in the brain was explored in
36    male SD rats exposed continuously to 200, 400, or 800 ppm for a month (Honma et al., 1980a, b).

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-106       DRAFT-DO NOT CITE OR QUOTE

-------
 1    The 800 ppm dose produced a decrease in ACh in striatum, and there was a dose-related increase
 2    in a peak containing glutamine, threonine, and serine in whole brain preparations. GAB A, NE,
 3    5-HT, and other amino acids were not affected.
 4          In a study from the same laboratory (Rosengren et al., 1986), Mongolian gerbils of both
 5    sexes were exposed to 60 or 300 ppm tetrachloroethylene for 3 months, followed by a 4-month
 6    solvent-free period. Changes in both S-100 and DNA concentrations in various brain regions
 7    were observed at the higher concentration, and decreased DNA in frontal cortex was  observed
 8    after exposure to 60 ppm.  The higher concentration also produced decreased brain but not body
 9    weight.  The results at 60 ppm were replicated in a follow-up study (Karlsson et al., 1987).
10          In a related study (Briving et al., 1986), Mongolian gerbils were exposed for 12 months
11    to tetrachloroethylene at 120 ppm.  At the end of exposure,  out of a total of 8 amino acids
12    assayed, taurine was significantly decreased in the two brain regions assessed (hippocampus and
13    cerebellum), and glutamine was elevated in hippocampus. y-Aminobutyric acid (GABA) levels
14    were unaffected, as was uptake of GAB A and glutamate.
15          Kyrklund and Haglid (1991) exposed pregnant guinea pigs to airborne
16    tetrachloroethylene continuously from day 33 through day 65 of gestation. The exposure was
17    continuous at 160 ppm except for 4 days at the beginning and end of the exposure period, when
18    it was reduced to 80 ppm.  In the control group there were three dams with litter sizes of four,
19    three and two pups, and in the exposed group there were three dams with litter sizes of two each.
20    The pup body weights differed between litters.  In the data analysis, three pups in the control
21    group were eliminated and the six pups in the treatment and control groups were assumed to be
22    independent, which is an invalid assumption. According to the authors' analysis, the offspring
23    had a slightly altered brain fatty acid composition, with a statistically significant reduced stearic
24    acid content in the tetrachloroethylene treatment group, which is consistent with the authors'
25    earlier findings in rats. This conclusion  might have been different if the investigators  had
26    grouped litters rather than pups as independent groups. The results suggest that
27    tetrachloroethylene could have reduced the litter size, but a much larger study would  be
28    necessary to establish reduced litter size as an effect of tetrachloroethylene.
29          Caucasan male and female NMRI mice were exposed to 9, 37, 75, or 150 ppm
30    continuously for 30 days, to 150 ppm for one of several exposure periods ranging from 5 to 30
31    days, or to 150 ppm tetrachloroethylene for 30 days with various recovery periods (Kjellstrand et
32    al., 1984). Other groups were exposed intermittently on several dosing and exposure regimens
33    that resulted in a TWA of 150 ppm for 30 days. Plasma BuChE levels, organ weights, liver
34    morphology, and motor activity were assessed.  BuChE was elevated after continuous exposure
35    to 37 ppm or greater.  Liver weight was increased at all doses following continuous exposure,
36    and body weight decreased at 37 ppm or above. Motor activity results following continuous

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-107      DRAFT-DO NOT CITE OR QUOTE

-------
 1    exposure were not reported. BuChE and liver weight were both elevated at a TWA of 150 ppm
 2    for 30 days, regardless of the length of the exposure pulse. This was true even for an hour's
 3    exposure (at 3,600 ppm) as well as at the lowest concentration (225 ppm).  All concentrations of
 4    intermittent exposure increased motor activity. A recovery period reversed the effects on BuChE,
 5    whereas liver weight was still slightly elevated at 150 days after cessation of exposure. Changes
 6    in liver morphology were detected following exposure to 9 ppm for 30 days and reversed after
 7    cessation of exposure.
 8           Tinston (1994) performed a multi-generation study of the effects on rats exposed to
 9    airborne concentrations of tetrachloroethylene. The details of the study are discussed  in Section
10    4.7.2.  The investigators observed several developmental effects. Of interest here were the signs
11    of CNS depression (decreased activity and reduced response to sound) observed for the first 2
12    weeks in both adult generations and when the exposure was resumed on day 6 postpartum in the
13    Fl generation (adults and pups).  These effects disappeared about 2 hrs after cessation of the
14    daily exposure.  Other overt signs of tetrachloroethylene poisoning among  the adults included
15    irregular breathing  and piloerection at both 300 and 1,000 ppm. These changes stopped
16    concurrently with cessation of exposure or shortly thereafter.
17
18    4.6.2.1.1. Summary of animal inhalation neurotoxicity studies. In order  to compare the animal
19    inhalation neurotoxicity studies with each other and to evaluate whether there is any relationship
20    across studies between the LOAEL of the administered dose and the duration of treatment, the
21    data were summarized (Table 4-6). In order to estimate the lowest concentration at which a
22    given effect occurs, the experiment showing that effect must have both a LOAEL (the lowest
23    concentration at which the effect occurred) and a NOAEL (the next lower concentration where
24    the effect did not occur). The experiments that meet this criteria are Mattsson et al. (1998), De
25    Ceaurriz et al. (1983), Wang et al. (1993), Honma et al. (1980 a, b), and Kjellstrand et al. (1984).
26           The total duration of exposure in these experiments is plotted in Figure 4-2 as  a function
27    of the LOAEL concentration in order to discover whether there is a systematic trend in this
28    relationship. The plot shows that there is no systematic trend. It also shows that the LOAEL
29    varies over a 22-fold range: from 37 ppm for 30 days for increased brain butyl cholinesterase in
30    mice observed by Kjellstrand et al. (1984) to 800 ppm for 13 weeks for alteration in the flash-
31    evoked potential in rats observed by Mattsson et al. (1998). Table 4-6 shows other observations
32    at comparatively low concentrations: decreased DNA in gerbils by Rosengren et al. (1986) and
33    Karlsson et al. (1987) at 60 ppm and increased motor activity in mice at 90 ppm, observed by
34    Kjellstrand et al. (1985). The LOAEL for these studies as a group is therefore in the range of 37
35    to 90 ppm, and the  effects at these levels are changes in neurotransmitter levels and increased
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-108      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Table 4-6. Summary of animal inhalation neurotoxicology studies
Subjects
F344 rats
Pilot study: male
10/dose
Follow-up study: males
and females
12/sex dose
NMRI mice, males
Swiss OF1 mice, males
10/dose
SD rats
pregnant females
13-21 litters/dose
males and female
Offspring assessed
CFY rats
pregnant females
15 litters/dose
male and female offspring
assessed
SD rats, males
8/dose
SD rats, males
10/dose
SD rats, males
5-6/dose
SD rats, males
5-6/dose
Mongolian gerbils
males and females
6/sex/dose
Effect
Changes in FEP, SEP, EEC
Increased amplitude and
latency in late component
of FEP
Increase in motor activity
Decrease in duration of
immobility
Decreased weight gain
Behavioral changes, more
extensive for late
pregnancy exposure
Decreased brain
acetylcholine
Transient decreases in
muscular strength and
exploratory behavior.
Latent increases in motor
activity in females at 100
days postnatally
Reduced brain weight,
DNA, protein
Decrease in brain RNA,
increase in brain
cholinesterase and increase
motor activity
Change in fatty acid
composition of cerebral
cortex
Neurotransmitter changes,
brain regions
Decrease in DNA, frontal
cortex
Decrease in brain weight
NOAEL/ LOAEL"
(ppm)
0, 800. 4 days, 6 hr/d
50, 200. 800.
13 wks, 6 hr/d, 5 d/wk
2Q, 3,600 1 hr
596. 649, 684, 820
4hrs
0, 100. 900 on
GD7-13oronGD 14-20
7 hr/d
0, 1,500 or 4.500 me/m3
GD 1-20 for 8 hrs/day
300. 600.
4 or 12 wks continuous (24
hr/d)
200, 4 days

320. 12 wks
continuous (24 hr/d),
30-day washout period
320. 4 wks continuous (24
hr/d)
200, 400. 800,
4 wks continuous
(24 hr/d)
6fl, 300, 12 wks,
continuous (24 hr/d)
16-week washout period
Authors
Mattssonetal. (1998)
Kjellstrandetal. (1985)
DeCeaurrizetal. (1983)
Nelson etal. (1980)
Szakmary et al. (1997)
Wang etal. (1993)
Savolainen et al. (1977a, b)
Kyrklund etal. (1990,
1988)
Honmaetal. (1980a,b)
Rosengren et al. (1986)
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    4-109      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
       Table 4-6. Summary of animal inhalation neurotoxicology studies (continued)
Subjects
Mongolian gerbils
gender unspecified
Mongolian gerbils
males and females
8/sex/dose
Mongolian gerbils
gender unspecified
6/dose
Mongolian gerbils
males 6/dose
Guinea pigs
pregnant females
3/litters/dose
males and female
Offspring assessed
NMRI mice,
males and females
3-8/sex dose
Males and females
10/sex dose
SD rats, multigeneration
study
28 litters/dose
Effect
Decrease in DNA, frontal
cortex
Decrease in brain weight
Taurine, glutamine changes
in brain regions
Decrease in brain weight,
change in fatty acids
Decreased brain long-chain
fatty acids
Decrease in brain stearic
acid in offspring after in
utero exposure13
Increase in butyl
cholinesterase
Increased motor activity
CNS depression in first 2
wksofFlandF2
generations, which ceased
2 hrs after daily exposures
NOAEL/LQAEL
(ppm)
60, 12 wks, continuous (24
hr/d)
120. 12 mos continuous (24
hr/d)
320, 12 wks continuous (24
hr/d)
120, 52 wks continuous (24
hr/d)
Maximum exposure 160.
GD 33 to 65 continuous
(24 hr/d)
9C, 32, 75, 150
4 wks continuous
(24 hr/d)
150. 4 wks intermittent-
(1, 2, 4, 8, or 16 hr/d)
0, 100. 300. 1,000
6 hr/d, 5 d/wk, except
during mating, 6 hr/d-7
d/wk
Authors
Karlssonetal. (1987)
Brivingetal. (1986)
Kyrklundetal. (1987)
Kyrklundetal. (1984)
Kyrklund and Hagid (199 1)
Kjellstrandetal. (1984)
Kjellstrandetal. (1984)
Tinston (1994)
 4
 5
 6
 7
 8
 9
10
a Experimental/observational NOAEL is underlined, LOAEL is double-underlined.
b Questionable findings because litter was not used as the unit of measure in analysis.
0 LOAEL for changes in liver weight.

FEP      = Flash-evoked potential
GD      = Gestational day
SEP      = Somatosensory-evoked potential
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      4-110      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
Animal Neurotoxicity Inhalation Studies
Duration vs. LOAEL
mnnn
IUUUU
i/r
^ mnn
3 IUUU
.c
cf
o ^nn
^= IUU
ro
3
Q
"~^ m
0) 1U
o
_i
-i









•














































1

•

i







10 100 1000
Log (LOAEL, ppm)
       Figure 4-2.  Summary of the relationship between LOAEL concentrations
       (ppm) and treatment duration (hours).

       LOAEL = Lowest-observed-adverse-effect level
motor activity.  Changes in fatty acid composition were observed at somewhat higher
concentrations (320 ppm).

4.6.2.2. Oral and Intraperitoneal Studies
       A study in male SD rats assessed the acute or short-term effects of tetrachloroethylene by
gavage on several screening tests (Chen et al., 2002). A single dose of 500 mg/kg to adult rats
produced changes on three different tests of pain threshold, locomotor activity, and seizure
susceptibility threshold following pentylenetetrazol infusion, whereas 50 mg/kg had statistically
significant effects on seizure threshold only. In the short-term study, young 45-50 gram rats
were dosed 5 days/week for 8 weeks with 5 or 50 mg/kg. Behavioral testing began 3 days after
the last dose. Locomotion was affected only at the high  dose, whereas both doses produced
effects on the other four endpoints. The 8-week exposure resulted in retarded weight gain in
both treated groups, which was about 10% at the end of the dosing period.
       The interpretation of these results is problematic. The tests were all observational in
nature, requiring scoring by the observer. The study by Chen et al. (2002) does not state whether
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-111      DRAFT-DO NOT CITE OR QUOTE

-------
 1    the observer(s) was blind to the treatment group of the animals, a condition that is essential for
 2    such tests to be valid. In fact, because there were differences in weight between control and
 3    treated rats, it would probably be easy to distinguish treated from control animals simply by
 4    looking at them. Further, the paper does not state whether all animals were tested by the same
 5    person for each task or, if not, whether there was any indication of inter-observer correlation.
 6    The potential effect of the difference in weight between the control and the treated groups on
 7    these measures is also unknown. Given that the difference between the control and the treated
 8    groups in response latency to painful stimuli is tenths or hundredths of a second with no dose
 9    response, these issues are of serious concern.
10          Various behavioral endpoints were assessed in 8-week-old ICR male mice at the
11    beginning of an experiment by Umezu et al. (1997).  Righting reflex was affected after single-
12    dose i.p. administration of tetrachloroethylene at 4,000 but not at 2,000 mg/kg or less, and ability
13    to balance on a wooden rod was decreased at 2,000 but not at 1,000 mg/kg or less. Response rate
14    on a fixed-ratio 20 (FR20) schedule—which requires 20 responses for each reinforcement—was
15    affected at 2,000 but not at 1,000 mg/kg or less 30 minutes after administration. In a procedure
16    in which a thirsty mouse was shocked every 20th lick of a water spout, mice dosed with 500
17    mg/kg but not with higher or lower doses received an increased number of shocks.  In an
18    FR20-FR20 punishment schedule, responding in the punishment condition was increased at
19    1,000 but not at 500 mg/kg or less. A puzzling aspect of the study is the mention in the methods
20    section of "breeding animals," with no further explanation. If the investigators bred their own
21    mice, there is no indication of how pups were assigned to treatment groups.
22          Moser et al. (1995) examined the effects of a number of potentially neurotoxic agents,
23    including tetrachloroethylene, on a neurotoxicity screening battery in adult female F344 rats
24    following either a  single gavage dose (acute exposure) or repeated gavage doses over 14 days
25    (subacute exposure). For the acute study, subjects were tested 4 and 24 hrs following exposure.
26    After acute exposure, a LOAEL of 150 mg/kg was identified for increased reactivity to being
27    handled 4 hrs  after dosing, with increased lacrimation, decreased motor activity, abnormal gate,
28    decreased response to an auditory stimulus,
29    decreased righting ability, and increased landing foot-splay at higher doses at 4 and/or 24 hrs
30    post-dosing.  A NOAEL was not identified.  In the subacute study, no endpoint was significantly
31    different from those of controls at doses of 50-1,500 mg/kg. This presumably represents
32    behavioral adaptation following repeated exposure to tetrachloroethylene.
33          Locomotor activity was monitored in NMRI mice gavaged with 5 or 320 mg/kg
34    tetrachloroethylene for 7 days beginning at 10 days of age (Fredriksson et al., 1993). Twelve
35    male pups from three or four litters were assigned to each treatment group. This study  design
36    does not conform to traditional developmental toxicity testing guidelines.  Locomotion, rearing,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-112      DRAFT-DO NOT CITE OR QUOTE

-------
 1    and total activity (vibration of the cage) were measured for 60 minutes at 17 and 60 days of age.
 2    A stastically significant increase in locomotor activity of treated mice in both dose groups was
 3    observed, and rearing behavior decreased as compared with controls for all three measures at 60
 4    days of age but not at 17 days of age in which testing followed shortly after the last dose.
 5           The persistent effects of subacute developmental exposures in this study raises some
 6    concerns.  Some caution in interpreting the results of the effect of tetrachloroethylene exposure is
 7    warranted for two reasons: (1)  the results at 320 mg/kg were no different than at 5 mg/kg,
 8    indicating no clear dose-response relationship between exposure and this effect, and (2) litter
 9    mates were used as independent observations in the statistical analysis.  This procedure can
10    increase the apparent a and result in an erroneous statistical result. For example, Holson and
11    Pearce (1992) demonstrated that for body weight, using three or four littermates as independent
12    observations, as in the above study, resulted in the nominal a increasing from 0.05 to a range of
13    0.23 to 0.38. Similar litter effects have been demonstrated for behavioral data (Buelke-Sam  et al.,
14    1985).
15           Fredriksson et al. completed a study that parametrically compared the effects of postnatal
16    dosing and resulting alternations in motor activity using both litter as the unit of measure and
17    their own within-litter randomization (Ericksson et al., 2005).  Their results were similar in both
18    the magnitude of effect across dose groups and in the variability within each dose group for both
19    experimental designs.  The authors' key assertion for using this randomization within a small
20    number of litters rather than the traditional  litter as the unit of measure is that it reduces the
21    overall number of animals needed to be generated to statistically determine an effect of chemical
22    exposure.
23           Locomotor activity was assessed in 6-week-old male Wister rats following i.p. doses of
24    100, 500, or 1,000 mg/kg tetrachloroethylene for 3 consecutive days, with activity being
25    monitored for at least 1 week following cessation of administration (Motohashi et al., 1993).
26    Animals were monitored 24 hrs/day, and locomotor activity (measured as change in electrical
27    capacitance of a circuit beneath the floor of the cage) was analyzed by time-series analysis and
28    spectral analysis.  All doses of tetrachloroethylene changed circadian rhythm in a dose-
29    dependent manner, with the increased activity at the start of the dark period delayed by
30    tetrachloroethylene exposure. Recovery took 3-5 days after cessation of exposure.
31           Operant performance on a fixed-ratio 40 schedule of reinforcement was assessed in adult
32    male SD rats gavaged with 160 or 480 mg/kg tetrachloroethylene immediately before testing
33    (Warren et al.,  1996).  The lower dose produced no effect on response rate over the 90-minute
34    session, whereas the higher dose produced  a transient  rate  decrease in three of six animals (with
35    recovery after 20 to 40 minutes) and induced a complete cessation of response in two of the six
36    animals.  Tetrachloroethylene concentrations increased rapidly after administration in blood,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-113      DRAFT-DO NOT CITE OR QUOTE

-------
 1    brain, fat, liver, and muscle. For the duration of the 90-minute period of testing, blood
 2    tetrachloroethylene levels were approximately linearly related to the administered dose, but brain
 3    tetrachloroethylene levels were similar for both dose groups. This study did not evaluate the
 4    persistent effects of exposure to tetrachloroethylene on cognitive performance.
 5           Table 4-7 presents a summary of the oral neurotoxicity animal studies. For the six oral
 6    neurotoxicity studies in rodents reviewed here, only one (Fredriksson et al., 1993) describes
 7    effects lasting more than 1 week. In that study the effect (increased motor activity) was the same
 8    at 5 and 320 mg/kg, and the results do not represent a clear dose-response relationship across two
 9    orders of magnitude of administered doses.  The lowest LOAEL occurring in the four remaining
10    studies is 100 mg/kg for delayed onset of circadian activity in rats (Motohashi et al., 1993). This
11    LOAEL is based on an i.p.-administered dose describing transient neurological  effects and is not
12    comparable to inhalation or ingestion LOAELs without pharmacokinetic modeling of an
13    appropriate dose metric. No information is available for irreversible neurological effects via the
14    oral route because no studies have evaluated the potential for neurotoxicity following chronic
15    oral exposure.
16
17    4.6.3. Summary of Neurotoxic Effects in Humans and Animals
18           Taken together, the animal and epidemiologic evidence is supportive of an association
19    between neurobehavioral deficits and tetrachloroethylene exposure.  The pattern of effects on the
20    visual system in humans may be consistent with decrements in visually mediated dysfunction, as
21    suggested by Echeverria et al.  (1995).  The test for VCS in humans is sensitive to neurological
22    dysfunction associated with many diseases  affecting the nervous system (NYS DOH, 2000).
23    Moreover, VCS deficits as well as color discrimination deficits are commonly present prior to
24    detectable pathology in the retina or optic nerve head, making this one of the earliest signs of
25    disease (Regan, 1989). Additionally, other organic solvents, as well as alcohol, induce effects on
26    memory and color vision (Altmann et al., 1995; Mergler et al., 1991; Hudnell et al., 1996a, b).
27    The consistency of these observations suggests construct validity for organic solvents as a class
28    because of their effects on visually mediated function. Hence, these observations, by analogy,
29    add support to an inference of tetrachloroethylene-induced neurobehavioral effects.
30           Studies of occupational (Seeber, 1989; Echeverria 1994, 1995)  and residential (Altmann
31    et al., 1995) exposures indicate that cognitive performance in humans exposed to
32    tetrachloroethylene is affected with effects on choice reaction times, visual-spatial information
33    processing, and other measures of cognitive performance.
34           The three epidemiological studies on dry cleaners chronically exposed to
35    tetrachloroethylene showed decrements in color vision at 7 ppm (Cavalleri et al., 1994, with a
36    follow-up of these workers [Gobba et al., 1998] showing greater loss in color discrimination in

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-114      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4-7.  Summary of oral neurotoxicity animal studies
Subjects
SD rats,
male
SD rats,
male
ICR mice,
male
8-10/dose
F344 rats,
female
NMRI male
mice,
post-natal
exposure
12 pups/
dose
(derived
from 3
litters)
Wistar rats,
male
n/dose?
Effect
Pain threshold, pain susceptibility,
weight gain decrement
Interpretation is unclear
Operant responses stopped
immediately after 480 mg/kg dose,
then 2/3 of animals recovered by 40
minutes.
Brain tetrachloroethylene
concentrations were the same at
both doses
NOAEL/LOAEL:
Righting reflex, 2,000/4,000
Balance, 1,000/2,000
Operant responses, 1,000/2,000
Punishment, 500/1,000
Increased reactivity, decreased
motor activity, decreased righting
ability, increased landing foot
splay, abnormal gait after one dose
No effect after repeated doses
Increased locomotion and
decreased rearing at day 60 in both
dose groups
No effect immediately after
treatment
Transient delay in circadian
activity, dose-related
NQAEL/LQAEL3 (mg/kg)
5, 50 mg/kg daily for 8 weeks
Gavage single dose at 0, 160^ 480
mg/kg
Single intraperitoneal doses at 0,
500. 1.000. 2,000, 4,000 mg/kg
Single doses: 150 mg/kg is
LOAEL
Repeated dosing for 14 days: 1.500
mg/kg is NOAEL
Gavage treatment 1, 320 mg/kg
daily for postnatal days 10-16
Intraperitoneal doses 0. 100, 500,
1,000 mg/kg-day for 3 days
Author
Chen et al.
(2002)
Warren et al.
(1996)
Umezu et al.
(1997)
Moser et al.
(1995)
Fredriksson et
al. (1993)
Motohashi et al.
(1993)
4
5
6
a Experimental/observational NOAEL is underlined, LOAEL is double-underlined.

n/dose    = Number of animals per dose not clearly defined
                 This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                      4-115      DRAFT-DO NOT CITE OR QUOTE

-------
 1    those who were subsequently exposed to a higher concentration and smaller loss in those
 2    exposed for lower concentrations), longer reaction times to visual stimuli at 8 ppm (Spinatonda
 3    et al., 1997) and decrements in cognitive function at 12 ppm (Seeber, 1989).
 4          Two studies of tetrachloroethylene exposure in residences near a dry cleaning facility
 5    (Altmann et al., 1995) and a day care facility (Schreiber et al., 2002) found decrements in several
 6    neurological parameters at lower exposures than did the studies of occupational exposures cited
 7    above.  This indicates that CNS effects can occur at a lower concentration than inferred from
 8    occupational studies of dry cleaners, which have been of exposures several-fold higher than
 9    those in these residential studies. LOAELs in the human studies of CNS effects ranged from 0.7
10    ppm to 41 ppm.
11          The lack of exposure-response relationships is a limitation; however, this lack may be a
12    reflection of poorer characterization of exposure in these studies.  Exposure-response
13    relationships are based only on an estimate of current exposure; historical exposure to
14    tetrachloroethylene is lacking and may be more important to an analysis of exposure response.
15    Moreover, most analyses use a concentration x time [C x t]) dose metric. Another metric, such
16    as peak concentration, may be more relevant of an exposure-response relationship.
17          Alcohol by itself cannot account for the observed deficits in neurobehavioral functions,
18    because statistical analyses of the epidemiologic observations accounted for this covariant.
19    However, effects from the interaction between tetrachloroethylene exposure and alcohol
20    consumption was not well investigated in these studies. Valic et al. (1997) showed greater
21    decrements in color vision among subject with both exposures as  compared with individuals with
22    solvent exposure only or with neither exposure.
23          No epidemiological studies investigating drinking water or other oral exposures to
24    tetrachloroethylene have explored the potential for neurotoxicity.
25          The research in animal models (rodents) on the effects of tetrachloroethylene on
26    functional endpoints consists almost exclusively of screening studies (functional observation
27    battery, motor activity) or effects on sensory system function, as assessed by evoked potentials.
28    Effects on motor activity and motor function have been observed  with some consistency
29    following either adult or developmental exposure. Changes in VEPs were also reported
30    following acute (4-day) and subchronic (13-week) exposure.  In addition, changes in brain DNA,
31    RNA, or protein levels and lipid composition were altered following inhalation, with changes
32    observed in cerebellum, hippocampus,  and frontal cortex.  The replication of these changes in
33    biochemical parameters and effects in brain weight in both rats and gerbils is pathognomonic.
34          Changes in neurotransmitters systems (Honma et al., 1980 a, b, Driving et al., 1986) and
35    circadian rhythm (Motohashi et al., 1993) in animal studies are consistent with neuroendocrine
36    alterations observed in humans (Ferroni et al., 1992). Operant tasks that test cognitive

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-116     DRAFT-DO NOT CITE OR QUOTE

-------
 1    performance have demonstrated performance deficits in rats and mice following acute
 2    tetrachloroethylene oral (Warren et al., 1996) and i.p. (Umezu et al., 1997) exposures. These
 3    findings in animal studies are consistent with observed effects on cognition and memory in
 4    humans. However, no studies to date have evaluated the persistent effects of tetrachloroethylene
 5    exposure on cognitive performance deficits in animal models. This is a clear data need that
 6    could help resolve the dose-response relationship in cognitive performance observed in both
 7    human occupational and residential studies.  The neurophysiological findings in animal studies,
 8    albeit at high doses (800 ppm), are consistent with the physiological dysfunction observed in
 9    visually mediated functions in humans.  In addition, the persistent changes in neurotransmitter
10    levels, regional DNA content, and brain weight in animal studies is consistent with neurological
11    effects in humans. Therefore, effects observed in human and rodent models exhibit a reasonable
12    degree of congruence.
13          The inhalation LOAEL for neurotoxic effects in humans is 0.2-41 ppm. For animals it is
14    37-90 ppm, with no apparent correlation between the LOAEL of administered concentration and
15    duration of treatment (see Section 4.6.2.1). Information for oral effects in humans is missing,
16    and the only animal data applicable to an oral exposure are from gavage administration of
17    tetrachloroethylene.  No information on long-term neurological effects in animals via the oral
18    route is available.
19
20    4.6.4. Mode of Action for Neurotoxic Effects
21          The MOA for the neurotoxic  effects of tetrachloroethylene is unknown; however, at
22    present,  the best surrogate for the dose metric for neurotoxicity is blood tetrachloroethylene.
23    There may be multiple mechanisms or MO As, which may differ for adult and developmental
24    exposure.  The acute effects of tetrachloroethylene  share much in common with those of other
25    solvents such as toluene, volatile anesthetics, and alcohols. There is emerging  evidence that such
26    agents act on the ligand-gated ion channel  superfamily in vitro (Shafer et al., 2005), particularly
27    on the inhibitory amino acids NMD A, nicotinic, and GABA receptors in vivo (Bale et al., 2005).
28    Volatile anesthetics and alcohol both interact with the glycine receptor (Yamakura et al., 1999;
29    Wick et al., 1998; Mihic, 1999). Affinities depend  on specific subunits of the receptor and are
30    correlated with behavioral effects on tests  such as loss of righting ability. Similarly, ethanol and
31    volatile anesthetics enhance GAB AA receptor function (Mihic,  1999).   Chronic effects of these
32    agents may also be dependent on the GABAA system (Grobin et al., 1998).
33          Other receptors, such as the dopaminergic/N-methyl-D-aspartate (NMDA) receptor, may
34    also be involved in the mediation of the effects of these agents, as may the glutamate kainate or
35    a-amino-3-hydroxy-5-methyl-4-isoxazole-4-propionic acid (AMPA) receptors (Harris et al.,
36    1995; Cruz et al., 1998). The solvents 1,1,1-trichloromethane and trichloroethylene enhanced

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-117     DRAFT-DO NOT  CITE OR QUOTE

-------
 1    neurotransmitter-activated currents at a ipl GABAA and a 1 glycine receptors (Beckstead et al.,
 2    2000).  It seems reasonable to speculate that tetrachloroethylene would also have as part of its
 3    MO A modulation of these systems, although specificity for different receptor subunits would
 4    undoubtedly differ from those for other nervous system depressants. Consistent with this
 5    hypothesis, glutamine levels were elevated in hippocampus following tetrachloroethylene
 6    administration (Driving et al., 1986), although GABA levels and uptake were unchanged.
 7           Tetrachloroethylene also affects the fatty acid composition of the brain following 30- or
 8    90-day exposure and persists for at least 30 days after cessation of exposure (Kyrklund et al.,
 9    1984, 1987, 1988, 1990).  It has long been known that the potency of an anesthetic is
10    proportional to its lipid solubility, from which it was inferred that anesthetics act on the lipid
11    bilayer of the plasma membrane.  However, this observation has not led to elucidation of the
12    mechanism of action of anesthetic agents or solvents. It is nonetheless interesting that
13    tetrachloroethylene produces changes in fatty acid composition of the brain.
14           Tetrachloroethylene effects on the nervous system are not explained simply by any direct
15    dosing  studies with metabolites. In particular, the pattern of neurotoxicity observed with DCA is
16    qualitatively different. DCA has been found to produce hind-limb paralysis, altered gait, muscle
17    weakness, and pathology in the spinal cord (Moser et al., 1999). In the one study that examined
18    the effects of chronic exposures to tetrachloroethylene on the histology and function of the
19    nervous system (Mattsson et al., 1998) there was no observed neuropathology or functional
20    deficits (i.e., hind-limb paralysis, altered gait, or muscle weakness). This discrepancy between
21    tetrachloroethylene and DCA could be explained by differences in target tissue dose available
22    from metabolized tetrachloroethylene and more direct exposure to DCA.
23
24    4.7. DEVELOPMENTAL/REPRODUCTIVE STUDIES
25    4.7.1. Human Studies
26           Adverse effects on reproduction and development assessed by epidemiologic
27    investigation include effects on fecundity (defined as reproductive potential and measured by
28    time to pregnancy); effects on sperm; the risk of adverse pregnancy outcomes such as
29    spontaneous abortion, stillbirth, congenital malformation, or low birth weight; and effects on
30    postnatal development (which in this evaluation includes the occurrence of childhood cancer).
31    Several of the adverse pregnancy outcome studies evaluated exposure during a critical window,
32    the first trimester of the pregnancy. In general, the epidemiologic studies evaluating effects on
33    reproduction and the developing fetus do not present quantitative information on level of
34    exposure to tetrachloroethylene. When information is available (identified in the discussion
35    below), an assertion of exposure to tetrachloroethylene in most cases was derived from self-
36    reported information provided by study subjects in mailed questionnaires or interviews. More
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-118      DRAFT-DO NOT CITE OR QUOTE

-------
 1    rarely, biological measures of exposure, such as tetrachloroethylene in blood or in urine, were
 2    available for a subset of these subjects.
 3          A number of studies found elevated risks of spontaneous abortions among women
 4    employed as dry cleaners (Doyle et al., 1997; Windham et al., 1991; Olsen et al., 1990;
 5    Lindbohm et al., 1990; Kyyronen et al., 1989; Bosco et al., 1987).  Two reports of the same
 6    study population do not note associations between spontaneous abortions and dry cleaning and
 7    laundry employment (McDonald et al., 1986, 1987). These reports are not inconsistent with the
 8    remaining body of literature due to possible bias. Approximately 25%  of the women
 9    hospitalized for a spontaneous abortion were not interviewed, and this may have introduced a
10    bias into the study if a decision to participate was related to solvent exposure. The studies
11    assessing spontaneous abortions among the wives of men exposed  to tetrachloroethylene
12    (Taskinen et al., 1989; Eskenazi et al., 1991a) were not remarkable due to the few numbers of
13    exposed cases.
14          The study by Doyle et al.  (1997) is the largest:  3,517 pregnant women who were
15    currently or previously employed in dry cleaning or laundry shops. The authors analyzed the
16    data by applying several different approaches and taking into account a number of important
17    covariates, which is a strength of this study. The findings were all  suggestive of an increased
18    risk of spontaneous abortions among pregnancies reported by women who were employed as dry
19    cleaners at any  time during pregnancy or three months before conception as compared with
20    unexposed pregnancies. In fact, lower 95% CIs for many of these  approaches were above a
21    relative risk of  1. Adding support was the observation that risk for pregnancies in dry cleaning
22    operators was larger than risks observed for pregnancies reported by women working in jobs in
23    laundry or nonoperator dry cleaning, suggesting the presence in this study of an exposure-
24    response association.
25          Doyle et al. (1997) is considered to carry greater weight than the other studies discussed
26    below due to its use of a pregnancy as the unit of analysis. Analyses that do not adjust for
27    previous pregnancy loss,  such as  those presented in McDonald et al. (1987), could lead to biased
28    estimates because a previous spontaneous abortion is a risk factor for a spontaneous abortion
29    with the current pregnancy. The  study design of Doyle et al.  (1997) minimizes the potential for
30    this type of bias because a woman with repeated pregnancy losses may be counted in both
31    exposed and unexposed categories, depending on her exposure status at the time of the
32    pregnancy.
33          Three other studies (Olsen et al., 1990; Bosco et al., 1987; Windham et  al., 1991)
34    examined the association between spontaneous abortions and occupational exposure to
35    tetrachloroethylene. Olsen et al. (1990) presented findings from a  four-country Nordic study of
36    spontaneous abortion, low birth weight, and congenital anomalies and observed a relative risk of
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-119      DRAFT-DO NOT CITE OR QUOTE

-------
 1    2.9 (95% CI = 1.0-8.4; eight exposed cases) for all data sets between spontaneous abortion and
 2    "high exposure" to tetrachloroethylene during the first trimester of pregnancy, primarily due to
 3    the large risk seen among subjects from Finland (OR = 4.5, 95% CI = 1.1-18.5, six exposed
 4    cases).  These analyses were based on 3,279 pregnancies among women dry cleaners and laundry
 5    workers linked to national registers of birth and reproductive failures.  Biological monitoring
 6    data were available for some of the subjects from Finland; blood tetrachloroethylene
 7    concentration ranged from 0.1 |imol/L to 2.6  |imol/L for cases (n = 4) and from 0.3 |imol/L to
 8    3.6 |imol/L for controls (n = 3; Kyyronen et al., 1989). Unfortunately, the number of subjects in
 9    Kyyronen et al. (1989) who were also included in Olsen et al. (1990) is not known.
10          The result for Finnish workers reported by Olsen et al. (1990) is of a similar magnitude as
11    that reported for essentially the same study subjects by two other Finnish investigators
12    (Kyyronen et al., 1989; Lindbohm et al., 1990). Overall, greater weight is placed on the Olsen et
13    al. (1990) findings due to the investigators' more systematic approach for evaluating an
14    exposure-effect association.
15          Bosco et al. (1987) observed a 4-fold higher history of prior spontaneous abortions
16    among women working in dry cleaning  shops than among these same women when they were
17    not employed outside their homes.  These findings were based on a small number of subjects and
18    were not statistically significant.  Mean  urinary TCA levels among women employed in dry
19    cleaning shops was 5 |ig/L, compared to 1.4 jig/L for women employed in shops that operated
20    only as an ironing service. Windham et al. (1991) reported a statistically significant elevated risk
21    of spontaneous abortions (OR = 4.7, 95% CI = 1.1-21.1) in tetrachloroethylene-exposed women
22    in analyses that adjusted for age, race, education, prior fetal loss, smoking, and number of hours
23    worked. This analysis was based on seven women identified with tetrachloroethylene exposure,
24    of which four were identified as having  exposure to trichloroethylene, for which the odds ratio
25    was also elevated (OR = 3.1, 95% CI = 0.9-10.4).  Both trichloroethylene and
26    tetrachloroethylene share a number of common or like metabolites, although human metabolism
27    via the P450 oxidative pathway is more extensive for trichloroethylene than for
28    tetrachloroethylene.
29          There is more limited evidence for reduced fecundity and effects on sperm with exposure
30    to tetrachloroethylene, but it is suggested in several studies (Rachootin and Olsen, 1983;
31    Eskenazi et al., 1991a, b; Sallmen et al., 1995). Sallmen et al. (1995) observed a lower
32    probability of achieving a clinically recognized pregnancy among women employed in dry
33    cleaning shops (incidence density ratio [IDR] = 0.44, 95% CI = 0.22-0.86, 11 women) in
34    analyses that adjusted for a number of other covariates. Furthermore,  tetrachloroethylene
35    exposure was associated with a decreased probability of pregnancy, although an exposure-
36    response pattern was not apparent (low exposure to tetrachloroethylene, IDR = 0.63, 95% CI =
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-120      DRAFT-DO NOT CITE OR QUOTE

-------
 1    0.34-1.17, based on 13 women; high exposure to tetrachloroethylene, IDR = 0.69, 95% CI =
 2    0.31-1.52, based on 7 women). Exposure was defined as frequency of tetrachloroethylene use,
 3    with no attention paid to level of exposure. Hence, exposure misclassification may partially
 4    explain the lack of an exposure-response relationship. Sallmen et al. (1995) examined these
 5    exposures as part of a larger evaluation of general organic solvent exposure for which a
 6    statistically significant association was noted (high level exposure, IDR = 0.41, 95% CI =
 7    0.27-0.62).
 8          Eskenazi et al. (1991a, b) reported a statistically significant reduced probability of
 9    pregnancy among highly exposed individuals. Eskenazi et al. (1991a) noted a lower per-cycle
10    pregnancy rate among wives of men who received higher-level exposure to tetrachloroethylene
11    (RR = 0.94, 95% CI = 0.85-1.04) as compared with wives of men who received lower-level
12    exposure.  The potential contribution of tetrachloroethylene exposure on time to conception was
13    small compared to the contribution observed from Hispanic ethnicity and smoking, which were
14    found to be stronger and statistically significant predictors of time to conception.
15          Eskenazi et al. (1991b) also found subtle spermatogenic effects among dry cleaners when
16    compared with laundry workers.  These effects were characterized as a greater proportion of
17    round sperm and a lower proportion of narrow sperm. Furthermore, tetrachloroethylene level
18    was a statistically significant predictor of decreased number of narrow sperm and of increased
19    numbers of spermatids with amplitude of lateral  head displacement, a measure of the unsteady
20    movement of the sperm  head about its average path of motion. Mean tetrachloroethylene
21    exposure was 1.2 ppm for dry cleaners and 0.01  ppm for laundry workers. More traditional
22    measures of semen quality, such  as number of sperm, concentration, volume, average percentage
23    of motile sperm or average percentage of abnormally shaped sperm, as well as the prevalence of
24    study subjects identified as azospermic or oligospermic, did not differ between these two job
25    titles. Moreover, it did not appear that the effects on semen parameters  or the per-cycle
26    pregnancy rate had a large impact on fertility rates.  Partners of these male dry cleaners did not
27    have fewer pregnancies  as compared with a national standard (Eskenazi et al.,  199la).
28          The findings from these studies are consistent with those observed by Rachootin and
29    Olsen (1983), whose study  subjects were couples seeking treatment for infertility. These
30    investigators noted that employment as a dry  cleaner was associated with hormonal disturbances
31    and delayed conception  in analyses that took into account the woman's age, education, residence,
32    and parity. Exposure to tetrachloroethylene was inferred but not documented.
33          Few epidemiologic  studies exist that evaluate other developmental toxicity endpoints
34    such as decreased birth weight, intrauterine growth restriction (IUGR; also known as  small for
35    gestation age [SGA]), and congenital anomalies. Many of the analyses were included in the
36    reports of spontaneous abortions discussed above.  Overall, no associations were noted in several
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-121      DRAFT-DO NOT CITE OR QUOTE

-------
 1    studies that assessed maternal or paternal exposure to tetrachloroethylene and increased
 2    incidences of stillbirths, congenital anomalies, or decreased birth weight (Olsen et al., 1990;
 3    Kyyronen et al., 1989; Taskinen et al., 1989; Windham et al., 1991).  These findings may be a
 4    reflection of the small numbers of exposed cases, or they may be attributable to exposure
 5    misclassification (biological marker data were available for only a few study subjects) or disease
 6    misclassification (which could be introduced from the grouping of several different outcomes
 7    into one category).
 8          The case-control study by Windham et al. (1991) observed a strong but imprecise
 9    association between IUGR and exposure to tetrachloroethylene (OR = 12.5, 95% CI not given in
10    the published paper and too few data for NCEA staff to calculate). This observation was based
11    on only one exposed case who also had exposure to trichloroethylene; an RR greater than 1 was
12    also observed for trichloroethylene exposure (OR = 4.2).
13          Studies of populations serviced by drinking water containing several contaminants,
14    including tetrachloroethylene and trichloroethylene, report  elevated risks such as adverse
15    pregnancy or postnatal outcomes attributed to living in a residence receiving contaminated water
16    effects (Lagakos et al., 1986; Bove et al., 1995; ATSDR, 1998).  Lagakos et al. examined the
17    relationship between several birth outcomes that were identified from questionnaires given to a
18    sample of residents from Woburn, MA.  This study was part of a larger study  evaluating  the
19    association between childhood leukemia among residents of this town and living in a residence
20    receiving drinking water from two wells contaminated with trichloroethylene,
21    tetrachloroethylene, and chloroform. The levels of these contaminants in the wells at the time
22    they were closed were 267 ppb (|ig/L) trichloroethylene, 21 ppb (|ig/L) tetrachloroethylene, and
23    12 ppb (|ig/L) chloroform.  The investigators observed statistically significant associations
24    between a residence receiving contaminated water and three outcomes: perinatal deaths  since
25    1970,  eye and ear anomalies, and CNS/oral cleft anomalies.
26          An analysis by Shawn and Robins (1986) of events  among residents of East Woburn, the
27    location of the contaminated wells, noted a statistically significant exposure-response trend only
28    for perinatal deaths. This analysis was presented in comments by the study authors to the study
29    by Lagakos et al. (1986) and was carried out to evaluate recall bias in that study, which these
30    authors concluded did not exist.
31          A case-control study of leukemia cases among children in Woburn, MA (MA DPH,
32    1997) noted a large risk between maternal exposure (e.g., living during the first trimester of
33    pregnancy in a residence that received contaminated water) and leukemia (ORadj = 8.3, 90% CI =
34    0.7-94.7, 10 exposed case). Risks increased significantly (p < 0.05) with increasing exposure
35    (never exposed, least exposed, most exposed). This study is more fully discussed in the section
36    reviewing the epidemiologic evidence on cancer effects.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-122      DRAFT-DO NOT CITE OR QUOTE

-------
 1          In a prevalence study, Bove et al. (1995) assessed the relationship between a number of
 2    birth outcomes, as identified from the birth certificate or from the New Jersey Birth Defects
 3    Registry, and residence in 75 towns for which monitoring data were available. Concentrations of
 4    trihalomethanes, trichloroethylene, and tetrachloroethylene, along with a wide range of other
 5    solvents, were identified from the monitoring data. Bove et al. (1992, 1995) observed risks
 6    above 1.5 between residence in a town with >10 ppb tetrachloroethylene detected in drinking
 7    water and oral cleft defects (OR = 3.5, 90% CI = 1.3-8.78, four exposed cases). No associations
 8    were reported for other birth outcomes such as CNS defects, neural tube defects, low birth
 9    weight, and small for gestational age and tetrachloroethylene exposure.
10          This analysis lacks information on risk factors for an individual. To address this
11    limitation, the investigators conducted a case-control study of oral cleft defects (Bove  et al.,
12    1992), where findings did not support the observations in the ecological study. The association
13    between tetrachloroethylene in water (>5 ppb) and oral clefts was not elevated (ORadj = 0.4, 95%
14    CI = 0-4.3, four exposed cases)  in the case-control analyses. Given the better design of the case-
15    control study and its ability to include information on individual study participants, this study
16    carries a greater weight than does the  1995 ecological study in the overall evaluation of the
17    relationship between tetrachloroethylene exposure and developmental effects.
18          An analysis by ATSDR (1998; results published by Sonnenfeld et al., 2001) examined
19    birth weight and gestational age among births of residents living in base family housing at Camp
20    Lejeune, NC. The residences received drinking water contaminated by solvents, including
21    trichloroethylene, tetrachloroethylene, and/or benzene. A large number of births (n = 6,117)
22    between 1968 and 1985 were identified from birth records and were classified as exposed to
23    tetrachloroethylene, i.e., the mother had resided at any time of pregnancy in the base housing,
24    specifically Tarawa Terrace, which had received tetrachloroethylene-contaminated drinking
25    water. Although quantitative information  on exposure is limited in this study (water samples
26    were collected on only three different occasions between 1982 and 1985), it is thought that the
27    well providing water to Tarawa Terrace was contaminated with tetrachloroethylene for as long as
28    30 years (ATSDR,  1998; Sonnenfeld et al., 2001). The highest concentrations of contaminants
29    measured in tap water from Tarawa Terrace were 215 ppb (|ig/L) tetrachloroethylene,  8 ppb
30    (ng/L) trichloroethylene (single  sample), and 12 ppb (|ig/L) 1,2,-dichloroethylene (single
31    sample). No information on level of tetrachloroethylene exposure was available prior  to 1982.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-123      DRAFT-DO NOT CITE OR QUOTE

-------
 1           The frequency of a birth of an SGA10 infant was slightly increased among women who
 2    were identified as tetrachloroethylene exposed (10.2%) as compared with those who were not
 3    (9%; OR = 1.2, 90% CI =1.0-1.3, 622 exposed births). The investigators also observed a
 4    smaller mean birth weight of exposed infants as compared with infants of mothers who lived in
 5    unexposed housing—a difference of 24 g, which was not considered to be of biological
 6    significance. Two susceptible groups were identified from this analysis: mothers 35 years or
 7    older and mothers with previous fetal deaths. For older mothers, the adjusted difference in mean
 8    birth weight between tetrachloroethylene-exposed and unexposed births was 205 g (90% CI =
 9    78-333), with a risk (OR) of 4 (95% CI = 1.6-10.2, 11  exposed births) between  exposure and
10    birth of an SGA infant.  For mothers with prior fetal losses, exposure to tetrachloroethylene in
11    drinking water was associated with a 60% higher risk for an infant that was identified as SGA
12    (95% CI = 1.2-2.1, 147 exposed births).
13           Inferences regarding developmental and reproductive effects from tetrachloroethylene are
14    limited due to small risks of low precision, the lack of a direct measure of tetrachloroethylene in
15    many studies, the small numbers of exposed cases, and possible biases such as recall or
16    misclassification bias. The epidemiologic evidence is strongest for spontaneous abortions among
17    exposed women.  A number of studies have reported an elevated risk of spontaneous abortions
18    and maternal exposure to tetrachloroethylene, primarily exposure received  through employment
19    as a dry cleaner (Doyle et al., 1997; Windham et al., 1991; Olsen et al., 1990; Lindbohm et al.,
20    1990; Kyyronen et al., 1989; Bosco et al., 1987). The epidemiologic evidence for infertility is
21    further suggestive of an association with tetrachloroethylene exposure (Rachootin and Olsen,
22    1983; Eskenazi et al., 1991a, b; Sallmen et al., 1995).  Any  conclusions of  effects on birth weight,
23    IUGR (SGA), or congenital anomalies and tetrachloroethylene exposure cannot  be drawn from
24    the available occupational studies, although drinking water  studies of exposures  to multiple
25    chemicals, including tetrachloroethylene, provide some limited evidence. There is very little
26    information about what exposure pattern (concentration and duration) is  associated with these
27    effects. Table 4-8 summarizes these studies.
28
29    4.7.2.  Animal Studies
30           Evaluation of the developmental and reproductive effects of tetrachloroethylene exposure
31    in animal models is based on several studies of in utero exposures to maternal animals during
32    specific periods of pregnancy. These studies include embryo explant in rats, multigeneration
             10 SGA is measured by comparing birth weight at specific gestational ages with a gestational-age-specific-
      birth-weight distribution. Live births of infants weights less than the 10th percentile are classified as SGA. Three
      standards were examined by Sonnenfeld et al. (2001), and the standard of Williams et al. (1982) provided the best fit
      to the data.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-124      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Table 4-8.  Developmental/reproductive studies in humans
Subjects
3,517 pregnancies in women
employees in dry cleaners
and laundries
3,279 pregnancies among
dry cleaners in four Nordic
countries. Occupational
records were linked to
national registries of birth
and reproductive failures.
53 women dry cleaners and
ironers whose pregnancy
outcomes were compared
with outcomes during time
as a housewife.
Case-control study of
spontaneous abortions
(1,361 cases, controls with
live births)
20 women in dry cleaning
shops, 90 unexposed control
women
Wives of 17 men exposed to
tetrachloroethylene dry
cleaners, wives of 32 male
laundry workers
34 male dry cleaners, 48
unexposed male laundry
workers
Case-control study of
infertility and delayed
conception: 1,069 infertile
case couples; 4,305 fertile
control couples
Effect
Spontaneous abortions
Spontaneous abortions
elevated among "high-
exposure" workers in all
countries, primarily
attributed to excesses in
Finland
Spontaneous abortions
fourfold higher than among
nonexposed housewives but
not statistically significantly
different
Spontaneous abortion excess
in 7 women, 4 of whom
were also exposed to
trichloroethylene
Lower probability of
clinically recognized
pregnancy
Lower per-cycle pregnancy
rate but no decrease in
fertility rate
Subtle sperm changes but no
change in clinical sperm
quality criteria
Delayed conception and
hormonal disturbances:
OR = 3 for exposure to dry-
cleaning chemicals (95% CI
= 1.2-7.4)
Exposure
Elevated risks by several
definitions of "exposed."
No tetrachloroethylene
measurements
Blood levels of TCA not
elevated in some of these
workers
TCA in dry cleaners was
5 (ig/L; TCA in ironers was
1.4jig/L

No dose-response with
tetrachloroethylene exposure
High vs. low
tetrachloroethylene exposure
categories
Mean exposure was 1 .2 ppm
Tetrachloroethylene
exposure inferred
Authors
Doyle et al.
(1997)
Olsen et al.
(1990)
Bosco et al.
(1987)
Windham et
al. (1991)
Sallmen et al.
(1995)
Eskenazi et
al. (1991a)
Eskenazi et
al. (1991b)
Rachootin,
and Olsen
(1983)
               This document is a draft for review purposes only and does not constitute Agency policy
    06/06/08                                  4-125      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
       Table 4-8. Developmental/reproductive studies in humans (continued)
Subjects
4,396 pregnancies among
residents of Woburn, MA.






80,938 live births and 594
fetal deaths among residents
in 75 New Jersey towns
Case-control study of
selected birth outcomes in
New Jersey; 49 cases of oral
cleft defects, 138 controls
1 1,798 births among women
living in United States
Marine Corp base housing
Effect
Statistically significant
positive association between
access to contaminated
water and
(a) perinatal deaths since
1970 and
(b) eye/ear birth anomalies
No association between
water access and
(a) incidence of
spontaneous abortion
(b) low birth rate
(c) perinatal deaths before
1970
(d) musculoskeletal birth
anomalies
(e) cardiovascular birth
anomalies
Oral cleft defects, OR = 3.5,
95% CI= 1.3-8.8) based on
four exposed cases
The association between
tetrachloroethylene in water
(>5 ppb) and oral cleft
defects was not elevated.
Excess of age (SGA) births
in women >35 years of age
among mothers with prior
fetal losses
Exposure
Tetrachloroethylene: 21
Mg/L
Trichloroethylene: 267 (ig/L
Chloroform: 12 (ig/L





>10 (ig/L in drinking water


Tetrachloroethylene: <215
Mg/L
Authors
Lagakos et al.
(1986)






Bove et al.
(1995)
Bove et al.
(1992)
ATSDR,
(1998), Son-
nenfeld et al.
(2001)
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
reproduction in rats, and an in vitro oocyte fertilization assay following in vivo exposure of adult
female rats.
       In an inhalation developmental toxicity study (Schwetz et al., 1975), Sprague-Dawley
rats and Swiss-Webster mice were exposed to airborne tetrachloroethylene at 300 ppm 7 hrs/day
on days 6-15 of gestation. Following laparohysterectomy on gestation days 21 or 18 (for rats
and mice, respectively), fetuses were weighed and measured, examined for external
abnormalities, and processed for the evaluation of either soft tissue or skeletal abnormalities.
Three other organic solvents were also tested with the same protocol; the concentration of all
agents was chosen to be approximately twice their threshold limit values.  Although the study
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   4-126      DRAFT-DO NOT CITE OR QUOTE

-------
 1    authors concluded that there was no significant maternal, fetal, or embryo toxicity for any of the
 2    solvents tested, the maternal and fetal data demonstrated a number of statistically significant
 3    differences from control values following gestational exposures to tetrachloroethylene in rats and
 4    mice.  In the rats, exposures to tetrachloroethylene produced slight but statistically significant
 5    maternal toxicity (4-5% reductions in mean maternal body weight gains) and embryotoxicity
 6    (increased resorptions; 9% in treated vs. 4% in controls).  In the mice, maternal toxicity consisted
 7    of a significant 21% increase in mean relative liver weight as compared with controls. The mean
 8    fetal weight in mice was significantly (9%) less than in the concurrent control, and the percent of
 9    litters with delayed ossification of the skull bones, delayed ossification of the sternebra, and
10    subcutaneous edema were significantly increased.
11          Szakmary et al. (1997) exposed CFY rats to tetrachloroethylene via inhalation throughout
12    gestation (i.e., gestation days 1-20) for 8 hrs/day at concentrations of 1,500, 4,500, or
13    8,500 mg/m3. In the same study, the study authors exposed C57B1 mice via inhalation on
14    gestation days 7-15 (i.e., during the period of organogenesis) to a concentration of 1,500 mg/m3
15    and New Zealand white rabbits during organogenesis (gestation days 7-20) to a concentration of
16    4,500 mg/m3. Maternal animals were killed approximately  1 day prior to expected delivery; a
17    gross necropsy was conducted, organ weights were recorded, blood was taken by aorta puncture
18    for hematology and clinical chemistry evaluations, ovarian corpora lutea were counted, and
19    uterine contents were examined (number and position of living, dead, or resorbed fetuses; and
20    fetal and placental observations and weights).  The numbers of litters available for evaluation
21    were as follows:  20 control and 21 or 22 per treated group in the rat, 77 control and 10 treated in
22    the mice, and 10 control and 16 treated in the rabbit.  One-half of the fetuses from each litter
23    were evaluated for visceral abnormalities, and the other half were evaluated for skeletal
24    development. The study authors reported that the organs of five dams and five embryos from
25    each group were also evaluated by routine histological methods. To evaluate the concentration
26    of tetrachloroethylene in maternal and fetal blood and in amniotic fluid, another subset of rats
27    (number not specified) was studied.  (For the 1,500 and 8,500 mg/m3 exposure levels, maternal
28    blood concentrations of tetrachloroethylene were 17.8+8.9 and 86.2+13.0 uL/mL, respectively.
29    Concentrations in the fetal blood were 66% and 30% of maternal blood concentrations, and
30    amniotic fluid concentrations were 33% and 20% of maternal blood concentrations.) In the rat,
31    at 4,500 and 8,500 mg/m3, maternal body weight gain during gestation was significantly
32    decreased  (37 and 40%, respectively), relative maternal liver mass was significantly increased
33    (10  and 6%, respectively), and serum aspartate amino transferase activity was increased (data not
34    provided)  as compared to controls. Percent pre-implantation loss was significantly increased
35    from controls by 133 and 117% at these exposure levels, while percent post-implantation loss
36    was increased non-significantly from controls by 80% in each group. Also, at 4,500 and

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-127     DRAFT-DO NOT CITE OR QUOTE

-------
 1    8,500 mg/m3, fetal weight was significantly decreased in 98.5 and 100% of all fetuses, the
 2    number of fetuses with skeletal retardation was significantly increased in 98.5% and 100% of
 3    fetuses, and the percent of fetuses with malformations was both significantly increased to 6.4%
 4    and 15.7% as compared to the control incidence of 2.0%.  Although the study authors judged the
 5    1,500 mg/m3 exposure level to be the NOAEL for the rat study, it is noted that there were
 6    concentration-dependent non-significant decreases in maternal body weight gain (13% lower
 7    than control), and increases in pre- and post-implantation loss (49% and 38% greater than control,
 8    respectively). The percent of weight-retarded fetuses increased to 3.4 times the control incidence,
 9    and the incidences of fetuses with skeletal retardation (48% increased) or total malformations
10    increased by 2.3 times the control incidence observed at the low-exposure level of 1,500 mg/m3.
11    Therefore, these findings are judged to be adverse consequences of treatment. The attribution of
12    these findings to treatment, and the designation of 1,500 mg/m3 as the study LOAEL is
13    consistent with the adverse developmental findings of Schwetz et al. (1975). In mice
14    (1,500  mg/m3) and rabbits (4,500 mg/m3), relative liver mass was significantly increased;
15    decreased maternal body weight gain was also observed in the rabbits.  In the mice, a
16    significantly increased number of fetuses with visceral malformations (details not specified) was
17    observed, while in the rabbits, two (of 16) does aborted, total resorption of four litters was
18    reported, and the percent of post-implantation loss was  significantly increased. The percent of
19    rabbit fetuses with malformations (details not provided  in the report) was also increased,
20    although not significantly.
21           An additional cohort of rats from the Szakmary  et al. study (15 litters/group at exposure
22    levels of 1,500 or 4,500 mg/m3 tetrachloroethylene) was allowed to deliver, and the offspring
23    (standardized to 8 pups/litter) were maintained on study to postnatal day 100. It was not clearly
24    specified in the report whether the daily inhalation exposures continued throughout the postnatal
25    period. Pre-weaning observations included weekly body weights, developmental landmarks
26    (pinna  detachment, incisor eruption, and eye opening), and functional  assessments (forward
27    movement, surface righting reflex, grasping ability, swimming ontogeny, rotating activity,
28    auditory startle reflex, and examination of stereoscopic  vision). After weaning, exploratory
29    activity in an open field, motor activity in an activity wheel, and development of muscle strength
30    were assessed. The study authors reported that adverse findings included a decreased survival
31    index (details not provided), a minimal decrease of exploratory activity and muscular strength in
32    treated offspring  (presumably at both exposure levels) that normalized by postnatal day 51, and
33    significantly increased motor activity on postnatal day 100 of females exposed to 4,500 mg/m3 of
34    tetrachloroethylene.
35           Nelson et al.  (1980) investigated developmental neurotoxicity in Sprague-Dawley rats by
36    exposing pregnant dams to tetrachloroethylene at concentrations of 100 ppm and 900 ppm during

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-128      DRAFT-DO NOT CITE OR QUOTE

-------
 1    either early pregnancy (gestation days 7 to 13) or late pregnancy (gestation days 14 to 20). They
 2    performed morphological examinations of the fetuses (gross, visceral, and skeletal) and
 3    behavioral testing and neurochemical analyses of the offspring.
 4          There were no alterations in any of the measured parameters in the 100 ppm groups. At
 5    900 ppm there were no skeletal abnormalities, but the weight gain of the offspring as compared
 6    with controls was depressed about 20% at postnatal weeks 3-5.  Developmental delay was
 7    observed in both the groups exposed in early and in late pregnancy. Offspring of the early
 8    pregnancy-exposed group performed poorly on an ascent test and on a rotorod test, whereas
 9    those in the late pregnancy group underperformed on the ascent test at only postnatal day 14.
10    However, later in development (days 21 and 25) their performance was higher than that of the
11    controls on the rotorod test. These pups were markedly more active in the open field test at days
12    31 and 32. Activity wheel testing on days 32 and 33 did not reveal statistically significant
13    changes.  Avoidance conditioning on day 34 and operant conditioning on days 40-46 failed to
14    suggest effects. Neurochemical analyses of whole brain (minus cerebellum) tissue in 21-day-old
15    offspring revealed significant reductions in acetylcholine levels at both exposure periods,
16    whereas dopamine levels were reduced among those exposed on gestation days 7-13.
17          All of the described effects in the 900 ppm group were statistically significant as
18    compared with controls. Unfortunately, none of the statistics for the 100 ppm treatments were
19    presented. The authors observed that more behavioral changes occurred in offspring exposed
20    during late pregnancy than in those exposed during early pregnancy.
21          Beliles et al. (1980) described an experiment in which male rats and mice were exposed
22    via inhalation to tetrachloroethylene concentrations of 100 and 500 ppm for 7 hrs/day for 5 days.
23    Sperm head abnormalities and abnormal sperm were evaluated at 1, 4, and  10 weeks after the last
24    dose. Rats were unaffected. At 4 weeks but not at 1 or 10 weeks after exposure there was a
25    significant increase (p < 0.05) in the percentage of mice with abnormal sperm heads (19.7%) for
26    animals inhaling 500 ppm.  For the 100 ppm and control groups the percentages were 10.3% and
27    6% (not statistically significant at the/? < 0.05 level), respectively. A positive control group
28    administered triethylene melanime was adversely affected (11.1%). The authors suggested that
29    the temporal appearance of the abnormal sperm heads  indicated that the spermatocyte and/or
30    spermatogonia were the stages most sensitive to the effects of inhaled tetrachloroethylene. In
31    this study the NOAEL was  100 ppm and the LOAEL was 500 ppm.
32          Hardin et al. (1981;  see also Beliles et al., 1980) found no developmental toxicity among
33    the fetuses from Sprague-Dawley rats or New Zealand White rabbits inhaling 500 ppm of
34    tetrachloroethylene for 7 hrs/day, 5 days/week.  Tetrachloroethylene was administered with and
35    without three-week pregestation exposures and with both full-term and terminal two-thirds-term
36    exposure.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-129      DRAFT-DO NOT CITE OR QUOTE

-------
 1          In a developmental toxicity study, Carney et al. (2006) investigated the effects of whole-
 2    body inhalation exposures to pregnant Sprague-Dawley rats at nominal concentrations of 0, 75,
 3    250, or 600 ppm (actual chamber concentrations of 0, 65, 249, or 600 ppm) tetrachloroethylene
 4    for 6 hrs/day, 7 days/week on gestation days (GD) 6-19. This study was conducted under Good
 5    Laboratory Practice (GLP) regulations according to current EPA and OECD regulatory testing
 6    guidelines. Maternal toxicity consisted of slight but statistically significant decreases in body
 7    weight gain during the first 3 days of exposure to 600 ppm, establishing a no-adverse-effect
 8    concentration of 249 ppm for dams. A slight, statistically significant decrease in gravid uterine
 9    weight at 600 ppm correlated with significant reductions in mean fetal body weight (9.4%) and
10    placental weight (15.8%) at GD 20 cesarean section. At >249 ppm, mean fetal and placental
11    weights were significantly decreased by 4.3% and 12.3% from control, respectively. A
12    significant increase in the incidence of incomplete ossification of the thoracic vertebral centra at
13    this exposure level was consistent with fetal growth retardation.  No treatment-related alterations
14    in fetal growth or development were noted at 65 ppm.  Therefore the LOAEL for this study is
15    249 ppm.
16          Saillenfait et al. (1995), using a rat whole embryo (day 10) culture system,  found
17    tetrachloroethylene-induced embryo toxicity, including mortality, malformations, and delayed
18    growth and differentiation. No adverse effect was produced at the 2.5 mM concentration, but
19    concentration-related trends of increasing toxicity occurred from 3.5 mM through  15 mM.
20    Statistical  tests for a concentration-related trend were not reported. The investigators found that
21    trichloroethylene produced similar effects, with potency somewhat less than that of
22    tetrachloroethylene. They also found that TCA and DCA caused a variety of abnormalities in
23    this culture system.
24          In a developmental toxicity screening study, timed-pregnant F344 rats were treated by
25    gavage with tetrachloroethylene doses of 900 or 1,200 mg/kg-day in corn oil vehicle on gestation
26    days  6-19 (Narotsky and Kavlock, 1995). There were 17 dams in each of the
27    tetrachloroethylene-treated groups and 21 in the control groups.  The dams were allowed to
28    deliver, and their litters were examined on postnatal days 1, 3, and 6.  At 1,200 mg/kg no live
29    pups were delivered on day 22 of gestation.  At 900 mg/kg-day there was maternal ataxia, and
30    weight gain was markedly less than in the controls. The number of pups per litter  was reduced
31    (p < 0.01) as compared with the controls at  day 22 of gestation. On postnatal day 6 the number
32    of pups per litter was reduced (p < 0.001) as compared with the controls. The investigators noted
33    that full-litter resorptions were not observed with other chemicals they tested in the presence of
34    maternal toxicity.  An increase in micro/anophthalmia was found in the offspring.  There was no
35    evaluation for skeletal  changes, and not all available pups were examined for soft tissue changes.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-130     DRAFT-DO NOT CITE OR QUOTE

-------
 1    Because of the high dose levels and limited evaluation of the soft tissue changes, the
 2    malformations described are of limited impact.
 3          A multigeneration study of the effects on rats of exposure to airborne concentrations of
 4    tetrachloroethylene was performed by Tinston (1994). Although this study has not been
 5    published,  it was submitted to EPA (Office of Prevention, Pesticides, and Toxic Substances and
 6    to the IRIS Office as a result of the data call-in for the IRIS update). It was conducted under
 7    good laboratory practice standards and received frequent quality assurance audits.  In this study,
 8    weanling male and female (Alpk:APfSD) rats (FO) were exposed to airborne tetrachloroethylene
 9    concentrations of 0, 100, 300, or 1,000 ppm 6 hrs/day, 5 days/week, for 11 weeks prior to mating
10    and then for 6 hrs/day during mating and through day 20 of gestation.  There were no exposures
11    from day 21 of gestation through day 5 postpartum.  One litter was produced in the first
12    generation (F1A).  The first-generation dams and their litters were exposed to
13    tetrachloroethylene from postnatal day 6 through 29, at which time parental animals for the
14    second generation were selected.   The second-generation parents (Fl) were then exposed 5
15    days/week during the 11-week pre-mating period. In the second generation, three litters were
16    produced:  F2A, F2B, and F2C. The F2A dams and litters were exposed from days 6 to 29
17    (control and 100 ppm) or days 7 to 29 (300 ppm). The 1,000 ppm exposure for the Fl dams
18    stopped after the F2A littering.
19          F2B litters were generated by mating the Fl parental males and females in the control,
20    300, and 1,000 ppm groups; the dams and F2B litters were not exposed to tetrachloroethylene
21    during lactation. An F2C litter was produced by mating Fl males exposed to 1,000 ppm with
22    unexposed females.  These  females and the F2C litters were killed on postnatal day 5 and
23    discarded without further examination.  Overall, the FO males were exposed for  19 weeks and the
24    Fl males were exposed up to 35 weeks. Postmortem evaluation in adults and selected weanlings
25    included organ weight and histopathology examination of liver, kidney, and reproductive organs;
26    sperm measures were not assessed.
27          Table 4-9 summarizes the results of the Tinston study. Signs of CNS depression
28    (decreased activity and reduced response to sound) were observed at 1,000 ppm for the first 2
29    weeks in both adult generations and again when the exposure was resumed on day 6 postpartum
30    in the F1 generation (adults and pups). Other signs of overt tetrachloroethylene  toxicity in the
31    adults included irregular breathing and piloerection at both 1,000 and 300 ppm and salivation
32    and tip-toe gait (in one Fl female) at 1,000 ppm.  These changes stopped with the cessation of
33    exposure or within approximately  30 minutes thereafter.
34          There were a number of changes relative to controls that were of minor biological
35    significance. One change was transient statistically significant reductions of mean body weights

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-131      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
             Table 4-9. Exposure concentrations (ppm) at which effects occurred in a
             two-generation study
Parameter
Clinical signs
(piloerection, irregular
breathing)
Behavioral effects
(decreased activity;
reduced response to sound)
Transient decreased body
weight gains
Decreased mean testes
weight
Increased liver and kidney
weights
Renal histopathology
Decreased pups born alive
(%)
Decreased mean % pup
survival days 1-5
Decreased mean % pup
survival days 5-22
Decreased mean male pup
weight day 1
Decreased mean female
pup weight day 1
Decreased mean male pup
weight day 29
Decreased mean female
pup weight day 29
Generation
FO
1,000, 300
1,000
1,000, 300

1,000
1,000







F1A

1,000

1,000


l,000b
1,000
l,000b
1,000C
1,000C
l,000b,
300b, 100
b,d
l,000b,
300b, 100d
Fl
1,000, 300
1,000
1,000, 300
1,000
1,000
1,000







F2A






1,000C
1,000C
l,000b
1,000C
l,000b


F2B






1,000C


1,000C
1,000C


F2Ca







1,000C
NA


NA
NA
 4
 5
 6
 7
 8
 9
10
11
12
     a Not exposed after delivery.
     V<0.05.

     d trend;? < 0.05.

     NA = Not applicable (pups terminated on day 5 postnatal)

     Source: Adapted from Tinston (1994).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-132      DRAFT-DO NOT CITE OR QUOTE

-------
 1    (originating from treated males and nontreated females) suggests the absence of male-mediated
 2    effects on reproductive outcome. Nevertheless, the alterations in testes weight cannot be
 3    discounted as a possible effect of treatment.
 4          In females, dystocia was noted in one FO dam at 100 ppm, two Fl dams at 300 ppm, and
 5    a total of four dams (two each FO and Fl) at 1,000 ppm; these dams were terminated without
 6    completion of delivery.  From the data for surviving dams and litters, it can be assumed that the
 7    difficulties in parturition were not associated with or attributable to alterations in mean gestation
 8    length or increased mean pup or litter weights. In fact, mean pup body weights showed a
 9    statistically significant decrease throughout the lactation period at 300 and 1,000 ppm for Fl A
10    litters and in early lactation for F2A and F2B litters. Additionally, mean Fl A male pup body
11    weight was significantly decreased (5% less than controls; p < 0.05)  at 100 ppm on postnatal day
12    29.  These postnatal day 29 mean body weight deficits in all treated groups were observed in the
13    animals selected as parents of the second generation, but by the second week of the Fl pre-
14    mating period, mean body weights were similar to those of controls for both 100 and  300 ppm
15    animals.
16          Mean litter size was decreased at 1,000 ppm for F2A and F2B litters.  Statistically
17    significant decreases in the number of live pups on postnatal day 1 (25% and 37% lower than
18    controls for F2A and F2B, respectively) are suggestive of either an adverse effect on fertilization
19    or on in utero survival. Early postnatal survival (i.e., on postnatal day 1 and between postnatal
20    days 1 and 5) was also compromised in F2A and F2B  pups at 1,000 ppm, with mean litter sizes
21    decreasing to 48% and 53% of those of controls, respectively. The number of dead pups and
22    litters with dead pups was  also increased,  although not significantly,  at 300 ppm for F2A litters.
23    Clinical observations data  for 1,000  ppm litters reported an increased incidence of F2A and F2B
24    pups that were found dead, were killed in extremis or were missing and presumed dead.  The
25    apparent increase in adverse survival findings at 300 and 1,000 ppm  in the second generation as
26    compared with the first generation could not be definitively attributed to any particular aspect of
27    study design or conduct (e.g., differences  in the duration of treatment), although it is noted that,
28    unlike the second generation (Fl) parental animals, the first generation (FO) rats were not
29    exposed to tetrachloroethylene during preconception and in utero development.
30          A deficiency of the Tinston study is that the pregnant rats were not exposed from
31    gestation day 21 through lactation day 6 or 7, and the exposure at the 1,000 ppm treatment level
32    stopped for the Fl dams at the littering of the F2B pups.  The F2B pups were not exposed
33    postnatally. A summary of the doses at which the effects were observed in the study  is presented
34    in Table 4-9.
35          In a study designed to examine the fertilizability of rat oocytes, female rats were exposed
36    to inhaled tetrachloroethylene at 12,000 mg/m3 (2 hrs/day, 5 days/week) for 2 weeks (Berger and

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-133      DRAFT-DO NOT CITE OR QUOTE

-------
 1    Horner, 2003). The percentage of extracted oocytes that were fertilized in vitro was reduced for
 2    tetrachloroethylene-treated females as compared with controls.
 O
 4    4.7.2.1. Summary of Animal Studies
 5          Table 4-10 summarizes the findings of the animal studies described in this section. The
 6    data show that inhalation of tetrachloroethylene by pregnant mice and rats during various periods
 7    of gestation resulted in fetal growth retardation and mortality in several studies and in delayed
 8    behavioral changes in the three studies that measured these effects (SzakmacZy et al 1997;
 9    Nelson et al. 1980; Tinston, 1994). Single studies have shown changes in brain acetyl choline
10    and dopamine, altered brain fatty acid composition, and altered sperm morphology. These
11    effects occurred at doses higher than 300 to 1,000 ppm in various studies.
12          The overall NOAEL for the animal developmental/reproductive inhalation studies is 100
13    ppm, based on Tinston (1994).  The overall LOAEL is 300 ppm, based on Tinston (1994) and
14    Schwetz et al. (1975), in which increased mortality and decreased body weight of the offspring
15    were observed. All of these studies used the inhalation route of exposure.
16
17    4.7.3. Summary of Human  and Animal Developmental/Reproductive Studies
18          Inferences regarding developmental and reproductive effects from tetrachloroethylene
19    exposure in humans are limited due to small risks of low precision, the lack of a direct measure
20    of tetrachloroethylene in many studies, the small  numbers of exposed cases, and possible biases,
21    particularly in studies where the information on birth outcome or exposure is obtained by
22    questionnaire or inferred by residence. The epidemiologic evidence is strongest for spontaneous
23    abortions among exposed women.  A number  of studies have reported an elevated risk of
24    spontaneous abortion and maternal exposure to tetrachloroethylene, primarily exposure received
25    through employment as a dry cleaner (Doyle et al., 1997; Windham et al., 1991; Olsen et al.,
26    1990; Lindbohm et al., 1990;  Kyyronen et al., 1989; Bosco et al.,  1987).
27          The epidemiologic evidence for infertility is further suggestive of an association with
28    tetrachloroethylene exposure  (Rachootin and Olsen, 1983; Eskenazi et al., 1991a, b; Sallmen  et
29    al., 1995). Strong conclusions about effects on birth weight—IUGR (SGA)—or congenital
30    anomalies and tetrachloroethylene exposure cannot be drawn from the available occupational
31    studies, although drinking water studies of exposures to multiple chemicals, including
32    tetrachloroethylene, provide some limited evidence.  There is very little information about what
33    exposure pattern (concentration and duration) is associated with these effects.
34          Inhalation of tetrachloroethylene by pregnant mice and rats during various fractions of
35    the gestation period has resulted in fetal growth retardation and mortality in several studies and
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-134      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4-10.  Summary of animal developmental/reproductive studies
       fortetrachloroethylene, in chronological order
   Subjects
                               Effects
        Concentration
    Authors
 SW Mice
                    Maternal toxicity, decreased fetal
                    weight, delayed ossification, 9%
                    decrease in birth weight
300 ppm on gestation days 6-15
Schwetz et al.
(1975)
 SD Rats
                    Maternal toxicity, increased
                    resorptions (fetal death)
300 ppm on days 6-15
Schwetz et al.
(1975)
 CFY Rats
                    Maternal toxicity (decreased
                    body weight gain, increased liver
                    weight and serum enzymes);
                    increased pre- and post-
                    implantation loss, skeletal
                    retardation, and total
                    malformations; decreased fetal
                    weight
1,500, 4,500, 8,500 mg/m3 on
gestation days 1-20

LOAEL= 1,500 mg/m3
Szakmary et al.
(1997)
 CFY Rats
                    Decreased postnatal survival,
                    minimal transient decreases in
                    exploratory activity and muscular
                    strength, and increased motor
                    activity in females on postnatal
                    day 100
1,500, 4,500 mg/m3 on gestation
days 1-20 (and perhaps
postnatally to PND 100)

LOAEL= 1,500 mg/m3
Szakmary et al.
(1997)
 C57B1 Mice
                    Maternal toxicity (increased liver
                    weight); visceral malformations
1,500 mg/m3 on gestation days
7-15

LOAEL= 1,500 mg/m3
Szakmary et al.
(1997)
NZW Rabbits
                    Maternal toxicity (decreased
                    body weight gain, increased liver
                    weight); abortions, total litter
                    resorptions, increased
                    postimplantation loss,
                    malformations
4,500 mg/m3 on gestation days
7-20

LOAEL = 4,500 mg/m3
Szakmary et al.
(1997)
 SDRats
                    Decreased weight gain,
                    behavioral changes (more
                    extensive for late pregnancy
                    exposure), decreased brain
                    acetylcholine
0, 100, 900 ppm on days
7-13 or on days 14-20

NOAEL = 100 ppm

LOAEL = 900 ppm
Nelson et al.
(1980)
 Mice
                    Abnormal sperm heads at 500
                    ppm but not at 100 ppm,
                    spermatogonia or spermatocyte
                    stage affected
0, 100, 500 ppm for 5 days

LOAEL = 500 ppm
Beliles et al.
(1980)
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                     4-135      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
            Table 4-10.  Summary of animal developmental/reproductive studies
            fortetrachloroethylene, in chronological order (continued)
Subjects
Rats, rabbits
SD rats
Rat embryo in
culture
F344 rats
SD rats, two-
generation
study
Rats
Effects
No developmental toxicity
Maternal toxicity (decreased
body weight gain; decreased
gravid uterine weight); fetal body
weight and placental weight
decrements, increased delays in
thoracic vertebral ossification
Mortality, malformations,
delayed growth and
differentiation
100% mortality at 1,200 mg/kg-
day, increased mortality and
micro/anophthalmia at 900
mg/kg-day; soft tissues not
examined
Increased death of F1A and F2A
and F2B pups, decreased body
weight
Reduced fertilizability of
extracted oocytes
Concentration
Exposures throughout gestation
NOAEL = 500 ppm
0, 75, 250 or 600 ppm (actual
concentrations: 0, 66, 249, 600
ppm), 6 hr/day, 7 days/week, on
gestation days 0-19
Maternal LOAEL = 600 ppm
Fetal LOAEL = 250 ppm
No effect at 2.5 mM, effects at
3.5 mM and higher
Gavage, 900, 1,200 mg/kg-day on
gestation days 6-19
0, 100, 300,
1,000 ppm
NOAEL = 100 ppm for
body weight reduction
12,000 mg/m3, 2 hrs/day, 5
days/week for 2 weeks
Authors
Hardin et al.
(1981)
Carney et al.
(2006)
Saillenfait et al.
(1995)
Narotsky and
Kavlock(1995)
Tinston(1994)
Berger and
Horner (2003)
 4
 5
 6
 7
 8
 9
10
11
12
13
14
    in delayed behavioral changes in the two studies that measured these effects.  Single studies have
    shown changes in brain acetyl choline and dopamine, altered brain fatty acid composition, and
    altered sperm morphology.  These effects occurred at doses higher than 300 to 1,000 ppm in
    various studies.
           The overall NOAEL for the animal developmental/reproductive inhalation studies is 100
    ppm, based on Tinston (1994).  The overall LOAEL is 300 ppm, based on Tinston (1994) and
    Schwetz et al. (1975), in which increased mortality and decreased body weight of the offspring
    were observed.  All studies used the inhalation route of exposure except for one gavage study
    (Fredriksson et al., 1993), which showed behavioral toxicity.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  4-136      DRAFT-DO NOT CITE OR QUOTE

-------
 1          The finding of spontaneous abortions in several human studies of dry cleaners is
 2    supported by the occurrence of reduced birth weight and mortality in several animal studies. The
 3    finding of low birth weight in the Camp Lejeune studies by ATSDR is supported by reduced
 4    birth weight in five animal studies (Schwetz et al., 1975; SzakmacZy et al., 1997; Nelson et al.,
 5    1980; Carney et al., 2006; and in the Fl generation but not the F2 generation of Tinston (1994).
 6    There are no human observations of behavioral changes to compare with the animal evidence of
 7    CNS effects. The subtle nonadverse effects on sperm seen in humans correspond to one report
 8    of abnormal sperm in mice.  The LOAEL for developmental/reproductive effects in animals is
 9    300 ppm, but no quantitative measures are available for human effects, although other dry
10    cleaner studies cited in previous sections have 8-hr TWA exposures of 10-20 ppm.
11          For oral exposures in humans there are only suggestive effects associated with
12    tetrachloroethylene exposure in drinking water, with no reliable data on exposures; therefore, this
13    information does not contribute to the determination of a LOAEL. There is little information on
14    developmental or reproductive effects in animals by the oral route of exposure.
15
16    4.7.4.  Mode of Action for Developmental Effects
17          Because of its lipid solubility, tetrachloroethylene can cross both the blood-brain barrier
18    and the placental barrier and, therefore, it can be present in all tissues, including the brain, during
19    development.
20          Peroxidation of the lipids of the cell membranes (Cojocel et al., 1989), alteration of
21    regulation  of fatty acid composition of the membrane (Kyrklund and Haglid, 1991), disturbances
22    in the properties of the nerve membrane (Juntunen, 1986), and progressively increased activity in
23    one or more of the phosphoinositide-linked neurotransmitters (Subramoniam et al., 1989) have
24    all been suggested as MO As for neurotoxic effects.  These mechanisms  could be involved during
25    development phases, as well as in adults.
26          The metabolite TCA may be the causative agent for developmental toxicity expressed as
27    morphological changes, lethality, or growth. Evidence in support of this speculative position is
28    presented in the following discussion.  TCA is a weak organic acid, as are many developmental
29    toxicants, such as ethylhexanoic acid and valproic acid. These materials accumulate to a greater
30    extent in the embryo/fetal compartment than in the mother, based on the pKa of the acid and the
31    pH gradient between the maternal plasma and the embryo compartments (O'Flaherty et al.,
32    1992). TCA could induce developmental toxicity by changing the intracellular pH or through
33    peroxisome proliferation. Ghantous et al. (1986) detected TCA in the amniotic fluid of pregnant
34    mice exposed to tetrachloroethylene via inhalation.
35          Smith et al. (1989) found that oral gavage doses of TCA (330, 800, 1,200, and 1,800
36    mg/kg-day) delivered on gestation days 6-15 to pregnant Long-Evans rats produced soft tissue
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-137      DRAFT-DO NOT CITE OR QUOTE

-------
 1    malformations, principally in the cardiovascular system.  Johnson et al. (1998) found cardiac
 2    defects in rat fetuses whose mothers received 2,730 ppm  TCA in drinking water during the
 3    period of cardiac development.  Saillenfait et al. (1995), using the rat whole embryo (day 10)
 4    culture system, found that both tetrachloroethylene and TCA induced embryo toxicity, including
 5    mortality, malformations, and delayed growth and differentiation. TCA produced a reduction in
 6    the first branchial arch as well as other morphological changes at a lower concentration (2.5 mM)
 7    than that at which tetrachloroethylene induced no adverse effect (3.5 mM).  TCA also induced a
 8    reduction of the yolk sac diameter at 1 mM.
 9
10    4.8. TOXIC EFFECTS IN OTHER ORGAN SYSTEMS
11           This section discusses effects in organ systems not covered previously.  It does not
12    include effects on the liver, kidney, or nervous system; nor does it include developmental or
13    reproductive effects. To be consistent with other sections of the document, effects in humans are
14    presented separately from those in animals. Immune effects and lymphoid cancer are the most
15    studied, and these are the predominant topics of the noncancer and cancer sections, respectively.
16
17    4.8.1.  Human Studies
18    4.8.1.1. Noncancer Effects
19    4.8.1.1.1. Immune-related effects in humans. Adverse  effects on the immune system resulting
20    from chemical exposure fall within the following principal domains: immunosuppression (host
21    resistance), immunostimulation, autoimmunity, and allergy-hypersensitivity. Various
22    immunologic measurements (e.g., T-cell counts, immunoglobulin (Ig) E levels, specific
23    autoantibodies) may provide evidence of altered immune response that may subsequently be
24    related to risk of clinically expressed diseases such as infections, asthma, or systemic lupus
25    erythematosus.  Tetrachloroethylene exposure via air or water may result in immune-mediated
26    organ-specific or systemic effects, as described in a case report of hypersensitivity pneumonitis
27    in a 42-year-old female dry cleaner worker (Tanois et al., 2004).  Another case report described
28    severe fatigue, weight loss, myalgia, arthralgia, cardiac arrhythmia, decreased T-cell count, high-
29    titer (1:160) antinuclear antibodies, and neurological symptoms that were linked to an unusual
30    chemical sensitivity to tetrachloroethylene in a municipal water supply (Rea et al.,  1991).
31
32    4.8.1.1.1.1. Tetrachloroethylene and immunologic parameters. Byers et al. (1988) provide
33    data pertaining to immune function from 23 family members of leukemia patients in Woburn,
34    Massachusetts.  In 1979, testing of the wells in this town  revealed that the water in two of the
35    wells was contaminated with a number of solvents, including  tetrachloroethylene (21 ppb) and
36    trichloroethylene (267 ppb; as cited in Lagakos et al.,  1986).  These wells had been in operation
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-138      DRAFT-DO NOT CITE OR QUOTE

-------
 1    from 1964 to 1979.  Byers et al. collected serum samples in May and June of 1984 and in
 2    November of 1985.  They determined the total lymphocyte counts and lymphocyte
 3    subpopulations (CDS, CD4, CDS), and the CD4/CD8 ratio in these samples, and in samples from
 4    a combined control group of 30 laboratory workers and 40 residents of Boston selected through a
 5    randomized probability area sampling process. The study authors also assessed the presence of
 6    autoantibodies (antismooth muscle, antiovarian, antinuclear, antithyroglobulin, and
 7    antimicrosomal antibodies) in the family member samples and compared the results with
 8    laboratory reference values.  The lymphocyte subpopulations were higher and the CD4/CD8
 9    ratio was lower in the Woburn family members compared to the controls in both of the samples
10    taken in 1984. In the 1985 samples, however, the subpopulation levels had decreased and the
11    CD4/CD8 ratio had increased; the values were no longer statistically different from the controls.
12    None of the family member serum samples had antithyroglobulin or antimicrosomal antibodies,
13    but 10 family member serum samples (43%) had antinuclear antibodies (compared to <5%
14    expected based on the reference value). Because the initial blood sample was taken in 1984, and
15    because of the considerable mixture of exposures that occurred in this setting, it is not possible to
16    determine the patterns at a time nearer to the time of the exposure, or to infer the exact role of
17    tetrachloroethylene in alterations of the immunologic parameters.
18          Andrys et al. (1997) examined immunologic parameters in 21 dry cleaning workers (20
19    women) and  16 office workers  in the dry cleaning plant (14 women) and compared them to
20    reference  values based on samples from blood donors and "healthy persons in the same region"
21    (n = 14-311,  depending on the  test). The mean age of the exposed workers and office controls
22    was 45.7 years and 31.9 years, respectively; no information was provided on the age or sex
23    distribution of the reference controls. The tests included measures of Ig, A, G, M, and E levels,
24    complement (C3 and C4) levels, phagocyte activity, C-reactive protein, a-macroglobulin, T-
25    lymphocytes, and a blast transformation test.  Several differences were observed between the
26    exposed workers and the office workers (e.g., higher levels of serum complement C3 and C4,
27    and of salivary IgA in the exposed), and between the exposed workers and the reference controls
28    (reduced T-lymphocytes, higher phagocytic activity, higher C3 levels in exposed).  However,
29    there were also many differences noted between the office workers and reference group
30    (including reduced T-lymphocytes in office workers). The lack of information about the
31    reference  group adds to the difficulty in interpreting these results.
32
33    4.8.1.1.1.2. Tetrachloroethylene andimmunosuppression.  In 1982, Lagakos et al. (1986)
34    conducted a telephone survey of residents of Woburn, Massachusetts, collecting information on
35    residential history and history of 14 types of medically diagnosed conditions. The survey
36    included 4,978 children born since 1960 who lived in Woburn before age 19. Completed

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-139      DRAFT-DO NOT CITE OR QUOTE

-------
 1    surveys were obtained from approximately 57% of the town residences with listed phone
 2    numbers. Lagakos et al. used information from a study by the Massachusetts Department of
 3    Environmental Quality and Engineering to estimate the contribution of water from the two
 4    contaminated wells to the residence of each participant, based on zones within the town
 5    receiving different mixtures of water from various wells, for the period in which the
 6    contaminated wells were operating. This exposure information was used to estimate a
 7    cumulative exposure based on each child's length of residence in Woburn. A higher cumulative
 8    exposure measure was associated with history of kidney and urinary tract disorders (primarily
 9    kidney or urinary tract infections) and with lung and respiratory disorders (asthma, chronic
10    bronchitis, or pneumonia).  There are no other human data that characterize the effects of
11    tetrachloroethylene-only exposure on immunosuppression, as measured by increased
12    susceptibility to infections.
13
14    4.8.1.1.1.3. Tetrachloroethylene and autoimmune disease. In the 1970s, recognition of a
15    scleroderma-like disease characterized by skin thickening, Raynaud's phenomenon, and
16    acroosteolysis and pulmonary involvement in workers exposed to vinyl chloride (Gama et al.,
17    1978) prompted research pertaining to the role of organic solvents in autoimmune diseases.
18    Exposure to the broad categories of solvents, organic solvents, or  chlorinated solvents has been
19    associated with a 2- to 3-fold increased risk of systemic sclerosis (scleroderma) in epidemiologic
20    studies summarized in a recent meta-analysis (Aryal et al., 2001) and in subsequent studies
21    (Garabrant et al., 2003; Maitre et al., 2004). Similar results were seen in studies of other
22    systemic autoimmune diseases including undifferentiated connective tissue disease (Lacey et al.,
23    1999), rheumatoid arthritis (Lundberg et al., 1994; Sverdrup et al., 2005), and antineutrophil-
24    cytoplasmic antibody (ANCA)-related vasculitis (Lane et al., 2003; Beaudreuil et al., 2005). In
25    contrast, there was little evidence of an association between solvent exposure and systemic lupus
26    erythematosus in two recent case-control studies (Cooper et al., 2004; Finckh et al., 2007).
27          As described in the preceding paragraph, the epidemiologic data in relation to the role of
28    solvents, as a broad category, in systemic autoimmune diseases, varies among these conditions.
29    Much more limited data is available pertaining to specific solvents, including tetrachloroethylene,
30    and risk of autoimmune diseases.  Case reports have been published describing a condition
31    similar to vinyl-chloride induced scleroderma in a man who worked as a presser in a dry cleaning
32    plant, and who also helped clean the tetrachloroethylene-containing drums on a weekly basis
33    (Sparrow, 1977), and a localized scleroderma in a man who had worked with tetrachloroethylene
34    as a metal degreaser, with workplace exposures reported to be between 10-25 ppm (Hinnen et al.,
35    1995; in German). Among 279 cases with connective tissue disease,  Goldman (1996) observed a
36    higher frequency of individuals who reported employment as a dry cleaner among systemic

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-140     DRAFT-DO NOT CITE OR QUOTE

-------
 1    sclerosis patients (4 of 33) compared with patients with other connective tissue diseases (1 of
 2    246; p < 0.01).  Similar patterns were seen with self-reported history of tetrachloroethylene
 3    exposure (3 of 33  systemic sclerosis patients compared with 2 of 246 other patients,/* < 0.01),
 4    but the author noted the difficulty in obtaining this type of information.
 5           One registry-linkage study from Sweden of rheumatoid arthritis (Lundberg et al., 1994)
 6    and three case-control studies of undifferentiated connective tissue disease (Lacey et al., 1999),
 7    scleroderma (Garabrant et al., 2003), and antineutrophil-cytoplasmic antibody (ANCA) related
 8    diseases (Beaudreuil et al., 2005) provide data concerning dry cleaning work or
 9    tetrachloroethylene exposure (Table 4-11). As expected in population-based studies, the
10    exposure prevalence is low, with approximately 4% of controls reporting work in dry cleaning
11    and 1% reporting exposure to tetrachloroethylene.  The observed associations are generally weak
12    (odds ratios for dry cleaning around 1.5 for the 3 large studies of women) and none of the
13    individual studies are statistically significant.  The results seen for the exposure to
14    tetrachloroethylene in the three studies that attempted this kind of assessment were more varied
15    (Lacey et al., 1999; Garabrant et al., 2003; Beaudreuil et al., 2005).  Only the study of ANCA-
16    related diseases resulted in an elevated odds ratio, but again this estimate was somewhat
17    imprecise (OR = 2.0, 95% CI = 0.6, 6.9; Beaudreuil et al., 2005). These studies are clearly
18    limited by the low prevalence of and difficulty in accurately characterizing occupational
19    exposure to tetrachloroethylene in population-based or clinical settings.
20
21    4.8.1.1.1.4.  Tetrachloroethylene and allergy and hypersensitivity. Allergy and hypersensitivity,
22    as assessed with measures of immune system parameters or immune function tests (e.g., asthma,
23    atopy) in humans, have not been extensively studied with respect to the effects of
24    tetrachloroethylene.
25           Th2 cytokines (e.g., interleukin-4) stimulate production of IgE and Thl cytokines (e.g.,
26    interferon-y) act to inhibit IgE production. Lehmann et al. (2001) examined IgE levels and
27    cytokine producing cells (interferon-y, tumor necrosis factor-a, and interleukin-4) in relation to
28    indoor levels of volatile organic compounds among children (age 36 months) selected from a
29    birth cohort study in Leipzip,  Germany. The hypothesis underlying this work is that a shift in
30    Thl to Th2 cytokine profile is a risk factor for IgE-mediated allergic disease in children (Tang et
31    al., 1994; Warner  et al., 1994). Enrollment into the birth cohort occurred between 1995 and
32    1996. The children in this allergy study represent a higher-risk group  for development of allergic
33    disease, with eligibility criteria that were based on low birth weight (between 1,500 and 2,500 g),
34    or cord blood IgE greater than 0.9 kU/1 with double positive family history of atopy.  These
35    eligibility criteria were met by 429 children; 200 of these children participated in the allergy
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-141      DRAFT-DO NOT CITE OR QUOTE

-------
o
Oi
      Table 4-11. Immune-related conditions in studies of dry cleaning or tetrachloroethylene exposure in humans"
oo  S3'

   I
   I
   §
   ***.
   £3'
  1
   TO'
il
   I
   TO
   Co
           Condition, location, diagnosis period,
         sample size, demographics, source of data
                                                                 Results
                                                              Authors
         Autommune diseases
Rheumatoid arthritis, Sweden (13 counties),
hospitalized 1981-1983, 896 male cases,
629 female cases; population comparison
(total 370,035 men, 140,139 women), ages
35-74. Registry linkage to 1960 and 1970
census occupation data
launderers and dry cleaning
  men: 1 exposed cases; OR = 0.8 (95% Cl = 0.1-5.0)
  women: 7 exposed cases; OR = 1.5 (95% Cl = 0.7-3.2)
                                                                                                          Lundberg et al. (1994)
Undifferentiated connective tissue disease,
Michigan and Ohio, diagnosed 1980-1991
(Michigan) 1980-1992 (Ohio). 205 cases,
2095 population controls. Women, ages 18
and older.  Structured interview (specific jobs
and materials; jobs held 3 or more months)
dry cleaning
 cases: 4.3%, controls 3.8% OR = 1.4 (95% Cl = 0.68, 2.8)
tetrachl oroethy 1 ene:
 cases: 0%, controls 1% OR = 0.00
                                                                                                          Lacey etal. (1999)
H I
O >
HH Oq
H TO
O
Scleroderma, Michigan and Ohio. Diagnosed
1980-1991 (Michigan), 1980-1992 (Ohio).
660 cases, 2227 population controls.
Women, ages 18 and older. Structured
interview (specific jobs and materials; jobs
held 3 or more months)
dry cleaning
 cases: 4.7%, controls 3.7% OR = 1.4 (95% Cl = 0.9, 2.2)
tetrachl oroethy 1 ene:
 self report cases: 1.1%, controls  1.0% OR = 1.4 (95% Cl =
    0.6, 3.4)
 expert review cases: 0.8%, controls 0.8% OR = 1.1 (95%
    CI = 0.4, 2.9)
                                                                                                          Garabrant et al. (2003)
ANCA-related diseases,13 France. Diagnosed
1999-2000. 60 patients, 120 hospital
controls, men and women (50% each), mean
age 61 years
tetrachl oroethy 1 ene
 cases:  8.3%, controls 4.1% OR = 2.0 (0.6-6.9)
                                                                                                          Beudreuil et al. (2005)
H
W

-------
       Table 4-11. Immune-related conditions in studies of dry cleaning or tetrachloroethylene exposure in humans*
       (continued)
    Condition, location, diagnosis period,
  sample size, demographics, source of data
                         Results
      Authors
 Allergy and hypersensitivity
 IgE levels, Germany, 1995-1996. 121
 children (ages 36 months), selected based on
 high risk profile for allergic diseases, blood
 sample and indoor air sampling (child's
 bedroom) of 26 volatile organic chemicals (4
 weeks around age 36 months)
no association between tetrachloroethylene measures and
total IgE or IgE-specific allergen antibodies
Lehmannetal. (2001)
 CD3 T-cell subpopulations from cord blood,
 Germany 1997-1999. 85 newborns, cord
 blood and indoor air sampling (child's
 bedroom) of 28 volatile organic chemicals (4
 weeks immediately after birth)
Tetrachloroethylene exposure associated with decreased
interferon-y cells, but no association with interleukin-4,
interleukin-2, or tumor necrosis factor-a
Lehmann et al. (2000)
 Exacerbation of asthmatic symptoms, Los
 Angeles, 1999-2000.  21 children (ages
 10-16 years), 3 month diaries, ambient
 levels and exhaled breath measures of 8
 volatile organic compounds and 8 criteria
 pollutant gases
Little evidence of an association between ambient
tetrachloroethylene exposure or exhaled tetrachloroethylene
measures and asthma symptoms
Delfino et al.
(2003 a, b)
a Includes case-control studies and cross-sectional studies, but does not include case reports.
b ANCA = antineutrophil-cytoplasmic antibody. Diseases included Wegener glomerulonephritis (n = 20), microscopic polyangiitis (n = 8), pauci-immune
 glomerulonephritis (n = 10), uveitis (n = 6), Churg-Strauss syndrome (n = 4), stroke (n = 4) and other diseases (no more than 2 each)

-------
 1    study described below, but complete data (IgE and volatile organic compound measurements)
 2    were available for only 121 of the study participants.
 3          Lehmann et al. measured 26 volatile organic compounds via passive indoor sampling in
 4    the child's bedroom for a period of 4 weeks around the age of 36 months. The highest exposures
 5    were seen for limonene (median 19.1 ug/m3), a-pinene (median 16.3 ug/m3), and toluene
 6    (median 13.3  ug/m3). The median exposure of tetrachloroethylene was 2.5 ug/m3 (0.87 ug/m3
 7    and 5.1 ug/m3 for the 25th and 75th percentiles, respectively).  The only strong correlation
 8    (r > 0.3) between tetrachloroethylene and the other volatile organic compounds measured was a
 9    correlation of 0.72 with trichloroethylene.  Blood samples were taken at the 36-month-study
10    examination and  were used to measure the total IgE and specific IgE antibodies directed to egg
11    white, milk, indoor allergens (house dust mites, cat, molds), and outdoor allergens (timothy -
12    perenial grass, birch- tree). There was no association between tetrachloroethylene exposure and
13    any of the allergens tested in this study, although some of the other volatile organic compounds
14    (e.g., toluene, 4-ethyltoluene) were associated with elevated total IgE levels and with
15    sensitization to milk or eggs.
16          Another study by Lehmann et al. (2002) examined the relationship between indoor
17    exposures to volatile organic compounds and T-cell subpopulations measured in cord blood of
18    newborns. The study authors randomly selected 85 newborns (43 boys and 42 girls) from a
19    larger  cohort study of 997 healthy, full-term babies, recruited between 1997 and 1999 in
20    Germany. Exclusion criteria included a history in the mother of an autoimmune disease or
21    infectious disease during the pregnancy. Twenty-eight volatile organic compounds were
22    measured via  passive indoor sampling in the child's bedroom for a period of 4 weeks after the
23    birth (a period which is likely to reflect the exposures during the prenatal period close to the time
24    of delivery). The levels were generally similar or slightly higher than the levels seen in the
25    previous study using samples from the bedrooms of the 36-month-old children. The highest
26    levels  of exposure were seen for limonene (median 24.3 ug/m3), a-pinene (median 19.3 ug/m3)
27    and toluene (median 18.3  ug/m3), and the median exposure of tetrachloroethylene was 3.4 ug/m3
28    (1.8 ug/m3 and 7.3 ug/m3 for the 25th and 75th percentiles, respectively). Flow cytometry was
29    used to measure the presence of CD3 T-cells obtained from the cord blood labeled with
30    antibodies against interferon-y, tumor necrosis factor-a, interleukin-2 and interleukin-4.
31    Tetrachloroethylene was the only one of the measured volatile organic compounds that was
32    associated with a reduced level of interferon-y. In the univariate analysis, the median percentage
33    of interferon-y cells was 3.6% and 2.6% in the groups that were below the 75th percentile and
34    above  the 75th percentile of tetrachloroethylene exposure, respectively.  The odds ratio between
35    high (above the 75th percentile) tetrachloroethylene exposure  and reduced (less than the 25th
36    percentile) levels of interferon-y cells was 2.9 (95% CI = 1.0-8.6), adjusting for family history of

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-144      DRAFT-DO NOT CITE OR QUOTE

-------
 1    atopy, gender and smoking history of the mother during pregnancy. There was no association
 2    between tetrachloroethylene exposure and interleukin-4 cells, but naphthalene and
 3    methylcyclopentane were associated with elevated levels of interleukin-4 cells.
 4          Delfino et al. (2003a, b) examined the exacerbation of asthmatic symptoms following
 5    exposure to volatile organic compounds that occurred due to variation in air quality over a 3
 6    month period in 1999-2000 in Los Angeles. This study included daily repeated exposures to
 7    ambient air pollutants and peak expiratory flow rates over a 3-month period in 21 children (17
 8    males and 4 females) of Hispanic origin ages 10-16 years; an additional child participated in the
 9    ambient air but not in the exhaled air portion of the study. Daily diaries were used to record
10    severity of symptoms and asthmatic episodes. Exposure metrics included exhaled breath
11    measures and ambient levels of eight volatile organic compounds (benzene, methylene chloride,
12    styrene, toluene, w,/?-xylene, o-xylene, />-dichlorobenzene, and tetrachloroethylene) and eight
13    criteria pollutant gases. An association between criteria air pollutants and subsequent symptoms
14    of asthma in children in the Los Angeles area suggest an increased risk of adverse health
15    outcomes with exposure to SO2 and NO2 (Delfino et al., 2003a). Although ambient levels of
16    tetrachloroethylene were associated with bothersome asthma symptoms (OR = 1.37, [95% CI =
17    1.09, 1.71]) per an interquartile range change), this association was reduced with the adjustment
18    for SO2 or NO2 (Delfino et al., 2003a). In the 21 children who partcipated in  the peak expiratory
19    flow measurements, the mean breath level of tetrachloroethylene was 4.40 ng/L (sd 10.77 ng/L),
20    the mean ambient level was 3.52 (sd 2.17) ng/L, and the correlation between the same-day
21    measures was 0.31 (p < 0.01; Delfino et al., 2003b).  There was little relation  between asthma
22    symptoms and exhaled breath levels of tetrachloroethylene. The mean exhalation of
23    tetrachloroethylene was 2.50 and 2.69 ng/L, respectively, in the two groups of asthma symptoms
24    (none or not bothersome; bothersome and more severe). Stronger associations were reported
25    between asthma symptoms and some of the other volatile organic chemicals,  specifically for
26    benzene, toluene, m,p-xy\Qne.
27          The limited available data from these studies (Lehmann et al., 2001; Lehmann et al.,
28    2002; Delfino et al., 2003a, b), provide weak evidence of an effect of tetrachloroethylene
29    exposure during childhood on allergic sensitization or exacerbation of asthma symptomology.
30    However, the observation of the association between increased tetrachloroethylene exposure and
31    reduced interferon-y in cord blood samples may reflect a sensitive period of development, and
32    points to the current lack of understanding of the potential immunotoxic effects of prenatal
33    exposures.
34
35    4.8.1.1.2. Endocrine system effects.  Only one study of endocrine system effects was found in
36    the published literature. Ferroni et al. (1992) observed  an increased serum concentration of

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-145      DRAFT-DO NOT CITE OR QUOTE

-------
 1    prolactin among tetrachloroethylene-exposed dry cleaners as compared with controls (12.1 + 6.7
 2    Hg/L vs. 7.4 + 3.1 |ig/L).  The median tetrachloroethylene concentration to which these workers
 3    were exposed was 15 ppm. The variance of serum prolactin concentration was wider for
 4    exposed subjects than for unexposed controls, and 4 of 41 subjects had serum prolactin
 5    concentrations above the upper normal limit (defined by Ferroni et al. as 25 |ig/L), compared
 6    with none of 23 controls.  The prevalence of abnormal concentrations, however, was not
 7    statistically significantly elevated. Positive correlations between the response  and
 8    tetrachloroethylene were not observed with either exposure duration or biomarker measures.
 9    The evaluation of prolactin was part of an overall assessment of neurobehavioral functioning
10    (see the discussion of this study in the section on neurobehavioral effects) for which these
11    investigators hypothesized a relationship between dopamine and serum prolactin concentration
12    (Ferroni et al., 1992; Mutti and Smargiassi, 1998).
13           Epidemiologic studies of other parameters of endocrine function are lacking.  The
14    endocrine system can be considered a potential target for tetrachloroethylene because another
15    like solvent, trichloroethylene, has been shown to induce endocrine system changes in both
16    humans (Goh et al., 1998; Chia et al., 1996,  1997) and experimental animals (Kumar et al., 2000).
17
18    4.8.1.2. Cancer
19           The body of literature reporting carcinogenic effects in humans associated with exposure
20    to tetrachloroethylene consists of cohort, proportional mortality, and case-control  studies. A
21    small number of  studies, including studies of cohorts involved in metal degreasing or in  aircraft
22    manufacturing/maintenance (Boice et al., 1999; Anttila et al.;  1995, Spirtas et al.,  1991), have
23    assessed tetrachloroethylene exposure explicitly. These  cohort studies present risks associated
24    with site-specific cancer mortality (Boice et  al., 1999; Spirtas  et al., 1991) or incidence (Anttila
25    et al., 1995) for a subcohort of the larger study population who were exposed to
26    tetrachloroethylene.  Additionally, a few case-control studies were able to examine the
27    relationship between cancer at specific sites  and tetrachloroethylene exposure (Vaughan et al.,
28    1997; Schlehofer et al., 1995; Pesch et al., 2000a; Heineman et al., 1994).
29           A larger body of evidence on cancer  exists for workers employed in dry cleaning. Dry
30    cleaners have potential exposures to a number of solvents, including tetrachloroethylene, which
31    has been in widespread use since the early 1960s (IARC, 1995). Information on the potential for
32    tetrachloroethylene exposure and concentration measurements is lacking for individual study
33    subjects in these  studies:  however, the cohort studies of Ruder et al. (1994, 2001) and Blair et al.
34    (1990, 2003) are  of individuals primarily exposed to tetrachloroethylene (Lynge et al., 1997).
35    The exposure assessment approach of Lynge et al. (2006), a case-control  study nested within a
36    cohort of dry-cleaning and laundry workers, relies on job title to increase sensitivity for

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-146      DRAFT-DO NOT  CITE OR QUOTE

-------
 1    tetrachloroethylene exposure identification, with dry cleaners identified as having greater
 2    potential for tetrachloroethylene exposure.  This study lacks information on tetrachloroethylene
 3    concentration for individual study subjects, is unable to classify job title for 20% of study
 4    subjects, and has a large percentage of cases from Sweden and Norway with job title provided by
 5    next-of-kin.
 6           Although several community-based drinking water studies are available (Aschengrau et
 7    al., 1993, 1998, 2003; Paulu et al., 1999; Fagliano et al., 1990; Cohn et al., 1994; MA DPH,
 8    1997; Vartiainen et al.,  1993; Lagakos et al., 1986), exposure in most of these studies was to a
 9    mixture of solvents, including tetrachloroethylene and trichloroethylene except for the studies by
10    Aschengrau et al. (1993, 1998, 2003)  and Paulu et al. (1999) that examined tetrachloroethylene
11    specifically.  Summary  tables of these analyses are presented in Appendix 4B.
12
13    4.8.1.2.1. Lymphoid cancer.  A number of epidemiologic studies including degreaser cohort
14    (tetrachloroethylene subcohorts), dry  cleaner and laundry worker cohort, case-control, and
15    community studies have examined tetrachloroethylene exposure and lymphoid cancer. Elevated
16    risks of lymphoid cancer incidence, specifically non-Hodgkin's lymphoma (NHL), were
17    observed in studies of degreasers exposed to tetrachloroethylene; a total of 15 cases were
18    observed in three available studies (Table 4B-2, Appendix 4B) versus 6.8 expected cases or
19    deaths (95% CI = 1.2-3.7).  Spirtas et al. (1991) observed a statistically significant elevated risk
20    for NHL for males and  females combined (standardized mortality ratio [SMR] = 4.0, 95% CI =
21    1.1-10.2, four deaths).  Two of the four deaths occurred among females, with females having the
22    highest risk.  The tetrachloroethylene  cohorts of Anttila et al.  (1994) and of Boice et al. (1999)
23    support the findings in Spirtas et al. (1991).  NHL risk is elevated—but not statistically
24    significantly—in both studies (routine exposed, SMR = 1.7, 95% CI = 0.7-3.3, eight deaths,
25    Boice et al., 1999; SIR = 3.8, 95% CI = 0.8-11.0, three cases, Anttila et al., 1994). Boice etal.
26    (1999)  also present an analysis of duration of exposure-response gradient for NHL mortality;
27    however, the inclusion of deaths with either routine or intermittent exposure in this analysis (a
28    total  of 16 deaths), with nonsolvents-exposed factory workers as referents prevents a comparison
29    to the excess NHL risk  reported for the eight deaths with routine exposure to tetrachloroethylene
30    (Table 8 in Boice et al., 1999). NHL  relative risk for subjects with >5 years duration of exposure
31    (intermittent or routine  exposure) to tetrachloroethylene was 1.6 (95% CI = 0.8-3.2) in Poisson
32    regression analysis using internal controls (factory workers not exposed to any solvent) and no
33    indication of a linear trend of increasing RR with increasing duration of exposure (p < 0.20).
34    The inclusion of subjects with differing exposure patterns in an analysis of exposure duration
35    likely introduces misclassification bias and does not diminish observations of routinely exposed
36    subjects.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-147      DRAFT-DO NOT CITE OR QUOTE

-------
 1          The three degreaser cohort studies provide only limited information on lymphoid tumors
 2    other than NHL due to few total numbers of observed deaths or incident cases for lymphoid
 3    neoplasms and, in general, a high proportion of cancers attributable to NHL. Furthermore, none
 4    of these cohort studies provide information on leukemia subtype. For example, Anttila et al.
 5    (1994) observed three cases of lymphoid tumors, and all three were attributable to NHL.
 6    Noteworthy, however, is the observation in one study of aircraft maintenance workers (Spirtas et
 7    al., 1991) of a large but imprecise risk for multiple myeloma and exposure to tetrachloroethylene
 8    (SMR =  17.0,  95% CI = 2.1-61.6) among females, which was not seen in another cohort of
 9    mostly male aircraft manufacturing workers (Boice et al., 1999).
10          Eight studies of incidence in dry cleaners and laundry workers in Scandanavian countries,
11    one study of leukemia incidence in Portland, Oregon,  and two studies of mortality of dry
12    cleaners  in the United States were available for review: Morton and Marjanovic (1984); Lynge
13    and Thygesen (1990); Andersen et al. (1999); Ruder et al. (2001); Cano and Pollan, (2001);
14    Travier et al. (2002); Blair et al. (2003) and Ji and Hemminki, (2005b, 2006). Observations in
15    several of these studies are summarized in  Table 4B-la and Table 4B-lb, Appendix 4B, while
16    others are discussed below. Anderson et al. (1999) examines cancer incidence by occupational
17    title in the  1970 census among citizens of Denmark, Finland, Norway, and Sweden, between
18    1971  and 1987-1991, depending on country. Site-specific lymphoma incidence, but not
19    leukemia subtype, is reported for launderers and dry cleaners. Further analysis of lymphoma
20    incidence in Swedish subjects, many of whom overlap with the larger cohort of Anderson et al.
21    (1999), was presented by Travier  et al.  (2002) who examined leukemia subtype and by Cano and
22    Pollan (2001)  who examined non-Hodgkin's lymphoma incidence.  Ji and Hemminki (2005b,
23    2006) expanded the Swedish  cohort, following launderers and dry cleaners identified on 1960,
24    1970, 1980, and 1990 censuses another 10  years to 2002. Lynge and Thygesen (1990) examine
25    lymphoma incidence among Danish dry cleaners and launderers who were identified with this
26    job title in the 1970 census and followed for 10 years to 1980.
27          Overall, association with Hodgkin's disease, NHL, or chronic lymphocytic leukemia are
28    suggested, although site-specific risks are not always elevated or statistically significant in all
29    studies nor were they elevated in both sexes (see Table 4B-la). Two studies examined
30    Hodgkin's disease (Anderson et al., 1999;  Travier et al., 2002) and both reported statistically
31    significant associations in females but not males: SIR =1.9 (95% CI = 1.1-2.9) in Andersen et
32    al. (1999), RR = 3.6 (95% CI = 1.2-11.1) in Travier et al. (2002). The elevated risk was
33    observed particularly among the subjects who were below 40 years of age in 1960 (Travier et al.,
34    2002). These  are subjects who used mainly tetrachloroethylene with possibly some
35    trichloroethylene.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-148      DRAFT-DO NOT CITE OR QUOTE

-------
 1          NHL risks in the four-country study of Andersen et al. (1999) were 1.5 (95% CI =
 2    1.0-2.1) for males and 1.0 (95% CI = 0.7-1.2) for females. Separate analyses of Swedish and
 3    Danish workers supported these observations: males, RR =1.4 (95% CI = 0.6-3.4); females,
 4    RR = 0.5 (95% CI = 0.2-1.6; Travier et al., 2002); males, RR = 1.9 (95% CI = 0.8-4.1; Cano and
 5    Pollan , 2001); males, SIR = 2.8 (95% CI = 0.9-6.5); females, SIR = 0.5 (95% CI = 0.1-1.5;
 6    Lynge and Thygesen,  1990).
 7          Three studies examine association with leukemia subtype (Ji and Hemminki, 2005b,
 8    2006; Travier et al., 2002; Morton and Marjanovic, 1984) and all report statistically significant
 9    association with chronic lymphocytic leukemia in females but not males. A noteworthy finding
10    in Travier et al. (2002) was a statistically significant elevated risk for leukemia among dry
11    cleaners or launderers in both the 1960 and 1970 Swedish census (RR =  1.8, 95% CI = 1.1-3.1),
12    mostly due to chronic lymphocytic leukemia (CLL); 6 of the 15 total leukemia cases, RR =1.9,
13    95% CI = 0.8-4.1; of which 5 of the 6 CLL cases were in females, RR = 2.9,  95% CI = 1.2-7.0.
14    Ji and Hemminki (2005b) also examined associations with leukemia subtypes and reported a
15    similar finding with chronic lymphocytic leukemia in females but not males (females: SIR =1.5,
16    95% CI= 1.1-2.1; males:  SIR = 0.9, 95% CI =  0.5-1.3).  Similarly Morton and Marjanovic
17    (1984) reported leukemia incidence, particularly lymphocytic leukemia incidence, was
18    statistically significantly higher in women laundry and dry cleaners in the Portland, Oregon area.
19    Age-standardized incidence rates (per 100,000) for women dry cleaners and laundry workers
20    compared to all women were: all leukemias, 23.7 compared to 6.7; lymphatic leukemia, 20.9
21    compared to 2.6. Many chronic lymphocytic leukemias and NHLs may arise from a common
22    cell type (Beers and Berkow, 1999, in Bukowski et al., 2003).
23          The update of the cohort mortality study by Ruder et al.  (2001) found only six deaths
24    attributable to cancer of the lymphatic and hematopoietic system.  The few deaths due to
25    lymphatic cancer greatly impact the statistical power in this study.  Blair et al. (2003), in an
26    updated mortality analysis of a cohort that is predominately female, present observed and
27    expected number of deaths for categories of lymphomas. Although based on a small number of
28    deaths in several categories, these  authors discuss this study as supporting an excess risk of
29    Hodgkin's disease (SMR = 2.0, 95% CI = 0.6-4.6, 5 cases) and this observation is consistent
30    with observations in the  Scandinavian studies discussed above.  Neither of the mortality studies
31    of United States dry cleaners examined leukemia subtype.
32          Ten case-control studies  examine site-specific lymphomas and occupational exposure to
33    tetrachloroethylene or job title of dry cleaner or launderer (Lynge et al., 2006; Mester et al.,
34    2006; Miligi et al, 2006; Kato et al., 2005; Fabbro-Peray et al., 2001; Seniori  Costantini et al.,
35    2001; Clevel et al. 1998; Blair et al., 1993; Scherr et al., 1992; and Hardell et al.,  1981; Table
36    4B-5, Appendix 4B).  Several studies examine cell type (Mester et al., 2006, Miligi et al., 2006,

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                 4-149     DRAFT-DO NOT CITE OR QUOTE

-------
 1    Clevel et al., 1998). The only study available on Hodgkin's disease observed a statistically
 2    significant elevated risk for male with a job title of dry cleaner or laundry worker (Seniori
 3    Costantini et al., 2001). Risks above 1.0 were observed in several studies of NHL and CLL
 4    although risks were not statistically significantly elevated, likely due to several factors discussed
 5    below (Mester et al., 2006 Miligi et al., 1999; 2006; Kato et al., 2005; Clevel et al., 1998; Blair et
 6    al., 1993; Scherr et al., 1992; Hardell et al., 1981).
 7          Overall, case-control  studies examining occupational exposure are quite limited for
 8    evaluating lymphoma and tetrachloroethylene for a number of reasons.  Reviewed case-control
 9    studies are typically population-based studies. More recent studies are of large number of cases
10    and controls compared to older studies.  These recent studies adopt procedures to blind
11    interviewers and apply more  refined exposure assessment methods, examining
12    tetrachloroethylene specifically as opposed to a grouping of dry cleaners and laundry workers
13    (Lynge et al.,  2006; Mester et al., 2006; and Miligi et al., 2006). However, the prevalence of
14    exposure to tetrachloroethylene or as a dry cleaner, or launderer, particularly long-duration
15    exposure, exposure, is low in these studies. A consequence of this is few exposed cases and
16    imprecise risks (wide confidence intervals) that reflect lower statistical power to examine
17    lymphoma and tetrachloroethylene exposure (Miligi et al., 2006; Mester et al., 2006; Seniori
18    Costantini et al., 2005). Four other aspects of population case-control studies are important to
19    consider in their interpretation. Risk is difficult to determine for low-prevalence jobs when
20    studying a regional population; associations may be nonlinear because categorical assignment of
21    duration and intensity are arbitrary and do not necessarily represent a linear dose relationship,
22    and duration and cumulative  exposure variables do not address age at first exposure, which also
23    affects cancer risk (NRC, 2005). Additionally, missing information either as a result of lower
24    participation rates or missing job history data, can introduce bias—particularly misclassification
25    bias.  For example, the lack of information to classify job title for roughly 20% of NHL cases
26    and controls in Lynge et al. (2006), in addition to a large number of next-of-kin interviews, limits
27    this study due to an increased potential for exposure misclassification and, as assessed using two
28    exposure groups (yes/no) in this study, results in risks close to  1.0 (no risk). Lynge et al. (2005)
29    is considered  a null or uninformative study for this reason rather than a study supporting no
30    association between tetrachloroethylene and NHL. Furthermore, quantitative exposure
31    information is missing from all studies and leads to substantial misclassification of exposure.
32          Four case-control  studies are available on childhood leukemia (acute lymphocytic
33    leukemia, ALL) and parental occupational exposure to tetrachloroethylene or to drinking water
34    contaminated with trichloroethylene, tetrachloroethylene, and other chlorinated solvents (Infante-
35    Rivard et al., 2005; Costas et al., 2002; Shu et al., 1999; Lowengart et al. 1987; Table 4B-5,
36    Appendix 4B). Many aspects discussed above for case-control studies examining occupational

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-150      DRAFT-DO NOT CITE OR QUOTE

-------
 1    exposure are found in ALL studies. While some studies appear consistent (Costas et al.,2002;
 2    Shu et al., 1999; Lowengart et al., 1987), these studies are insensitive for assessing association,
 3    or lack thereof, between ALL and tetrachloroethylene exposure because observations are based
 4    on few exposed cases.  Other studies are needed to clarify the role of tetrachloroethylene and
 5    ALL.
 6          Four studies examine drinking water exposure and lymphoma: a case-control study by
 7    Aschengrau et al. (1993), and the ecological analyses by Cohn et al. (1994); Fagliano et al.,
 8    1990; and Vartiainen, 1985 (see Table 4B-13 and Appendix 4B). In a study by Aschengrau et al.
 9    (1993), where tetrachloroethylene was identified as the putative exposure, an elevated risk of
10    leukemia was observed for those most exposed (90th percentile of exposure; with no latency,
11    OR =8.3; 95%CI= 1.5-45.3; considering a latency period of 5 years, OR = 5.9; 95% CI =
12    1.4-24.9).  An exposure-response relationship is suggested; the crude unadjusted risk for high-
13    level exposure (exposure at the 90th percentile; OR = 6.0, 95% CI = 0.6-32.0) is larger than the
14    unadjusted risk for any exposure (OR =1.8, 95% CI = 0.6-4.3).
15          Moreover, the case-control study of Costas et al. (2002) and the ecologic studies by
16    Fagliano et al. (1990) and Cohn et al. (1994) provide some evidence of an association between
17    NHL or  leukemia and drinking water that includes trichloroethylene and tetrachloroethylene.
18    The actual level of exposure to tetrachloroethylene and other solvents in these studies is not
19    known, and in the case of Costas et al. (2002), trichloroethylene was measured in the well water
20    at concentrations an order of magnitude higher than tetrachloroethylene. Each of these solvents
21    is hypothesized to be metabolized and bioactivated to TCA (see metabolism discussion). Thus,
22    exposure to tetrachloroethylene can be considered to add to the level of these metabolites
23    generated through trichloroethylene exposure.
24          The  classification of lymphoid neoplasms, specifically lymphomas, has recently
25    undergone a revision, primarily on the basis of new findings from molecular biology, genetics,
26    and immunology, which have changed older concepts of lymphoid cancer, making them obsolete
27    (Herrinton,  1998). Although lymphomas have been classified in the past into distinct categories
28    (e.g., leukemia, reticulosarcoma/lymphosarcoma), lymphomas can  share common  biological
29    properties (Weisenburger, 1992) and differentiation pathways. For example, advances in
30    molecular biology have blurred the distinction between lymphoid leukemia and lymphoma
31    (Herrinton,  1998). Few studies assessing tetrachloroethylene exposure have included analyses of
32    diagnostic subcategories and no studies  have examined cellular or molecular markers.
33    In 1994, the International Lymphoma Study Group published the revised European-American
34    Lymphoma (REAL) classification, and this system has been adopted by WHO (Harris et al.,
35    2000a).  Modifications in the REAL include the grouping of lymphatic leukemia with NHL
36    (Herrinton,  1998; Miligi et al.,  1999); the REAL/WHO Classification considers lymphomas and

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-151     DRAFT-DO NOT CITE OR QUOTE

-------
 1    lymphoid leukemias of the same cell type as one disease, with different clinical presentation or
 2    stages (Harris et al., 2000b). This implies that the classification system used in many of the
 3    epidemiologic studies is imprecise for both diseases. The resulting bias related to disease
 4    misclassification affects observed risk estimates by masking underlying associations, biasing
 5    risks towards the null or RR close to 1.0. Hence,  the elevated risks observed for different
 6    categories of lymphoid tumors in individual studies are noteworthy because of study
 7    insensitivities and, additionally, these risks may not be inconsistent with the pathogenesis of
 8    disease and with an etiology associated with tetrachloroethylene.
 9          Rats exposed chronically to tetrachloroethylene for 2 years developed increased
10    incidences of mononuclear cell leukemia (MCL) or large granular lymphocyte leukemia (JISA,
11    1993; NTP, 1986a).  Section 4.8.2.2.1 describes these observations and their interpretation as a
12    human cancer hazard. Large granular lymphocyte (LGL) cells exist in humans that are
13    morphologically, biochemically, and functionally similar to the cells involved in MCL in the
14    F344 rat (Stromberg, 1985). In humans, clonal disorders of LGLs represent a biologically
15    heterogeneous spectrum of lymphoid malignancies thought as originating either from mature
16    T-cell or natural killer (NK) cells (Sokol and Loughran, 2006). LGL disorders can clinically
17    present as an indolent (chronic) or aggressive diseases (Sokol and Loughran, 2006). The
18    indolent form of LGL leukemia is a disease of the elderly, with a median age at diagnosis of 60
19    years.  A number of clinical conditions have been seen in patients with LGL leukemia.  These
20    include the following: red cell aplasia and aplastic anemia; other lymphoproliferative disorders
21    such as NHL, Hodgkin's lymphoma, multiple myeloma, hairy cell leukemia, and B-cell
22    lymphoprliferative disorders; and  autoimmune diseases such as rheumatoid arthritis and systemic
23    lupus erythematousus (Rose and Berliner, 2004).  The etiology of LGL disorders is not known
24    (Rose and Berliner, 2004; Sokol and Loughran, 2006). Several possible etiologies have been
25    proposed including chronic activation of T-cell by a viral antigen or autoantigen in which case
26    LGL leukemia could be considered as an autoimmune disorder (Sokol  and Loughran, 2006).
27          Existing epidemiologic studies of tetrachloroethylene exposure are simply not able to
28    inform human relevance examinations of rat MCL.  Lymphoid tumor pathobiology in rats and
29    humans, its historical and current classification, and epidemiology, including observations in
30    tetrachloroehylene-exposed populations, have bearing on examination  of the human relevance of
31    rat mononuclear cell leukemia.  Important to any  examination are the changes in diagnostic and
32    classification criteria of human lymphoid tumors  and lack of data on molecular markers in the
33    tetrachloroethylene epidemiologic studies, as discussed above.  Diagnostic and classification
34    criteria may not be uniform across studies and hinders comparison of consistency  within
35    epidemiologic studies of lymphoid cancers and tetrachloroethylene exposure and,  also, between
36    human and rat lymphoid tumor observations. Furthermore, adoption of consensus nomenclatures

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-152      DRAFT-DO NOT CITE OR QUOTE

-------
 1    of human lymphoid tumors, i.e., the WHO scheme, for rats will facilitate cross-species
 2    comparisons, as was recently conducted by the hematopathology subcommittee of the Mouse
 3    Models for Human Cancers Consortium (Morse et al., 2002).
 4
 5    4.8.1.2.2. Esophageal cancer. Both cohort and case-control studies support an association
 6    between tetrachloroethylene and excess risk of esophageal cancer. An overall excess in the
 7    number of observed deaths was seen in the recent updates of the dry cleaner mortality studies
 8    (Ruder et al., 2001; Blair et al., 2003):  a total of 31 deaths as compared with 13.7 expected
 9    deaths (RR = 2.3, 95% CI =  1.5-3.2; Table 4B-lb, Appendix 4B). This finding is the same as
10    that of Wartenberg et al. (2000), who examined a  slightly different set of studies for esophageal
11    cancer. The recent PMR study by Walker et al. (1977) provides further support (aged <65  years
12    at time of death, PMR= 1.7, 95% CI= 1.1-2.5). Both Ruder et al. (2001) and Blair et al. (2003)
13    reported a similar magnitude of risk among workers employed before 1960 and after 1960. No
14    clear picture of increasing risk was seen in either study between duration of exposure and
15    esophageal cancer risk. Except for Boice  et al. (1999), studies of the degreasers do not present
16    risks for esophageal cancer (Table 4B-2, Appendix 4B).  Boice reported a risk of 1.5 (95% CI =
17    0.5-3.2), based on six deaths.
18          The cohort and PMR studies cannot directly address possible effects due to smoking or
19    alcohol, which are risk factors for the squamous cell histologic type of esophageal cancer.  It is
20    not known whether elevated risk may reflect smoking and alcohol effects. Data from the
21    National Health Interview Survey (Nelson et al., 1994) suggest that the prevalence of smoking
22    among dry cleaners and laundry operators as equal to that of "blue collar" workers.  Moreover,
23    Ruder et al. (2001) and Blair et al. (2003)  note that the magnitude of the risks for several
24    smoking-related cancers was greater than  could be explained by smoking alone, suggesting a
25    further contribution from another risk factor, such as occupational exposure.
26          The incidence of esophageal cancer is generally higher for nonCaucasan males than for
27    Caucasan males (Blot and McLaughlin, 1999;  Brown et al.,  2001).  In contrast, Ruder et al.
28    (2001) observed similar SMRs for esophageal  cancer across all race-sex groupings
29    (supplementary table at http://www.cdc.gov/niosh/dc-mort.html), providing further support for
30    occupational exposure as a risk factor.  For these reasons, the observations in Ruder et al. (2001)
31    and Blair et al. (2003) together suggest that the excess esophageal cancer risks seen in the dry
32    cleaner studies cannot be entirely due to smoking, alcohol, or some other factor that may be
33    as soci ated with race.
34          Additionally, in the case-control study  by Vaughan et al. (1997), the OR for cumulative
35    (ppm-yr) tetrachloroethylene exposure and squamous cell esophageal cancer, which were
36    adjusted for effects of both smoking and alcohol consumption, although imprecise (large

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-153     DRAFT-DO NOT CITE OR QUOTE

-------
 1    confidence intervals), were significantly elevated (Table 4B-6, Appendix 4B).  In fact, the
 2    magnitude of risk for tetrachloroethylene exposure after adjustment in the statistical analysis for
 3    smoking and alcohol consumption was larger for the effect of tetrachloroethylene exposure when
 4    compared with the crude or unadjusted OR, suggesting that the association between occupational
 5    exposure and esophageal cancer may be underestimated in those studies that could not control
 6    for these factors.
 7          Lacking information to classify job title for 25% and 19% of cases and controls,
 8    respectively, Lynge et al. (2006) provides little weight for informing an examination of the
 9    presence or absence of association between tetrachloroethylene and esophageal cancer.
10
11    4.8.1.2.3. Cervical cancer.  There is some evidence for an excess in risk for cervical cancer
12    mortality.  The total number of observed and expected numbers of cancers in the
13    tetrachloroethylene cohort mortality studies was 31 observed deaths versus 19 expected (RR =
14    1.6, 95% CI = 1.1-2.3; Table 4B-lb, Appendix 4B). An association with dry cleaning is
15    supported by an exposure-response trend in Ruder et al. (2001), although RRs for cervical cancer
16    mortality in Blair et al. (2003) did not appear to differ between subjects with medium/high
17    exposure and those with little or no exposure to tetrachloroethylene. The number of female
18    subjects in studies of workers exposed to tetrachloroethylene as a degreasing agent is few, with a
19    consequence of limited  statistical power (Table 4B-2, Appendix 4B). Data availability on
20    socioeconomic and lifestyle factors in the dry cleaner studies precludes an evaluation of these
21    factors.
22
23    4.8.1.2.4. Suggestive evidence of cancer at other sites. More limited are the findings of excess
24    risks from cancers of the bladder,  lung, pancreas, and small bowel.
25
26    4.8.1.2.4.1. Bladder cancer. The recent updates of two cohort mortality studies of dry cleaners
27    with tetrachloroethylene as the primary exposure (Blair et al., 2003; Ruder et al., 2001) provide
28    some evidence for an excess risk for bladder cancer mortality (Table 4B-lb, Appendix 4B).
29    Ruder et al. (2001) observed statistically significant differences in bladder cancer risk among the
30    entire cohort (SMR = 2.2, 95% CI = 1.1-4.1, 10 observed deaths). No deaths were observed
31    among subjects employed after 1960,  a date corresponding with greater usage of
32    tetrachloroethylene.  The magnitude of bladder cancer risk in Blair et al. (2003), on the other
33    hand, was similar regardless of level of exposure (little or no exposure vs. medium/high
34    exposure). The Nordic  incidence  studies are consistent with the mortality studies (Table 4B-la,
35    Appendix 4B). The study of Lynge et al. (2006), who examine bladder cancer incidence and
36    tetrachloroethylene exposure using a nested case-control approach in a cohort of Nordic dry

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-154       DRAFT-DO NOT CITE OR QUOTE

-------
 1    cleaners and laundry workers, contributes little weight to causal evidence. The limitations of this
 2    study preclude drawing conclusions on whether this study is suggestive of, or lack of, an
 3    association between tetrachloroethylene and site-specific cancers such as bladder cancer.  This
 4    study is uninformative due to a high percentage of study subjects whose job title could not be
 5    classified (for example, 16% of bladder cancer cases and controls) and potential bias resulting
 6    from the large number of next-of-kin interviews.
 7           Of the studies of workers exposed to tetrachloroethylene as a degreasing agent, only the
 8    study by Boice et al. (1999) reports data for bladder cancer, based on two deaths (Table 4B-2,
 9    Appendix 4B). Several case-control studies of bladder cancer (Table 4B-7, Appendix 4B) have
10    examined a job title as a dry cleaner or laundry worker and present risks adjusted for a number of
11    factors, including cigarette smoking, a known risk factor for this cancer. These studies also
12    provide weak support for an association: RRs ranged from  1.3 to 2.8, although increased risks
13    were generally not statistically significant.
14           Two population case-control studies examined tetrachloroethylene exposures specifically,
15    Pesch et al. (2000b) and Aschengrau et al. (1993). Pesch et al. (2000b) examined occupational
16    exposure to tetrachloroethylene using a job exposure and job task exposure matrix. Urothelial
17    cancer cases (a category that includes cancer of the urinary bladder, ureter, and renal pelvis)
18    were histologically confirmed. A statistically significant excess risk (OR) was reported in both
19    exposure assessment methods for males with substantial exposure to tetrachloroethylene in
20    analyses that adjusted for age, study center, and smoking.
21           Aschengrau et al. (1993) reported an adjusted OR of 4.9 (95% CI = 0.7-25.1) with high
22    exposure to tetrachloroethylene in drinking water (Table 4B-13, Appendix 4B).  Strengths of this
23    study are the use of exposure modeling to reconstruct tetrachloroethylene delivery to a home and
24    adjustment of ORs for sex, age at diagnosis, vital status, educational level, and smoking.
25
26    4.8.1.2.4.2. Lung cancer. Lung cancer risk was elevated in the mortality studies by Blair et al.
27    (2003) and Ruder et al. (2001), where the total number of observed deaths was 144,  with  106.6
28    deaths expected (RR = 1.4, 95% CI = 1.1-1.6; Table 4B-lb, Appendix 4B); in the incidence
29    study by Lynge and Thygesen (1990; RR = 1.2, 95% CI = 0.9-1.6) 60 cases among males and
30    females, 49.4 expected cases (Table 4B-la, Appendix 4B); and in the degreaser studies, a total of
31    51 cases observed with 43 expected; RR = 1.2, 95% CI = 0.9-1.6; Table 4B-2, Appendix 4B).
32    The possible effect of smoking cannot be examined in the cohort studies.
33           Two case-control studies (Pohlabeln et al. 2000; Brownson et al., 1993) examined the
34    association between occupational risk factors and lung cancer among nonsmokers (Table 4B-8,
35    Appendix 4B). Both studies reported an association between lung cancer (among nonsmokers)
36    and dry cleaning work. Additionally, the case-control study by Paulu et al. (1999) provides

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-155      DRAFT-DO NOT CITE OR QUOTE

-------
 1    further support for an association between tetrachloroethylene exposure and lung cancer (Table
 2    4B-13, Appendix 4B). This study examined oral (drinking water) exposure to
 3    tetrachloroethylene, and the statistical analysis adjusted for both active and passive smoking.
 4
 5    4.8.2. Animal Studies
 6          In addition to the toxic effects in animals already mentioned (in liver, kidney, nervous
 7    system, and developmental/reproductive system), effects have also been reported on cardiac
 8    function, immunosuppression, and cancer at other sites.
 9
10    4.8.2.1. Noncancer Effects
11    4.8.2.1.1  Whole animal toxicity. Hayes et al. (1986) administered tetrachloroethylene to SD
12    rats in drinking water at doses of 0, 14, 400, and 1,400 mg/kg-day for 90 days and  observed the
13    body weights of all animals weekly throughout the experiment;  and the liver, kidney, and brain
14    weights of all animals at necropsy. They also performed measurements of hematological and
15    serum chemistry parameters on 10 animals per group at the end of the period.  They found
16    significant (p < 0.05)  decrements in body weight gain in males at 1,400 mg/kg-day and in
17    females at 400 and 1,400 mg/kg-day. No effects that could be attributed to administered dosing
18    were observed in hematology, serum chemistry,  urinalysis, mortality, or organ weights.  The
19    body-weight-gain decrements in this  experiment signify a LOAEL of 400 mg/kg-day and a
20    NOAEL of 14 mg/kg-day.
21
22    4.8.2.1.2. Cardiac toxicity. Kobayashi et al. (1982) treated animals using intravenous injections
23    of tetrachloroethylene. In the animals examined (rabbits, cats, and dogs), tetrachloroethylene
24    enhanced the vulnerability of the ventricles to epinepherine-induced arrhythmias.  The threshold
25    doses were 10,  24, and 13 mg/kg in rabbits, cats and dogs, respectively.
26          Cardiac effects of some tetrachloroethylene metabolites  have been examined in animals.
27    As mentioned in Section 4.6.4, Smith et al. (1989) and Johnson  et al. (1998) observed cardiac
28    anomalies in rat fetuses after exposure of pregnant rats to TCA.  Epstein  et al. (1992) saw cardiac
29    defects in rat fetuses after exposure to DC A. This work indicated a developmental LOAEL of
30    1,900 mg/kg-day DCA.  DCA has also been shown to concentrate in rat myocardial
31    mitochondria (Kerbey et al., 1976). More research into cardiac  toxicity resulting from exposures
32    to tetrachloroethylene and its metabolites is needed to fully characterize possible adverse cardiac
33    effects.
34
35    4.8.2.1.3. Immunotoxicity. The animal evidence for immunotoxicity following exposure to
36    tetrachloroethylene is also very limited. These studies consist of a mixed solvent exposures and
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-156     DRAFT-DO  NOT CITE OR QUOTE

-------
 1    some inhalation and oral studies in which experimental animals were dosed with
 2    tetrachloroethylene alone.
 3          Immune systems parameters were altered in a mouse study (female B6C3F1)
 4    administered tetracholroethylene (maximum concentration 6.8 ppm) along with a mixture 24
 5    contaminants frequently found in ground water near superfund sites. Exposure lasted 14 or 90
 6    days and mice were sacrificed to assess immune system parameters. Evidence of
 7    immunosuppression was seen, with a dose related decrease in antibody response to sheep red
 8    blood cells and decreased host resistance to following challenge to Plasmodium yoelli.  There
 9    were no changes in lymphocyte number, T-cell subpopulations, NK cell activity, or in challenge
10    listeria monocytgens or PYB6 tumor cells. While these findings may be attributed to
11    B-cell/humoral immunity these  effects cannot be attributed to tetrachloroethylene alone
12    (Germolecetal., 1989).
13          In another inhalation study, mice were given a single exposure of 170 mg/m3
14    tetrachloroethylene (50 ppm) for 3 hrs and then challenged with Klebsiella pneumoniae; an
15    increase in streptococcal pneumonia was observed (Aranyi et al., 1986). Interpretation of the
16    significance of these findings is confounded with a high degree of mortality in the control group.
17          In a study by Hanioka et al. (1995), atrophy of the spleen and thymus was observed in
18    rats receiving 2,000 mg/kg/d tetrachloroethylene via corn oil gavage for 5 days.  No effect was
19    seen in the 1,000 mg/kg/d group. In a 14-day corn oil gavage (1,000 mg/kg/d) study of
20    tetrachloroethylene, no effects were observed on thymus and spleen weights of adult rats at dose
21    that produced liver toxicity (Berman et al., 1995). Another study employed 3 daily ip doses of
22    tetrachloroethylene to mice (Schlichting et al., 1992).  No effects were observed on ex vivo
23    natural killer cell activity or humoral responses of T-cells to exogenous mitogens.
24          A series  of experiments  in the lupus-prone MRL +/+ mice examined the effect of
25    trichloroethylene on the expression of features of lupus (Griffin et al., 2000a; Gilbert et al., 2004).
26    Activation and expansion of CD4+ T-cells has been demonstrated, through a mechanism that
27    appears to be mediated through the CYP2E1 metabolism of trichloroethylene and inhibition of
28    FasL expression on the surface of the CD4+ T-cells (Griffin et al., 2000b; Blossom et al., 2004;
29    Blossom and Gilbert, 2006).  Trichloroethylene exposure via drinking water was also shown to
30    induce an autoimmune hepatitis, characterized by mononuclear infiltration around the portal vein,
31    in the MRL +/+ mice (Griffin et al., 2000c). This evidence of immunological alterations
32    following trichloroethylene exposure provides suggestive evidence  for perchloroethylene, a
33    halogenated solvent which shares some common metabolites with trichloroethylene.  To date,
34    similar studies have not been conducted with other solvents, so the extent to which these findings
35    pertain to tetrachloroethylene is not known.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-157      DRAFT-DO NOT CITE OR QUOTE

-------
 1          Additional data from inhalation, oral, and dermal exposures of different durations are
 2    needed to assess the potential immunotoxicity of tetrachloroethylene along multiple dimensions,
 3    including immunosuppression, autoimmunity, and allergic sensitization. This lack of data taken
 4    together with the concern that other structurally related solvents have been associated with
 5    immunotoxicity contributes to uncertainty in the database for tetrachloroethylene.
 6
 7    4.8.2.2.  Cancer Effects
 8    4.8.2.2.1. Mononuclear cell leukemia in rats. NTP (1986a) reported that the chronic inhalation
 9    administration of tetrachloroethylene to male and female F344/N rats at concentration levels of 0,
10    200, and 400 ppm caused positive trends in the incidence of MCL in both sexes. The incidence
1 1    data are  shown in Table 5-8 (Chapter 5). For males, there was a statistically significant trend (p
12    = 0.004), and for females the trend was marginally significant (p = 0.053). Pairwise comparisons
13    of tumor incidences in dosed and control groups of males (life table analysis) disclosed
14    statistically significant increases in both the low- and high-dose groups.  Analysis  of the data for
15    female rats revealed a significant increase in the low-dose group and a marginally  significant
16    increase in the high-dose group.  Interpretation of these data is somewhat clouded  by the fact that
17    overall incidences of MCL in the concurrent chamber control groups were high relative to
18    historical chamber control groups at the performing laboratory (males 28/50 [56%] vs. 1 17/250
19    [47%]; females 18/50 [36%] vs. 73/249 [29%]). The concurrent control group rates were also
20    higher than the NTP program historical rate for untreated control groups (males 583/1,977
21    [29%]; females 375/2,021 [18%]).
22          Because of these factors, NTP conducted supplemental analyses of the progression of the
23    disease,  the effect of tetrachloroethylene on the time of onset  of advanced MCL, and the
24    contribution of MCL to early deaths in control and dosed animals.  The results of these
25    supplemental analyses showed the following:
28       •  In both males and females, tetrachloroethylene produced a dose-related increase in the
29          severity of MCL.
30
31       •  Tetrachloroethylene exposure significantly shortened the time to onset of MCL in female
32          rats.
33
34       •  Although there was no remarkable effect of tetrachloroethylene exposure on survival of
35          female rats, there was an increased incidence of advanced MCL in female rats that died
36          before the scheduled termination of the study. Thus, a more appropriate statistical
37          analysis was conducted in which only the incidences of advanced MCL in rats were
38          considered.  Significantly positive trends and significant increases in the incidences of
39          advanced MCL were observed in both male and female rats in the high-dose groups.
40
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-158      DRAFT-DO NOT CITE OR QUOTE

-------
 1          In 1987, the U.S. EPA's Science Advisory Board took exception to the use of these
 2    special analyses because they did not represent generally accepted approaches to evaluating
 3    increased incidences of MCL.  According to the NTP report, however, the interpretation of MCL
 4    incidences in the tetrachloroethylene study was based on standard methods of data evaluation
 5    (NTP, 1986a).  The special analyses were conducted to support rather than to establish the
 6    interpretation.
 7          The Japan bioassay (JISA, 1993) in F344/DuCrj rats exposed for 104 weeks at
 8    concentrations of 0, 50, 200, and 600 ppm also reported clearly significant trends in MCL in
 9    males. In females, MCL showed a marginally significant trend with dose. These data are also
10    shown in Table 5-8 (Chapter 5). In this assay, the control incidences for both males and females
11    were also higher than those for historical controls (for males, study control incidence was 11/50,
12    [22%], whereas for the historical controls it was 147/1,149 [13%]). This higher incidence in
13    concurrent controls versus historical controls also occurred in the NTP assay. The historical
14    control incidence data for the Japan rat studies are shown in Table 5-10 (Chapter 5). The Japan
15    bioassay report did not include an analysis of the tumor latency.
16
17    4.8.2.2.1.1. Discussion of issues associated with rat mononuclear cell leukemia (MCL).
18    Under the conditions of the bioassay s, a carcinogenic effect of tetrachloroethylene in male and
19    female rats was evidenced by significant increases of MCL in both sexes.  However, the
20    reliability of MCL in the rat in predicting  human carcinogenic risk associated with
21    tetrachloroethylene exposure has been questioned for several reasons, such as high spontaneous
22    background incidences, use of special supplemental analyses to aid in data interpretation, and the
23    relevance of MCL in F344/N rats to human hazard.  Some of the issues have been reviewed by
24    Caldwell (1999) and others.
25
26    4.8.2.2.1.2. Background.  Lymphohematopoietic neoplasms, leukemias, and lymphomas
27    represent uncontrolled proliferation or clonal expansion of bone marrow or lymphoid cells that
28    can no longer differentiate to mature blood cells (U.S. EPA,  1997a; Sawyers et al., 1991; Nowell,
29    1991). Like other cancers, they are thought to develop via a multi-step process in the
30    transformation of a normal cell to  a malignant cell.  Although rodent models—especially mouse
31    models—are considered to be highly relevant for understanding most aspects of hematopoiesis
32    (Bagby,  1994), several differences exist between humans and rodents. Differences in cell
33    composition, number, and proliferation rates  and organization of the stem-cell compartment
34    likely influence the hematotoxic and carcinogenic effects observed in human and rodent systems
35    following exposure to carcinogenic chemicals (U.S. EPA, 1997b).
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-159      DRAFT-DO NOT CITE OR QUOTE

-------
 1          In adult humans, hematopoiesis occurs in the bone medullary spaces, with extramedullary
 2    hematopoiesis occurring in the spleen, liver, and lymph nodes only under stress conditions. In
 3    rodents, however, hematopoietic cells are commonly found in the spleen and, to some extent, the
 4    liver. The primary type of lymphohematopoietic cancer induced by chemicals in humans is
 5    myeloid leukemia; however, the induction of lymphoma has been associated with
 6    immunosuppressive agents.  In contrast, lymphohematopoietic tumors in rats and mice originate
 7    primarily in lymphoid tissue.
 8
 9    4.8.2.2.1.3.  Issues. The usefulness of increased incidences of MCL in predicting human
10    carcinogenic risk associated  with exposure to tetrachloroethylene has been questioned on several
11    grounds, and these issues are discussed below.
12          MCL is a recognized as a common, spontaneously occurring neoplasm in F344 rats, and
13    its rate of appearance in historical control groups is highly variable. High-incidence MCL occurs
14    only in the F344 rat strain and not in mice.  For this reason, Caldwell (1999), for example, has
15    stated that marginal increases in incidences are of questionable biological significance.  High and
16    variable control incidence is also an issue with liver tumors in mice, and it has always been a
17    source of uncertainty in risk  assessments.
18          Although the occurrence in a single strain indicates a genetic susceptibility of these rats
19    to spontaneous MCL, the occurrence in humans of a similar genetic susceptibility is by no means
20    ruled out. Humans are more genetically heterogeneous than are inbred rat strains, and some
21    individuals could potentially possess the  same inherited susceptibility that is exhibited in F344
22    rats.
23          Another issue is the pathobiology of MCL. Some scientists believe it is too poorly
24    understood to allow the tumors to be used to determine human health risk. However, MCL is a
25    relatively well defined  and well understood rodent neoplasm that is characterized by infiltration
26    of pleomorphic blastlike mononuclear cells in numerous organs. The disease per se, which is
27    splenic in origin but later infiltrates the liver, lung, bone marrow, lymph nodes, and other organs,
28    is readily and unequivocally diagnosed by standard histopathological techniques. MCL has also
29    been described as large, granular, lymphocyte leukemia and is known to be a rapidly
30    progressing and fatal neoplasm whose incidence is age related. The tumor is transplantable; its
31    eti ol ogi cal factor i s unknown.
32          The  similarities and differences in the tissue of origin,  precursor cell line, and pathologic
33    characteristics of rat MCL and human lymphoid cancers have  been reviewed by Caldwell (1999)
34    Ishmael and Dugard (2006) and EPA (U.S. EPA, 1997b).  Both diseases result from abnormal
3 5    development and maturation of lymphocytes. It is known that the tissue of origin in rats is the
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-160      DRAFT-DO NOT CITE OR QUOTE

-------
 1    spleen and that lymphomas and leukemias in humans originate in the bone marrow.  The
 2    precursor cell line is not known for either rat MCL or the human lymphoid cancers.
 3          Large granular lymphocyte cells exist in humans that are morphologically, biochemically,
 4    and functionally similar to the cells involved in MCL in the F344 rat (Stromberg, 1985). The
 5    pathological characteristics of rat MCL are similar in some respects to one of the human T-cell
 6    leukemias (Caldwell, 1999), and some investigators believe MCL serves as a model for T-cell
 7    leukemia (Stromberg, 1985). However, discounting a rodent neoplasm simply because it has no
 8    exact human counterpart is not a scientifically defensible position. Strict site concordance is not
 9    a requirement for relevancy in extrapolation of hazard potential. For example, many aromatic
10    amines are probable bladder carcinogens in humans but are likely to produce Zymbal gland
11    tumors in rats, for which there is no analogous organ in humans.
12          The specific mechanism of leukemogenesis in rats is not understood, but neither is it well
13    understood in humans.  A possible link to MO A for tetrachloroethylene-induced MCL in rats
14    comes from early reports of toxicity of cysteine  S-conjugates, where S-(l,2,-dichlorovinyl)-L-
15    cysteine, the trichloroethylene metabolite,  was implicated in induction of aplastic anemia and
16    marked biochemical alteration of DNA in bone marrow, lymph nodes, and thymus in calves
17    (Bhattacharya and Schultze, 1971, 1972).
18          As discussed elsewhere, the GSH conjugate of tetrachloroethylene is hydrolyzed in the
19    kidney to the comparable cysteine S-conjugate, a compound that can be cleaved to form a
20    mutagen. Humans as well as rodents activate the conjugate via FMO3, CYP3 A and/or the beta
21    lyase pathway. Thus, the possibility exists that the tetrachloroethylene S-conjugate
22    S-(l,2,2-trichlorovinyl)-L-cysteine may be involved in inducing leukemia in rats and may have
23    the potential to produce blood dyscrasias in humans as well.  However, a recent report of a study
24    in which TCVC was given to two calves did not find that it produced bone marrow injury in
25    these animals at dose levels comparable to those of DCVC that caused bone marrow toxicity in
26    calves in the same study (Lock et al., 1996).
27
28    4.8.2.2.1.4.  Summary and conclusions regarding the leukemia finding in rats. Leukemia
29    incidences were significantly increased in both male and female rats, in spite of high
30    spontaneous background incidences. Although caution is recommended with regard to
31    interpreting results for species that have high spontaneous incidences for any specific tumor, it is
32    also important to remember that high spontaneous incidences of lymphohematopoietic cancer are
33    not unique to rodent strains—e.g., high incidences of leukemia occur in certain genetically
34    susceptible humans as well (U.S.  EPA, 1997a).  In addition to causing increased tumor
35    incidences, tetrachloroethylene caused a dose-related increase in severity of MCL in both sexes
36    and shortened the time to tumor in female rats in one of the chronic bioassays.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-161     DRAFT-DO NOT  CITE OR QUOTE

-------
 1          The principal type of chemically induced lymphohematopoietic cancer in humans is
 2    myeloid leukemia, with the exception being the lymphohematopoietic cancers induced by
 3    immunosuppressive agents, which are usually associated with development of lymphomas.
 4    There is some epidemiologic evidence that occupational exposure to tetrachloroethylene is
 5    associated with NHL; thus, lymphohematopoietic cancers are observed in both rats and humans.
 6    Although the specific etiology of NHL is unknown at the present time, it is thought likely to be
 7    related to imbalances or disturbances in the immune system. Two potent immunosuppressive
 8    chemotherapeutic agents, cyclosporin and azathioprine, are associated with NHL.
 9          If a chemical produces a significant increase in MCL in the F344 rat, the finding cannot
10    be ignored.  The observation of a significant increase of MCL in rats signals that the chemical
11    may possibly cause similar or different types of tumors in humans.
12
13    4.8.2.2.2. Tumors at other sites  in animal bioassays. In the NTP inhalation study, an elevated
14    incidence of rare brain gliomas in rats was observed. In males in the control and the mid- and
15    high-tetrachloroethylene concentration groups, the incidences were 1/50, 0/50, 4/50, respectively,
16    and there was a significantly positive dose-related trend by the life table test but not by the
17    incidental tumor trend test.  In females the incidence was 1/50, 0/50, 2/50. These are rare tumors
18    in NTP rat bioassays; the historical control incidence for males and females combined in the
19    laboratory was 2/247 (0.8%),  and in the overall NTP program it was 6/1,971 (0.3%). Because
20    these tumors had not been observed in the previous NTP studies of trichloroethylene (NTP,
21    1990b) or pentachloroethane (NTP,  1983), and because they appeared in the untreated groups,
22    NTP concluded that they were not related to tetrachloroethylene exposure.
23          On the other hand, the tumors in the high-dose males occurred slightly earlier (88, 96,
24    102, and 103 weeks) than in the control group (99 weeks); in the high-dose females they
25    occurred more definitively earlier (75 and 78 weeks in the high-dose group vs. 104 weeks in the
26    control group). In addition, a Fisher's exact test of the significance of combined male and
27    female incidences in the tetrachloroethylene-treated animals shows significance with respect to
28    both lab and NTP program historical data, whereas the control incidence is not significant with
29    respect to either of the historical  data sets.  Therefore, although the data showing that
30    tetrachloroethylene is causing brain gliomas in rats is not strong, it is suggestive.
31          The incidence of interstitial testicular tumors in male F344 rats treated with
32    tetrachloroethylene was significantly higher than in controls in the study. However, it is
33    common in control rats (the NTP program historical control incidence was 1,729/1,949 [89%]),
34    and the incidence in treated animals was not higher than in historical laboratory or historical
35    program controls.  Also, the combined incidence of hyperplasia and tumors was not significant
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-162      DRAFT-DO NOT CITE OR QUOTE

-------
 1    with respect to that of the study controls.  For these reasons, the NTP concluded that the
 2    marginally higher incidence was not related to tetrachloroethylene exposure.
 3          Hemangioendotheliomas in the liver and spleen in male mice were observed in the Japan
 4    bioassay. These were mentioned in Section 4.4.2.2 in connection with liver cancer, and the data
 5    are given in Section 5.2.2.  These tumors were not observed in the NTP studies.
 6
 7    4.8.3.  Summary of Immunotoxicologic Effects in Humans and Animals and Potential Mode
 8    of Action
 9          The epidemiologic evidence pertaining to tetrachloroethylene exposure in relation to
10    immune-related conditions, is limited (Table 4-11). Estimated associations based on population-
11    based case-control studies of various systemic autoimmune diseases have low statistical power
12    (and thus are highly imprecise) because of the relatively low prevalence of occupational
13    exposure to this chemical in the general population. Exposure misclassification is likely in these
14    studies, and in general would be expected to result in an attenuation of the observed effect. To
15    date, no relevant occupational cohorts have examined the risk of developing these diseases.  The
16    few studies that have  been conducted related to allergy and hypersensitivity have used direct
17    measures of tetrachloroethylene and other compounds in ambient or exhaled air samples, but
18    these studies do not provide much evidence of an adverse effect relating to tetrachloroethylene
19    exposures. However, the only study of asthma severity was quite small (n = 21), and our
20    understanding of the impact of changes in specific T-cell subsets (i.e., interferon-y) is currently
21    limited.
22          The immune system is clearly crucial to the prevention of disease caused by infectious
23    agents. Alterations in immune function may also contribute to the development of non-
24    infectious diseases including cancer, autoimmune diseases, and hypersensitivity disorders. Many
25    immunosuppressive agents are human carcinogens (Tomatis et al., 1989), and as described in the
26    previous section, inhibition of the natural immune surveillance could play a role in the
27    hepatocarcinogenic properties of tetrachloroethylene.  The numerous immune-mediated activities
28    of relevance to the pathogenesis of a variety of disease include the binding and processing of
29    antigens by B-cells and T-cells, alteration and loss of tolerance to self-antigens, defects in
30    apoptosis which may  effect the clearance of cells, and the secretion of pro- and anti-
31    inflammatory cytokines (Seliger, 2005; Ayensu et al., 2004).
32          Binding of reactive compounds to cellular macromolecules has been proposed as an
33    important step in the pathogenesis of several diseases, both for cancer (Hinson and Roberts,
34    1992) and for chemically induced autoimmune disease (Uetrecht et al.,  1988).  The modification
35    of proteins may lead to more immunoreactive products, and these may lead to the development
36    of autoantibodies and the cellular damage seen in alcoholic liver disease and in autoimmune

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-163      DRAFT-DO NOT CITE OR QUOTE

-------
 1    diseases (e.g., systemic lupus erythematosus, rheumatoid arthritis; Kurien et al., 2006). Reactive
 2    metabolites of tetrachloroethylene have been shown to bind irreversibly to cellular
 3    macromolecules in vitro (e.g., Costa and Ivanetich, 1980) and in vivo (Pegg et al., 1979;
 4    Schumann et al., 1980).  Binding occurs proportionally to the amount metabolized, and
 5    metabolism is proportional to toxicity (e.g., Buben and O'Flaherty, 1985).  Several published
 6    studies have demonstrated formation of trichloroacylated protein adducts, for example, in liver
 7    and kidney of rats (Birner et al., 1994) and in plasma of rats and humans (Pahler et al., 1999)
 8    following exposures to tetrachloroethylene.  Another example is the detection of
 9    trichloroacetylated protein adducts formed in the liver of MRL-lpr/lpr and MRL +/+ mice treated
10    with tetrachloroethylene (Green et al., 2001). These strains are "lupus-prone" mice, because of
11    their genetic susceptibility to the development of lupus-like disease. Further studies designed to
12    identify the adducted proteins  may help to elucidate an MO A for tetrachloroethylene-induced
13    autoimmune response, which,  in turn, may be related to cancer-causing activity (not clear how
14    this autoimmune response is related to cancer).
15           Apoptosis, a form of natural cell death differing from necrosis, is essential for proper
16    functioning of the immune system and clearance of tumor cells. Apoptosis also plays an
17    important role in the pathogenesis of autoimmune diseases. Genetic or environmental exposures
18    leading to increased apoptosis or to decreased clearance of apoptotic debris may stimulate the
19    production of autoantibodies directed against intracellular antigens (Gaipl et al., 2006). The
20    production of free radicals, increased lipid peroxidation, and increased apoptosis was
21    demonstrated in a recent study using human lung adenocarcinoma cells treated with
22    tetrachloroethylene (Chen et al., 2002b).  Alternatively, the inhibition of apoptosis of CD4+
23    T-cells may also effect the development of autoimmune disorders, as demonstrated by the recent
24    studies of trichloroethylene  metabolites (Blossom and Gilbert, 2006). Thus, with respect to
25    autoimmune diseases, as well  as neurodegenerative and other diseases, the strict regulation of
26    apoptosis signalling is crucial  (Ethell and Buhler, 2003; Schattenberg et al., 2006).
27
28    4.9. SUSCEPTIBLE POPULATIONS
29           Variation in response among segments of the population may be due to age, genetics, and
30    ethnicity, as well as to differences in lifestyle, nutrition, and disease status. These could be
31    potential risk factors that play  an important role in determining an individual's susceptibility and
32    sensitivity to chemical exposures.  Studies on tetrachloroethylene toxicity and MOA in relation
33    to some of these risk factors are discussed below.
34
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-164      DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.9.1. Life Stages
 2           Individuals of different life stages are physiologically, anatomically, and biochemically
 3    different. Early and later life stages differ greatly from mid-life stages in body composition,
 4    organ function, and many other physiological parameters that can influence the absorption,
 5    distribution, metabolism, and elimination of chemicals and their metabolites from the body (ILSI,
 6    1992).  The limited data on tetrachloroethylene exposure suggest that these subpopulations—
 7    particularly individuals in early life stages—may have greater susceptibility than does the
 8    general population. This section presents and evaluates the pertinent published literature
 9    available to assess how individuals of differing life stages may respond differently to
10    tetrachl oroethy 1 ene.
11
12    4.9.1.1.  Life Stage-Specific Exposures
13           Section 2.2 describes the various exposure routes of concern for tetrachloroethylene. For
14    all postnatal life stages, the primary exposure routes of concern include inhalation (see Section
15    2.2.1) and contaminated water (see Section 2.2.2).  In addition, certain exposure pathways to
16    tetrachloroethylene are unique to early life stages, such as through placental transfer or via breast
17    milk ingestion, or may be increased during early or later life stages.  In utero, there is biological
18    plausibility of transfer of tetrachloroethylene  across the human placental barrier as seen in
19    rodents (Ghantous et al., 1986; Szakmary et al., 1997).  Fetal blood concentrations have been
20    modeled for human exposure (Gentry et al., 2003).
21           For infants, a unique exposure route of concern is ingestion of breast  milk (see Section
22    2.2.4).  The breast milk of one woman was found to contain  10 mg/L tetrachloroethylene 1 hr
23    following a visit to her spouse working at a dry cleaning establishment, dropping to 3 mg/L after
24    24 hrs (Bagnell and Ellenberger, 1977).  Tetrachloroethylene has also been measured in the
25    breast milk of a woman living in an apartment located in a building housing a dry cleaning
26    facility (Schreiber et al., 2002; NYS DOH, 2005b). PBPK models have been used to estimate
27    the dose a nursing infant might receive from an exposed mother's breast milk (Gentry et al.,
28    2003; Schreiber, 1993).  Using different exposure scenarios, Schrieber predicted that breast milk
29    concentrations could range from 1.5 ug/L for a typical residential scenario, 16-3,000 ug/L for a
30    residential scenario near a dry cleaner, to 857-8,440 ug/L for an occupational scenario.
31    Assuming that a 7.2 kg infant ingests 700 mL of breast milk per day, Schreiber estimated dose to
32    the infant could range from 0.0001 to 0.82 mg/kg/day (Schreiber, 1993). Therefore, the dose an
33    infant could receive through breast milk may exceed the previous EPA RfD level (0.01 mg/kg-
34    day). Byczkowski and Fisher (1995) refined  the approach used by Schrieber (1993) and found
35    that with the same exposure conditions, the results predicted lower doses to the infant (0.0009-
36    0.202 mg/kg/day).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-165      DRAFT-DO NOT CITE OR QUOTE

-------
 1          For infants on formula, ingestion of contaminated water may be of concern. Taking into
 2    account tetrachloroethylene volatilization in boiling water, Letkiewicz et al. (1982) estimated
 3    that 22% of formula-fed infants received fluids contaminated with tetrachloroethylene levels
 4    found in the water supply.  Data showed that about 11% (0.5  x 22%) of formula-fed infants
 5    could receive an increased exposure as compared with adults on a mg/kg basis through drinking
 6    contaminated water.
 7          Dairy products have been found to have elevated concentrations of tetrachloroethylene
 8    (see Section 2.2.3), and children ingest larger quantities of dairy products compared to adults
 9    (NRC, 1993). Therefore, there may be concern for ingestion of contaminated dairy products in
10    early life stages, although this exposure route for tetrachloroethylene has not been well
11    characterized for any life stage.
12          Inhalation exposures may be increased for both early and later life stages compared to
13    adults, since children and the elderly have increased ventilation rates per kg body weight
14    compared to adults (NRC, 1993; U.S. EPA, 2006) and since they spend the majority of their time
15    indoors (NRC, 1993; U.S. EPA, 2002), where increased concentrations of tetrachloroethylene
16    have been  found (U.S. EPA, 2001b). Section 2.2.1 describes increased indoor air concentrations
17    measured inside apartments containing dry cleaned clothing (Thomas et al., 1991), in  apartments
18    above or adjacent to dry cleaners (Altmann et al., 1995; Chien, 1997; Garetano and Gochfeld,
19    2000; McDermott et al., 2005; Schreiber et al., 1993, 2002), and in daycare centers adjacent to
20    dry cleaners (NYS DOH, 2005b, c). In addition, inhalation may also occur during showering or
21    bathing as dissolved tetrachloroethylene in the warm tap water is volatilized (Rao and Brown,
22    1993).
23         Dermal exposures may be increased for both early and later life stages compared to adults,
24    since infants have increased skin area per kg body weight  (NRC, 1993) and the elderly
25    experience changes in permeability (U.S. EPA, 2006). Dermal exposure may occur in an
26    occupational setting from direct handling of tetrachloroethylene or in a residential setting from
27    showering, bathing, or swimming in contaminated water (Rao and Brown, 1993; U.S. EPA 2001).
28    (see Section 2.2.2) While dermal exposure is generally not considered a major route  of exposure,
29    this route of exposure is not well characterized for early life stages (prenatal or postnatal), or
30    later life stages.
31
32    4.9.1.2.  Early Life Stage Effects
33          Although limited data exist on tetrachloroethylene toxicity as it relates to early life stages,
34    there is enough information to discuss the qualitative differences.  In addition to the evidence
35    described below, Section 4.7 contains  information on both human and animal evidence for
36    reproductive and developmental outcomes such as spontaneous abortion/fetal loss, low birth

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-166      DRAFT-DO NOT CITE OR QUOTE

-------
 1    weight, IUGR, SGA, congenital abnormalities, sperm quality, developmental delays, and
 2    behavioral changes. Together, Sections 4.4 on liver toxicity, 4.5 on kidney toxicity, 4.6 on
 3    neurotoxicity, and 4.8 on toxic effects in other organ systems characterize a wide array of
 4    postnatal developmental effects.
 5
 6    4.9.1.2.1. Differential effects in early life stages.  Preconception exposure has been associated
 7    with altered semen quality in occupationally exposed humans (Eskenazi et al., 1991b; Rachootin
 8    and Olsen, 1983; Sallmen et al., 1995), as well as in mice, but not in rats (Beliles et al., 1980).
 9    Additionally, reduced testes weight was seen in rats after inhalation exposure (Tinston, 1994).
10          A number of human studies have shown spontaneous abortion among women employed
11    as laundry workers (Hemminki et al., 1980) or dry cleaners (Ahlborg, 1990; Bosco et al., 1987;
12    Doyle et al., 1997; Olsen et al., 1990; Kyyronen et al., 1989), married to men employed as dry
13    cleaners (Taskinen et al., 1989; Eskenazi et al., 1991a), exposed to other occupational  solvents
14    (Windham et al., 1991; Lindbohm et al.,  1990), or living in residences receiving contaminated
15    water (Lagakos et al.,  1986; Bove et al., 1995; ATSDR, 1998). However, another study
16    population of women working as laundry or dry cleaning workers did not experience
17    spontaneous abortion (McDonald et al., 1986, 1987). Reduced fertility has been seen in women
18    and men exposed occupationally (Eskenazi et al, 1991a, b; Sallmen et al.,  1995; Rachootin and
19    Olsen, 1983).
20          In the literature concerning animal studies, there is evidence of increased pre- and post-
21    implantation loss (Szakmary et al., 1997) and increased resorption of rodent pups after maternal
22    inhalation (Schwetz et al., 1975; Szakmary et al., 1997), reduction in litter size after maternal
23    gavage (Narotsky and Kavlock, 1995), and litters with dead pups (Tinston, 1994).  However,
24    fetal loss was not seen in other animal studies (Carney et al., 2006; Hardin et al., 1981). In vitro
25    studies show decreased fertilized oocytes (Berger and Horner, 2003), and increased mortality,
26    malformations, and delayed growth and differentiation of embryos (Saillenfait et al., 1995) when
27    exposed to tetrachloroethylene.
28          After residential exposure to contaminated water, birth outcomes related to in utero
29    exposure in humans include perinatal death, birth defects (eye and ear anomalies, and CNS/oral
30    cleft anomalies; Lagakos et al., 1986), and IUGR (Windham et al., 1991) or SGA (Sonnenfeld et
31    al., 2001).  Also, childhood leukemia has been associated with in utero exposure to
32    tetrachloroethylene due to maternal ingestion of contaminated water (MA DPH, 1997; see
33    Section 4.9.1.2.4). The study population reported in Sonnenfeld et al., (2001) is currently being
34    further examined to determine any association between maternal ingestion of contaminated water
35    and the incidences of birth defects (e.g., neural tube defects and oral clefts; ATSDR, 2003).
36    However, other human studies have not shown effects after occupational exposure for other birth

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-167      DRAFT-DO NOT CITE OR QUOTE

-------
 1    outcomes, such as stillbirths, congenital anomalies, or decreased birth weight (Bove et al., 1992;
 2    Olsen et al., 1990, Kyyronen et al., 1989, Taskinen et al., 1989, Windham et al., 1991).  In
 3    addition, preconception or prenatal exposure may lead to other latent outcomes such as an
 4    increased risk for schizophrenia as seen in a large prospective study of parental occupational
 5    exposure to tetrachloroethylene (Perrin et al., 2007).
 6          Birth outcomes in animals exposed to tetrachloroethylene in utero include skeletal
 7    retardation and malformations, decreased body weight and weight gain, altered brain fatty acid
 8    composition, and developmental delay.  Skeletal retardation and malformations were increased
 9    in rodent pups after maternal inhalation exposure 7 days per week (Carney et al., 2006;
10    Szakmary et al., 1997), but no birth defects were seen in other animal studies using a similar
11    dose but for 5 days per week (Hardin et al., 1981).  Exposure resulted in decreased fetal or pup
12    body weights (Carney et al., 2006; Szakmary et al., 1997; Tinston,  1994), along with reduction in
13    weight gain (Nelson et al., 1980). Also, Kyrklund and Haglid (1991) noticed slightly altered
14    brain fatty acid composition after maternal exposure during pregnancy was also noted (Kyrklund
15    and Haglid, 1991). Developmental delay was seen in rat offspring after maternal exposure
16    during pregnancy (Nelson et al., 1980).  In addition, cardiac anomalies have  been seen in rats
17    exposed to the metabolites TCA (Smith et al., 1989; Johnson et al., 1998) and DC A (Epstein et
18    al.,  1992; see Sections 4.6.2, 4.7.2, and 4.8.2).
19          Neurotoxicological effects in children have been reported after low exposure levels to
20    tetrachloroethylene (see Section 4.6 and Table 4-4). While other neurotoxic effects are seen in
21    adults (see Table 4-5), decreased VCS has been the main observation in children, including in
22    those who resided in an apartment building with a dry cleaning establishment (Schreiber et al.,
23    2002; NYS DOH, 2005a). Children who attended a day care center adjacent to a dry cleaner
24    were too young to take a visual exam given to the adult workers that demonstrated decreased
25    VCS.  Other neuropsychological tests conducted on the children attending this day care center 5
26    weeks after exposure ceased did not consistently find any effect on cognition or behavior (NYS
27    DOH, 2005b). A follow-up evaluation of a different set of children attending the same day care
28    center 4 to 5 years after exposure showed no residual changes in VCS or  color vision, although
29    these children were not tested immediately after exposure (NYS DOH, 2005c). A case study
30    reported reduced VCS in a 2V2 year old boy after prenatal exposure to tetrachloroethylene (Till et
31    al., 2003). Sections 4.6.2 and 4.7.2 discuss studies of postnatal neurological effects in animals
32    after prenatal exposure.  Altered brain biochemistry was seen in the offspring of exposed rodents
33    (Kyrklund & Haglid, 1991; Nelson et al., 1980), and the offspring showed signs of
34    developmental delay (Nelson et al, 1980), altered motor activity (Szakmary et al., 1997; Tinston,
35    1994), decreased muscular strength (Szakmary et al., 1997), and short-term reduced response to
36    sound in pups (Tinston, 1994).  In addition, young animals have been directly exposed

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-168      DRAFT-DO NOT CITE OR QUOTE

-------
 1    postnatally to tetrachloroethylene.  One gavage study on young 45-50 gram rats showed
 2    behavioral and locomotor effects (Chen et al., 2002a), and another gavage study on 10-day old
 3    mice showed increased locomotor activity and decreased rearing behavior (Fredriksson et al.,
 4    1993).  Following i.p. dosing, 8-week-old male mice showed effects on the righting reflex and
 5    balancing (Umezu et al.,  1997), and 6-week-old rats showed effects on locomotor activity
 6    (Motohashi et al., 1993).  Both human and animal evidence supports an association between
 7    neurodevelopmental effects and tetrachloroethylene exposure.
 8           Section 4.8.1.1.1 and Table 4-10 describe studies  relating tetrachloroethylene to immune
 9    response in children. Lehmann et al. (2002) examined cord blood samples for T-cell
10    subpopulations and associated them with indoor exposure to VOCs measured 4 weeks after birth
11    (likely to reflect late-prenatal exposures);  however, another study examining indoor exposure to
12    VOCs and allergic sensitization and cytokine secretion in 3-year-old children at high risk for
13    development of allergic disease (low birth weight, high cord blood IgE, family history of atopy)
14    found no association between tetrachloroethylene exposure and any of the allergens tested in this
15    study (Lehmann et al. 2001).  In a study of inhalation exposure, Delfmo et al. (2003a, b)
16    measured the concentration of ambient air pollutants, including tetrachloroethylene, and
17    correlated it with subsequent symptoms of asthma in children in the Los Angeles area.  The
18    results suggest an increased risk with exposure to tetrachloroethylene (Delfmo et al.,  2003a).
19    However, another analysis  of the data examined the amount of tetrachloroethylene and other
20    volatile organic compounds in exhaled breath of asthmatic children (Delfmo et al., 2003b).
21    Although there was a significant correlation between ambient and exhaled concentrations, the
22    investigators did not find any association with exhalation concentrations and asthma  symptoms
23    or ambient air concentrations and asthma  symptoms, although the OR for exhaled breath was
24    larger than for ambient air exposure (OR = 1.94, 95% CI = 0.8-4.7; Delfmo et al., 2003b). An
25    18-year-old without personal or family history of bronchial asthma developed respiratory
26    symptoms (cough, dyspnea, altered forced expiratory volume) after maintaining dry cleaning
27    machines (Boulet, 1988). The limited, available data from these studies provide weak evidence
28    of an effect of tetrachloroethylene exposure during childhood on allergic sensitization or
29    exacerbation of asthma symptomology. However, the observation of the association between
30    increased tetrachloroethylene exposure and reduced interferon-y in cord blood samples may
31    reflect a sensitive period of development,  and points to our current lack of understanding of the
32    potential immunotoxic effects of prenatal  exposures.
33           Other postnatal health effects after tetrachloroethylene exposure have been seen in
34    children.  In one case study with inexact exposure information, tetrachloroethylene vapors off-
35    gassing from dry-cleaned fabrics were implicated in causing the death of a 2-year-old boy after
36    sleeping in a room with curtains that had been incorrectly dry cleaned (Gamier et al., 1996).

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-169      DRAFT-DO NOT CITE  OR QUOTE

-------
 1    Bagnell and Ellenberger (1977) reported that a child suffered from obstructive jaundice and
 2    hepatomegaly after consuming contaminated breast milk, with conditions improving when
 3    breastfeeding was discontinued.  In the one case of a child's direct ingestion of
 4    tetrachloroethylene, a 6-year-old boy who swallowed 12-16 g tetrachloroethylene lost
 5    consciousness and lapsed into a coma (Koppel et al., 1985). This 6-year-old also experienced
 6    drowsiness, vertigo, agitation, and hallucinations, but he later recovered.  Follow-up testing on
 7    the boy was not reported; therefore, any potential long-term effects of the exposure are unknown
 8    (see Section 2.2.5).
 9
10    4.9.1.2.2. Toxicokinetics and tetrachloroethylene in early life stages.  Chapter 3 describes the
11    toxicokinetics of tetrachloroethylene.  Early life stage-specific information regarding absorption,
12    distribution, metabolism, and excretion needs to be considered for a child-specific and chemical-
13    specific PBPK model.  To adequately address the risk to infants and children, age-specific
14    parameters for these values should be used in PBPK models that can approximate the internal
15    dose a infant or child receives based on a specific exposure level (Byczkowski and Fisher, 1994;
16    Clewell et al., 2004; Gentry et al., 2003; Rao and Brown, 1993; see Section 3.5).
17           As discussed in Section 3.1, exposure may occur via inhalation, ingestion, and skin
18    absorption.  In addition, prenatal exposure may result in absorption via the transplacental route.
19    Exposure via inhalation is proportional to the ventilation rate, duration of exposure, and
20    concentration of expired air, and children have increased ventilation rates per kg body weight
21    compared to adults, with an increased alveolar surface area per kg body weight for the first two
22    years (NRC, 1993). For lipophilic compounds such as tetrachloroethylene, percent adipose
23    tissue, which varies with age (NRC, 1993), will affect absorption and retention of the absorbed
24    dose. It is not clear to  what extent dermal absorption may be different for pregnant women and
25    children compared to adults, given their increased surface areas and thinner outer skin layers.
26           The distribution of tetrachloroethylene to specific organs will depend on organ blood
27    flow and the lipid and water content of the organ (NRC, 1993), which may vary between life
28    stages.  Rodent studies demonstrate that tetrachloroethylene crosses the placental barrier when
29    pregnant dams are exposed (Ghantous et al., 1986; Szakmary et al., 1997), and in humans it has
30    been shown that during lactation, tetrachloroethylene distributes to breast milk (NYS DOH,
31    2005b; Schreiber, 1993; Sheldon et al., 1985). However, a noticeable difference exists between
32    the milk/blood partition coefficients for rats (12) and for humans (2.8; Byczkowski and Fisher,
33    1994), reflecting the higher fat content of rat milk.
34           Tetrachloroethylene can also cross the blood-brain barrier during both prenatal and
35    postnatal  development; this may occur to a greater extent in younger children. Based on the
36    modeled dose of tetrachloroethylene to the brain after a showering/bathing scenario, a study by

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-170      DRAFT-DO NOT CITE OR QUOTE

-------
 1    Rao and Brown (1993) showed that for a given set of exposures, the younger a person is, the
 2    greater the estimated concentration of tetrachloroethylene in the brain. Modeling showed that
 3    after a 30-minute bathing scenario, a 3-year-old child accumulated higher brain tissue
 4    concentrations of tetrachloroethylene as compared with a 10-year-old and an adult. An autopsy
 5    conducted on the previously mentioned 2-year-old boy found dead after exposure to dry-cleaned
 6    curtains  revealed the highest levels of tetrachloroethylene in the brain, 77 mg/kg. Levels in his
 7    blood, heart, and lungs were 66 mg/L, 31 mg/kg, and 46 mg/kg, respectively (Gaillard et al.,
 8    1995; Gamier et al., 1996).
 9          Animal studies provide clear evidence that tetrachloroethylene distributes widely to all
10    tissues of the body (see Section 3.2) but it is not clear whether distribution may vary
11    differentially with lifestage.
12          Section 3.3.3 describes the production of CYP enzymes involved in the metabolism of
13    tetrachloroethylene.  Expression of these enzymes changes during various stages of fetal
14    development (Hakkola et al., 1996a, 1996b, 1998) and during postnatal development (Tateishi et
15    al., 1997). One study modeled the role of the age-dependent development of CYP2E1 in
16    oxidative metabolism (TCA) in the mother and lactating infant (Vieira et al., 1996). A number
17    of other  human studies suggest that CYP2B6 may also play a role in the metabolism of
18    tetrachloroethylene (White et al., 2001), although this enzyme was not detected in placental or
19    fetal liver samples (Hakkola et al., 1996a, b), and differences between a group of 10 perinatal
20    and infant patients showed significantly lower CYP2B6 protein expression in placental hepatic
21    microsomes as compared with an adult group (Tateishi et al., 1997).
22          The major processes of excretion of tetrachloroethylene and its metabolites are discussed
23    in Section 3.3 and 3.4. Tetrachloroethylene or its metabolites have been measured in blood
24    (NYS DOH, 2005a; Popp  et al., 1992), exhaled breath (Delfmo et al, 2003b; Schreiber et al.,
25    2002; NYS DOH, 2005a), and urine of children (NYS DOH, 2005b; Schreiber et al., 2002; Popp
26    etal., 1992).
27
28    4.9.1.2.3. Toxicodynamics and tetrachloroethylene in early life stages. Toxicodynamic
29    responses to chemical exposures can change throughout different life stages. TCA, a metabolite
30    of tetrachloroethylene, is hypothesized to be the causative agent for developmental toxicity
31    expressed as morphological changes, lethality, and/or growth. TCA could accumulate to a
32    greater extent in the embryo/fetal compartment than in the mother, based on the pKa of the acid
33    and the pH gradient between the maternal plasma and the embryo compartments (O'Flaherty et
34    al., 1992). TCA could induce developmental toxicity by changing the intracellular pH or
35    through  peroxisome proliferation. Ghantous et al. (1986) detected TCA in the amniotic fluid of
36    pregnant mice exposed to tetrachloroethylene via inhalation (see Section 4.7.4).

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-171      DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.9.1.2.4. Susceptibility to cancer in early life stages. The epidemiologic and experimental
 2    animal evidence is limited regarding susceptibility to cancer from exposure to
 3    tetrachloroethylene during early life stages. The human epidemiological evidence is summarized
 4    above for cancer in the liver (see Section 4.4.1.2), kidney (see Section 4.5.1.2), and other organ
 5    systems (see Section 4.8.1.2).  The animal research is summarized above for cancer in the liver
 6    (see Section 4.4.2.2), kidney (Section 4.5.2.2), and other organ systems (see Section 4.8.2).
 7          Few studies have examined cancer in children after exposure to tetrachloroethylene;
 8    however, those few have found evidence for concern for leukemia (see Section 4.8.1.2.1). One
 9    case-control study of children residing in Woburn, Massachusetts diagnosed with leukemia
10    examined exposure to drinking water contaminated with multiple solvents, including
11    tetrachloroethylene (Costas et al., 2002;  MA DPH,  1997). This study reported a large but
12    imprecise association and a dose-response relationship between maternal exposure during
13    pregnancy and childhood leukemia in the offspring when compared with exposure prior to
14    pregnancy or postnatal exposure to the infant via lactation (Costas et al., 2002; MA DPH,  1997),
15    and altered immune response was found in family members of the cases (Byers et al., 1988; see
16    Section 4.8.1.1.1). However, it is difficult to uniquely identify tetrachloroethylene as the
17    causative agent given the higher concentrations of trichloroethylene reported in these studies.
18    Similarly, in another case-control study of childhood leukemia, paternal exposure to chlorinated
19    solvents have been associated with increased risk (Lowengart et al., 1987), although this and
20    other case-control studies do not show an increased risk from paternal (Lowengart et al., 1987;
21    Shu et al.,  1999) or maternal (Infante-Rivardet al., 2005; Shu et al.,  1999) occupational exposure
22    to tetrachloroethylene,  possibly due to the relatively small sample size. Another study
23    population is currently being further examined to determine any association between maternal
24    ingestion of contaminated water and the incidence of childhood cancers (ATSDR,  2003).  One in
25    vitro study of human mononuclear cord blood cells exposed to tetrachloroethylene found that
26    pathways involved in cancer induction were affected through altered gene expression of
27    inflammatory responses, tumor and metastatis progression, and the apoptotic process  (Diodovich
28    et al., 2005). Leukemia has also presented in adult humans after tetrachloroethylene exposure
29    (see Section 4.8.1.2.1, Appendix 4B, and Tables 4B-5 and 4B-13), and MCL has been seen in
30    exposed adult rats in spite of high spontaneous background incidences (see Section 4.8.2.4).
31    While interspecies extrapolation of rat MCL to humans has been questioned (see Section
32    4.8.2.4.1),  the findings in humans suggest relevance for the leukemia findings in animals.
33    However, no data are available on leukemia risk in young animals exposed to tetrachloroethylene.
34
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-172     DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.9.1.3. Later Life Stages
 2           Few studies examine the effects of tetrachloroethylene exposure in elderly adults. One
 3    study found elevated blood tetrachloroethylene levels (310-1770 ug/L) and urine trichloroacetic
 4    acid levels (22-1650 ug/L) in an elderly couple living above a dry cleaning facility (Popp et al.,
 5    1992).  Another residential study examined two individuals over the age of 60 years and found
 6    that the mean scores of VCS were lower than the 12th percentile of all control subjects (Schreiber
 7    et al., 2002).  These studies suggest that older adults may experience increased exposure to
 8    tetrachloroethylene and resulting increased VCS deficits than younger adults. However, there is
 9    no further evidence for elderly individuals exposed to tetrachloroethylene beyond these two
10    studies.
11
12    4.9.2. Other Susceptibility Factors
13           Aside from age, many other factors may affect susceptibility to tetrachloroethylene
14    toxicity.  A partial list of these factors includes gender, genetic polymorphisms, pre-existing
15    disease status, nutritional status, diet, and previous or concurrent exposures to other chemicals.
16    The toxicity that results due to changes in multiple factors may be  quite variable, depending on
17    the exposed population and the type of exposure. Qualitatively, the presence of multiple
18    susceptibility factors will increase the variability that is seen in a population response to
19    tetrachl oroethy 1 ene toxi city.
20
21    4.9.2.1. Health and Nutritional Status
22           It is known that kidney diseases can affect the clearance of chemicals from the body, and
23    therefore poor kidney health may lead to increased half-lives for tetrachloroethylene and its
24    metabolites. Similarly, liver disease may change the metabolic profiles in the liver, thus
25    potentially altering tetrachloroethylene metabolism.
26           Co-exposure to a-tocopherol (vitamin E) along with tetrachloroethylene resulted in
27    decreased rat (Costa et al., 2004) and mouse (Ebrahim et al., 1996, 2001) liver cell toxicity. A
28    similar protective effect was also seen with co-exposure to 2-deoxy-D-glucose in mice (Ebrahim
29    et al., 1996, 2001) and taurine in mice (Ebrahim et al., 2001).  However, no associations were
30    found for blood levels of vitamin E and beta-carotene in rats (Toraason et al., 2003; see Sections
31    4.3 and 4.4.4.4.3).
32
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-173      DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.9.2.2.  Gender
 2          In humans, it has not been determined whether there is a gender difference in response to
 3    exposure to tetrachloroethylene. However, because gender also affects metabolic capabilities,
 4    which vary throughout development, it is important to consider sex-specific changes.
 5          In the case of tetrachloroethylene, there is some indication that tetrachloroethylene
 6    metabolism is different between males and females. One PBPK model found gender-specific
 7    differences that were small (although significant) in tetrachloroethylene blood concentrations but
 8    considerable (2-fold at age 40) with regard to TCA blood concentration levels (Clewell et al.,
 9    2004; see Section 3.5.2 and Figure 3-3.).  Opdam and Smolders (1986) exposed six human
10    subjects to concentrations ranging from 0.5-9 ppm and found alveolar concentrations in male
11    subjects to be only slightly less than those in females (see Figures 3-6a, b). It is not known
12    whether gender variation of beta lyase activity (see Section 3.3.3.2.3), the most important
13    activator of toxic products in the conjugation pathway, exists in humans as it does in rats, with
14    metabolism in males being faster than in females (Volkel et al., 1998), although there seems to
15    be little gender difference in the concentrations of metabolites in blood, regardless of age
16    (Sarangapani et al., 2003).
17          Ferroni et al. (1992) evaluated neurological effects of tetrachloroethylene exposure
18    among female dry cleaners and concluded that tetrachloroethylene exposure in dry cleaning
19    shops may impair neurobehavioral performance and affect pituitary function.  The pituitary is
20    controlled in part by hypothalamic dopamine, which is important to neurotransmission. Study
21    participants were tested during the proliferation phase of menstruation which may better  capture
22    changes in prolactin secretion but also may potentially confound findings if there are individual
23    differences in severity of menstruation and in the timing  of test session relative to the day of
24    menstruation (U.S. EPA, 2004; see Section 4.6.1.2.5).
25          In a study of aircraft maintenance employees, Spirtas et al.  (1991) observed an increased
26    risk for NHL in females compared to males (see Section  4.8.1.2.1). Although quantitative
27    exposure information on tetrachloroethylene was not obtained in this study, differences in
28    exposure potential and level of exposure may explain the difference in risk between women and
29    men. Differences in physiological parameters may also explain the observed gender difference
30    in risk.
31          The studies by Pesch et al. (2000a) and Dosemeci et al. (1999) suggest that there may be
32    gender differences in risk to renal cell carcinoma with occupational exposure to
33    tetrachloroethylene; in both studies the risks were higher in males than in females (see Section
34    4.5.1.2). In a rat inhalation study, tubule cell hyperplasia was  observed in eight males at various
35    doses, but in only one female at high dose. Also, renal tubule  adenomas and adenocarcinomas
36    were observed only in males; however, chronically induced tetrachloroethylene  neoplastic

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-174     DRAFT-DO NOT CITE OR QUOTE

-------
 1    kidney lesions do not exhibit sex specificity (NTP, 1986a).  In a rat gavage study, there was no
 2    gender difference for toxic nephropathy (NCI, 1977).  A marked gender difference was seen
 3    between male and female rats in the severity of acute renal toxicity with male rats being more
 4    affected than female rats (Lash et al., 2002), but otherwise no gender variation was observed for
 5    chronic nephrotoxicity not associated with alpha-2u-globulin nephropathy  (see Sections 4.5.2.2
 6    and 4.5.4.3.3).
 7          In the liver, male rats showed an increased incidence of spongiosis hepatitis as compared
 8    with females, but there was no gender difference in hepatocellular adenomas and carcinomas;
 9    however, the spleen showed increased effects in males versus females (JISA, 1993; see Sections
10    4.4.2.1 and 4.4.2.2).
11
12    4.9.2.3. Race/Ethnicity
13          One residential study found that buildings with >100 |ig/m3 tetrachloroethylene were
14    more likely in minority neighborhoods (OR = 6.7; 95% CI = 1.5-30.5; NYS DOH, 2005a).  In
15    addition to possible increased exposure, different racial or ethnic groups may express metabolic
16    enzymes in different ratios and proportions due to genetic variability.
17          In a follow-up study  on the mortality of a cohort of dry cleaners, bladder cancer was
18    elevated among Caucasian men and women, and kidney cancer was elevated among black men
19    and women; however, these  associations were not strongly related to duration or estimated level
20    of exposure to tetrachloroethylene (Blair et al., 2003). One  study found that following
21    tetrachloroethylene exposure, TCA concentration in the urine of six Asian subjects was no
22    different from the levels found in six Caucasians; however,  this  study was confounded by
23    significant differences in alcohol consumption between the  Caucasian and Asian populations
24    (Jang and Droz, 1997).
25          Eskenazi et al. (1991a) noted a slightly lower per-cycle pregnancy rate among wives of
26    men who received higher level exposure to tetrachloroethylene,  but the potential contribution of
27    tetrachloroethylene exposure to time to conception was small when compared with the
28    contribution observed from Hispanic ethnicity and smoking, which were found to be stronger
29    and statistically significant predictors of time to conception.
30
31    4.9.2.4. Genetics
32          Human  variation in response to tetrachloroethylene exposure may be associated with
33    genetic variation. For example, in a study of six adults, Monster et al. (1979) found that the
34    mean coefficient of interindividual variation for tetrachloroethylene uptake was 17%.  Human
35    genetic polymorphisms in metabolizing enzymes involved in biotransformation of
36    tetrachloroethylene are now  known to exist (U.S. EPA, 1991; IARC, 1995; Lash and Parker,
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-175     DRAFT-DO NOT CITE OR QUOTE

-------
 1    2001).  Section 3.3.3.1.5 discusses CYP isoforms and genetic polymorphisms, Section 3.3.3.2.1
 2    covers GST isoenzymes and polymorphisms, and Section 3.3.4 describes differences in
 3    enzymatic activity.
 4          Reitz et al. (1996) examined tetrachloroethylene metabolism in seven adult human liver
 5    samples and found a fivefold difference in the rate of tetrachloroethylene metabolism between
 6    the 50th and 99th percentiles. Opdam (1989) found a 2-fold spread in tetrachloroethylene blood
 7    concentrations in a study population of nine adult human subjects. In this study, the amount of
 8    fat and the blood concentrations seemed to be positively correlated but could not be confirmed;
 9    the author suggested that if the subjects had a wider range of body fat levels (range in this study
10    was only 7-22 kg), a larger amount of interindividual variation would be expected.
11    Computer modeling was used to examine the toxicokinetic variability of tetrachloroethylene
12    (Bois et al.,  1996; Chiu and Bois, 2006).  However,  whether CYP or GSH polymorphisms
13    account for interindividual variation in tetrachloroethylene metabolism among humans, and thus
14    differences in susceptibility to tetrachloroethylene-induced toxicities, is not known.
15
16    4.9.3.  Multiple Exposures and Cumulative Risks
17          When considering health risks, it is important to consider the cumulative impact of
18    effects that may be due to multiple routes of exposure.  EPA  published Framework for
19    Cumulative Risk Assessment (U.S. EPA,  2003c) to address these issues. A human aggregate
20    exposure model developed by McKone and Daniels (1991) incorporated likely exposures from
21    air, water, and soil media through inhalation, ingestion, and dermal contact.  They asserted that
22    the aggregate exposure may be age dependent, but did not present any data for persons of
23    differing life stages.
24           The limited data summarized by the ATSDR in its draft interaction profile on
25    tetrachloroethylene, trichloroethylene, 1,1-dichloroethane,  and  1,1,1-trichloroethane suggest that
26    additive joint action is plausible (ATSDR, 2001). Co-exposure to other pollutants, including
27    trichloroethylene and methylchloroform which produce some of the same metabolites and
28    similar health effects as tetrachloroethylene, is likely to occur in occupational settings as well as
29    in non-occupational sources such as in ground water contamination (e.g., Bove et al., 1995;
30    Lagakos et al., 1996; MA DPH, 1997; ATSDR, 1998; Sonnenfeld et al., 2001). However, no
31    evidence was among available studies indicates greater-than-additive effects for liver and kidney
32    toxicity.
33          Due to the effects that many chemicals have on inducing and/or repressing metabolic
34    enzymes as well as on organ systems, co-exposures  may alter the way in which
35    tetrachloroethylene is metabolized and cleared from the body. Inhibition or induction of the
36    enzymes responsible for tetrachloroethylene metabolism can—and likely does—alter

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-176      DRAFT-DO NOT CITE OR QUOTE

-------
 1    susceptibility to toxicity (U.S. EPA, 1985a; IARC, 1995; Lash and Parker, 2001).  Numerous
 2    environmental pollutants and therapeutic agents have the potential to induce or inhibit
 3    tetrachloroethylene-metabolizing enzymes. For example, tetrachloroethylene metabolism is
 4    increased by inducers of CYP enzymes such as toluene, phenobarbital, and pregnenolone-
 5    16-a-carbonitrile, whereas CYP inhibitors such as SKF 525A, metyrapone, and carbon monoxide
 6    decrease tetrachloroethylene metabolism (Moslen et al., 1977; Ikeda and Imanura, 1973; Costa
 7    and Ivanetich, 1980). Likewise,  tetrachloroethylene exposure may increase the effects of
 8    exposures to other chemicals or stressors.  For instance, adverse effects due to exposure to
 9    chlorinated solvents and alcohol  may be increased because tetrachloroethylene may induce
10    shared metabolic enzymes (see Section 3.3.4).
11          The acute effects of tetrachloroethylene share much in common functionally with those
12    of other solvents (e.g., toluene, volatile anesthetics, and alcohols) such as changes in reaction
13    time, nerve conduction velocity,  and sensory deficits. There is emerging evidence that such
14    agents act on the ligand-gated ion channel superfamily in vitro (Shafer et al., 2005), particularly
15    on the inhibitory amino acids NMD A, nicotinic, and GAB A receptors in vivo (Bale et al., 2005).
16    Other organic solvents induce effects on memory and color vision (Altmann et al., 1995; Mergler
17    et al., 1991; Hudnell et al., 1996a, b).  The consistency of these observations suggests a common
18    MO A of organic solvents to altered vision pattern. Hence,  a concern exists for neurobehavioral
19    effects from interaction or competitive inhibition between tetrachloroethylene and exposures
20    with similarly hypothesized MOAs.
21          The interaction between tetrachloroethylene, trichloroethylene, and 1,1,1-trichloroethane
22    (methylchloroform) was modeled in rats (Dobrev et al., 2001) and in computer models for
23    humans (Dobrev et al., 2002) and were shown to compete for metabolic capacity.  The
24    interaction between tetrachloroethylene and trichloroethylene showed a less-than-additive effect
25    on the liver and kidney through inhibition of TCA formation (Pohl et al., 2003). Similarly, when
26    exposed to tetrachloroethylene, rat liver cells had increased toxicity when co-exposed to
27    peroxidation drugs such as cyclosporine A, valproic acid, and amiodarone (Costa et al., 2004),
28    and n-hexane and ethylbenzene inhibited the metabolism of tetrachloroethylene in rats (Skowron
29    etal., 2001).
30          Alcohol and smoking are generally regarded as confounders, although the additive or
31    interactive effects of these exposures along with tetrachloroethylene are not well characterized.
32    Exposure to alcohol changes the metabolic profiles in the liver,  thus increasing metabolism to
33    TCA through the CYP2E1 pathway (Pastino et al., 2000). Alcohol by itself cannot account for
34    the observed deficits in neurobehavioral functions, because statistical analyses of the
35    epidemiologic observations accounted for this covariate. Meskar et al. (2001) also contended
36    that alcohol induces CYP2E1, which in turn has the ability  to activate other compounds such as

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-177      DRAFT-DO NOT CITE OR QUOTE

-------
 1    halogenated solvents.  This could potentially cause higher toxicity of tetrachloroethylene for
 2    those who use alcohol, if the oxidative metabolism leads to a proto-oxidant. Valic et al. (1997)
 3    showed greater decrements in color vision among subjects with both exposures as compared with
 4    individuals with solvent exposure only or with neither exposure (see Sections 4.6.1.3 and 4.6.3).
 5           Regarding esophageal  cancer, occupational observations suggest that the magnitude of
 6    the risks for several smoking-related cancers among dry cleaners was greater than could be
 7    explained by smoking alone, suggesting a further contribution from another risk factor, such as
 8    occupational exposure (Blair et al., 2003; Ruder et al., 2001; see Section 4.8.1.2.2).
 9
10    4.9.4.  Uncertainty of Database for Susceptible Populations
11           There is a need to better characterize the implications of tetrachloroethylene exposures to
12    susceptible populations. A number of areas where the data base is currently insufficient are
13    identified below.
14
15    4.9.4.1. Uncertainties of Exposure
16           A number of uncertainties exist regarding exposure to subpopulations.  Further evaluation
17    of the effects of multiple routes and pathways of exposures and aggregate risks is needed.
18    Similarly, the effects due to co-exposures to other compounds with similar or different MO As
19    need to be evaluated. An estimate of multiple exposures is needed to know where along the
20    dose-response curve to place an incremental exposure to tetrachloroethylene. The size of any
21    increased risk will be different in dissimilar regions of the dose-response curve. This means that
22    a dose that is safe for an unexposed population will not necessarily be a safe dose if background
23    and other exposures are considered.  Until quantitative conclusions can be made for each
24    susceptibility factor, it will be very hard to consider the impacts of changes in multiple
25    susceptibility factors.
26           Although there is more information on early life exposure to tetrachloroethylene than on
27    other potentially susceptible populations, there remain a number of uncertainties regarding
28    children's susceptibility. For example, it is not clear to what extent tetrachloroethylene may pass
29    through the placenta in humans,  as shown in a rodent study (Ghantous et al., 1986).  Also, there
30    is limited information that evaluates nonoccupational exposures to tetrachloroethylene (e.g., in
31    homes and automobiles) for all susceptible populations or on additional exposures that may
32    modify an individual's exposure to tetrachloroethylene. Improved PBPK modeling  and
33    validation of these models will aid in determining how variations in metabolic enzymes affect
34    tetrachloroethylene metabolism.
35           Although inhalation is believed to be of most concern for tetrachloroethylene, the
36    pathways of exposure as well as the MO A for children are not well characterized. Inhalation
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-178      DRAFT-DO NOT CITE OR QUOTE

-------
 1    exposures may occur when tetrachloroethylene vapors are released from treated clothing or the
 2    clothing worn by occupationally exposed individuals, as well as when vapors are exhaled in the
 3    breath of exposed workers (ATSDR, 1997; Aggazzotti et al.,  1994a, b).  Dry-cleaned garments
 4    transported in an automobile may also lead to unexpectedly high levels of exposure to children
 5    who sit in the rear seats of cars, nearest to where most items are stored (Park et al., 1998; Chien,
 6    1997).  Inhalation exposure may also occur during showering or bathing as dissolved
 7    tetrachloroethylene in the warm tap water becomes volatilized (Rao and Brown, 1993; see
 8    Section 2.2.1).
 9          Although there is more information on early life exposure to tetrachloroethylene than on
10    other potentially susceptible populations, a number of uncertainties remain regarding children's
11    susceptibility. For example, though demonstrated in a study of placental transport by Ghantous
12    et al. (1986), it is not clear to what extent tetrachloroethylene may pass through the human
13    placenta. Also, there is limited information that evaluates nonoccupational exposures to
14    tetrachloroethylene (e.g., in homes and automobiles) for all susceptible populations or on
15    additional exposures that may modify an individual's exposure to tetrachloroethylene. Improved
16    PBPK modeling and validation of these models will aid in determining how variations in
17    metabolic enzymes affect tetrachloroethylene metabolism.
18          Although ingestion of tetrachloroethylene through breast milk may be a significant
19    pathway of exposure for some infants (see Sections 2.2.4 and 3.2), it has been suggested that if
20    these infants live adjacent to or in close proximity of dry cleaning facilities, the dose received
21    through ingestion of breast milk will become insignificant when compared with the inhalation
22    exposure and subsequent dose (Schreiber, 1997).
23          Certain foods have been found to be contaminated with tetrachloroethylene (see  Section
24    2.2.3).  Because children consume a high level  of dairy due to their need for calcium for bone
25    growth, the lipophilicity of tetrachloroethylene may pose a higher concern for children than for
26    adults.  Dairy intake is generally highest during infancy and decreases throughout life (NRC,
27    1993).
28          It is not clear to what extent dermal absorption is possible for children.  Although an
29    infant's skin has similar permeability to adults, a premature infant may have increased
30    permeability (Guzelian et al.,  1992).  Also, an infant has approximately double the ratio of
31    surface area to body weight compared to adults (NRC, 1993), which could imply increased
32    exposure during bathing and swimming, which has already been modeled for adults by Rao and
33    Brown (1993).
34          It is not known to what extent tetrachloroethylene is absorbed by a child and to which
35    organs the chemical and its metabolites may be distributed. A validated PBPK model is needed
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-179      DRAFT-DO NOT CITE OR QUOTE

-------
 1    that contains physiologic parameter information for infants and children, including the effects of
 2    maternal inhalation exposure and the resulting concentration in breast milk.
 O
 4    4.9.4.2.  Uncertainties of Effects
 5          More studies specifically designed to evaluate effects in early and later life stages are
 6    needed in order to more fully characterize potential life stage-related tetrachloroethylene toxicity.
 7    Because the neurological effects of tetrachloroethylene constitute the most sensitive endpoints of
 8    concern  for noncancer effects, it is quite likely that the early  life stages may be more susceptible
 9    to these outcomes than are adults.  Life stage-specific neurotoxic effects, particularly in the
10    developing fetus, need further evaluation. It is important to consider the use of age-appropriate
11    testing for assessment of these and other outcomes, both for cancer and noncancer outcomes.
12          The reduction in fertility seen in some studies (Eskenazi et al., 1991a, b; Rachootin and
13    Olsen, 1983; Sallmen et al., 1995) occurs by an unknown mechanism.  Altered sperm quality is
14    one possibility (Beliles et al., 1980; Eskenazi et al., 1991b), as is spontaneous abortion/fetal loss
15    occurring early in gestation without maternal knowledge of the pregnancy, thereby being
16    misclassified as infertility (see Section 4.7.1).
17          Data specific to the carcinogenic effects of tetrachloroethylene exposure during the
18    critical periods of development of experimental animals and  humans also do not exist. The
19    perinatal period, which encompasses the end of pregnancy and the early postnatal period, may be
20    the most susceptible window for exposure for tetrachloroethylene across species (Beliles, 2002).
21    Several of the adverse pregnancy outcome studies evaluated  exposure during a critical window,
22    the first trimester of the pregnancy, a critical window for exposure (MA DPH, 1997; Kyyronen
23    et al., 1989; Ahlborg, 1990; Taskinen et al.,  1989). Exposure during another developmental
24    period may not result in certain outcomes that occur from exposure during this critical window.
25    Alternately, another window of exposure may result in a different outcome than occurs during
26    the first trimester.
27
28    4.9.5. Conclusions on Susceptibility
29          There is some evidence that certain subpopulations may be more susceptible to exposure
30    to tetrachloroethylene. These subpopulations include early and later life stages, health and
31    nutrition status, gender, race/ethnicity, genetics, and multiple exposures and cumulative risk.
32          Cancer outcomes of concern for perinatal exposure are not well characterized in either
33    the human epidemiological or the experimental animal literature. Data-derived noncancer
34    outcomes of concern in early life stages are spontaneous abortion/fetal loss, mortality, and
35    neurological impairment.  As described above, the evidence for spontaneous abortion following
36    prenatal  exposures to tetrachloroethylene is well characterized in humans, and fetal loss is well
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-180      DRAFT-DO NOT CITE OR QUOTE

-------
 1    characterized in experimental animals. However, the human epidemiological data that support
 2    this conclusion do not provide information on the maternal dose to tetrachloroethylene that may
 3    have resulted in spontaneous abortion. Further, data from the experimental animal studies
 4    suggest that this finding may be a high-dose effect.  Together, this evidence suggests that
 5    reference values that are established on the basis of more sensitive neurological endpoints should
 6    mitigate potential risk of fetal lethality.
 7           Regarding postnatal hazard, the correlation of tetrachloroethylene exposure to childhood
 8    mortality is based on a single report in which a 2-year-old toddler died following exposure to
 9    dry-cleaned curtains (Gamier et al., 1996). Ambient dose was not well characterized, although
10    tissue levels of tetrachloroethylene indicated the possibility of bioaccumulation in the brain. Yet,
11    it is noted that early postnatal mortality was  not observed in animal studies. The overall
12    confidence in this endpoint, observed in a single individual, is minimal relative to predicting risk
13    for the broader population.
14           Likewise, evidence of neurobehavioral impairment in children is based on a minimal data
15    set,  consisting of only four children who resided in an apartment building with a dry cleaning
16    establishment and who demonstrated visual  system impairment (Schreiber et al., 2002).
17    Confidence in the results of this study for the assessment of risk to the broader population of
18    children is minimal due to the size of the studied population (i.e., only four individuals).  Given
19    the lack of information on these factors in referent children, it is difficult to evaluate the possible
20    contribution of other factors that may have contributed to the observed visual system impairment
21    in children (Storm and Mazor, 2004).  The lack of associations reported in Storm and Mazor
22    does not contradict the findings of Schreiber et al., given the differences in study protocol and
23    procedures (Hudnell and Schreiber, 2004). Decreased VCS has been observed in children who
24    resided in an apartment building with a dry cleaning establishment (Schreiber et al., 2002; NYS
25    DOH, 2005a). Children who attended a day care center adjacent to a dry cleaner did not
26    consistently show any effect on cognition or behavior when assessed 5 weeks after exposure
27    ceased  (NYS DOH, 2005b), and no effect on VCS was seen 5 years after exposure (NYS DOH,
28    2005c). Moreover, the findings of Till et al. (2001a, b, 2005) and Laslo-Baker et al. (2004),
29    although not definitive, further suggest that the developing fetus is susceptible to maternal
30    organic solvent exposures (see  Section 4.6.1).
31           Other subpopulations with potential for susceptibility to tetrachloroethylene include the
32    elderly, diminished health status, gender, race/ethnicity, and multiple/cumulative exposures.
33    There is suggestive evidence that there may  be greater susceptibility for exposures to the elderly.
34    Diminished health status (e.g., impaired kidney liver or kidney) will likely affect an individual's
35    ability to metabolize tetrachloroethylene, whereas certain nutrients may have a protective effect
36    on exposure.  Gender and race/ethnic differences in susceptibility are likely due to variation in

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-181      DRAFT-DO NOT CITE OR QUOTE

-------
 1    physiology and exposure, and genetic variation likely has an effect on the toxicokinetics of
 2    tetrachloroethylene. Multiple and cumulative exposures are likely to cause competition in
 3    metabolic capacity.  Future research should better characterize possible susceptibility for certain
 4    life stages or subpopulations.
 5
 6    4.10. SUMMARY OF HAZARD IDENTIFICATION
 7    4.10.1. Description of Effects and Exposure Levels at Which They Occur
 8           In the previous sections of this document the effects in each organ system were discussed
 9    pairwise in three categories:  humans/animals,  noncancer/cancer, and inhalation/oral.  The
10    summaries in Sections 4.4.3, 4.5.3, 4.6.3, and 4.7.3 pertain to each of the organ systems
11    individually. In this section, the same effects are integrated across organ systems, with the
12    primary subdivision being humans/animals, the secondary subdivision being noncancer/cancer,
13    and later  subdivisions being the exposure route and organ system. Section 4.10.2 summarizes
14    the potential modes of action. The dose levels, effect levels, and concentrations discussed here
15    are those  observed in the studies and are not corrected for continuous exposure or for human
16    equivalency. In Chapter 5, these corrections are made before deriving the RfCs and RfDs.
17
18    4.10.1.1.  Summary of Effects in Humans
19    4.10.1.1.1. Human noncancer effects.  The epidemiologic evidence indicates that the primary
20    targets of tetrachloroethylene noncancer toxicity  are the CNS, kidneys, liver, and developing
21    fetus.  The epidemiologic evidence supporting these inferences is derived primarily from studies
22    of tetrachloroethylene-exposed dry cleaners—with two studies reporting neurobehavioral effects
23    in residents living in housing located in close proximity to a dry cleaning facility—and from
24    studies reporting effects to the developing fetus in populations exposed to drinking water
25    contaminated with tetrachloroethylene and other  solvents.  In the drinking water studies, several
26    of the contaminants are congeners (e.g., tetrachloroethylene and trichloroethylene), which are
27    metabolized in the body to TCA and DCA.
28           The epidemiologic database is primarily composed of studies of a prevalence or cross-
29    sectional  design. Although a cohort study, by definition, is able to identify that exposure has
30    indeed occurred before disease, the available epidemiologic studies, including  those of a cross-
31    sectional  design, support a causal role of tetrachloroethylene.
32           In most cases, the number of study subjects was not large; however, the issue of sample
33    size affects the power of the study to detect underlying risk. Hence, observed effects are
34    considered noteworthy if chance and bias are minimized.  Furthermore, the number of studied
35    individuals in the epidemiologic studies of tetrachloroethylene is not any smaller than that of
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-182      DRAFT-DO NOT CITE OR QUOTE

-------
 1    epidemiologic studies for many chemicals identified in the U.S. EPA's IRIS. In fact, it is a rare
 2    case when hazard inferences are based on a large human population.
 3           Studies have adopted a number of methods to infer exposure to tetrachloroethylene.
 4    Exposure was ascertained in many cases indirectly by questionnaire, by job title, or by the
 5    subject living in a residence receiving drinking water containing tetrachloroethylene.  There is
 6    higher confidence of exposure potential to tetrachloroethylene in those studies using the
 7    occupational title of dry cleaner because tetrachloroethylene is the solvent of choice.
 8    Atmospheric montoring of dry cleaning facilities show 8-hr TWA concentrations in the range of
 9    10-20 ppm, with short-term exposures many times this value. However, the analyses that
10    combined dry cleaners with laundry workers carry more uncertainty than do studies whose
11    analyses included only dry cleaners, because laundry workers have a lower probability for
12    exposure to tetrachloroethylene. More rarely,  studies incorporated biological measures such as
13    tetrachloroethylene excretion in breath or urinary TCA.
14           When looked for,  exposure-response gradients have not  been observed, generally, and
15    this is another uncertainty associated with the inferences regarding a causal  association.
16    However, exposure misclassification may partially explain the lack of exposure-response
17    associations, hence, the lack of an exposure-response gradient does not diminish the observed
18    associations between tetrachloroethylene exposure and adverse  effects. Moreover, observed
19    effects cannot be considered to arise from confounding; investigators have taken great effort to
20    take into account the effects of smoking, age, and other factors through matching exposed
21    subjects with like controls or through statistical analysis of the data. This synthesis of the
22    epidemiologic evidence places greatest weight on those studies  where confounding has been
23    adequately controlled and identifies those studies where confounding may be a possible
24    explanation for observed results.  It is not possible to examine residual confounding or effects
25    not explained by variables adjusted for in the study's design or statistical analysis. Residual
26    confounding is an issue for both epidemiologic and toxicologic  studies and may explain
27    observed study findings.
28           In general, observations from the epidemiologic studies  are consistent with the biology of
29    tetrachloroethylene.   Tetrachloroethylene is lipophilic and would distribute to organs rich in
30    lipids, e.g., the CNS. The liver and kidney are considered target organs due to their ability to
31    metabolize tetrachloroethylene. Moreover, systemic effects, specifically to  the kidney, liver,
32    CNS,  and the developing  organism, have been observed in experimental animals. Thus, humans
33    do not appear as an exception to the systemic toxicity elicited by tetrachloroethylene. The
34    pattern of effects seen with tetrachloroethylene exposure is similar to that seen with other
35    solvents,  such as trichloroethylene.  These findings together support a causal role of
36    tetrachloroethylene in the development a number of systemic effects in humans.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-183      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2    4.10.1.1.2. Human cancer effects.  Overall, the epidemiologic evidence considered as a whole
 3    has associated tetrachloroethylene exposure with excess risks for a number of site-specific
 4    cancers. Studies of tetrachloroethylene and cancer showed positive associations between
 5    exposure and cancer of the lymphoid system, esophagus, and cervix, with more limited evidence
 6    for cancer of the bladder, kidney, and lung. For both lymphoid and esophageal cancer, excess
 7    risk was observed in studies of dry cleaners and in degreasers, populations who have exposure to
 8    tetrachloroethylene and other solvents.  In both cases, average risks were doubled as compared
 9    with those of referents.  Studies of drinking water exposure also support an association between
10    lymphoid cancer and tetrachloroethylene and other solvents, as do case-control studies that
11    assessed employment as a dry cleaner or laundry  worker.  Chance and confounding by smoking
12    are unlikely explanations for the observed excesses is risks.  Furthermore, the finding of elevated
13    risk for lymphohematopoietic system cancer incidence in a Swedish cohort of subjects who
14    developed suspected solvent-related disorders from organic solvent exposures supports the
15    findings of the tetrachloroethylene studies (Berlin etal., 1995). EPA judged that these data,
16    though limited and not consistently observed across all studies, suggested an association between
17    lymphoma and tetrachloroethylene.
18          For esophageal cancer, indirect evidence suggests that esophageal risk in these studies is
19    larger than that expected due to smoking.  Blair et al. (2003) stated that if the magnitude of the
20    difference in smoking for dry cleaners and the general population is in the range of 10% or less,
21    confounding from smoking in their study is unlikely to result in an RR greater than 1.2,  a finding
22    similar to that of Kriebel et al. (2004). Hence, the finding of a doubling in risk strongly suggests
23    occupational exposure as a contributing  (etiologic) factor. Observations from one case-control
24    study that was able to adjust for the effects of smoking support the cohort study findings.
25          The epidemiologic evidence also is suggestive of excess risks for cervical cancer, based
26    on observations  in dry cleaner and laundry worker cohorts, with few cases in the degreaser
27    studies.  Unfortunately, information is not available on possibly confounding factors such as
28    socioeconomic and lifestyle factors. Associations with kidney, bladder, and lung cancers and dry
29    cleaning employment or, more specifically, with tetrachloroethylene, were reported in recent
30    updates of the American and Nordic cohorts and in case-control studies. Conclusions are more
31    uncertain for these sites, because they are based either on heterogenous observations between
32    differing study designs or on a small number of available studies.  Overall, EPA judged these
33    findings as suggestive of an association.
34          Other reviews of the tetrachloroethylene epidemiologic evidence have concluded that
35    "little consistent evidence existed for an association with a specific cancer such as kidney"
36    (McLaughlin and Blot, 1997) or that there is "limited evidence" for cancers of the cervix,

                This  document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-184     DRAFT-DO NOT CITE OR QUOTE

-------
 1    esophagus, bladder, or kidney or for NHL (IARC 1995; Weiss 1995; Lynge et al., 1997; Ulm et
 2    al., 1996).  The Institute of Medicine (IOM) reviewed a similar—but not the same—body of
 3    epidemiologic literature as EPA (IOM, 2002). For example, the large four-country Nordic
 4    cohort of dry cleaners and laundry workers in Andersen et al. (1999) was not considered, nor was
 5    the most recent update of the cohort of American dry cleaner and laundry workers in Blair et al.
 6    (2003). The IOM committee concluded limited or suggestive evidence of an association between
 7    bladder and kidney cancers and tetrachloroethylene and dry cleaning solvents. No conclusions
 8    were presented on the epidemiologic evidence on esophageal and lung cancers and
 9    tetrachloroethylene, given that the committee could not reach a consensus opinion. Some
10    committee members believed that the overall evidence was limited by potential confounding
11    from smoking in cohort studies, whereas other committee members considered, for esophageal
12    cancer, the lack of other smoking-related cancers in cohort studies or, for lung cancer, the
13    presence of exposure-response relationships as supportive  of an conclusion of limited/suggestive
14    evidence. For other cause-specific cancers, the committee concluded that there was inadequate
15    or insufficient evidence to determine whether an association existed.
16          U.S. EPA's analysis is similar to those of the IOM  committee's on kidney and bladder
17    cancer, and, like some committee members, EPA considered the evidence on esophageal and
18    lung cancers as suggestive of an association.  U.S. EPA's conclusions on lymphoma are
19    supported, in part, by  observations from studies not considered by the IOM committee.
20          Mundt et al. (2003) reviewed a body of epidemiologic studies similar to U.S. EPA's and
21    presented conclusions as to whether an association was "likely" or "not likely." The authors
22    reported that little support existed on which to base a conclusion that tetrachloroethylene was a
23    strong occupational risk factor, but that "because of a number of positive findings suggested
24    from some of these epidemiological studies, one cannot definitely rule out the possibility that
25    associations between PCE [tetrachloroethylene] and some  cancers exist in humans." This
26    conclusion is consistent with conclusions in this  assessment, although it is expressed differently.
27          Although epidemiologists acknowledge that using guidelines to assess causation is an
28    imperfect process, some find that the aspects developed by A.B. Hill (1965) are helpful in
29    making these difficult judgments. Making this determination may precede  an understanding of
30    the underlying mechanisms and involves consideration of several aspects that would be
31    characteristic of a cause-and-effect relationship (Hill, 1965; Rothman and Greenland, 1998).
32
33       1. Strength of the observed association.  The finding of large and precise risks increases
34          confidence that the association is likely not due to chance, bias, or other factors. For
35          tetrachloroethylene, observed risks are generally modest, 2-fold or less. The observed
36          risks for esophageal cancer are not thought to be attributable to smoking or alcohol,
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-185      DRAFT-DO NOT CITE OR QUOTE

-------
 1           although insufficient data exist on socioeconomic factors important to cervical cancer to
 2           assess their impact on observed elevated risks for these site-specific cancers.

 3       2.  Consistency of the observed association. An inference of causality is strengthened when
 4           a pattern of elevated risks is observed across several independent studies. Excess risks
 5           for lymphoid cancers are seen in studies of dry cleaners, degreasers, and populations
 6           exposed to drinking water containing tetrachloroethylene—and in some cases
 7           trichloroethylene—and for esophageal cancer in studies of dry cleaners and degreasers.
 8           Excess risks for the other site-specific cancers are less consistently observed across these
 9           populations.
10
11       3.  Specificity of the observed association.  Traditionally, specificity has been described in
12           terms of one cause, one disease (Hill, 1965). This implies that one factor is associated
13           with the observed effect and no other effects are associated with the putative factor.
14           Tetrachloroethylene causes cancer at several sites in rats and mice; hence, there is no
15           expectation that tetrachloroethylene would be associated with only one human cancer.
16           Furthermore, many agents cause cancer at multiple  sites, and many cancers have multiple
17           causes.  Specificity has little meaning in this case, and therefore the lack of specificity
18           does not detract from the weight of the overall epidemiologic evidence.
19
20       4.  Temporal relationship of the observed association.  Causal relationships have temporality,
21           i.e., the cause precedes the effect.  Associations between tetrachloroethylene exposure
22           and several forms of cancer are established primarily by cohort and case-control studies,
23           in which the temporal relationship is well described. Many drinking water studies are
24           ecologic or prevalence studies, in which knowledge of the temporal relationship is
25           lacking. The exceptions are those studies assessing exposure to residents of Cape Cod,
26           MA, Woburn, MA, and Camp Lejeune, NC.  For this reason, the conclusions place
27           greater weight on the cohort and case-control studies.
28
29       5.  Biological gradient (exposure-response relationship). A clear exposure-response
30           relationship often suggests cause and effect.  For tetrachloroethylene, biological gradients
31           are only sporadically observed, though most studies identify exposure only as a
32           dichotomous variable (yes/no), or the number of site-specific cancers is often too small  to
33           identify biological gradients. For esophageal cancer, the dry cleaner studies of
34           tetrachloroethylene exposure (Blair et al., 2003;  Ruder et al., 2001) showed no clear
35           picture of exposure response relationships. Exposure response analyses in the drinking
36           water studies collectively suggest that greater exposure to drinking water contaminated
37           with tetrachloroethylene—and in a smaller number  of studies, with chlorinated solvents
38           both tetrachloroethylene and trichloroethylene—is associated with lymphoid cancer,
39           particularly leukemia and NHL (Aschengrau et al.,  1993; MA DPH 1997; Fagliano et al.,
40           1990; Cohnetal., 1994).
41
42       6.  Biological plausibility.  The mechanistic studies (discussed in another section in this
43           assessment) investigating tetrachloroethylene carcinogenic  effects in rats or mice and
44           their relevance to humans indicate that carcinogenesis is complex and likely involves
45           multiple mechanisms. Overall, the MOA for site-specific cancer is not know at this time.


                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-186      DRAFT-DO NOT CITE OR QUOTE

-------
 1       7.  Coherence. Coherence means that the causal interpretation of the data should not
 2           seriously conflict with generally known facts about the natural history and biology of the
 3           disease.  The strongest associations between tetrachloroethylene and human cancer are
 4           for the lymphopoietic system and esophagus, with more limited evidence for cervix,
 5           bladder, kidney, and lung.  Several of these organ systems are also targets for noncancer
 6           toxicity. The associations between cervical and esophageal cancer have no suitable
 7           animal counterparts.
 8
 9       8.  Experimental evidence (from human populations).  Experimental evidence (e.g., a
10           "natural experiment" that measures effects with exposure and in the absence of exposure)
11           is seldom available from human populations and exists only when conditions of exposure
12           are altered to create a kind  of quasi-experiment.  There are few data to evaluate this
13           criterion. The only study that does present information notes that childhood leukemia
14           cases appeared to be more evenly distributed throughout Woburn, MA, after closure of
15           the two wells contaminated with trichloroethylene and tetrachloroethylene (MA DPH,
16           1997).
17
18       9.  Analogy. The pattern of effects associated with tetrachloroethylene, particularly cancers
19           of the lymphoid system, cervix, kidney, and pancreas, has similarities to that of several
20           other chlorinated solvents and to mixed-solvent exposures.
21
22           Together, the evidence on tetrachloroethylene partially fulfills several of these criteria
23    and is suggestive of a cause-and-effect relationship between tetrachloroethylene and human
24    cancer. The body of human evidence is not sufficient to regard tetrachloroethylene as a known
25    human carcinogen.
26
27    4.10.1.1.3.  Susceptibility.  Many of tetrachloroethylene's metabolites are formed through the
28    enzyme system that also metabolizes ethanol  and other drugs and environmental pollutants.
29    Exposures to these chemicals can alter tetrachloroethylene's toxicity, not only altering the
30    pharmacokinetics of tetrachloroethylene but also the pharmacodynamics for toxicity. For
31    tetrachloroethylene's effects on the nervous system, kidney, and liver, the available limited data
32    suggest that a joint effect that reflects the addition of all exposures is plausible. In addition,
33    susceptibility to tetrachloroethylene's toxicity may vary among individuals because of both
34    intrinsic factors, (age, sex,  and genetic  factors, including metabolic polymorphisms) and
35    acquired factors (disease status, nutritional status).
36
37    4.10.1.2.  Summary of Effects in Animals
38    4.10.1.2.1.  Animal noncancer effects.  Tetrachloroethylene exposure in animals results in
39    toxicity to the liver, kidney, and nervous system and also causes developmental and reproductive
40    effects. These are all sites  of high  metabolic activity, and the CNS is also a lipid accumulation
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-187      DRAFT-DO NOT CITE OR QUOTE

-------
 1    site.  The immune system is potentially affected, but there are very few studies of these effects,
 2    and none of them are in intact animals. No information is available on the effects of
 3    tetrachloroethylene on the endocrine system in animals.  The effects have been discussed in
 4    several other review documents and are described in Sections 4.4.2.1, 4.5.2.1, 4.6.2.1, 4.6.2.2,
 5    and 4.7.2.1.
 6           In the liver, several measures of toxicity have been observed, such as increased liver
 7    weight, infiltration of fat, necrosis, peroxisome proliferation, polyploidy of hepatocytes, and
 8    increased triglycerides. In kidney, increased weight, hyperplasia, hyaline droplets,  and protein
 9    cast formation in tubules have been observed. In the CNS, alteration of brain neurotransmitter
10    levels, increased motor activity, and delayed reaction times to visual stimuli have been observed.
11    Fetal growth retardation, increased fetal mortality, and behavioral changes occurring after birth
12    to animals exposed in utero have been observed. Section 4.10.1.3 describes the doses at which
13    these effects occurred.
14
15    4.10.1.2.2. Animal cancer effects. Carcinogen bioassays in rats and mice have shown benign
16    and malignant tumors at various sites, as summarized in Sections 4.4.2.2, 4.5.2.2, and 4.8.2.
17    Data on the incidence and dose levels at which these effects occur are presented in  Section 5.3.2
18    and Tables 5-6,  5-8, and 5-9 (Chapter 5).  One study that used the oral route found liver
19    adenomas and carcinomas in mice. Two inhalation bioassays, both of which used both mice and
20    rats, found tetrachloroethylene-induced excess incidence of hepatocellular adenomas and
21    carcinomas (mice) and mononuclear  cell leukemia (rats). One of these studies found
22    hemangioendothelioma (mice) and the other found brain glioma, kidney tubular cell tumors, and
23    testicular interstitial cell tumors, all in male rats. Brain and kidney tumors are rare  in unexposed
24    animals, but they were found to be slightly elevated above control levels in only one of the two
25    inhalation studies.
26           Testicular tumors are extremely common in control animals, and the statistically
27    significant elevation in one of the bioassays was not considered by the investigators as related to
28    tetrachloroethylene exposure; however, the testicular tumors and kidney tumors are consistent
29    with exposure of rats to trichloroethylene, a structural analogue of tetrachloroethylene.
30    Mononuclear cell leukemias in rats were elevated in both inhalation bioassays. As  discussed in
31    Section 4.8.2.4.1, this is a relatively common tumor in nonexposed animals, with a possible site
32    concordance with  human lymphoid cancer, but because the mechanism of formation for both
33    human and animal hematopoietic  neoplasms is not understood in all of its complexity, they are of
34    possible relevance to humans.
35
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-188      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
      4.10.1.3.  Summary of Effect Levels
             Table 4-12 summarizes the lower ranges of air concentrations and oral doses at which
      effects occur in each of the organ systems discussed in this document.  The lowest of these air
      concentrations is 0.7 ppm (mean), which is associated with neurological effects observed in
      residents living above dry cleaning facilities (Altmann et al., 1995) in Germany. This is close to
      the mean concentration measured by Schreiber et al. (2002) for a similar exposure situation in
      the United States (0.4 ppm).  The lowest concentration showing effects in animals is 9 ppm,
      where liver toxicity was observed in mice which are more sensitive than the rat test strains.

             Table 4-12. Summary of low-effect levels of exposure to tetrachloroethylene
Organ System
Liver
Kidney
Neurological
Developmental,
reproductive
Other organs
Humans
Inhalation
(ppm)
12-16 (Table 4-1)
1.2 and 8. 8
(Table 4-3)
0.3 (Table 4-5)
1.2 (Table 4-8)
Exposure uncertain0
Oral
(ug/kg/day)

-
-
6, uncertain
(Table 4-8)
-
Animals
Inhalation
(ppm)
9-50, mice
(Table 4-2)
100 for cancer in
micea
100 ppm, miceb
37-90, mice and
gerbils (Table 4-6)
100, rats
(Table 4-9)
No conclusiond
Oral
(mg/kg/day)
100 (Table 4-2)
386 in mice for
cancer3
No studiesb
No chronic study
(Table 4-7)
-
-
12
13
14
15
16
17
18
19
20
21
22
23
24
25
     a See Section 4.4.2.2.
     b See Section 4.5.2.1.
     c See Section 4.8.1.
     d See Section 4.8.2.

     - = No studies available
            For subchronic oral exposures, the lowest dose for which adverse effects occurred in
     animals is 100 mg/kg-day.  These data come from the Buben and O'Flaherty (1985) gavage
     study, where the exposure duration was 6 weeks. There are no reliable data for humans exposed
     orally.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-189      DRAFT-DO NOT CITE OR QUOTE

-------
 1    4.10.2. Characterization of Cancer Hazard
 2              Tetrachloroethylene is "Likely to be a human carcinogen by all routes of exposure"
 3    within the framework of the 2005 carcinogen risk assessment guidelines (U.S. EPA, 2005b).
 4    This conclusion is based on reported associations in epidemiologic studies between
 5    tetrachloroethylene exposure and site-specific cancers and by the induction of site-specific
 6    tumors in rodents given tetrachloroethylene by oral gavage and inhalation. Several metabolites
 7    of tetrachloroethylene also  are considered rodent carcinogens. Metabolites from the oxidative
 8    pathway, TCA and DCA, produce liver tumors in mice, and DCA also induces liver tumors in
 9    rats. Metabolites from the GST pathway have not been tested in a standardized 2-year bioassay.
10    This hazard characterization is discussed in more detail in Section 4.10.2.2. The context for this
11    statement is described in the following section.
12
13    4.10.2.1. Background
14          As specified in the guidelines, the descriptor "Likely to be carcinogenic to humans"
15    expresses the conclusion regarding the weight of evidence for carcinogenic hazard potential, and
16    it is presented only in the context of a weight of evidence narrative.  Although the term "likely"
17    can have a probabilistic connotation in other contexts, its use as a weight of evidence descriptor
18    does not correspond to a quantifiable probability of whether the chemical is carcinogenic.  The
19    five recommended standard hazard descriptors are as follows:
20          "Carcinogenic to humans"
21          "Likely to be carcinogenic to humans"
22          "Suggestive evidence of carcinogenic potential"
23          "Inadequate information to assess carcinogenic potential"
24          "Not likely to be carcinogenic to humans"
25
26          These descriptors are not unlike those used by the IARC, NTP, and other health agencies
27    that weigh carcinogenicity evidence. If there are no or insufficient pertinent data, then the
28    descriptors "Inadequate information to assess carcinogenic potential" or "Suggestive evidence of
29    carcinogenic potential"  are used. If the evidence is stronger, as is  the case with
30    tetrachloroethylene, the descriptor "Likely to be carcinogenic to humans" is used; convincing
31    evidence, usually conclusive demonstration of causality in epidemiological studies, would
32    support "Carcinogenic to humans."  On the other hand, if the conclusion is negative (i.e., strong,
33    consistent and compelling information indicating the absence of human health hazard), the agent
34    would be described as "Not likely to be carcinogenic to humans."  Thus, going down the list of
35    descriptors from "Carcinogenic to humans" to "Inadequate information to assess carcinogenic
36    potential" indicates a decrease in the level of evidence or of a human health hazard. In summary,
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-190      DRAFT-DO NOT CITE OR QUOTE

-------
 1    use of the weight of evidence descriptor "Likely to be carcinogenic to humans" for
 2    tetrachloroethylene is intended to communicate that the available information indicates the
 3    presence of a human health hazard.
 4          The weight-of-evidence conclusion represented by the top three levels of evidence is
 5    related to but distinct from the quantitative dose-response assessment/conclusions in that the
 6    judgment that an agent is a human carcinogen does not guarantee adequate data to quantitatively
 7    estimate human risk. Notably, evaluation of an agent that is judged a likely human carcinogen
 8    may offer data conducive to estimating human risk. Indeed, dose-response assessments are
 9    generally completed for agents considered "Carcinogenic to humans" and "Likely to be
10    carcinogenic to humans."  Section 5.4 provides the dose-response analyses for
11    tetrachloroethylene.
12
13    4.10.2.2. Hazard Characterization for Tetrachloroethylene
14          Overall, the epidemiologic evidence considered as a whole has associated
15    tetrachloroethylene exposure with excess risks for a number of site-specific cancers.  Lymphoid
16    cancer is now recognized as a combination of NHL, Hodgkin's disease, lymphosarcoma,
17    multiple myeloma,  and lymphatic leukemia.  Cohort studies of dry cleaner and laundry workers
18    and of degreasers suggest excess risks of lymphoid cancers, as do case-control studies of
19    drinking water exposure and occupational exposure.  Exposure to a number of solvents is likely
20    in most of the case-control studies; however, these solvents have a qualitatively similar profile of
21    metabolites, although quantitative differences are expected. One study of exposure only to
22    tetrachloroethylene in drinking water reported a statistically significant association, based on a
23    small number of exposed cases, between leukemia and a residence receiving tetrachloroethylene-
24    contaminated water (Aschengrau et al., 1993).
25          Both cohort and case-control studies of dry cleaning workers support an association
26    between tetrachloroethylene and excess risk of esophageal  cancer.  Recent updates of dry
27    cleaners and laundry worker cohorts (Ruder et al., 2001; Blair et al., 2003) carry great weight in
28    this evaluation because dry cleaners are predominately exposed to tetrachloroethylene and a
29    statistically significant elevated mortality from this cancer continued to be observed. Little
30    weight is given to the Lynge et al. (2006) study due to potential biases that likely dampen their
31    observations and because of these biases it is considered a null study. No clear patterns are seen
32    in the tetrachloroethylene studies for either level or duration of exposure and response.
33          The possibility that other exposures such as smoking and alcohol consumption may
34    potentially confound the associations observed in Blair et al. (2003) and Ruder et al.  (2001)
35    cannot be directly addressed. Indirect evidence suggests that the esophageal risk in these studies
36    is larger than that expected due to smoking. Moreover, the case-control study by Vaughan et al.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-191       DRAFT-DO NOT CITE OR QUOTE

-------
 1    (1997) provides support for an association with tetrachloroethylene; a statistically significant
 2    association was observed between tetrachloroethylene exposure and esophageal cancer after
 3    adjustment for smoking, alcohol, and socioeconomic status.  Support by analogy is derived from
 4    the finding of excess esophageal cancer incidence in a cohort that was occupationally exposed to
 5    trichloroethylene (Hansen et al., 2001).
 6          More deaths from cervical cancer were observed among American and Nordic female dry
 7    cleaners or laundry workers than were expected.  The observation of exposure-response trends in
 8    the studies that presented this information (Blair et al., 1990; Ruder et al., 1994, 2001) support an
 9    association with dry cleaning.  Lack of data on socioeconomic status—a proxy for exposure to
10    the human papilloma virus, a known risk factor for cervical cancer—indicates great uncertainty
11    for asserting this association with tetrachloroethylene exposure.
12          There is also some support, albeit less than for the sites above, for an association between
13    dry cleaning occupations and other cancers, specifically,  cancers of the kidney, bladder, and lung.
14    These findings are based on heterogenous observations of differing study designs, on a small
15    number of available studies, or on small  numbers of study subjects.
16          An open question in the dry cleaner studies is the  specificity of exposure to
17    tetrachloroethylene. Elevated mortality for cancer of the  esophagus and cervix were observed in
18    two cohorts that were considered to have primarily tetrachloroethylene exposures.  However,
19    individuals who may have had exposures to other dry cleaning solvents were also included in
20    these studies.  There are only three studies of cancer incidence or mortality among degreasers
21    exposed to tetrachloroethylene, and they are of a  small number of subjects with
22    tetrachloroethylene exposure and, consequently, of few site-specific cases.  These studies are
23    only now collectively beginning to provide insight on associations between tetrachloroethylene
24    exposures and site-specific cancers.
25          In rodents, hepatocellular carcinomas in both male and female B6C3F1 mice have been
26    observed following inhalation and oral gavage exposure,  and the same tumor response was
27    observed in male and female Crj :BDF1 mice after inhalation exposure.  MCL, a common tumor
28    site in treated  and untreated F344 rats, was significantly increased in both males and females in
29    inhalation bioassays carried out in both Japan and the United States. Malignant liver
30    hemangiosarcomas and splenic hemangioendotheliomas were also observed in male mice in the
31    Japan bioassay. In the U.S. inhalation study in F344 rats, a small excess incidence of rare renal
32    tubule cell carcinoma and adenoma was  observed in males. Testicular interstitial cell tumors, a
33    common tumor in treated and untreated F344 rats, were significantly elevated in the U.S.
34    bioassay, and  an elevation of rare brain glioma incidence was also observed in these rats.
35          The major metabolite of tetrachloroethylene in humans and rodents, TCA, is carcinogenic
36    by gavage in male mice, and another metabolite in rodents, DCA, is also carcinogenic by gavage

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-192      DRAFT-DO NOT CITE OR QUOTE

-------
 1    in male mice.  The MO A of tetrachloroethylene or its metabolites in the likely causation of
 2    cancer is not known. Extensive testing of tetrachloroethylene showed that it does not damage
 3    DNA except in a few studies of conditions where the GSH metabolites would be generated, and
 4    it induces chromosome aberrations in some studies.  Several of the known or putative oxidative
 5    metabolites are mutagenic.  Metabolism through kidney GSH conjugation produces
 6    trichlorovinyl  GSH and trichlorovinyl cysteine, which were mutagenic in the Salmonella test but
 7    which have not been tested in mammalian genotoxicity assays. The latter metabolite reacts with
 8    beta lyase in the kidney to produce reactive thiol compounds.  Other metabolites, including
 9    reactive sulfoxides, can be also be produced by FMO3 or CYP3 A metabolism of TCVC. This is
10    a plausible MOA for the rare rat kidney tumors observed in one bioassay. However, the MO As
11    for human tumors and the mice liver tumors is still unknown.  Therefore, there is little
12    mechanistic basis for choosing a low-dose extrapolation model.
13
14    4.10.3.  Mode-of-Action Summary
15          The MOA for tetrachloroethylene-induced carcinogenesis is not yet fully characterized,
16    completely tested, or understood.  The database for hepatocarcinogenesis is especially limited
17    with regard to chemical-specific studies.  The available evidence points to multiple MO As being
18    involved. Furthermore, although there is some evidence for common MOAs, there is also
19    evidence indicating differences in the potential MOA across organ systems.
20          Tetrachloroethylene exposure has been associated with peroxisome proliferation in
21    rodent liver and kidney.  Compelling insight into the hypothesized MOA by which certain
22    chemicals induce proliferation of peroxisomal organelles and possibly cancer—specifically, the
23    focus has been on liver cancer—was disclosed by the discovery of the PPAR receptors, a class of
24    nuclear receptors closely related to the thyroid hormone and retinoid receptors that were first
25    shown to be activated by peroxisome proliferators by Issemann and Green (1990). To date, three
26    known subtypes of PPAR have been described in mammals:  PPAR gamma, PPAR-5, and
27    PPAR-a.
28          Evidence exists to support PPAR-a as being the specific receptor that is necessary for
29    transient cell proliferation and its role in hepatocarcinogenesis has been the subject of several
30    investigations, although most studies have explored the potent agonist Wy-14,643 (Lee et al.,
31    1995; Peters et al., 1997a; Corton et al., 2000). Activation of the steroid-like PPAR receptor
32    regulates transcription of the genes.  The PPAR target genes encode enzymes involved in
33    peroxisomal and mitochondrial beta-oxidation and ketone body synthesis as well as certain P450
34    4A enzymes, fatty-acid binding  proteins, apolipoproteins, lipoprotein lipase, malic enzyme, and
35    phosphoenolpyruvate carboxykinase (Issemann et al.,  1993; Desvergne and Wahli, 1995; Reddy
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-193      DRAFT-DO NOT CITE OR QUOTE

-------
 1    et al., 1986). The PPAR genes are expressed in a wide range of tissues, and PPAR occurs across
 2    species.
 3          Several recent studies have expanded the scientific understanding of the PPAR-a mode of
 4    action proposed by Klaunig et al (2003; see Caldwell et al., 2008).  First, Yang et al (2007)
 5    demonstrated that PPAR-a activation in hepatocytes induces peroxisome proliferation but not
 6    liver tumors. The approach entailed targeting expression of PPAR-a to hepatocytes by placing
 7    the VP16 PPAR-a transgene gene under control of the liver enriched activator protein (LAP)
 8    promoter. LAP-VP16 PPAR-a transgenic mice showed a number of PPAR-a -mediated effects:
 9    decreased serum triglycerides and free fatty acids, peroxisome proliferation, enhanced
10    hepatocyte proliferation, and induction of cell-cycle and PPAR-a target genes. However,
11    compared with wild-type mice  exposed to Wy-14,643, the extent of hepatomegaly was reduced
12    and no hypertrophy or eosinophilic cytoplasms was seen in LAP-VP16 PPAR-a mice.  Also in
13    contrast with wild-type mice exposed to Wy-14,643, no evidence of non-parenchymal cell
14    proliferation was observed in the LAP-VP16 PPAR-a transgenic mice. Moreover, at one year of
15    age no evidence  of preneoplastic hepatic lesions or hepatocellular neoplasia was observed in
16    LAP-VP16 PPAR-a transgenic mice.  As noted by the authors, PPAR-a activation only in mouse
17    hepatocytes is sufficient to induce peroxisome proliferation and hepatocyte proliferation but
18    "... is not sufficient to induce liver tumors."
19          Secondly, Ito et al. (2007) found that DEHP, a proposed robust example of PPAR-a
20    agonism-induced hepatocarcinogenesis, yields liver tumors in a 2-year study in PPAR-a knock-
21    out mice. This study demonstrates the limitations, cited by the FIFRA SAP,  of drawing
22    conclusions from the one-year bioassays of high doses of Wy-14,643 referenced above (e.g.,
23    Peters 1997).  It  supports the view that knock-out mouse bioassays should be carefully
24    characterized and conducted for 2 years to assess whether PPAR-a activation is indeed necessary
25    for induction of liver cancer. Thus, although a weak peroxisome proliferator, chemical-specific
26    data supporting the hypothesis  that PPAR-a activation plays a prominent or essential role in
27    tetrachloroethylene tumor induction are lacking.  Critical review of the scientific literature
28    reveals significant data gaps regarding the relationship between the PPAR-a  activation and
29    neoplasia induced by peroxisome proliferators as a group and tetrachloroethylene specifically. If
30    PPAR-a does play a role in tetrachloroethylene-induced tumorigenesis, available information
31    suggests relevance to humans cannot be ruled out.
32          Although accumulation of alpha-2u-globulin has been suggested as an MOA leading to
33    nephropathy that culminates in the formation of renal tumors, the available data do not support
34    this MOA for tetrachloroethylene. Indeed, the available data suggest that alpha-2u-globulin
35    accumulation following tetrachloroethylene exposure occurs only at doses higher than those used
36    in the carcinogenicity bioassays.  In addition, tetrachloroethylene does not meet all the criteria to

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-194      DRAFT-DO NOT CITE OR QUOTE

-------
 1    suggest that alpha-2u-globulin accumulation is the MOA.  Therefore, an important role for
 2    alpha-2u-globulin accumulation in tetrachloroethylene-induced renal tumors is highly unlikely.
 3           The role of genotoxicity in tetrachloroethylene liver cancer, an effect that is thought to be
 4    related to products of CYP metabolism, is uncertain. The available data suggest that several of
 5    the chloroacid metabolites are mutagenic. In particular, tetrachloroethylene oxide, the primary
 6    metabolite hypothesized to be formed during CYP metabolism, is a known bacterial mutagen; it
 7    has not been tested in mammalian systems, although genotoxicity in such tests could be
 8    anticipated based on the expected DNA reactivity of the epoxide moiety. GSH-derived
 9    intermediates also exhibit genotoxicity. The glutathione conjugation of tetrachloroethylene in
10    the kidney leads sequentially to S(l,2,2-trichlorovinyl)glutathione and
11    S(l,2,2-trichlorovinyl)cysteine—TCVG and TCVC. TCVC can be further processed by beta-
12    lyase to yield an unstable thiol, 1,2,2-trichlorovinylthiol, that may give rise to  a highly reactive
13    thioketene which can form covalent adducts with cellular nucleophiles including DNA. TCVC
14    can also undergo FMO3 or P450 oxidation to reactive intermediates; additionally, sulfoxidation
15    of both TCVC and its N-acetylated product occurs, resulting in reactive metabolites (Ripp et al,
16    1997, 1999; Werner et al., 1996). While most of these intermediates have not been characterized
17    for mutagenic potential, TCVG, TCVC and NAcTCVC are clearly mutagenic in Salmonella tests.
18    In addition, tetrachloroethylene exhibited mutagenicity in Salmonella in the few studies of
19    conditions that could generate GSH-derived metabolites and, following in vivo exposures,
20    induces SSB and DNA binding in kidney. A mutagenic MOA is therefore likely to play a role in
21    the development of tetrachloroethylene-induced renal cancer. A mutagenic MOA could also
22    play a role in the formation of other tumors such as brain gliomas and  MCL in rats, if sufficient
23    concentration of the potentially genotoxic metabolites arising from the initiation of GSH
24    conjugation occurs at target sites. In the kidney, the conjugates are concentrated after being
25    transported from the liver, in addition to being generated on site in that target tissue.  Further
26    processing by beta lyase, FMO3 or CYPs yields mutagenic products in what could be
27    sufficiently high target tissue concentrations.  Beta lyase is also found  in the brain and other
28    tissues.
29           In summary, cancers resulting from tetrachloroethylene exposures are likely due to
30    multiple MO As that vary from target tissue to target tissue. The MO As for tetrachloroethylene-
31    induced cancers are not yet well understood. The potential MO As for cancer discussed in this
32    section are summarized in Table 4-13 below. The implications of these potential MO As to risk
33    extrapolation to concentrations lower than those producing effects in animal bioassays are
34    explored in Table 4-14.  These considerations underlie the discussion in Chapter 5  of the
35    uncertainties in modeling risk at low concentrations.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-195      DRAFT-DO NOT CITE OR QUOTE

-------
Table 4-13. Summary of potential modes of action for cancer
Organ, tumor type
Kidney
adenocarcinoma in
male rats.
Potential MOA
Mutagenicity. Via the GSH pathway in
liver and kidney, glutathione and
cysteine conjugates are produced.
Identification of the urinary
mercapturate metabolite in humans and
rodents shows that this pathway is
operative in both species.
Tubular cell necrosis and
nephrotoxicity followed by hyperplasia
and neoplastic transformation.
Accumulation of arpha-2u-globulin, a
male rat specific protein, in hyaline
droplets of tubule cells leading to a
specific pattern of nephropathy
(involving tubule degeneration,
compensatory hyperplasia, neoplastic
transformation).
PPAR-a receptor activation in kidney.
Evidence for MOA
Reactive thiol and
sulfoxide compounds
are formed in kidney
from TCVG/TCVC
that are mutagenic.
In vitro kidney damage
from processing of
GSH metabolites to
cytotoxic species.
Hyaline droplets
observed in short-term
(42-day) experiments
BUT ONLY at high
dose (1,000 mg/kg-
day).
Peroxisome
proliferation occurs in
kidney, but no
evidence supports
causal association with
tumorigenesis
Limitations/evidence against MOA
Several metabolites (TCVG/TCVC,
NAcTCVC) are mutagenic in bacteria
but are not characterized in other
systems; other GSH metabolites have
not been identified and characterized.
Mutagenicity is commonly assumed to
contribute to cancer.
Relationship between non-genotoxic
kidney necrosis and carcinogenicity has
not been studied.
Tetrachloroethylene-induced renal
nephropathy also occurs in female rats
and mice of both sexes. Hyaline
droplets are not observed in 28-day
experiments at lower concentrations
(400 ppm) that induce kidney tumors.
alpha-2u-globulin protein not
identified. Two features characteristic
of this MO A do not occur: (1)
mineralization of tubules in the chronic
bioassay; (2) male-rat-only
nephrotoxicity.
Peroxisome proliferation is greater in
mouse kidney than rat kidney; however
male rats develop kidney tumors.
Weight of evidence
Experimental evidence
for this pathway and
identification of its
urinary mercapturate
metabolites supports
the MOA.
Inadequate
tetrachloroethylene-
specific data exist to
support this potential
MOA.
Kidney tumors are not
specific to the male rat.
This MOA only occurs
at concentrations
greater than those
necessary for tumor
induction.
Little evidence to
support this MOA.

-------
o
Oi
Table 4-13. Summary of potential modes of action for cancer (continued)
This document is a draft for review purposes only and does not constitute Agency policy
06/08 4-197 DRAFT: DO NOT CITE OR QUOTE

Organ, tumor type
Liver hepatocellular
carcinoma in male and
female mice.
Liver and spleen
hemangiosarcoma in
male mice.
Blood mononuclear
cell leukemia in male
and female rats.
Brain glioma in male
rats. (Rare)
Potential MOA
PPAR-a receptor activation
Mutagenicity produced by P450
pathway in liver.
Necrosis of liver cells leads to
compensatory hyperplasia, excess
mutation rate and neoplastic
transformation.
Mutagenicity of P450 and/or GSH-
derived reactive intermediates.
Mutagenicity. There is a potential for
circulating genotoxic metabolites, or
their precursors produced by liver
metabolism, to be further processed to
reactive products in other tissues.
Perc incorporation into brain cell
membranes, disruption of membrane
ion channels, interference with
neurotransmitters.
Evidence for MOA
PPAR-a is found in
human liver, testis,
pancreas.
Tetrachloroethylene
oxide and DCA are
mutagenic and other
metabolites
(trichloroacetyl
chloride) are capable of
binding to DNA.
No evidence, but
plausible hypothesis.
Tetrachloroethylene
oxide and DCA are
mutagenic and other
metabolites
(trichloroacetyl
chloride) are capable of
binding to DNA.
DCVC, a metabolite of
trichloroethylene,
induced blood
dyscrasias (aplastic
anemia, DNA alteration
in bone marrow, lymph
nodes and thymus in
calves.
Perc induces persistent
changes in fatty acid
composition of brain in
rodents.
Limitations/evidence against MOA
Perc1 only weakly activates PPAR-a.
There are no studies in PPAR-a null
mice evaluating PERC-induced tumors.
Genotoxicity is commonly assumed to
contribute to carcinogenesis.
Cytotoxicity and compensatory
hyperplasia are not observed at
bioassay doses.
Genotoxicity is commonly assumed to
contribute to carcinogenesis.
TCVC, the analogous perc metabolite,
did not produce bone marrow injury in
calves in a single study.
No information about cell membrane
changes at low doses or longer than
acute exposures. Connection with
carcinogenesis is speculative.
Weight of evidence
Some, but not
conclusive, evidence
for this MOA
Some, but not
conclusive, evidence
for this MOA
Some evidence against
this MOA.
No studies have
explored this MOA for
tetrachloroethylene-
induced
hemangiosarcomas.
No further studies have
explored this MOA.
No studies have
explored this MOA.

-------
       Table 4-13. Summary of potential modes of action for cancer (continued)
Organ, tumor type
Brain glioma in male
rats. (Rare)
(continued)
Potential MOA
Mutagenicity of metabolites generated
in situ.
Evidence for MOA
Beta lyase has been
observed in rat brain
tissue. Therefore a
potential exists for
metabolizing
TCVG/TCVC to
reactive compounds.
Limitations/evidence against MOA
Pathway from genotoxic metabolites to
brain cancer is not defined.
Weight of evidence
Little evidence for this
MOA.
1 Perc = Tetrachloroethylene

-------
o
Oi
Table 4-14.  Quantitative Implications of different modes of action: candidate modeling approaches
Organ
Kidney
adenocarcimoma in
male rats.
Liver hepatocellular
carcinoma in male and
female mice.
Liver and spleen
hemangiosarcoma in
male mice.
Potential MOA
Mutagenicity of GSH-
derived metabolites.
Necrosis, hyperplasia,
neoplastic
transformation.
Alpha-2ji in hyaline
droplets in male rats.
PPAR-a activation.
PPAR-a receptor
activation, peroxisome
proliferation,
hyperplasia, neoplastic
transformation.
Mutagenicity of P450
metabolites.
Necrosis, hyperplasia,
neoplastic
transformation.
Mutagenicicity (e.g.,
of GSH metabolites).
Weight of evidence
Evidence for MOA.
No direct evidence.
Not likely to occur at
bioassay doses.
Little evidence.
Some, but not
conclusive, evidence
for this MOA.
Some, but not
conclusive, evidence
for this MOA.
Evidence against this
MOA. Probably not
occurring.
No studies have
explored this MOA.
Inference about low-
dose shape
Linear no threshold.
Non-linear model
resulting from
distribution of
thresholds for
individuals in a
population.
MOA does not occur in
humans.
Too little evidence to
inform modeling
approach.
Receptor binding shape
could be linear, sub-
linear or supra-linear,
depending on binding
constants.
Linear no threshold.
Too little evidence to
inform modeling
approach .
Linear no threshold.
Candidate
models
POD, linear
extrapolation.
Log probit or
Log logistic.


Model for
receptor binding
induced by gene
activation.
(Kohnetal.,
1993).
POD, linear
extrapolation.

POD, linear
extrapolation.
Limitations of extrapolation
procedure
Model is commonly used. As
implemented in the current BMDL
software, upper confidence limits on
dose are not calculated.
Variability of thresholds is not the
same in humans as in animals. BMDS
implementation of models is not
believed to be accurate.


Binding data to inform modeling
approach not available.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.

Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
oo S3'



   I

   I

   §
   ***.

   £3'
  1
   TO'
   l
   2
   •••2
   I
   TO
H I

O >
HH Oq
H TO
O
H
W

-------
       Table 4-14. Quantitative Implications of different modes of action:  candidate modeling approaches (continued)
Organ
Blood mononuclear cell
leukemia in male and
female rats.
Brain glioma in male
rats (rare tumor).
Potential MOA
Mutagenic GSH
metabolites.
Perc1 incorporation
into cell membranes.
Mutagenicity of
metabolites generated
in situ.
Weight of evidence
Only two
conflicting,
inconclusive reports.
No evidence.
Little evidence but
plausible.
Inference about low-
dose shape
Linear no threshold.
Linear no threshold.
Linear no threshold.
Candidate
models
POD, linear
extrapolation.
POD, linear
extrapolation.
POD, linear
extrapolation.
Limitations of extrapolation
procedure
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
Model is commonly used. As
implemented, upper confidence limits
on dose are not given.
1 Perc = Tetrachloroethylene

-------
 1    4.10.4. Rationale for Selection of Dose Metric
 2    4.10.4.1. Liver
 3           There are several possible choices to consider as the dose metric for tetrachloroethylene-
 4    induced liver toxicity and carcinogenicity. First is administered dose or exposure concentration.
 5    Tetrachloroethylene hepatotoxicity is associated, however, with cytochrome P450 metabolism
 6    occurring in the liver.  Several investigators have reported hepatotoxicity in rodent studies to be
 7    directly related to metabolism. Because liver toxicity, including carcinogenicity, is generally
 8    considered to be caused by metabolites rather than by the parent compound, choosing the
 9    specific chemical  species responsible for adverse effects in the liver would be the preferred
10    choice over administered dose/exposure concentration, particularly since tetrachloroethylene
11    metabolism is nonlinear with dose of parent compound, with the percent metabolized decreasing
12    with increasing dose.
13           TCA is considered a key product of this P450 oxidation pathway.  TCA is the major
14    urinary metabolite from tetrachloroethylene biotransformation, and it is the principal metabolite
15    in the systemic circulation.  TCA, like the parent compound, also causes liver toxicity and
16    carcinogenicity in mice. Therefore, the second plausible dose metric for use with the liver target
17    tissue is the concentration and AUC for TCA.
18           The MOA for tetrachloroethylene-induced liver toxicity and carcinogenicity is not clear,
19    however, and whether TCA is the sole contributory metabolite to tetrachloroethylene-induced
20    hepatotoxicity and cancer is unknown.  Other possible P450 oxidation products, such as DCA,
21    are also associated with liver toxicity when administered directly. In addition, it is not known
22    whether reactive intermediates such as tetrachloroethylene oxide and trichloroacetyl chloride are
23    involved in tetrachloroethylene-induced liver toxicity.
24           Hepatic toxicity correlates better with metabolism than with administered dose.  In other
25    words, a better linear relationship exists between metabolism and hepatotoxicity than between
26    administered dose and hepatotoxicity. Because of the uncertainty about which metabolite
27    species are involved in causing liver toxicity and the degree to which they are involved, the most
28    appropriate dose metric is considered to be total metabolism.  Production of the putative
29    metabolites is then considered to be directly proportional to the total amount of
30    tetrachloroethylene metabolized, a reasonable assumption.
31
32    4.10.4.2. Kidney
33           More than one choice was considered for the kidney target organ dose metric. The most
34    simplistic dose metric is administered dose or exposure concentration.  However, renal toxicity,
35    including kidney cancer, is associated with metabolism. It is  specifically associated with GSH-
36    dependent metabolism, although P450 metabolism could potentially contribute to renal toxicity.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-201       DRAFT-DO NOT CITE OR QUOTE

-------
 1    It is generally accepted that the interorgan GSH-dependent pathway, which also occurs
 2    completely in the kidney target organ, results in the production and accumulation of mutagenic
 3    metabolites.  Unfortunately, the measurements of GSH-dependent metabolism are from in vitro
 4    studies or are of urinary excretion products, and are not representative of the toxic species in
 5    vivo.
 6          The total production of the thioketene reactive intermediate divided by the volume of the
 7    kidney has been proposed as the dose metric for use in the PBPK model for kidney target organ.
 8    In order to use this dose metric, however, several assumptions must be made. One assumption
 9    might be that all GSH conjugate formed in the liver is transported to the kidney. Excretion of N-
10    acetyl TCVC is the measurement used to represent flux through the pathway. Clearance of
11    TCVC would be modeled, and production of toxic metabolites would be assumed to be
12    proportional to overall flux.  Unfortunately, the flux through the beta lyase and FMO3/CYP3A—
13    or sulfoxide-producing branches of the pathway—has not been measured in vivo.
14          Better methods are needed to quantitate the reactive species that are generated during
15    tetrachloroethylene metabolism, particularly in the beta lyase, FMO3 and CYP3A sections of the
16    pathway, to improve the usefulness of data in development and validation of PBPK models.  The
17    amounts of N-acetyl  TCVC excreted in urine represent only a portion of the flux through the
18    overall GSH-dependent pathway.  This excretion of mercapturate does not represent the
19    processing of TCVC and also N-acetyl TCVC to the reactive and toxic products important to
20    toxicity.  The fraction of overall flux represented by the excretory product is simply unknown,
21    and how it is related to the fraction processed through the beta lyase branch of the path is also
22    unknown. Several products formed in the pathway are unstable and reactive, and, therefore, they
23    are difficult to quantitate. Because the quantitative information about toxic metabolites from the
24    GSH-dependent pathway is not available, and there is no way of knowing whether the
25    measurement of excretory mercapturate is proportional to production of the toxic species
26    produced in this pathway, the administered exposure and total metabolites are both considered as
27    dose metrics  (see Section 5.3.3.2). Production of the putative metabolites, then, is considered to
28    be directly proportional to the total amount of tetrachloroethylene metabolized.
29
30    4.10.4.3. Hematopoietic Target Organ
31          Tetrachloroethylene causes mononuclear cell leukemia in rats. Although the specific
32    mechanism of leukemogenesis in rats is not understood, neither is it well understood in humans.
33    Whether the parent compound, a metabolite, or several metabolites are involved in the
34    tetrachloroethylene induction of the leukemia is not known. In the case of the
35    tetrachloroethylene congener trichloroethylene, the comparable DCVC conjugate metabolite has
36    been associated with causing adverse effects to the hematopoietic system. A possible link to

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-202      DRAFT-DO NOT CITE OR QUOTE

-------
 1   MO A for tetrachloroethylene-induced MCL in rats comes from early reports of toxicity of
 2   cysteine S-conjugates where DCVC, the trichloroethylene metabolite, was implicated in
 3   induction of aplastic anemia and marked biochemical alteration of DNA in bone marrow, lymph
 4   nodes, and thymus in calves (McKinney et al., 1957; Schultze et al., 1959; Bhattacharya and
 5   Schultze, 1971, 1972). To the contrary, however, the single study of TCVC, the
 6   tetrachloroethylene conjugate, in calves, did not result in the adverse effects observed in studies
 7   of exposures to DCVC. Therefore, because considerable uncertainty surrounds the identification
 8   of the causative chemical species, the administered exposure and total metabolites are both
 9   considered as dose metrics (see Section 5.3.3.2).  Production of the putative metabolites is
10   considered to be directly proportional to the total amount of tetrachloroethylene metabolized.
11
12   4.10.4.4.  Central Nervous System
13          As discussed in Section 4.6.4, the best surrogate for internal dose is blood
14   tetrachloroethylene concentration.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    4-203      DRAFT-DO NOT CITE OR QUOTE

-------
 1                                       APPENDIX 4A:
 2        CONSISTENCY OF TETRACHLOROETHYLENE AND TRICHLOROACETIC
 3                            ACID HEPATOCARCINOGENICITY
 4
 5
 6          TCA, a metabolite of tetrachloroethylene, is associated with hepatocarcinogenicity in
 7   male and female mice (Bull et al., 1990, 2002; Daniel et al., 1993; DeAngelo et al., 2008;
 8   Herren-Freund et al., 1995; Ferreira-Gonzalez, 1987; Pereira, 1996), as is tetrachloroethylene
 9   (NCI, 1977; NTP, 1986; JISA, 1992).
10          There has been some suggestion that TCA does not account for all of the toxicity
11   observed with tetrachloroethylene exposure (Buben and O'Flaherty, 1985; Clewell et al., 2005).
12   The purpose of this investigation was to compare the incidence of hepatocarcinogenicity
13   observed with tetrachloroethylene exposure to that observed with TCA exposure, in order to
14   examine whether the TCA that is expected to be generated by tetrachloroethylene can account
15   for tetrachloroethylene's hepatocarcinogenicity.  This was carried out by pooling the separate
16   TCA studies, fitting a time-to-tumor model to the TCA data, and comparing the incidence of
17   hepatocellular tumors expected based on the TCA studies with that observed in the
18   tetrachloroethylene bioassays.
19
20   4A.1. METHODS
21          Table 4A-1 summarizes data from the available TCA studies considered for carrying out
22   dose-response modeling. As detailed below, a number of these TCA studies lack information for
23   a complete comparison of hepatocarcinogenicity between tetrachloroethylene and TCA.
24
25   4A.1.1. Response Data
26          EPA generally emphasizes combining hepatocellular adenomas and carcinomas in
27   developing cancer risk values, for three reasons: (1) Hepatocellular adenomas develop from the
28   same cell lines as carcinomas and can progress to carcinomas; (2) Adenomas are often
29   distinguished from carcinomas only on the basis of size; and (3) histopathologic  decision criteria
30   may vary between laboratories or over time.
31          However, most of the TCA studies either did not consider adenomas or did not report
32   combined incidence of adenomas and carcinomas.  Lacking data on adenomas, the studies that
33   only provided carcinoma incidence may under-represent hepatocellular tumor incidence. For
34   studies not reporting combined incidence of adenomas and carcinomas, there could be some
35   double-counting of animals when the separate totals of adenomas and carcinomas are added
36   together.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-204      DRAFT-DO NOT CITE OR QUOTE

-------
 1          Therefore, for the purposes of this analysis, only the chronic data of DeAngelo et al.
 2    (2008) and the chronic data of Pereira (1996) were considered further for comparing with the
 3    tetrachloroethylene bioassays in male and female mice, respectively.  The DeAngelo et al. study
 4    was conducted at the lowest TCA levels of all the available studies, with exposures spanning
 5    about 6 to 60 mg/kg-day; these exposure levels span the range of TCA equivalents in the
 6    tetrachloroethylene bioassays.  Comparison of the Pereira data with the responses of the female
 7    mice in the tetrachloroethylene bioassays is limited by the availability of carcinoma data only,
 8    and by the study being conducted only through Week 82, not through Week 104.
 9          Table 4A-3 provides the hepatocellular adenoma or carcinoma incidence data from the
10    two tetrachloroethylene bioassays considered in this assessment, NTP (1986) and JISA (1993)
11    (for convenience, the studies will be referred to in the remainder of this appendix as the NTP and
12    JISA studies). For comparison across data sets, all incidences were normalized by converting
13    each to extra risk, [P(d)-P(0)]/[l-P(0)].
14
15    4A.1.2. Exposure Level Conversions
16          TCA bioassay exposures were generally reported in terms of water concentration, in
17    mg/L or mmol/L. Table 4A-1 provides the exposure levels as reported by each set of authors.
18    Some reports provided mg/kg-day equivalents. TCA exposures in mg/kg-day for the Pereira
19    (1996) study were interpolated from the other TCA studies which reported exposures in mg/kg-
20    day (see Table 4A-2).
21          The Reitz et al. (1996) PBPK model was used to estimate total metabolites corresponding
22    to the bioassay exposures in the NTP and JISA studies. Then it was assumed that 60% of the
23    total metabolites were TCA, as assumed in the model of Gearhart et al. (1993).  Although it is
24    possible that the extent of metabolism to TCA may be dose dependent, as for trichloroethylene-
25    induced TCA, there were insufficient data to characterize a dose dependency of TCA formation
26    for tetrachloroethylene.
27          The estimates of TCA induced by tetrachloroethylene exposure are internal doses, while
28    the exposures in the TCA bioassays were administered doses. Because orally administered TCA
29    has been estimated to be 95% absorbed in mice, the tetrachloroethylene-induced TCA estimates
30    were adjusted by dividing by 0.95, in order to approximate administered TCA exposures that
31    would be compatible with the dose-response modeling of the TCA drinking water studies.
32
33    4A.1.3. Dose-Response Model
34          The TCA data sets for male  and female mice were fit separately.  The male mice TCA
35    were modeled using the multistage model (BMDS 1.4.1;  U.S.EPA,  2007), given by:
36
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-205      DRAFT-DO NOT CITE OR QUOTE

-------
 1                 P(d) = 1 - exp[(-q0 - qi x d - q2 x d2 x... q6 x d6 ),
 2   where d = exposure level.
 3          The TCA data set for female mice was fit using a multistage-Weibull model because the
 4   only available data were limited to two time points less than the 104-week length of the
 5   tetrachloroethylene bioassays; this model provided a means of including both time points in the
 6   same analysis and facilitated extrapolation to 104 weeks. The multistage-Weibull model is
 7   given by:
 8
 9                 P(d,t) = 1- exp[(-q0 - qi x d - q2  x d2 x... q6  x d6 ) x tz,
10
11   where:
12          d = exposure level
13          t = time to observation of the  tumor
14          q;, z = parameters estimated in fitting the model
15
16   Time of scheduled sacrifice was input as the time to observation of each tumor. All tumors were
17   taken to be incidental to the death of affected animals. The software used was Tox_Risk (see
18   Section 5.4.4.1).
19          For comparison with the observed tetrachloroethylene data, model predictions were also
20   adjusted to estimate extra risk.
21
22   4A.2. RESULTS
23   4A.2.1. Trichloroacetic Acid (TCA), Male Mice
24          Figure 4A-2 provides the result of fitting a multistage model to the DeAngelo et al. data.
25   The responses at the control and low dose levels did not follow a monotonically increasing
26   pattern (the low-dose response was lower than the control), but a nearly linear one-stage model
27   provided an adequate fit (p = 0.15; model output included with Figure 4A-2).
28
29   4A.2.2. Trichloroactic Acid (TCA) Data, Female Mice
30          The evaluation of the model fit of the female mouse TCA data (Pereira, 1996) followed
31   the same steps as for the male mice.  The hepatocellular tumor data for female mice exposed to
32   TCA in drinking water are shown in Table 4A-2 and Figure 4A-2. These data include groups of
33   animals evaluated at three exposure levels plus control at two time points, for a total of eight
34   groups. The response at the high dose (463 mg/kg-day) was very similar for both time points, at
35   25 - 28%. A one-stage model also provided the best fit to these data,  with the two highest doses
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-206      DRAFT-DO NOT CITE OR QUOTE

-------
 1    at week 82 fitting least well.  Because the fit at the lower doses was relatively good, no other
 2    attempts were made to refine the dose-response model for the TCA female mouse data.
 O
 4    4A.2.3. Comparison of Tetrachloroethylene Hepatocellular Tumor Data With Predictions
 5    Based on Trichloroacetic Acid Data
 6          For the male mice, the extra risk of adenomas or carcinomas observed following 104
 7    weeks of inhalation exposure to tetrachloroethylene in the two bioassays is provided in Table
 8    4A-4, for comparison with the predicted extra risk of adenomas or carcinomas from the TCA-
 9    based dose-response modeling for male mice (based on the data of DeAngelo et al., 2008).  For
10    each male exposure group in the tetrachloroethylene bioassays, the observed proportion
11    responding is higher than that predicted using the TCA drinking water study, by 2- to 12-fold.
12    Mitigating factors to investigate further include possible differences in histopathology protocols
13    between laboratories and adequacy of the assumptions used to derive the TCA-equivalents
14    corresponding to the tetrachloroethylene exposure levels. Comparison between the
15    tetrachloroethylene and TCA studies for the male mice at Week 104 suggests concordance, but
16    "inconclusive" appears to be a plausible conclusion as well.
17          For the female mice, the extra risk of carcinomas observed following 104 weeks of
18    inhalation exposure to tetrachloroethylene in the two bioassays is provided in Table 4A-5, for
19    comparison with the predicted extra risk of carcinomas from the TCA-based dose-response
20    modeling for female mice (based  on the data of Pereira, 1996). The female mouse TCA model
21    appears to agree with the lack of carcinomas in female mice at the lower two exposures in the
22    JISA study, at approximately 3 and  11 mg/kg-day, but underestimates the observed incidence at
23    30 mg/kg-day of 29% by 90-fold. In contrast, the TCA model underpredicts the observed
24    carcinomas in both exposed groups of female mice in the NTP study by more than 200-fold.
25    In addition to the mitigating factors mentioned above, note that the tetrachloroethylene bioassays
26    were conducted at exposures associated with lower TCA levels than were used in the female
27    mouse TCA study. That is, for JISA female mice, the highest bioassay exposures were
28    associated with 28 mg/kg-day TCA (NTP) and 30 mg/kg-day, and the lowest exposure level in
29    the TCA study was approximately 47 mg/kg-day.  Consequently, there is a degree of
30    extrapolation beyond the TCA data set that may impact the predictions. Also note that the
31    tetrachloroethylene bioassays do not have sufficient resolution (let alone statistical power) to
32    detect response levels as low as those predicted by the TCA model in this range of exposures
33    (bioassays with 50 animals/group cannot provide estimates below 1/50 or 2%).
34
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  4-207      DRAFT-DO NOT CITE OR QUOTE

-------
 1    4A.3.  DISCUSSION AND CONCLUSIONS
 2          This analysis suggests that TCA may not explain the incidence of carcinomas observed in
 3    the available tetrachloroethylene bioassays, at least at TCA levels near 2 mg/kg-day in male mice
 4    and 20 mg/kg-day in female mice.  Otherwise, these data are inconclusive.
 5          As mentioned earlier in this appendix, some of the assumptions made in quantifying the
 6    dose-response relationships may have contributed to an overestimate of TCA's carcinogenicity.
 7    The Pereira study involved planned sacrifices at the reported time points, while the
 8    tetrachloroethylene studies did not. This would have led to earlier detection of tumors in the
 9    TCA studies relative to the tetrachloroethylene bioassays due to the detection of some tumors
10    before they may have become fatal, and, therefore, a slightly higher estimate of carcinoma
11    incidence in the time-to-tumor model.  Therefore, dose-response estimates based on the female
12    TCA study may contribute to overestimating risk, all else being equal.
13          And as mentioned earlier, another uncertainty is the use of PBPK estimates of TCA
14    levels resulting from inhalation  or oral exposure to tetrachloroethylene.  Another interpretation
15    of tetrachloroethylene-induced TCA levels has been provided by Clewell et al. (2005), who
16    provided TCA levels corresponding to the bioassay levels in the NTP bioassay, but not the JISA
17    bioassay.  These levels were 16.3 mg/kg-day for the low-dose males, 30.6 mg/kg-day for the
18    high-dose males, 16.9 mg/kg-day for the low-dose females, and 31.6 mg/kg-day for the high-
19    dose females. These levels differ from those estimated here by no more than 15%, which does
20    not explain the differences in response levels compared in this analysis.  Given the current state
21    of the science, the impact of this source of uncertainty is not well understood.
22          The differing results from the other TCA studies underscore the need to consider the joint
23    incidence of adenoma and carcinomas, which could have a substantial impact on this analysis.
24    The relative time courses of and correlation between adenomas and carcinomas in the TCA
25    bioassays are less clear, because relevant data were not included in the TCA reports.  This is
26    perhaps the most uncertain part of this analysis. Additional information should be obtained from
27    the original investigators for further evaluation if possible, perhaps in a meta-analysis.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-208      DRAFT-DO NOT CITE OR QUOTE

-------
       Table 4A-1.  Trichloroacetic Acid (TCA) drinking water studies in male mice: incidence of hepatocellular
       adenomas and carcinomas
Source
Bull et al.
(1990)a
Bull et al.
(2002)
Herren-Freund
etal. (1987)
Ferreira-
Gonzalez et al.
(1995)
DeAngelo et al.
(2008)
Weeks of
exposure
37
52
52
61
104
104
TCA
exposure,
g/L
2
0
1
2
0
0.5
2
0
5
0
4.5
0
0.05
0.5
Equivalent TCA
exposure
(mg/kg-day)
330
0
170
330
0
NR
NR
0
NR
0
NR
0
8
68
N
11
35
11
24
20
20
20
22
22
16C
11
56
48
51
Incidence of
adenomas
0
0
2
1
0
5
6
2
8
NR
NR
10
10
20
Incidence of
carcinomas
3
0
2
4
0
3
3
0
7
3C
8
26
14
32
Incidence of
adenomas or
carcinoma
3
0
NR
NR
0
6
8
2
NR
NR
NR
31
21
36
Proportion
responding with
carcinomas
0.27
0.0
0.18
0.17
0.0
0.15
0.15
0.0
0.32
0.19
0.73
0.55
0.44
0.71
a Cumulative TCA exposures were provided in g/kg for the mice evaluated at 52 weeks. Those exposures were converted to mg/kg-day by multiplying by
  (1,000 mg/g)/(7 days/week * 52 weeks).
0 Estimated from the reported proportion responding by selecting the smallest group size and incidence value consistent with the precision of the reported
  proportion.
NR = not reported

-------
            Table 4A-2.  Trichloroactic acid (TCA) drinking water study in female
            mice—incidence of hepatocellular adenomas and carcinomas


Weeks of
exposure
52



82




TCA
exposure,
g/La
0.0
0.3
1.1
3.3
0.0
0.3
1.1
3.3
Equivalent
TCA
exposure1"
(mg/kg-day)
0
47
189
463
0
47
189
463



N
40
40
19
20
90
53
27
18


Incidence of
adenomas
1
3
3
2
2
4
3
7


Incidence of
carcinomas
0
0
0
5
2
0
5
5

Incidence of
adenomas or
carcinomas
1
3
3
NR
NR
4
NR
NR
Proportion
responding
with
carcinomas
0.0
0.0
0.0
0.25
0.022
0.0
0.19
0.28
     a Exposure concentration was reported in mmol/L.
     b Estimated by interpolating exposures in Table 4A-1.

     NR = not reported
     Source:  Adapted from Pereira (1996).
1
2
3
4
       Table 4A-3.  Incidence of hepatocellular adenomas and carcinomas in
       B6C3F1 mice exposed to tetrachloroethylene in two inhalation bioassays

Sex
Male





Female






Bioassay
NTP
(1986)

JISA
(1993)


NTP
(1986)

JISA
(1993)


Administered
exposures,
(ppm)
0
100
200
0
10
50
250
0
100
200
0
10
50
250
Cumulative liver tumor incidence at week 104

Adenomas
12
8
19
7
13
8
26
3
6
2
3
3
7
26

Carcinomas
7
25
26
7
8
12
25
1
13
36
0
0
0
14
Adenomas
or
carcinomas
17
31
41
13
21
19
40
4
17
38
3
3
7
33

Total at risk8
49
47
50
46
49
48
49
45
42
48
50
47
48
49
5
6
7
a Animals dying before the first appearance of a hepatocellular tumor, but no later than week 52, were omitted from
  the totals because these animals were presumed not to have adequate time on study to develop tumors.

           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                    4-210      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
             Table 4A-4. Comparison of cumulative hepatocellular tumor incidence in
             male mice exposed for 104 weeks to tetrachloroethylene in chronic inhalation
             bioassays, to predictions based on trichloroacetic acid (TCA) exposure via
             drinking water

Study
NTP
(1986)
JISA
(1993)


N
49
47
50
46
49
48
49
Tetrachloro-
ethylene
exposure
(ppm)
0
100
200
0
10
50
250

Total
metabolites
(mg/kg-day)
0
27
41
0
3
14
36
Tetrachloro-
ethylene-induced
TCAa
(mg/kg-day)
0
17
26
0
2
9
23

Observed
proportion
responding1"'0
0.0
0.479
0.724
0.0
0.203
0.158
0.744

Predicted
proportion
responding1"
0.0
0.139
0.204
0.0
0.017
0.076
0.182

Observed/
predicted
-
o
J
4
.
12
2
4
 6
 7
 8
 9
10
11
12
13
14
15
16
17
      Estimated using PBPK model of Reitz et al. (1996) and adjusted for use with the TCA dose-response model by
      dividing by 0.95 to approximate a drinking water exposure to TCA (see Section 4A. 1.2).
      Extra risk.
      Calculated from Table 4A-1.
            Table 4A-5. Comparison of cumulative hepatocellular carcinoma incidence
            in female mice exposed for 104 weeks to tetrachloroethylene in chronic
            inhalation bioassays, to predictions based on trichloroacetic acid (TCA)
            exposure via drinking water



Study
NTP
(1986)

JISA
(1993)





N
45
42
48
50
47
48
49

Tetrachloro-
ethylene
exposure
(ppm)
0
100
200
0
10
50
250


Total
metabolites
(mg/kg-day)
0
31
45
0
4
18
47
Tetrachloro-
ethylene-
induced
TCAa
(mg/kg-day)
0
20
28
0
3
11
30


Observed
proportion
responding1"'0
0.0
0.287
0.728
0.0
0.0
0.0
0.286


Predicted
proportion
responding1"
0.0
0.0014
0.0029
0.0
0.00002
0.00041
0.0031



Observed/
predicted
.
210
260
.
-
-
90
18
19
20
21
22
23
     a Estimated using PBPK model of Reitz et al. (1996) and adjusted for use with the TCA dose-response model by
      dividing by 0.95 to approximate a drinking water exposure to TCA (see Section 4A. 1.2).
     b Extra risk.
     c Calculated from Table 4A-2.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   4-211      DRAFT-DO NOT CITE OR QUOTE

-------
                                       Multistage Model with 0.95 Confidence Level
 1
 2
 5
 6

 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
                   0.8
                   0.7
                   0.6
                   0.5
                   0.4
                   0.3
                          Multistage
                     BMDL
                                        MD
        0          10


16:4806/092008
                                               20
                                                    30

                                                 dose
40
50
60
      Figure 4A-1. Multistage dose-response fit of male mouse hepatocellular

      tumor incidence associated with exposure to trichloroacetic acid in drinking

      water; data from DeAngelo et al. (2008).
         The form of  the probability function is:
                            ound +  (1-background)*[1-EXP(

                      -betal*dose^l)]
         Dependent  variable = hep a  c

         Independent variable = mg kg d


       Total number of observations  = 3

       Total number of records with  missing value

       Total number of parameters in model = 2

       Total number of specified parameters = 0

       Degree of polynomial = 1
       Maximum number of iterations  =  250
       Relative Function Convergence has been set to:  le-008
       Parameter Convergence has been  set to: le-008

           This document is a draft for review purposes only and does not constitute Agency policy

06/06/08                                               DRAFT-DO NOT CITE OR QUOTE

                                           4-212

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
"_/
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68









Default Initial Parameter Values
Background = 0.489396






Asymptotic Co


Background

Background

Beta (1)






Interval
Variable
Conf. Limit
Background
Beta (1)


1

-0.54










0.

* - Indicates that this





Model
Full model
Fitted model
Reduced model

AIC:




Dose Est

0 . 0000 0 .
5 . 6000 0 .
58.0000 0.

Chi^2 = 2.07


Benchmark Dose

Specified effect

Risk Type

Confidence level

BMD

BMDL

BMDU

Taken together, (
interval for the





Beta(l) = 0.00926311


rrelation Matrix of Parameter Estimates

Beta(l)

-0.54

1



Parameter Estimates

95.0% Wald Confidence

Estimate Std. Err. Lower Conf. Limit Upper

0.493963 * *
00878738 * *

value is not calculated.



Analysis of Deviance Table

Log (likelihood) # Param's Deviance Test d.f. P-value
_
_
_






. Prob.

4940
5183
6960

d.f.


102.285 3
103.323 2 2.07524 1 0.1497
106.011 1 7.4518 2 0.02409

210. 645


Goodness of Fit
Scaled
Expected Observed Size Residual

27.662 31 56 0.892
24.877 21 48 -1.120
35.497 36 51 0.153

= 1 P-value = 0.1499


Computation

=

=

_

=

=

=

6.53618
BMD

0.1

Extra risk

0. 95

11. 99

6.53618

43.6117

, 43.6117) is a 90 % two-sided confidence

           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                               DRAFT-DO NOT CITE OR QUOTE
                                          4-213

-------
15:1906/03/2005
    1


  0.8


  0.6


  0.4


  0.2


    0
1
2
3
4
5
6
                                             Incidental Graph
                                  tca_fem_carc.ttd - TCA, females, hep. care.
                                          Model: Multistage Weib
                     Dose (mg/kg/day)=47
                     Dose (mg/kg/day)=189
                     Dose (mg/kg/day)=463
                     HoelWalburg(189)
                     Hoel Walburg (463)
                                             H
                                                                 H
                  20
                                     40
  60

Time (wks)
80
100
120
      Figure 4A-2.  Multistage-Weibull dose-response fit of female mouse
      hepatocellular carcinoma incidence associated with exposure to
      trichloroacetic acid in drinking water; data in Table 4A-2. Multistage-
                                                             -5
      Weibull model parameters: q0 = 3.39 x 10  , q2 = 6.10 x  10°, z = 1.9.
          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                             DRAFT-DO NOT CITE OR QUOTE
                                         4-214

-------
 1                       APPENDIX 4B:  HUMAN STUDIES OF CANCER
 2
 3           The body of literature assessing carcinogenic effects associated with exposure to
 4    tetrachloroethylene is composed of cohort, proportionate mortality, and case-control studies. A
 5    small number of studies, including studies of cohorts involved in metal degreasing or in aircraft
 6    manufacturing/maintenance (Boice et al., 1999; Anttila et al., 1995; Spirtas et al.,  1991), have
 7    assessed tetrachloroethylene exposure explicitly. These cohort studies present risks associated
 8    with site-specific cancer mortality (Boice et al., 1999; Spirtas et al., 1991) or incidence (Anttila
 9    et al., 1995) for a subcohort of the larger study population who were exposed to
10    tetrachloroethylene. Additionally, a few case-control studies were able to examine the
11    relationship between cancer at specific sites and tetrachloroethylene exposure (Vaughan et al.,
12    1997; Schlehofer et al., 1995; Pesch et al., 2000a; Heineman et al., 1994).
13           A larger body of evidence on cancer exists for workers employed as dry cleaners. Dry
14    cleaners have potential exposures to a number of solvents, including tetrachloroethylene, which
15    began to be widely used in the early 1960s (IARC,  1995). Two cohorts, Ruder et  al. (1994,
16    2001) and Blair et al.  (1990, 2003), are of individuals primarily exposed to tetrachloroethylene
17    (Lyngeetal., 1997).
18           Last, several community-based drinking water studies are available (Aschengrau et al.,
19    1993, 1998; Paulu et al., 1999; Fagliano et al., 1990; Cohn et al., 1994; MA DPH, 1997;
20    Vartiainen et al., 1993; Lagakos et al., 1986). Exposure in most of these studies was to a mixture
21    of solvents, including tetrachloroethylene and trichloroethylene, although the studies by
22    Aschengrau et al. and Paulu et al. examined tetrachloroethylene specifically.
23           Observations from these epidemiologic studies are summarized in Tables 4B-la through
24    4B-13.  Table 4B-la presents results  of incidence studies of dry cleaners and laundry workers
25    and Table 4B-lb) those of mortality studies of dry cleaners. Three studies of aircraft
26    maintenance/manufacturing workers  or metal degreasing workers identified workers (a
27    subcohort) exposed to tetrachloroethylene. Tables 4B-3 through 4B-13 present results of case-
28    control studies of specific organ sites (e.g., lymphoma, liver, kidney, esophagus).  A few studies
29    assessed exposure to tetrachloroethylene, with the majority of studies evaluating the relationship
30    between site-specific  cancer and employment in dry cleaning.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                              DRAFT-DO NOT CITE OR QUOTE
                                               4-215

-------
Table 4B-la. Standardized incidence ratios (95% confidence intervals) in cohort studies of dry cleaners and
laundry workers
Site
Bladder
Breast
Cervix
Esophagus
Hodgkin's
disease
Kidney
Leukemia
Ji and Hemminki
(2005 a, b; 2006);
Ji et al. (2005)
Male





0.9b
(0.7-1.1)
61 obs/
0.8C
(0.6-1.1)
47 obs/
Ji and Hemminki
(2005 a, b; 2006); Ji
et al. (2005)
Female





1.4b
(1.1-1.7)
92 obs/
1.3C
(1.0-1.6)
80 obs/
Andersen et al.
(1999) a
Male
1.1
(0.9-1.5)
62 obs/
54.4 exp


0.8
(0.3-1.7)
7 obs/8.5 exp
0 obs/4.0 exp
1.0
(0.7-1.5)
24 obs/
23.3 exp
0.7
(0.03-1.2)
12 obs/
17.8 exp
Andersen et al.
(1999) a
Female
0.9
(0.7-1.2)
57 obs/
64.0 exp
0.9
(0.8-0.97)
634 obs/
7 12 exp
1.2
(1.01-1.4)
155 exp/
131.4 exp
1.0
(0.5-1.6)
14 obs/ 14. 4 exp
1.9
(1.1-2.9)
19obs/10.1exp
0.9
(0.7-1.2)
57 obs/64.8 exp
0.9
(0.7-1.2)
46 exp/50.9 exp
Lynge and
Thygesen
(1990)
Male
0.6
(0.2-1.3)
6 obs/9.7 exp




1.5
(0.6-3.3)
6 obs/4 exp
0.7
(0.1-2.6)
2 obs/2.8 exp
Lynge and
Thygesen (1990)
Female
0.9
(0.4-1.7)
8obs/9.1 exp
0.8
(0.7-1.1)
94obs/110.7exp
0.8
(0.6-1.2)
34 obs/40.3 exp


0.6
(0.2-1.4)
5 obs/8.6 exp
0.8
(0.2-1.7)
5 obs/6.7 exp

-------
o
Oi
oo 3s



   I

   I

   §
   ***.

   £3'
  1
   TO'

 Table 4B-la.  Standardized incidence ratios (95% confidence intervals) in cohort studies of dry cleaners and

laundry workers (continued)
Site
Primary liver
Lung
Multiple
myeloma
Non-
Hodgkin's
Pancreas
Prostate
Rectum
Ji and Hemminki
(2005 a, b; 2006);
Ji et al. (2005)
Male

1.4
(1.2-1.5)
247 obs/
1.0
(0.7-1.4)
29 obs/
1.0
(0.8-1.3)
59 obs/



Ji and Hemminki
(2005 a, b; 2006); Ji
et al. (2005)
Female

1.3
(1.1-1.5)
156 obs/
0.9
(0.6-1.2)
3 lobs/
0.9
(0.7-1.1)
64 obs/



Andersen et al.
(1999) a
Male
1.3
(0.6-2.3)
1 1 obs/
8.7 exp
1.2
(1.1-1.5)
141 obs/
113.7 exp
1.4
(0.8-2.3)
14 obs/
10.1 exp
1.5
(0.96-2.1)
27 obs/
18.5 exp
1.4
(0.98-2.0)
35 exp/
24.8 exp
1.1
(0.9-1.3)
118 obs/
11.3 exp
0.9
(0.6-1.3)
33 obs/
37.1 exp
Andersen et al.
(1999) a
Female
1.3
(0.9-1.9)
28obs/21.2exp
1.2
(1.08-1.4)
172 obs/ 148. 3 exp
0.9
(0.6-1.3)
3 lobs/ 34.8 exp
1.0
(0.7-1.2)
55 obs/57.9 exp
1.0
(0.8-1.3)
85 obs/83.3 exp

1.0
(0.8-1.2)
106 obs/
110.4 exp
Lynge and
Thygesen
(1990)
Male
1.2b
(0.3-3.5)
3 obs/2.5 exp
1.1
(0.8-1.7)
28 obs/24.5 exp
3.3
(0.9-8.5)
4 obs/ 1.2 exp
2.8
(0.9-6.5)
5 obs/ 1.8 exp
2.4
(1.1-4.5)
9obs/3.8exp
1.5
(0.7-2.6)
11 obs/7.6 exp
1.4
(0.6-2.6)
9 obs/6.5 exp
Lynge and
Thygesen (1990)
Female
2.7b
(1.5-4.5)
14 obs/5.2 exp
1.3
(0.9-1.8)
32 exp/24.9 obs
1.1
(0.2-3.1)
3 obs/2.8 exp
0.5
(0.1-1.5)
3 obs/6 exp
1.4
(0.7-2.4)
13 obs/9.3 exp

0.7
(0.4-1.3)
11 obs/15.5 exp
to
   §
   I
   TO
H I

O >
HH Oq
H TO
O
H
W

-------
o
Oi
oo S3'

   I
   I
   §
   ^.
   Co'
   1
   TO'
to
00 2
   I
   TO
       Table 4B-la.  Standardized incidence ratios (95% confidence intervals) in cohort studies of dry cleaners and
       laundry workers (continued)
Site
Skin
Stomach
Ji and Hemminki
(2005 a, b; 2006);
Ji et al. (2005)
Male


Ji and Hemminki
(2005 a, b; 2006); Ji
et al. (2005)
Female


Andersen et al.
(1999) a
Male
1.0
(0.8-1.3)
60 obs/
60.7 exp
1.3
(0.9-1.7)
47 obs/
37.6 exp
Andersen et al.
(1999) a
Female
0.9
(0.8-1.04)
174 obs/
196.1 exp
1.0
(0.8-1.2)
89obs/93.7exp
Lynge and
Thygesen
(1990)
Male
1.1
(0.6-1.9)
14 obs/12.7 exp
1.3
(0.5-2.7)
7 obs/5.3 exp
Lynge and
Thygesen (1990)
Female
0.7
(0.5-1.1)
23obs/32.1exp
1.3
(0.6-2.3)
11 obs/8.6 exp
a Anderson et al. (1999), a study of cancer incidence and occupation in Nordic populations (Norway, Sweeden, finland, and Denmark), includes
some subjects in these previously-published and earlier studies: McLaughlin et al. (1987), Lynge and Thygesen (1990) and Chow et al. (1995).
Cano and Pollan (2001) and Travier et al. (2002) present observations from Sweedish subjects included in Anderson et al. (1999).
b Ji et al. (2005) updates the analysis of McLaughlin et al.(1987) who also presented kidney cancer incidence rate for dry cleaners and laundry
workers.
0 Ji and Hemminki (2005b) present standardized incidence ratios for these leukemia subtypes. For males, chronic lymphatic leukemia, 0.9 (0.5-
1.3), 19 obs.; acute myelogeneous leukemia, 0.6 (0.2-1.2), 7 obs.; chronic myelogenous leukemia, 0.9 (0.3-1.9), 5 obs.; polycythemia vera, 1.0
(0.4-2.0), 7 obs. For females, chronic lymphocytic leukemia 1.5 (1.1-2.1), 32 obs.; acute myelogenous leukemia, 1.4 (0.8-2.0), 20 obs.; chronic
myelogenous leukemia, 0.3 (0.03-0.9), 2 obs.: polycythemia vera, 1.7 (0.9-2.7), 14 obs.
H I
O ^
HH Oq
H TO
O
        obs = Observed
        exp = Expected
H
W

-------
1
2
Table 4B-lb.  Standardized mortality ratios in cohort studies of dry cleaners
Site
Total mortality
Bladder
Breast
Buccal
Cervix
Esophagus
Hodgkin's disease
Kidney
Larynx
Leukemia
Liver/biliary passages
Lung
Lymphatic-hematopoietic
Multiple myeloma
Non-Hodgkin's
Ruder et al. (2001)a
1.0
(0.9-1.1)
302 obs/299 exp
Oobs
0.8
(0.3-1.7)
6 obs/7.7 exp

1.9
(0.5-4.8)
4 obs/2. 1 exp
2.7
(0.9-6.2)
5 obs/1.9 exp

1.7
(0.2-6.3)
2 obs/1.2 exp


0.2b
(0.0-1.3)
1 obs/5 exp
1.2
(0.7-1.8)
19 obs/17.3 exp
1.1
(0.4-2.4)
6 obs/5. 6 exp

1.4b
(0.6-2.9)
7 obs/5 exp
Blair et al. (2003)
1.0
(1.0-1.1)
23 5 lobs/23 51 exp
1.3
(0.7-2.4)
12 obs/9.2 exp
1.0
(0.8-1.3)
68 obs/68 exp
1.1
(0.5-2.0)
10obs/9.1exp
1.6
(1-2.3)
27 obs/16.9 exp
2.2
(1.5-3.3)
26obs/11.8exp
2.0
(0.6-4.6)
5 obs/2. 5 exp
1.0
(0.4-2.0)
8 obs/8 exp
1.7
(0.6-3.7)
6 obs/3.5 exp
0.8
(0.4-1.4)
12 obs/15 exp
0.8
(0.4-1.5)
10 obs/12.5 exp
1.4
(1.1-1.6)
125 obs/89.3 exp
1.0
(0.7-.3)
39 obs/39 exp
0.8
(0.3-1.6)
7 obs/16 exp
0.9
(0.5-1.6)
12 obs/13.2 exp
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   4-219      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4B-lb.  Standardized mortality ratios in cohort studies of dry cleaners
       (continued)
Site
Pancreas
Prostate
Rectum
Skin
Stomach
Ruder et al. (2001)a
0.8
(0.2-2.4)
3 obs/3.8 exp
0.7C
(0.1-2.4)
2 obs/3.1 exp
Oobs


Blair et al. (2003)
1.1
(0.7-1.5)
28 obs/25.5 exp
1.0
(0.6-1.6)
17 obs/17 exp
1.3
(0.2-2.1)
15 obs/11.5 exp
0.8
(0.2-2.1)
4 obs/5 exp
0.9
(0.6-1.4)
20 obs/22.2 exp
4
5
6
7
 a Standard mortality ratios (SMRs) for subjects whose employment begun after 1960, e.g., the
tetrachloroethylene subcohort.
b SMR for the PCE subcohort for liver cancer and for non-Hodgkin's lymphoma are not presented in
Ruder et al. (2001). The SMRs in this table are for the entire cohort.
0 Ruder reports SMR for the larger category of "male genital organ cancer."
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    4-220      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4B-2. Standardized incidence and mortality ratios in cohort studies of
       tetrachloroethylene-exposed degreasers
Site
Total mortality
Bladder
Breast
Buccal
Cervix
Esophagus
Hodgkin's disease
Kidney
Larynx
Leukemia
Liver/biliary
passages
Lung
Lymphatic-
hematopoietic
Multiple myeloma
Non-Hodgkin's
Boice et al.
(1999)a
(mortality)
0.9
(0.9-1.0)
476 obs/528.9 exp
0.7
(0.1-2.5)
2 obs/2.9 exp
1.2
(0.3-3.0)
4 obs/3.5 exp
0.6
(0.1-2.0)
2 obs/3.6 exp
(0-7.8)
0 obs/0.5 exp
1.5
(0.5-3.2)
6 obs/4.1 exp
0 obs/0.6 exp
0.7
(0.1-2.5)
2 obs/2.9 exp
0.7
(0-3.6)
1 obs/1.6 exp
1.1
(0.4-2.6)
5 obs/4.6 exp
2.1
(0.8-4.2)
7 obs/3.4 exp
1.1
(0.8-1.4)
46 obs/42.6 exp
1.1
(0.6-1.9)
14 obs/12.4 exp
0.4
(0-2.3)
1 obs/2.5 exp
1.7
(0.7-3.3)
8 obs/4.7 exp
Anttila et al.
(1995)
(incidence)




3.2
(0.4-11.6)
2 obs/1.6 exp


1.8
(0.2-6.6)
2 obs/1.1 exp



1.9
(0.6-4.5)
5 obs/2.6 exp
1.4
(0.3-4.0)
3 obs/2.2 exp
(0-9.8)
0 obs/0.4 exp
3.8
(0.8-11.0)
3 obs/0.8 exp
Spirtas et al. (1991)
Males
(mortality)













Oobs
1.9
(0.2-6.9)
2 obs/1.1 exp
Females
(mortality)













17.1
(2.1-61.6)
2 obs/0. 1 exp
9.7
(1.2-35.0)
2 obs/0. 2 exp
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  4-221      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
            Table 4B-2. Standardized incidence and mortality ratios in cohort studies of
            tetrachloroethylene-exposed degreasers (continued)
Site
Pancreas
Prostate
Rectum
Skin
Stomach
Boice et al.
(1999)a
(mortality)
1.5
(0.7-2.8)
10 obs/6.7 exp
1.1
(0.6-2.0)
12 obs/10.5 exp
1.2
(0.3-3.6)
3 obs/2.4 exp
lb
(0.1-3.4)
2 obs/2.1 exp
1.4
(0.6-2.9)
7 obs/4.9 exp
Anttila et al.
(1995)
(incidence)
3.1
(0.6-9.0)
3 obs/1 exp




Spirtas et al. (1991)
Males
(mortality)





Females
(mortality)





4
5
     a Standard mortality ratios (SMRs) are for subjects routinely exposed to tetrachloroethylene.
     b SMR is for melanoma of the skin.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   4-222      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4B-3.  Case-control studies: liver cancer
Reference
Lynge et al.
(2006)
Lynge et al.
(1995)
Suarez et al.
(1989)
Hernberg et al.
(1988)
Austin et al.
(1987)
Hardell et al.
(1984)
Hernberg et al.
(1984)
Stemhagen et
al. (1983)
Houten and
Sonnesso
(1980)
Exposure
assessment
Occupation
Occupation
Death
certificate
Mail survey
andlH
Interview
Mail survey
Mail survey
andlH
Interview
Interview
Exposure
classification
DC, other in
dry cleaning,
or
unclassifiable
DC
DC services
DC operators
Solvent
exposure
(tetrachloro-
ethylene, TCE,
CC14)
DC
Solvent
exposure
Solvent
exposure
DC, laundry,
and other
garment
services
DC
Number of
subjects
(cases/controls)
DC, 11/95
Other, 2/22
Unclassifiable,
23/121
17/85
1,742/1,742
377/385
80/146
102/200
126/175
265/530
102/
Odds ratio
(95% CI)
DC. 0.8 (0.4-
1.5)
Other in DC,
0.4(0.1-1.9)
Unclassifiable,
1.1(0.6-2.1)
No exposed
cases
1.0 (0.4-2.2)
PLC
0.6 (0.2-1.8)
0.6(0.3-1.2)
male, PLC
3.4(1.3-8.6)
female, PLC
no exposed
cases
1.8(1.0-3.4)
PLC
2.1(1.1-4.0)
HCC
2.3 (0.8-7.0)
PLC
2.5(1.0-6.1)
PLC
2.3(0.9-6.1)
HCC
Too small to
assess; 2 cases
in DC
Exposure
response

-


-

-


3
4
5
DC = Dry cleaner
PLC = Primary liver cancer
HCC = Hepatocellular cancer
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    4-223      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Table 4B-4. Case-control studies: kidney cancer
Reference
Lynge et al.
(2006)
Parent et al.
(2000)
Pesch et al.
(2000a)
Dosemeci et
al. (1999)
Delahunt et
al. (1995)
Lynge et al.
(1995)
Mandel et al.
(1995)
Schlehofer et
al. (1995)
Auperin et al.
(1994)
Exposure
assessment
Occupation
Questionnaire,
interview
Questionnaire,
interview
Occupation
from
Interview to
JEM
Survey
Occupation
Interview
Interview
Occupation
Exposure
classification
DC, other in dry
cleaning, or
unclassifiable
DC/laundry worker
Tetrachloroethylene
JEM
males
females
JTEM
males
females
Tetrachloroethylene
(male)
Tetrachloroethylene
(female)
Tetrachloroethylene
(both)
DC
DC
DC industry (male)
DC solvents (male)
DC solvents
(female)
Iron/steel industry
(male)
Tetrachloroethylene
Metal industry
DC
Number of
subjects
(cases/controls)
DC, 29/196
Other, 9/34
Unclassifiable,
43/241
142/533
population and
1,900 cancer
935/4,298
273/462
165/225
438/687
914/12,756a
16/80
1,732/2,309
277/286
196/347
Odds ratio
(95% CI)
DC. 0.7 (0.4-
1.1
Other in DC,
1.2(0.5-2.5)
Unclassifiable,
0.8(0.5-1.2)
1.7(0.6-4.7)
1.4(1.0-2.0)
0.7 (0.3-2.2)
1.3 (0.7-2.3)
2.0 (0.5-7.8)
1.1 (0.7-1.7)
0.8(0.3-2.1)
1.1 (0.7-1.6)
1.9(0.3-13.9)
0.7 (0.2-3.6)
0.9 (0.3-2.4)
1.4(1.1-1.7)
1.6(1.0-2.7)
1.6(1.2-2.2)
2.5 (1.2-5.2)
1.6(1.1-2.5)
Too few
exposed
Exposure
response

None with
exposure
duration
None

-
-
None
-
-
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   4-224      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4B-4.  Case-control studies: kidney cancer
Reference
Mellemgaard
etal. (1994)
McCredie
and Stewart
(1993)
Partanen et
al. (1991)
Harrington et
al. (1989)
Sharpe et al.
(1989)
Asal et al.
(1988)
Exposure
assessment
Interview
Interview
Mail survey
Questionnaire
Mail survey
Occupation
Exposure
classification
DC (male)
DC (female)
Solvents (male)
Solvents (female)
Iron/steel industry
(male)
DC industry
Solvent exposure
Iron/steel industry
DC, solvents
Iron/metalwork
Solvent exposure
DC
Organic solvents
Degreasing solvents
DC (male)
DC(female)
Metal degreasing
Number of
subjects
(cases/controls)
368/396
489/523
338/338
54/54
164/161
315/313 + 336
(hospital +
population)
Odds ratio
(95% CI)
2.3 (0.2-27.0)
2.9(0.3-33.0)
1.5(0.9-2.4)
6.4(1.2-23.0)
1.4(0.8-2.4)
2.7(1.1-6.7)
1.5(1.1-2.1)
2.1 (1.0-4.4)
Too few
exposed
1.9(0.9-3.8)
1.0(0.2-4.9)
No exposed
cases
1.7(0.9-3.2)
3.4 (0.9-12.7)
0.7 (0.2-2.3)
2.8 (0.8-9.8)
1.7(0.7-3.8)
Exposure
response

Trend in
years
worked
for
iron/steel


-

3
4
5
6
7
a Controls were subjects with primary tumors from sites other than the urinary tract.

DC = Dry cleaner
JEM = Job exposure matrix
JTEM = Job task exposure matrix
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    4-225      DRAFT-DO NOT CITE OR QUOTE

-------
o
Oi
Table 4B-5. Case control studies:  lymphoid cancer


Reference

Exposure
assessment

Exposure
classification
Number of
subjects
(cases/controls)

Odds ratio
(95% CI)

Exposure
response
Hodgkin's Disease
Senior!
Costantini et al. (2001);
Miligietal. (1999)
Interview


Occupation: DC,
laundry worker,
presser
1,250/1,779


2.5 (0.3-24.6)a [Males]
3.5 (1.5-8.2)a [Females]




Non-Hodgkin's lymphoma
Lynge et al. (2006)




Mester et al. (2006)
(Hodgkin's and non-
Hodgkin's)

Miligi et al. (2006),
Seniori
Costantini et al. (2001),
Miligietal. (1999)
(non-Hodgkin's and
chronic lymphatic
leukemia)







Country-wide pension
database (for Danish and
Finnish subjects) or from
blinded interview (for Norse
and Swedish subjects)
Interview



Interview













DC, other in dry
cleaning, or
unclassifiable


Job title: DC,
launderer and
presser

Tetrachloroethylene
using JEM;
Occupation: DC,
launderer and
presser









DC, 42/219
Other, 8/48
Unclassifiable
52/255

710/710



1,428 /1,520
(Miligi et al.,
2006):
1,250/1,779
(Seniori
Costantini et
al. (2001)







DC, 1.0 (0.7-1. 4)b
Other in DC, 0.7 (0.3-1. 6)b
Unclassifiable, 0.9 (0.6-1. 4)b


Duration of employment:
1-10 yr, 0.8 (0.3-2.5)c
>10yr, 3.4(0.6-18.5)c
Allyrs, 1. 3(0.5-3 .2)c
Tetrachloroethylene, exposure
intensity:
Very low/low, 0.6 (0.3-l-2)d
Medium/high, 1.2 (0.6-2.5)d

Duration of medium/high
exposure:
<15yr, 1.3 (0.5-3.3)d
>15 yr, 3 cases-ORnot
presented
Tetrachloroethylene :
Diffuse NHL 1.9 (0.7-5. 7)d
DC, 1.6 (0.3-9. l)c [Males]
DC, 0.7 (0.3-1. 5)c [Females]









Linear trend,
p'0.72












oo S3'


   I

   I

   §
   ***.

   £3'
  1
   TO'

to
   I
   TO
H I

O >
HH Oq
H TO
O
H
W

-------
o
Oi
Table 4B-5. Case control studies: lymphoid cancer (continued)
Reference
Kato et al. (2005)
Fabbro-Peray et al.
(2001)
Blair etal. (1993)
Scherretal. (1992)
Hardell etal. (1981)
(Hodgkin's and non-
Hodgkin's)
Exposure
assessment
Interview
Interview
Interview
Interview
Survey
Exposure
classification
Occupation/Dry
Cleaning fluids
Dry cleaning
solvents
Occupation/JEM
Laundry/Garment
Services
Occupation: DC,
launderer, leather
product fabrication
Trichloroethylene,
Tetrachloroethy lene ,
styrene, benzene
Number of
subjects
(cases/controls)
376/463
445/1025
622/1,245
303/303
169/338
Odds ratio
(95% CI)
1.6 (0.5-5. l)e
1.0(0.6-1.6)f
2.0
(0.97-4.3)
1.6(0.6-4.0)
4.6
(1.9-11.4)
Exposure
response
-
—


<
Leukemia / Lymphoma
Senior!
Costantini et al. (2001)
Clavel etal. (1998)
Malone etal. (1989)
Interview
Self-administered
questionnaire
Interview
Occupation: DC,
launderer and
presser
Occupation: DC,
launderer
Occupation: DC,
launderer
383/269
226/425
427/1,683
All leukemia,
3. 3 (0.3-32.4) a [Males}
Hairy Cell Leukemia,
3.0(0.2-49.2)8
Chronic lymphocytic leukemia,
1.1 (0.6-2.0)h [All respondents]
0.9 (0.4-1. 8)h [Self-respondents]

—

oo S3'

   I
   I
   §
   ^.
   Co'
  1
   TO'

to
   I
   TO
H I
O
H
W

-------
o
Oi
Table 4B-5. Case control studies: lymphoid cancer (continued)


Reference

Exposure
assessment

Exposure
classification
Number of
subjects
(cases/controls)

Odds ratio
(95% CI)

Exposure
response
Leukemia, Childhood (Acute lymphocytic leukemia)
Infante-Rivard et al.
(2005)








Costas et al. (2002)















Interview; questionnaire on
maternal occupation. JEM
used to assign chemical-
specific exposures.






Interview; questionnaire to
parents which included
information on use of public
drinking water in the home.
Hydraulic mixing model used
to infer drinking water
containing TCE and other
solvents delivery to residence








Tetrachloroethylene









Residence receiving
water from Wells G
andH













790/790









19/37















Exposure period:
2 years before pregnancy up to
birth, 1.0(0.4-2.3)'
During pregnancy, 0.8 (0.3-2.3)1
Any exposure,
0.9 (0.4-2.2)1
Lowest exposure rank, 1.0 (0.4-
2.6)1
Highest exposure rank, 0.6 (0.5-
12.8)1
Exposure period of mother:
From 2 years before conception
to case diagnosis,
Never exposed, 1.0
Least exposed, 5.0 (0.8-33.5)1
Most exposed, 3.6 (0.5, 24.8)1

During 2 years before
conception,
Never exposed, 1.0
Least exposed, 2.5 (0.4-15.2)1
Most exposed, 2.8 (0.3-26.4)1
From birth to diagnosis,
Never exposed, 1.0
Least exposed, 1.8 (0.3-1 0.8)1
Most exposed, 0.9 (0.2-4.6)1
-

























oo S3'



   I


   I

   §
   ***.

   £3'
  1
   TO'

to
   I
   TO
H I
O
H
W

-------
o
Oi
        Table 4B-5. Case control studies:  lymphoid cancer (continued)
Reference

Shu etal. (1999)
Lowengart et al. (1987)
Exposure
assessment

Interview: questionnaire to
assess parental occupation
using job-industry title and
self-reported exposure
history.
Interview; questionnaire
sought information on
occupation and occupational
exposures of both parents.
Exposure
classification

Tetrachloroethylene
Tetrachloroethylene
Number of
subjects
(cases/controls)

1842/1986
123/123
Odds ratio
(95% CI)
During pregnancy
Never exposed, 1.0
Least exposed, 3.5 (0.2, 58. 1)1
Most exposed, 14.30 (0.9-224)1
Exposure period of mother:
Anytime, 0.4 (0.1-1. 4)k
Preconception, 1.4 (0.2-8.6)k
During pregnancy, 1.3 (0.2-8. 4)k
Postnatal, 0.4 (0.1-1. 5)k
Exposure period of father:
One year before pregnancy,
00(^ = 0.39)'
During pregnancy,
00(^ = 0.39)'
After delivery,
00(0.19-00)'
Exposure
response
Linear trend,
/K0.05


oo S3'

   I
   I
   §
   ***.
   £3'
   1
   TO'

to
   I
   TO
   Co
H I
O >
HH Oq
H TO
O
H
W
a Adjusted for age
b RR from logistic regression models adjusting for matching criteria (country, sex, five-year age group and five-year calendar period at the time of
diagnosis of the case).
0 Adjusted for smoking and alcohol consumption.
d Adjusted for sex, age, education, and area.
e Adjusted for age at index date, family history of hematologic cancer, college education, surrogate status, interview year, body mass index 10
years before interview, pain-relieving drugs use, antibiotic use, household pesticide products use, and work duration involving pesticide exposure.
f Adjusted for age, gender, area, and education level.
B Conditional logistic regression with the baseline category of unexposed workers, adjusted for smoking and farming as an occupation.
h Adjusted for race, age, education, sex, and study site
1 Adjusted for maternal age and level of schooling.
J Adjusted using a composite covariate to control for socio-economic status, maternal age at birth of child and breast-feeding.
k Adjusted for maternal education, race, and family income.

-------
1 Discordant pair analysis

DC = Dry cleaner
JEM = Job exposure matrix
RR = Relative risk

-------
1
2
Table 4B-6. Case-control studies: esophagus
Reference
Lynge et
al. (2006)
Parent et
al. (2000)
Vaughan et
al. (1997)
Exposure
assessment
Occupation
Occupation-
interview
Occupation-
interview
Exposure
classification
DC, other in dry
cleaning, or
unclassifiable
Solvents
Dry cleaner
Tetrachloroethy lene ,
probable exposure
Cumulative
exposure:
-1-29 ppm-y
-30+ ppm-y
Number of
subjects
(cases/controls)
DC, 8/86
Other, 5/31
Unclassifiable
18/108
99/533
315/313+336
(hospital +
population)
Odds ratio
(95% CI)
DC, 0.8 (0.3-
1.7)
Other in DC, 1.2
(0.4-3.6)
Unclassifiable,
2.0 (0.9-4.6)
1.6" (0.8-2.5)
Too few
exposed
subjects
precluded an
analysis of this
occupational
group.
6.4a (0.6-68.9)
11.9s (1.1-
124.0)
0 exposed cases
Exposure
response


None
a Odds ratio for squamous cell carcinoma, adjusted for a number of covariates including smoking and
alcohol consumption.
4
5
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   4-231      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
        Table 4B-7.  Case-control studies:  bladder cancer
Reference
Lynge et
al. (2006)
Pesche et
al. (2000b)a
Teschke et
al. (1997)
Silverman
etal.
(1989a, b)
Smith et al.
(1985)
Exposure
assessment
Occupation
Questionnaire-
interview
Occupation-
interview
Occupation-
interview
Occupation
Exposure
classification
DC, other in dry
cleaning, or
unclassifiable
Tetrachloroethylene
- substantial exp
JEM
males
females
ITEM
males
DC/laundry worker
DCb, ironer, presser
- male, white
- male,
non-white
DC/laundry worker
Number of
subjects
(cases/controls)
CD, 93/292
Other, 12/52
Unclassifiable,
57/234
1,035/4,298
105/139
(population
controls)
2,100/3,874
126/383
103/1,869
Odds ratio
(95% CI)
DC, 1.4(1.1-
1.9)
Other in DC,
1.1(0.6-2.1)
Unclassifiable,
1.2(0.8-1.8)
1.4(1.0-1.9)
0.7 (0.2-2.5)
1.8(1.1-3.7)
2.3d
(0.4-13.9)
2.8(l.l-7.4)c
1.3 (0.9-2.0),
nonsmoker;
3.0 (1.8-5.0),
former smoker;
2.9 (2.4-6.5),
current smoker
Exposure
response


Increased risk
with
>10 years of
exposure
Duration of
employment
showed no
effect
Duration of
exposure
showed no
effect
3
4
5
6
7
a Cases are histological confirmed urothelial cancers, a category comprised of urinary bladder, ureter, and renal
pelvis.  Odds ratio and 95% CI adjusted for age, study center, and smoking.
b A prior susceptible occupation.
0 Odds ratio adjusted for smoking and employment in other high-risk occupations.
d Odds ratio adjusted for sex, age, and cigarette smoking.
     DC = Dry cleaner
                 This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                       4-232      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 4B-8. Case-control studies: lung
Reference
Pohlabeln
etal.
(2000)
Brownson
etal.
(1993)
Exposure
assessment
Occupation-
interview
Occupation-
interview
Exposure
classification
DC/laundry
worker
DC
Number of
subjects
(cases/controls)
650a/l,542
429b/ 1,021
Odds ratio
(95% CI)
Females, 1.83
(0.98-3.4)
Ever, 1.8
(1.1-3.0)
> 1.1 25 years,
2.9(1.5-5.4)
Exposure
response
"
Trend with
duration of
employment
4
5
6
a Cases are nonsmokers.
b Cases are either nonsmokers or ex-smokers.
DC = Dry cleaner
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                    4-233      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
       Table 4B-9. Case-control studies: pancreas
Reference
Lin and
Kessler
(1981)






Partanen et
al. (1991)

Kernan et
al. (1999)










Exposure
assessment
Occupational
history
obtained by
interview





Job title and
industrial
branch
Occupation
as coded on
death
certificate








Exposure
classificatio
n
DC/gasoline








DC,
launderer,
presser
Tetrachloro-
ethylene










Number of
subjects
(cases/controls)
61I6T








625/1,700


63,0977
252,386










Odds ratio
(95% CI)
2.8(1.1-7.1)








2.4(0.3-17.2)b


High intensity of
exposure:
BF,0.8C (0.6-1.2)
BM, 1.2C (0.9-1.5)
WF, lc (0.9-1. 2)
WM, 0.9C (0.8-1.0)
High probability
of exposure:
BF,0.7d (0.5-1.1)
BM, ld (0.7-1. 4)
WF, 0.8C (0.6-1.1)
WM, 0.9C (0.8-1.0)
Exposure
response
Risk increased
with increasing
duration of
exposure.
>10 years
exposure,
OR = 5.1
(95% CI:
1.2B20.0)
-


-











 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
a Cases and controls selected from hospital admissions.  Control subjects were free of cancer and were
  selected at random from contemporaneous admissions.
b Odds ratio (OR) adjusted for age, gender, smoking, alcohol consumption, and diabetes.
0 OR adjusted for age, metropolitan status, region of residence, and marital status.

BF = Black female
BM = Black male
DC = Dry cleaner
WF = White female
WM = White male
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-234      DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
       Table 4B-10. Case-control studies: intestineCsmall bowel or colon
Reference
Fredriksson et al.
(1989)
Colon
Kaerlev et al.
(2000)
Adeno-carcinoma
of the small bowel
Exposure
assessment
Occupational
survey-
Questionnaire
Questionnaire-
interview
Exposure
classification
DC
Trichloroethylene
among DC
DC industry
Number of
subjects
(cases/controls)
312/623
79/2,649
(population+
hospital
controls)
Odds ratio
(95% CI)
2.0(0.5-7.1),
females
7.4
(1.1-47.0)
6.5b(1.6-
26.9)
Exposure
response

Positive trend
with duration
of employment
(industry and
occupation)
 4
 5
 6
 7
 8
 9
10
11
12
13
a Odd ratio adjusted for sex, country, and year of birth.

BF = Black female
BM =Black male
DC = Dry cleaner
WF =White female
WM =White  male
       Table 4B-11. Case-control studies: brain
Reference
Heineman
etal. (1994)
Exposure
assessment
Occupation
from
interview to
JEM
Exposure classification
Organic solvents
Tetrachloroethylene
Trichloroethylene
MeCl2
Number of
subjects
(cases/controls)
300/320
Odds ratio
(95% CI)
1.3 (0.9-1.8)
1.2(0.8-1.6)
1.1 (0.8-1.6)
1.3 (0.9-1.8)
Exposure
response
Positive
trend for
methylene
chloride
14
15
16
17
18
19
JEM= Job exposure matrix
       Table 4B-12. Case-control studies: breast
Reference
Band et al.
(2000)
Exposure
assessment
Occupational
survey-
questionnaire
Exposure
classification
Dry cleaner/
Laundry worker
Number of
subjects
(cases/controls)
1,018/1,025
Odds ratio
(95% CI)
1.8a (0.4-7.7),
pre-menopausal
4.9b (1.3-18.7), post-
menopausal
Exposure
response

20
21
22
a Ever employed as dry cleaner or laundry worker.
b Usually employed as dry cleaner or laundry worker.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    4-235      DRAFT-DO NOT CITE OR QUOTE

-------
1
2
Table 4B-13. Community-based studies
Outcome
Hodgkin's
disease
Leukemia




Multiple
myeloma
Non-
Hodgkin's
lymphoma
(NHL)
Bladder

Kidney
Exposure
assessment
Residence in town
VOCs other than
THMs in town water
Tetrachloroethylene
in town water
Residence in town;
VOCs in water
Residence in town
Distribution of
drinking water with
tetrachloroethylene
Residence in town
Tetrachloroethylene
in town water
Residence in town
Distribution of
drinking water with
tetrachloroethylene
Solvents in town
water (including
tetrachloroethylene,
TCE)
Distribution of
drinking water with
tetrachloroethylene
Reference
Vartiainen
(1985)
Fagliano et
al.
(1990)
Cohn et al.
(1994)
Lagakos et
al. (1986)
Vartiainen
(1985)
Aschengrau
etal. (1993)
Vartiainen
(1985)
Cohn et al.
(1994)
Vartiainen
(1985)
Aschengrau
etal. (1993)
Mallin
(1990)
Aschengrau
etal. (1993)
Relative risk
0.8 (0.3-1.7)
Hausjavi, Finland
1.4 (0.7-2.5)
Hattula, Finland
1.5 (1-2.2) f
1.0 (0.7-1. 5) m
1.4(1. 1-1.9) f
1.1 (0.8-1.4) m
2.2 (1.5-2.9)
1.2 (0.8-1.7)
Hansjavi, Finland
0.7(0.4-1.1)
Hattula, Finland
2.1(0.9-5.2)a'b
8.3(1. 5-45. 3)b'c
0.7 (0.3-1.3)
Hausjavi, Finland
0.6 (0.2-1.3)
Hattula, Finland
1.1 (0.9-1.3) f
1.1 (0.9-1.3) nf
0.6(0.3-1.1)
Hausjavi, Finland
1.4 (1.0-2.0)
Hattula, Finland
1.4 (0.7-2.9)3
4.0 (0.7-25. l)b'c
1.7m
2.6 f
1.1 (OA-2.^b
Exposed
cases
6
11
28
25
56
63
20
33
19
7
2
7
6
87
78
14
31
13
4
Not in
paper
6
Comments

Exposure response for
females
Exposure response for
females
Exposure response not
consistent

Risk for 90th percentile
higher than risk for
ever exposed

High-grade NHL;
2.7 (1.3-5.6) f
0.4 (0.1-1. 8) m

Risk for 90th percentile
higher than risk for
ever exposed


               This document is a draft for review purposes only and does not constitute Agency policy
    06/06/08                                  4-236      DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2
        Table 4B-13.  Community-based studies (continued)

Outcome
Liver



Colon-
rectum

Lung



Breast








Exposure
assessment
Residence in village



Distribution of
drinking water with
tetrachloroethylene
Distribution of
drinking water with
tetrachloroethylene

Distribution of
drinking water with
tetrachloroethylene







Reference
Vartiainen et
al. (1993)


Paulu et al.
(1999)

Paulu et al.
(1999)


Aschengrau
etal. (1998)

Paulu et al.
(2002)
Aschengrau
et al. (2003)
Vieira et al.
(2005)

Relative risk
0.7 (0.3-1.4)
Hausjavi, Finland
0.6 (0.2-1.3)
Hattula, Finland
1.1(0.7-1.5)"' b
l(0.3-2.8)b'c

l.l(Q.l-l.lfb
3.7(1.0-11.7)b'c


l.l(Q.l-l.lfb
1.1 (0.3-3.8)b'c



2.2(l.l-4.3)b'c'd
1.3(0.7-2.6)b'c'e
1.0 (0.4-2.5/

Exposed
cases
7

6

44
5

33
5


36
4



28
18
10


Comments
-



Risk for 90th percentile
higher than risk for
ever exposed
Exposure response
with tetrachloro-
ethylene exposure
level
Risk for 90th percentile
higher than risk for
ever exposed




None

 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
  Resided in home receiving tetrachloroethylene-contaminated drinking water.
  Odds ratio adjusted for sex, age at diagnosis, vital status at interview, educational level, and occupational
  exposure to solvents.  The usual number of cigarettes smoked and history of cigar and pipe use, living with a
  smoker, and occupational history associated with lung cancer were additionally controlled in the lung cancer
  analyses. All multivariate models of breast cancer included adjustment for age of diagnosis, vital status at
  interview, family history of breast cancer, age at first live birth or stillbirth, prior history of breast cancer/disease,
  and occupational exposure to solvents.
  Resided in home receiving water at a high relative delivered dose of tetrachloroethylene, defined by Aschengrau
  et al. (1993, 1998, 2004) and Paulu et al. (1999) as a level above the 90th percentile.
  Analyses were conducted among the present study population.
  Analyses were conducted among the present study population and the study population reported in Aschengrau et
  al. (1998).
  Odds ratio adjusted for age at diagnosis or index year, family history of breast cancer, personal history of breast
  cancer before current diagnosis or index year, age at first live-birth or still birth, vital status at interview, and
  occupational exposure to solvents.  The tetrachloroethylene exposure  variable, personal delivered dose or FDD,
  was inferred from a model that extended the relative delivered dose (ROD) model to include information on tap
  water consumption and bathing habits.
f = Female
m = Male
TCE = Trichloroethylene
THMs = Trihalomethanes
VOCs = Volatile organic compounds
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-237       DRAFT-DO NOT CITE OR QUOTE

-------
  1                                       REFERENCES FOR CHAPTER 4
  2
  3
  4     Aggazzotti, G; Fantuzzi, G; Predieri, G; et al. (1994a) Indoor exposure to perchloroethylene (PCE) in individuals
  5     living with dry-cleaning workers.  Sci Total Environ 156:133-137.
  6
  7     Aggazzotti, G; Fantuzzi, G; Righi, E; et al.  (1994b) Occupational and environmental exposure to perchloroethylene
  8     (PCE) in dry cleaners and their family members. Arch Environ Health 49(6):487-93.
  9
10     Ahlborg, G, Jr. (1990) Pregnancy outcome  among women working in laundries and dry cleaning shops using
11     tetrachloroethylene. Am J Ind Med 17:567-575.
12
13     Ahn, YS; Zerban, H; Grobholz, R; et al. (1992) Sequential changes in glycogen content, expression of glucose
14     transporters and enzymic patterns during development of clear/acidophilic cell tumors in rat kidney.  Carcinogenesis
15     13:2329-2334.
16
17     Albers, JW; Wald, JJ; Werner, RA; et al. (1999) Absence of polyneuropathy among workers previously diagnosed
18     with solvent-induced toxic encephalopathy. J Occup Environ Med 41:500-509.
19
20     Alden, CL. (1985) Species, sex, and tissue specificity in toxicologic and proliferative responses. Toxicol Pathol
21     13:135-140.
22
23     Alden, WW; Repta, AJ. (1984) Exacerbation of cisplatin-induced nephrotoxicity by methionine. Chem Biol Interact
24     48:121-124.
25
26     Altmann, L; Bottger, A; Wiegand, H (1990) Neurophysiological and psychophysical measurements reveal effects of
27     acute low-level organic solvent exposure in humans. Int Arch Occup Environ Health 62:493-499.
28
29     Altmann, L; Weigand, H; Bottger, A; et al.  (1992) Neurobehavioral and neurophysiological outcomes of acute
30     repeated perchloroethylene exposure. Applied Psychology: An International Review 41:269-279.
31
32     Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological  outcome of chronic
33     low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
34
3 5     Amler, RW; Mueller, PW; Schuyltz, MG. (1998) Biomarker of kidney function for environmental health field
36     studies.
37
3 8     Andersen, A; Barlow, L; Engeland, A; et al. (1999) Work-related cancer in the Nordic countries. Scand J Work
39     Environ Health 25(suppl 2): 116.
40
41     Andrys, C; Hanovcova, I; Chylkova, V; et al. (1997) Immunological monitoring of dry-cleaning shop workers--
42     exposure to tetrachloroethylene. Cent EurJ Public Health 5:136-142.
43
44     Anger, WK; Liang, YX; Nell, V; et al. (2000) Lessons learned: 15 years of the WHO-NCTB:  a review.
45     Neurotoxicology 21:837-846.
46
47     Anna, CH; Maronpot, RR; Pereira, MA; et  al. (1994) Ras proto-oncogene activation in dichloroacetic acid-,
48     trichloroethylene-and tetrachloroethylene-induced liver tumors in B6C3F1 mice.  Carcinogenesis  15:2255-2261.
49
50     Antti-Poika, M. (1982a) Prognosis of symptoms in patients with diagnosed chronic organic solvent intoxication. Int
51     Arch Occup Environ Health 51:81-89.
52
53     Antti-Poika, M. (1982b) Overall prognosis  of patients with diagnosed chronic organic solvent intoxication. Int Arch
54     Occup Environ Health 51:127-138.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-238      DRAFT-DO NOT CITE OR QUOTE

-------
  1     Anttila, S; Hirvonen, A; Husgafvel-Pursiainen, K; et al. (1994) Combined effect of CYP1 Al inducibility and
  2     GSTM1 polymorphism on histological type of lung cancer.  Carcinogenesis 15:1133-1135.
  3
  4     Anttila, A; Pukkala, E; Sallmen, M; et al. (1995) Cancer incidence among Finnish workers exposed to halogenated
  5     hydrocarbons. JOccup Environ Med 37:797-806.
  6
  7     Aranyi, C; O'Shea, W J; Graham, J A; Miller, F J (1986) The effects of inhalation of organic chemical air
  8     contaminants on murine lung host defenses. Fundam Appl Toxicol 6:713-720.
  9
10     Arlien-Soborg, P. (1982) Solvent neurotoxicity.  Boca Raton, FL: CRC Press.
11
12     Aryal, B K; Khuder, S A; Schaub, E A (2001) Meta-analysis of systemic sclerosis and exposure to solvents. Am J
13     Ind Med 40:271-274.
14
15     Asal, NR; Geyer, JR; Risser, DR; et al. (1988) Risk factors in renal cell carcinoma. II. medical history, occupation,
16     multivariate analysis, and conclusions. Cancer Detect Prev  13:263-279.
17
18     Aschengrau, A; Seage, GR.  (2003) Essentials of epidemiology for public health. Boston: Jones and Bartlett.
19
20     Aschengrau, A; Ozonoff, D; Paulu, C; et al. (1993) Cancer risk and tetrachloroethylene-contaminated drinking water
21     in Massachusetts [see comments]. Arch Environ Health 48:284-292.
22
23     Aschengrau, A; Paulu, C; Ozonoff, D. (1998) Tetrachloroethylene-contaminated drinking water and the risk of
24     breast cancer.  Environ Health Perspect 106 Suppl 4:947-953.
25
26     Aschengrau, A; Rogers, S; Ozonoff, D. (2003) Perchloroethylene-contaminated drinking water and the risk of breast
27     cancer: additional results from Cape Cod, Massachusetts, USA. Environ Health Perspect 111:167-173.
28
29     Ashby, J. (1996) Alpha 2 mu-globulin nephropathy in white ravens. Environ Health Perspect 104:1264-1267.
30
31     Ashby, J. Brady, A; Elcombe, CR; et al.  (1994) Mechanically-based human hazard assessment of peroxisome
32     proliferator-inducedhepatocarinogenesis. Hum Exp Toxicol 13(suppl2):Sl-117.
33
34     ATSDR (Agency for Toxic  Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
3 5     (Update). Prepared by Sciences, International, under subcontract to Research Triangle Institute. ATSDR. Atlanta,
36     GA.
37
3 8     ATSDR (Agency for Toxic  Substances and Disease Registry). (1998) Volatile organic compounds in drinking water
39     and adverse pregnancy outcomes. Atlanta, GA.
40
41     ATSDR (Agency for Toxic  Substances and Disease Registry). (2001) Interaction Profile:  1,1,1-Trichloroethane, 1,1-
42     Dichloroethane, Trichloroethylene, and Tetrachloroethylene (Draft).  Available online at
43     http://www.atsdr.cdc.gov/interactionprofiles/ip02.html. Accessed November 12, 2004.
44
45     ATSDR (Agency for toxic substances and disease registry)  (2003) Survey of specific childhood cancers and birth
46     defects among children whose mothers were pregnant while living at U.S. Marine Corps Base Camp Lejeune, North
47     Carolina, 1968-1985.
48
49     Auboeuf, D; Rieusset, J; Fajas, L; et al. (1997) Tissue distribution and quantification of the expression of mRNAs of
50     peroxisome proliferator-activated receptors and liver X receptor-alpha in humans: no alteration in adipose tissue of
51     obese and NIDDM patients. Diabetes 46:1319-1327.
52
53     Auperin, A; Benhamou, S; Ory-Paoletti, C; et al. (1994) Occupational risk factors for renal cell carcinoma: a case-
54     control study.  Occup Environ Med 51:426-428.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-239       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Austin, H; Delzell, E; Grufferman, S; et al. (1987) Case-control study of hepatocellular carcinoma, occupation, and
  2     chemical exposures. JOccupMed 29:665-669.
  3
  4     Ayensu, W K; Tchounwou, P B; McMurray, R W (2004) Molecular and cellular mechanisms associated with
  5     autoimmune diseases. Int J Environ Res Public Health 1:39-73.
  6
  7     Bagby, GCJ. (1994) Hematopoiesis. In: Stamatoyannopoulos, G; Nienhuis, AW; Majerus, PW; et al., eds. The
  8     molecular basis of blood diseases.  Philadelphia: W.B. Saunders; pp. 7-103.
  9
10     Baggetto, LG. (1992) Deviant energetic metabolism of glycolytic cancer cells. Biochimie 74:959-974.
11
12     Bagnell, PC; Ellenberger, HA. (1977) Obstructive jaundice due to a chlorinated hydrocarbon in breast milk. Can
13     MedAssoc 1117:1047-1048.
14
15     Bale, A S; Adams, T L; Bushnell, P J;  Shafer, T J; Boyes, W K (2005) Role of NMD A, nicotinic, and GAB A
16     receptors in the steady-state visual-evoked potential in rats. Pharmacol Biochem Behav 82:635-645.
17
18     Band, PR; Le, ND; Fang, R; et al. (2000) Identification of occupational cancer risks in British Columbia: a
19     population-based case-control study of 995 incident breast cancer cases by menopausal status, controlling for
20     confounding factors. JOccup Environ Med 42:284-310.
21
22     Barbier, O; Duran-Sandoval, D; Pineda-Torra, I; et al. (2003) Peroxisome proliferator-activated receptor-alpha
23     induces hepatic expression of the human bile acid glucuronidating UDP-gllucuronosyltransferase 2B4 enzyme. J
24     Biol Chem 278(35):32852-32860.
25
26     Barrass, N; Stewart, M; Warburton, S; et al. (1993) Cell proliferation in the liver and thyroid of C57B1/10J mice
27     after dietary administration of chlordane.  Environ Health Perspect 101 Suppl 5:219-223.
28
79
       Baylin, SB; Herman, JG; Graff, JR; et al. (1998) Alterations in DNA methylation: a fundamental aspect of neoplasia
30     Adv Cancer Res 72:141-196.
31
32     Beaudreuil, S; Lasfargues,  G; Laueriere, L; El, G Z; Fourquet, F; Longuet, C; Halimi, J M; Nivet, H; Buchler, M
33     (2005) Occupational exposure in ANCA-positive patients: a case-control study. Kidney Int 67:1961-1966.
34
3 5     Beckstead, MJ; Weiner, JL; Eger, El; et al. (2000) Glycine and gamma-aminobutyric acid(A) receptor function is
36     enhanced by inhaled drugs of abuse. Mol Pharmacol 57:1199-1205.
37
3 8     Beliles, RP. (2002) Concordance across species in the reproductive and developmental toxicity of
3 9     tetrachloroethylene. Toxicol Ind Health 18:91-106.
40
41     Beliles, RP; Brusick, DJ; Mecler, FJ. (1980) Teratogenic-mutagenic risk of workplace contaminants:
42     trichloroethylene, perchloroethylene and carbon disulfide. Prepared by Litton Bionetics, Inc. NTIS Publication No.
43     PB-82 185-075, NIOSH Contract Report No. 201-77-0047.  Available from: National Technical Information Service,
44     Springfield, VA.
45
46     Bellinger, D; Hu, H; Titlebaum, L; et al. (1994) Attentional correlates of dentin and bone lead levels in adolescents.
47     Arch Environ Health 49:98-105.
48
49     Bentley, P; Calder, I; Elcombe, C; et al. (1993) Hepatic peroxisome proliferation in rodents and its significance for
50     humans.  Food Chem Toxicol 31:857-907.
51
52     Bergamaschi, E; Mutti, A; Bocchi, MC; et al. (1992) Rat model of perchloroethylene-induced renal dysfunctions.
53     Environ Res 59:427-439.
54
55


                    This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-240       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Berger, T and Horner, CM (2003) In vivo exposure of female rats to toxicants may affect oocyte quality. Reprod
  2     Toxicol 17:273-281.
  3
  4     Berlin, K; Edling, C; Persson, B; Ahlborg, et al. (1995) Cancer incidence and mortality of patients with suspected
  5     solvent-related disorders.  Scand J Work Environ Health 21:362-367.
  6
  7     Berman, E; Schlicht, M; Moser, VC; MacPhail, RC. (1995) A multidisciplinary approach to toxicological
  8     screening: I. Systemic toxicity. J Toxicol Environ Health 45:127-143.
  9
10     Bernard, A; Lauwerys, R. (1995) Low-molecular-weight proteins as markers of organ toxicity with special reference
11     to Clara cell protein.  Toxicol Lett 77:145-151.
12
13     Bhattacharya, RK; Schultze, MO. (1971) Properties of DNA isolated from tissues of calves treated with S-(l,2-
14     dichlorovinyl)-L-cysteine. II. Primer-template activity for bacterial DNA polymerases. Arch Biochem Biophys
15     145:575-582.
16
17     Bhattacharya, RK; Schultze, MO. (1972) Properties of DNA treated with S-(l,2-dichlorovinyl)-L-cysteine and a
18     lyase.  Arch Biochem Biophys 153:105-115.
19
20     Bhunya SP, Behera BC. (1987). Relative genotoxicity of trichloroacetic acid (TCA) as revealed by different
21     cytogenetic assays: bone marrow chromosome aberration micronucleus and sperm-head abnormality in the mouse.
22     Mutat Res 188: 215-221.
23
24     Bhunya, SP; Jena GB. (1996) The evaluation of clastogenic  potential of trichloroacetic acid (TCA) in chick in vivo
25     test system. Mutat Res 367: 254-259.
26
27     Birner, G; Richling, C; Henschler, D; et al. (1994) Metabolism of tetrachloroethene in rats: identification of N
28     epsilon- (dichloroacetyl)-L-lysine and N epsilon-(trichloroacetyl)-L-lysine as protein adducts. Chem Res Toxicol
29     7:724-732.
30
31     Birner, G; Rutkowska, A; Dekant, W.  (1996) N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and 2,2,2-
32     trichloroethanol: two novel metabolites of tetrachloroethene in humans after occupational exposure.  Drug Metab
33     Dispos 24:41-48.
34
3 5     Blair, A; Stewart, PA; Tolbert, PE; et al.  (1990) Cancer and  other causes of death among a cohort of dry cleaners.
36     BrJIndMed 47:162-168.
37
3 8     Blair, A; Linos, A; Stewart, PA; et al. (1993) Evaluation of risks for non-Hodgkin's lymphoma by occupation and
39     industry exposures from a case-control study. Am JIndMed 23:301-312.
40
41     Blair, A; Petralia, SA; Stewart. PA. (2003) Extended mortality follow-up of a cohort of dry cleaners. Ann Epidemiol
42     13(l):50-56.
43
44     Blossom, SJ; Gilbert, KM. (2006) Exposure to a metabolite of the environmental toxicant, trichloroethylene,
45     attenuates CD4+ T cell activation-induced cell death by metalloproteinase-dependent FasL shedding. Toxicol Sci
46     92:103-114.
47
48     Blossom, SJ; Pumford, NR; Gilbert, KM. (2004) Activation and attenuation of apoptosis of CD4+ T cells following
49     in vivo exposure to two common environmental toxicants, trichloroacetaldehyde hydrate and trichloroacetic acid. J
50     Autoimmun 23:211-220.
51
52     Blot, WJ; McLaughlin, JK. (1999) The changing epidemiology of esophageal cancer. Semin Oncol  26:2-8.
53
54     Bogen, KT; McKone,  TE. (1988) Linking indoor air and pharmacokinetic models to assess tetrachloroethylene risk.
55     Risk Anal 8:509-520.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-241        DRAFT-DO NOT CITE OR QUOTE

-------
  1    Boice, JD; Marano, DE; Fryzek, JP; et al. (1999) Mortality among aircraft manufacturing workers. Occup Environ
  2    Med 56:581-597.
  3
  4    Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol
  5    70(6):347-55.
  6
  7    Bonse, G; Henschler, D. (1976) Chemical reactivity, biotransformation, and toxicity of polychlorinated aliphatic
  8    compounds. CRC Crit Rev Toxicol 4:395-409.
  9
10    Bonse, G; Urban,  T; Reichert, D; et al. (1975) Chemical reactivity, metabolic oxirane formation and biological
11    reactivity of chlorinated ethylenes in the isolated perfused rat liver preparation.  Biochem Pharmacol 24:1829-1834.
12
13    Bosco, MG; Figa-Talamanca, I; Salerno, S. (1987) Health and reproductive status of female workers in dry cleaning
14    shops. Int Arch Occup Environ Health 59:295-301.
15
16    Boulet, LP (1988) Increases in airway responsiveness following acute exposure to respiratory irritants. Reactive
17    airway dysfunction syndrome or occupational asthma?. Chest 94:476-481.
18
19    Bove, FJ; Fulcomer, MC; Klotz, JB; et al. (1992) Public drinking water contamination and birthweight, and selected
20    birth defects: a case-control study.  New Jersey Department of Health, New Brunswick, NJ.
21
22    Bove, FJ; Fulcomer, MC; Klotz, JB; et al. (1995) Public drinking water contamination and birth outcomes [see
23    comments].  AmJEpidemiol 141:850-862.
24
25    Bowler, RM; Mergler, D; Huel, G; et al. (1991) Neuropsychological impairment among former microelectronics
26    workers. Neurotoxicology 12:87-103.
27
28    Bowman, KJ. (1982) A method for quantitative scoring of the Farnsworth Panel D-15. Acta Opthalmol 60:907-916.
29
30    Breslow, NE; Day, NE. (1994) Statistical method in cancer research. Vol. 2: The design and analysis of cohort
31    studies. New York:  Oxford University Press.
32
33    Briving, C; Jacobson, I; Hamberger, A; et al.  (1986) Chronic  effects of perchloroethylene and trichloroethylene on
34    the gerbil brain amino acids and glutathione.  Neurotoxicology 7:101-108.
35
36    Broadwell, DK; Darcey, DJ; Hudnell, HK; et al. (1995) Work-site clinical and neurobehavioral assessment of
37    solvent-exposed microelectronics workers. Am J Ind Med 27:677-698.
38
39    Brodkin, CA; Daniell, W; Checkoway, H; et al. (1995) Hepatic ultrasonic changes in workers exposed to
40    perchloroethylene. Occup Environ Med 52:679-685.
41
42    Bronzetti, G; Galli, A; Corsi, C; et al. (1984)  Genetic and biochemical investigation on chloral hydrate in vitro and
43    in vivo. Mutat Res 141: 19-22.
44
45    Brown, RP; Delp, MD; Lindstedt, SL; et al. (1997) Physiological parameter values for physiologically based
46    pharmacokinetic models. Toxicol Industrial Health 13:407-484.
47
48    Brown, LM; Hoover, R; Silverman, D; et al. (2001) Excess incidence of squamous cell esophageal cancer among
49    US Black men: role of social class and other risk factors. Am J Epidemiol 153:114-122.
50
51    Brownson, RC; Alavanja, MC; Chang, JC. (1993) Occupational risk factors for  lung cancer among nonsmoking
52    women: a case-control study in Missouri (United States).  Cancer Causes Control 4:449-454.
53
54    Bruschi, SA; Bull, RJ. (1993) In vitro cytotoxicity of mono-,  di-, and trichloroacetate and its modulation by hepatic
55    peroxisome proliferation.  FundamAppl Toxicol 21:366-375.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                       4-242       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Buben, JA; O=Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
  2    and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
  O
  4    Buelke-Sam, J; Kimmel, CA; Adams, J; et al. (1985) Collaborative behavioral teratology study: results.
  5    Neurobehav Toxicol Teratol 7:591-624.
  6
  7    Bukowski, JA; Huebner, WW; Schnatter, AR; et al. (2003) An analysis of the risk of B-lymphocyte malignancies in
  8    industrial cohorts. J Toxicol Environ Health Part A 66:5 81 -5 97.
  9
10    Bull, RJ. (2000) Mode of action of liver tumor induction by trichloroethylene and its metabolites, trichloroacetate
11    and dichloroacetate. Environ Health Perspect 108 Suppl 2:241-259.
12
13    Bull, RJ. (2004) Trichloroethylene and liver tumors in mice. Symposium on new scientific research related to the
14    health effects of TCE. Washington, DC.
15
16    Bull, RJ; Sanchez, IM; Nelson, MA; etal. (1990) Liver tumor induction in B6C3F1 mice by dichloroacetate and
17    trichloroacetate.  Toxicology 63:341-359.
18
19    Bull, RJ; Orner, GA; Cheng, RS; et al. (2002) Contribution of dichloroacetate and trchloroacetate to liver tumor
20    induction in mice by trichloroethylene. Toxicol App Pharmacol 182(l):55-65.
21
22    Bull, RJ; Sasser,  LB; Lei, XC. (2004) Interactions in the tumor-promoting activity of carbon tetrachloride,
23    trichloroacetate, and dichloroacetate in the liver of male B6C3F1 mice. Toxicol 199:169-83.
24
25    Bursch, W; Lauer, B; Timmermann-Trosiener, I; et al. (1984) Controlled death (apoptosis) of normal and putative
26    preneoplastic cells in rat liver following withdrawal of tumor promoters.  Carcinogenesis 5:453-458.
27
28    Byczkowski, JZ; Fisher, JW. (1994) Lactational transfer of tetrachloroethylene in rats. Risk Anal 14:339-349.
29
30    Byczkowski, JZ; Fisher, JW. (1995) A computer program linking physiologically based pharmacokinetic model
31    with cancer risk assessment for breast-fed infants. Comput Methods Programs Biomed 46:155-163.
32
33    Byers, VS; Levin, AS; Ozonoff, DM; et al. (1988) Association between clinical symptoms and lymphocyte
34    abnormalities in a population with chronic domestic exposure to industrial solvent- contaminated domestic water
35    supply and a high incidence of leukaemia.  Cancer Immunol Immunother 27:77-81.
36
37    Cai, SX; Huang,  MY; Chen, Z; et al. (1991) Subjective symptom increase among dry-cleaning workers exposed to
3 8    tetrachloroethylene vapor.  Ind Health 29:111-121.
39
40    Cal EPA (California Environmental Protection Agency). (2001) Public health goal for tetrachloroethylene in
41    drinking water. Prepared by Office of Environmental Health Hazard Assessment, California EPA. Available online
42    at http://www.oehha.org/water/phg/pdf/PCEAug2001.pdf.
43
44    Caldwell, DJ. (1999) Review of mononuclear cell leukemia in F-344 rat bioassays and its significance to human
45    cancer risk: A case study using alkyl phthalates.  Regul Toxicol Pharmacol 30:45-53.
46
47    Caldwell, JC; Keshava, N; Evans, MV. (2008). Difficulty of mode of action determination for trichloroethylene: An
48    example of complex interactions of metabolites and other chemical exposures. Environ Mol Mutagen 49(2): 142-154.
49
50    Campagna, D; Mergler, D;  Huel, G; et al.  (1995) Visual dysfunction among styrene-exposed workers. Scand J Work
51    Environ Health 21:382-390.
52
53    Campagna, D; Gobba, F; Mergler, D; et al. (1996) Color vision loss among styrene-exposed workers
54    neurotoxicological threshold assessment.  Neurotoxicology 17:367-373.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-243      DRAFT-DO NOT CITE OR QUOTE

-------
  1     Cano, MI; Pollan, M. (2001) Non-Hodgkin=s lymphomas and occupation in Sweden.  Int Arch Occup Environ
  2     Health 74:443-449.
  O

  4     Capone, JP. (1994) Mechanisms of peroxisome proliferator-induced carcinogenesis: historical perspectives and
  5     current status. Prepared by Eastern Research Group, Inc., Lexington, MA EPA Contract No. 69-D9-0133.
  6     Available from: National Center for Environmental Assessment, Office of Research and Development, U.S. EPA,
  7     Washington, DC.
  8
  9     Carney, EW; Thorsrud, BA; Dugard, PH; et al.  (2006) Developmental toxicity studies in Crl:CD (SD) rats following
10     inhalation exposure to trichloroethylene and perchloroethylene. Birth Defects Research (Part B) 77:405-412
11
12     Carter, JH; Carter, HW; Deddens, JA; et al. (2003) A 2-year dose-response study of lesion sequences during
13     hepatocellular carcinogenesis in the male B6C3F1 mouse given the drinking water chemical dichloroacetic acid.
14     Environ Health Perspect 111 (1): 5 3 -64.
15
16     Cattley, RC; Glover, SE. (1993) Elevated 8-hydroxydeoxyguanosine in hepatic DNA of rats following exposure to
17     peroxisome proliferators: relationship to  carcinogenesis and nuclear localization. Carcinogenesis 14:2495-2499.
18
19     Cattley, RC; Popp, JA. (1989) Differences between the promoting activities of the peroxisome proliferator WY-
20     14,643 and phenobarbital in rat liver. Cancer Res 49:3246-3251.
21
22     Cattley, RC; Roberts, RA. (2000) Peroxisome proliferators and carcinogenesis: editorial perspectives. Mutat Res
23     448:117-119.
24
25     Cattley, RC; Marsman, DS; Popp, JA. (1991) Age-related susceptibility to the carcinogenic effect of the peroxisome
26     proliferator WY-14,643 in rat liver.  Carcinogenesis 12:469-473.
27
28     Cattley, RC; DeLuca, J; Elcombe, C; et al. (1998) Do peroxisome proliferating compounds pose a
29     hepatocarcinogenic hazard to humans? Regul Toxicol Pharmacol 27:47-60.
30
31     Cavalleri, A; Gobba, F; Paltrinieri, M; et al. (1994) Perchloroethylene exposure can induce colour vision loss.
32     Neurosci Lett 179:162-166.
33
3 4     Chang, LW; Daniel, FB; DeAngelo, AB. (1992) Analysis of DNA strand breaks induced in rodent liver in vivo,
3 5     hepatocytes in primary culture, and a human cell line by chlorinated acetic acids and chlorinated acetaldehydes.
36     Environ MolMutagen 20:277-288.
37
38     Channel, SR; Latendresse, JR; Kidney, JK; et al. (1998) A subchronic exposure to trichloroethylene causes lipid
39     peroxidation and hepatocellular proliferation in male B6C3F1 mouse liver. Toxicol Sci 43:145-154.
40
41     Charbonneau, M; Short, BB; Lock, EA; et al. (1987) Mechanism of petroleum-induced sex-specific protein droplet
42     nephropathy and renal cell proliferation in Fischer 344 rats: Relevance to humans. In: Hemphill, DD, ed.  Trace
43     substances in environmental health Vol. 21. Columbia, MO:  University of Missouri; pp. 263-273.
44
45     Chawla, A; Schwarz, EJ; Dimaculangan, DD; et al. (1994) Peroxisome proliferator-activated receptor (PPAR)
46     gamma: adipose- predominant expression and induction early in adipocyte differentiation. Endocrinology 135:798-
47     800.
48
49     Chen, HH; Chan, MH; Fu, SH. (2002a) Behavioural effects of tetrachloroethylene exposure in rats: acute and
50     subchronic studies. Toxicology 170:201-209.
51
52     Chen, SJ; Wang, JL; Chen, JH;  et al. (2002b) Possible involvement of glutathione and p53 in trichloroethylene- and
53     perchloroethylene-induced lipid peroxidation and apoptosis in human lung cancer cells. Free  Radic Biol Med
54     33:464-472.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-244       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Chevalier, S; Roberts, RA. (1998) Perturbation of rodent hepatocyte growth control by nongenotoxic
  2     hepatocarcinogens: mechanisms and lack of relevance for human health (review). Oncol Rep 5:1319-1327.
  3
  4     Chia, SE; Ong, CN; Tsakok, MF; et al. (1996) Semen parameters in workers exposed to trichloroethylene.  Reprod
  5     Toxicol 10:295-299.
  6
  7     Chia, SE; Goh, VH; Ong, CN. (1997) Endocrine profiles of male workers with exposure to trichloroethylene. Am J
  8     IndMed 32:217-222.
  9
10     Chien, Y-C. (1997) The influences of exposure pattern and duration on elimination kinetics and exposure
11     assessment of tetrachloroethylene in humans. PhD thesis, Rutgers University, New Brunswick, NJ.
12
13     Chiu, WA; Bois, FY. (2006) Revisiting the population toxicokinetics of tetrachloroethylene 30. Arch Toxicol
14     80:382-385.
15
16     Chow, WH; McLaughlin, JK; Malker, HS; et al. (1995) Esophageal cancer and occupation in a cohort of Swedish
17     men. Am J IndMed 27:749-757.
18
19     Chu, S; Huang, Q; Alvares, K; et al. (1995) Transformation of mammalian cells by overexpressing H2O2-generating
20     peroxisomal fatty acyl-CoA oxidase.  Proc Natl Acad Sci U S A 92:7080-7084.
21
22     Clevel, J; Manderau, L; Conso; F; et al. (1998) Occupational exposure to solvents and hairy cell leukaemia. Occup
23     EnvironMed 55:59-64.
24
25     Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
26     pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79(2):381-393.
27
28     Clewell, HJ; Gentry, PR; Kester, JE; et al. (2005) Evaluation of physiologically based pharmacokinetic models in
29     risk assessment: an example with perchloroethylene. Crit Rev Toxicol (accepted for publication).
30
31     Cimini, A; Cristiano, L; Bernardo, A; et al. (2000) Presence and inducibility of peroxisomes in a human
32     glioblastoma cell line. Biochim Biophys Acta 1474:397-409.
33
34     Cohn, P; Klotz, J; Bove, F; et al. (1994) Drinking water contamination and the incidence of leukemia and non-
35     Hodgkin's Lymphoma.  Environ Health Perspect 102:556-561.
36
37     Cojocel,  C; Beuter, W; Muller, W; et al. (1989) Lipid peroxidation: a possible mechanism of trichloroethylene-
3 8     induced nephrotoxicity. Toxicology 5 5:131 -141.
39
40     Coler HR; Rossmiller, HR. (1953) Tetrachloroethylene exposure in a small industry.  AMA Arch Ind Hyg Occup
41     Med 8:227-233.
42
43     Cooper, GS; Parks, CG; Treadwell, EL; et al. (2004) Occupational risk factors for the development of systemic
44     lupus erythematosus. J Rheumatol 31:1928-1933.
45
46     Gorton, JC; Fan, LQ; Brown, S; et al. (1998) Down-regulation of cytochrome P450 2C family members and positive
47     acute-phase response gene expression by peroxisome proliferator chemicals. Mol Pharmacol  54:463-473.
48
49     Gorton, JC; Anderson,  SP; Stauber, A. (2000) Central role of peroxisome proliferator-activated receptors in the
50     actions of peroxisome proliferaters.  Annu Rev Pharmacol Toxicol 40:491-518.
51
52     Costa, AK; Ivanetich, KM. (1980). Tetrachloroethylene metabolism by the hepatic microsomal cytochrome P-450
53     system. Biochem Pharmacol 29(20):2863-2869.
54
55     Costa, C; Barbara, M; Catania, et al. (2004) Cytotoxicity evaluation after coexposure to perchloroethylene and
56     selected peroxidant drugs in rat hepatocytes. Toxicol In Vitro 18:37-44.

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                       4-245       DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2     Costantini, AS; Miligi, L; Kriebel, D; et al. (2001) A multicenter case-control study in Italy on
  3     hematolymphopoietic neoplasms and occupation. Epidemiology 12:78-87.
  4
  5     Costas K; Knorr RS; Condon SK. (2002) A case-control study of childhood leukemia in Woburn, Massachusetts: the
  6     relationship between leukemia incidence and exposure to public drinking water. Sci Tot Environ 300:23-35.
  7
  8     Cruz, SL; Mirshahi, T; Thomas, B; et al. (1998) Effects of the abused solvent toluene on recombinant N-methyl-D-
  9     aspartate and non-N-methyl-D-aspartate receptors expressed inXenopus oocytes. J Pharmacol Exp Ther 286:334-
10     340.
11
12     Czirjak, L; Pocs, E; Szegedi, G. (1994) Localized scleroderma after exposure to organic solvents. Dermatology
13     189:399-401.
14
15     Czirjak, L; Kumanovics, G.  (2002) Exposure to solvents in female patients with scleroderma. Clin Rheumatol
16     21:114-118.
17
18     Dallas, CE; Chen, XM; O=Barr, K; et al. (1994) Development of a physiologically based pharmacokinetic model for
19     perchloroethylene using tissue concentration-time data. Toxicol Appl Pharmacol 128-159.
20
21     Daniel, J. (1963) The metabolism of 36-Cl-labeled trichloroethylene and tetrachloroethylene in the rat. Biochem
22     Pharmacol 12:795-802.
23
24     Daniel, FB; DeAngelo, AB; Stober, JA; et al.  (1992) Hepatocarcinogenicity of chloral hydrate, 2-
25     chloroacetaldehyde, and dichloroacetic acid in the male -6C3F1 mouse. Fundam Appl Toxicol 19:159-168.
26
27     Daniel, FB; Meier, JR; DeAngelo, AB. (1993) Advances in research on carcinogenic and genotoxic by-products of
28     chlorine disinfection: chlorinated hydroxyfuranones and chlorinated acetic acids. Ann 1st Super Sanita 29:279-291.
29
30     Daniell, WE; Claypoole, KH; Checkoway, H; et al. (1999) Neuropsychological function in retired workers with
31     previous long-term occupational exposure to solvents. Occup Environ Med 56:93-105.
32
33     Davidson, IW; Beliles, RP. (1991) Consideration of the target organ toxicity of trichloroethylene in terms of
34     metabolite toxicity and pharmacokinetics. Drug Metab Rev 23:493-599.
35
36     DeAngelo, AB. (2000) Response to AEpigenetic mechanisms of chemical carcinogenesis@ by James E. Klaunig,
37     Lisa M. Kamendulis, and Xu Yong. Hum Exp Toxicol 19:561-562.
38
39     DeAngelo, AB; Daniel, FB; McMillan, L; et al. (1989) Species and strain sensitivity to the induction of peroxisome
40     proliferation by chloroacetic acids.  Toxicol Appl Pharmacol 101:285-298.
41
42     DeAngelo, AB; Daniel, FB; Stober, JA; et al.  (1991) The carcinogenicity of dichloroacetic acid in the male B6C3F1
43     mouse.  Fund Appl Toxicol 16:337-347.
44
45     DeAngelo, AB; Daniel, FB; Most, BM; et al. (1996) The carcinogenicity of dichloroacetic acid in the male Fischer
46     344 rat. Toxicology 114:207-221.
47
48     DeAngelo, AB; Daniel, FB; Most, BM; et al. (1997) Failure of monochloroacetic acid and trichloroacetic acid
49     administered in the drinking water to produce liver cancer in male  F344/N rats. J Toxicol Environ Health 52:425-
50     445.
51
52     DeAngelo, AB; George, MH; House, DE. (1999) Hepatocarcinogenicity in the male B6C3F1 mouse following a
53     lifetime exposure to dichloroacetic  acid in the drinking water: dose-response determination and modes of action. J
54     Toxicol Environ Health 58:485-507.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                       4-246       DRAFT-DO NOT CITE OR QUOTE

-------
  1     DeAngelo, AB; Daniel, FB; Wong, DM; et al. (2008). The induction of hepatocellular neoplasiaby trichloroacetic
  2     acid administered in the drinking water of the male -6C3F1 mouse. J Tox and Env Health 71:1053-1065.
  3
  4     De Ceaurriz, J; Desiles, JP; Bonnet, P; et al. (1983) Concentration-dependent behavioral changes in mice following
  5     short-term inhalation exposure to various industrial solvents.  Toxicol Appl Pharmacol 67:383-389.
  6
  7     Degrassi, F and Tanzarella C. (1988). Immunofluorescent staining of kinetochores in micronuclei: A new assay for
  8     the detection of aneuploidy. Mutat. Res. 203: 339-345.
  9
10     Dekant, W. (2001) Chemical-induced nephrotoxicity mediated by glutathione S-conjugate formation. Toxicol Lett
11     124:21-36.
12
13     Dekant, W; Vamvakas, S; Berthold, K; et al. (1986) Bacterial beta-lyase mediated cleavage and mutagenicity of
14     cysteine conjugates derived from the nephrocarcinogenic alkenes trichloroethylene, tetrachloroethylene and
15     hexachlorobutadiene. Chem Biol Interact 60:31-45.
16
17     Delahunt, B; Bethwaite, PB; Nacey, JN.  (1995) Occupational risk for renal cell carcinoma. A case-control study
18     based on the New Zealand Cancer Registry. Br J Urol 75:578-582.
19
20     de la Iglesia, FA; Gough, AW; Sigler, RE; et al. (1997) Alpha 2u-globulin nephropathy and ravens: do ravens of a
21     different feather flock together? Environ Health Perspect 105:903-904.
22
23     Delfino, RJ; Gong, H, Jr.; Linn, WS; et al. (2003a) Asthma symptoms in Hispanic children and daily ambient
24     exposures to toxic and criteria air pollutants. Env Health Perspect 111(4):647-656.
25
26     Delfino, RJ; Gong, H; Linn, WS; et al. (2003b) Respiratory symptoms and peak expiratory flow in children with
27     asthma in relation to volatile organic compounds in exhaled breath and ambient air. Journal of Exposure Analysis
28     and Environ Epidemiol 13:348-363.
29
30     Deonandan, R; Campbell, K; Ostbye, T;  et al. (2000) A comparison of methods for measuring socio-economic
31     status by occupation or postal area. Chronic Dis Can21:114-118.
32
33     Desvergne, B; Wahli, W. (1995) PPAR: A key factor in nutrient/gene interactions. In: Baeujerle, P., ed. Inducible
34     transcription. Boston: Birkhauser; pp.  142-176.
35
36     Dietrich, DR. (1997) Doubting nongenotoxic mechanisms of renal cancer: comparing apples and oranges in the
3 7     alpha2u-globulin hypothesis. Environ Health Perspect  105:898-902.
38
39     Diodovich, C; Ferrario, D; Casati, B; Malerba, I; Marafante, E; Parent-Massin, D; Gribaldo, L (2005) Sensitivity of
40     human cord blood cells to tetrachloroethylene: cellular and molecular endpoints. Arch Toxicol 79:508-514.
41
42     Dobrev, ID; Andersen, ME; Yang, RS. (2001) Assessing interaction thresholds for trichloroethylene in combination
43     with tetrachloroethylene and 1,1,1-trichloroethane using gas uptake studies and PBPK modeling. Arch Toxicol
44     75(3):134-144.
45
46     Dobrev, ID; Andersen, ME; Tang RS. (2002) In silica toxicology: simulating interaction thresholds for human
47     exposure to mixtures of trichloroethylene, tetrachloroethylene, and 1,1,1-trichloroethane. Env Health Perspect
48     110(10):1031-1039.
49
50     Doherty, AT; Ellard, S; Parry, EM; et al. (1996) An investigation into the activation and deactivation of chlorinated
51     hydrocarbons to genotoxins in metabolically competent human cells.  Mutagenesis 11:247-274.
52
53     Donoghue, AM; Dryson, EW; Wynn-Williams, G. (1995) Contrast sensitivity in organic-solvent-induced chronic
54     toxic encephalopathy. J Occup Environ Med 37:1357-1363.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-247      DRAFT-DO NOT CITE OR QUOTE

-------
  1     Dosemeci, M; Cocco, P; Chow, WH. (1999) Gender differences in risk of renal cell carcinoma and occupational
  2     exposures to chlorinated aliphatic hydrocarbons.  Am JIndMed 36:54-59.
  3
  4     Doyle, P; Roman, E; Beral, V; et al. (1997) Spontaneous abortion in dry cleaning workers potentially exposed to
  5     perchloroethylene. Occup EnvironMed 54:848-853.
  6
  7     Dreessen, B.; Westphal, G.; Hunger, I; et al. (2003). "Mutagenicity of the glutathione and cysteine S-conjugates of
  8     the haloalkenes l,l,2-trichloro-3,3,3-trifluoro-l-propene and trichlorofluoroethene in the Ames test in comparison
  9     with the tetrachloroethene-analogues." Mutat Res 539(1-2): 157-66.
10
11     Dreyer, C; Krey, G; Keller, H; et al. (1992) Control of the peroxisomal beta-oxidation pathway by a novel family of
12     nuclear hormone receptors. Cell 68:879-887.
13
14     Dybing, E; Mikalsen, S-O; Huttunen, J; et al. (1995) Peroxisome proliferation, genotoxicity and carcinogenicity.
15     IARC Technical Report 24:55-85.
16
17     Eacho, PI; Larder, TL; Brodhecker, CA. (1991) Hepatocellular DNA synthesis in rats given peroxisome
18     proliferating agents: comparison of WY-14,643 to clofibric acid, nafenopinandLY171883.  Carcinogenesis
19     12:1557-1561.
20
21     Ebrahim, AS; Babakrishnan, K; Sakthisekaran, D. (1996) Perchloroethylene-induced alterations in glucose
22     metabolism and their prevention by 2-deoxy-D-glucose and vitamin E in mice. J Appl Toxicol 16:339-348.
23
24     Ebrahim, AS; Babu, E; Thirunavukkarasu, C; et al. (2001) Protective role of vitamin E, 2-deoxy-D-glucose, and
25     marine on perchloroethylene induced alterations in ATPases. Drug Chem Toxicol 24(4):429-437.
26
27     Echeverria,  D; Heyer, N; Checkoway, H; et al. (1994) Behavioral investigation of occupational exposure to solvents:
28     Perchloroethylene among dry cleaners, and styrene among reinforced fiberglass laminators.  Final Report. Report
29     prepared for the Centers for Disease Control and Prevention under Grant No. 5 R01 OHo2719-03. Battelle Centers
3 0     for Public Health Research and Evaluation.
31
32     Echeverria,  D; White, RF; Sampaio, C. (1995) A behavioral evaluation of PCE exposure in patients and dry
33     cleaners: a possible relationship between clinical and preclinical effects. J Occup Environ Med 37:667-680.
34
35     Elcombe, CR. (1985) Species differences in carcinogenicity and peroxisome proliferation due to trichloroethylene: a
36     biochemical human hazard assessment. Arch Toxicol Suppl 8:6-17.
37
38     Elcombe, CR; Mitchell, AM. (1986) Peroxisome proliferation due to di(2-ethylhexyl) phthalate (DEHP): species
39     differences and possible mechanisms.  Environ Health Perspect 70:211-219.
40
41     Elcombe, CR; Rose, MS; Pratt, IS. (1985) Biochemical, histological, and ultrastructural changes in rat and mouse
42     liver following the administration of trichloroethylene: possible relevance to species differences in
43     hepatocarcinogenicity.  Toxicol ApplPharmacol79:365-376.
44
45     Elfarra, AA.; Krause RJ. (2007) S-(l,2,2-trichlorovinyl)-L-cysteine sulfoxide, a reactive metabolite of S-(l,2,2-
46     Trichlorovinyl)-L-cysteine formed in rat liver and kidney microsomes, is a potent nephrotoxicant.  J Pharmacol Exp
47     Ther 321(3): 1095-101.
48
49     Epstein, DL; Nolen, GA; Randall, JL; et al. (1992) Cardiopathic effects of dichloroacetate in the fetal Long-Evans
50     rat.  Teratology 46:225-235.
51
52     Ericksson, P; von Rosen, D; Viberg, H; et al. (2005) Developmental toxicology in the  neonatal mouse: the use of
53     randomly selected individuals as statistical unit compared to the litter in mice neonatally exposed to PBDE 99.
54     Toxicologist84(S-l):219-220.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-248       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Eskenazi, B; Fenster, L; Hudes, M; et al. (1991a) A study of the effect of perchloroethylene exposure on the
  2     reproductive outcomes of wives of dry-cleaning workers. Am J Ind Med 20:593-600.
  3
  4     Eskenazi, B; Wyrobek,  AJ; Fenster, L; et al. (1991b) A study of the effect of perchloroethylene exposure on semen
  5     quality in dry cleaning workers.  Am J Ind Med 20:575-591.
  6
  7     Ethell, DW; Buhler, LA. (2003) Fas ligand-mediated apoptosis in degenerative disorders of the brain. J Clin
  8     Immunol 23:439-446.
  9
10     Fabbro-Peray, P; Daures, JP; Rossi,  JF. (2001) Environmental risk factors for non-Hodgkin=s lymphoma: a
11     population-based case-control study in Languedoc-Roussillon, France.  Cancer Causes Control 12:201-212.
12
13     Fagliano, J; Berry, M; Bove, F; et al. (1990) Drinking water contamination and the incidence of leukemia: an
14     ecologic study. Am J Public Health 80:1209-1212.
15
16     Fan, LQ; Coley, J; Miller, RT; et al. (2003) Opposing mechanisms of NADPH-cytochrome P450 oxidoreductase
17     regulation by peroxisome proliferators. Biochem Pharmacol. 65:949-959.
18
19     Ferreira-Gonzalez, A; DeAngelo, AB; Nasim, S; et al. (1995) Ras oncogene activation during hepatocarcinogenesis
20     inB6C3Fl male mice by dichloroacetic and trichloroacetic acids.  Carcinogenesis 16:495-500.
21
22     Ferroni,  C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
23     perchloroethylene. Neurotoxicology 13:243-247.
24
25     Finckh, A; Cooper, GS; Chibnik; et  al. (2007) Occupational exposures and risk of systemic lupus erythematosus.
26     Arthritis Rheum 54:3648-3654.
27
28     Flavell, DM; Pineda, TI; Jamshidi, Y; et al. (2000) Variation in the PPAR-a gene is associated with altered function
29     in vitro and plasma lipid concentrations in Type II diabetic subjects. Diabetologia 43:673-680.
30
31     Flower, DR; North, AC; Attwood, TK. (1993). Structure and sequence relationships in the lipcalins and related
32     proteins. Protein Sci 2:753-761.
33
34     Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
3 5     solvents.  Int Arch Occup Environ Health 52:1-9.
36
37     Fredriksson, M; Bengtsson,  NO; Hardell, L; et al. (1989) Colon cancer, physical activity, and occupational
38     exposures. A case-control study. Cancer 63:1838-1842.
39
40     Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult  mice orally exposed to tri- and
41     tetrachloroethylene as neonates. Toxicol Lett 66:13-19.
42
43     Frenette, B; Mergler, D; Bowler, R.  (1991) Contrast-sensitivity loss in a group of former microelectronics workers
44     with normal visual acuity. Optom Vis Sci 68:556-560.
45
46     Gaillard, Y; Billault, F; Pepin, G. (1995) Tetrachloroethylene fatality: case report and simple gas chromatographic
47     determination in blood and tissues. Forensic Sci Int 76:161-168.
48
49     Gaipl, US; Kuhn, A; Sheriff, A; et al. (2006) Clearance of apoptotic cells in human SLE. Curr Dir Autoimmun
50     9:173-187.
51
52     Galloway, SM; Armstrong, MJ; Reuben, C; et al. (1987) Chromosome aberrations and sister chromatid exchanges in
53     Chinese  hamster ovary cells: evaluations of 108 chemicals.  Environ Mol Mutagen 10 Suppl 10:1-175.
54
55     Gama, C; Meira, JB. (1978) Occupational acro-osteolysis. J Bone Joint Surg Am 60:86-90.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-249       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Canning, AE; Brunk, U; Dallner, G. (1983) Effects of dietary di(2-ethylhexyl)phthalate on the structure and function
  2     of rat hepatocytes.  Biochim Biophys Acta 763:72-82.
  3
  4     Garabrant, DH; Lacey, J; Laing, TJ; et al. (2003) Scleroderma and solvent exposure among women. Am J Epidemiol
  5     157/6:493-500.
  6
  7     Garetano, G; Gochfeld, M. (2000) Factors influencing tetrachloroethylene concentrations in residences above dry-
  8     cleaning establishments. Arch Environ Health 55:59-68.
  9
10     Gamier, R; Bedouin, J; Pepin, et al. (1996) Coin-operated dry cleaning machines may be responsible for acute
11     tetrachloroethylene poisoning: report of 26 cases including one death.  JToxicol ClinToxicol 34:191-197.
12
13     Geller, AM; Hudnell, HK. (1997) Critical issues in the use and analysis of the Lanthony Desaturate Color Vision
14     test. Neurotoxicol Teratol 19:455-465.
15
16     Geller, AM (2001) A table of color distance scores for quantitative scoring of the Lanthony
17     Desaturate color vision test. Neurotoxicol Teratol 23:265-267.
18
19     Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
20     tetrachloroethylene. Am JIndMed 21:661-671.
21
22     Gentry, PR; Covington, TR; Clewell, HJ. (2003) Evaluation of the potential impact of pharmacokinetic differences
23     on tissue dosimetry in offspring during pregnancy and lactation. Regul Toxicol Pharmacol 38:1-16.
24
25     George, MH; Moore, T; Kilburn, S; et al. (2000) Carcinogenicity of chloral hydrate administered in drinking water
26     to the male F344/N rat and male B6C3F(1) mouse. Toxicol Pathol 28:610-618.
27
28     Germolec, DR; Yang, RS; Ackermann, MF; et al. (1989) Toxicology studies of a chemical mixture of 25
29     groundwater contaminants. II. ImmunosuppressioninB6C3Fl mice. Fundam Appl Toxicol  13:377-387.
30
31     Ghantous, H; Danielsson, BR; Dencker, L; et al. (1986) Trichloroacetic acid accumulates in murine amniotic fluid
32     after tri- and tetrachloroethylene inhalation. Acta Pharmacol Toxicol (Copenh) 58:105-114.
33
34     Gilbert, KM; Whitlow, AB; Pumford, NR. (2004) Environmental contaminant and disinfection by-product
3 5     trichloroacetaldehyde stimulates T  cells in vitro. Int Immunopharmacol 4:25-36.
36
37     Oilman, AG; Rail, TW; Nies, AS; et al. (1990) The pharmacological basis of therapeutics. New York: McGraw-
38     Hill; pp. 244-573.
39
40     Glinghammar, B; Skogsberg, J; Hamsten, A; et al. (2003) PPARdelta activation induces COX-2 gene expression and
41     cell proliferation in human hepatocellular carcinoma cells. Biochem Biophys Res Commun 308(2):361-368.
42
43     Gobba, F; Righi, E; Fantuzzi, G; et al. (1998) Two-year evolution of perchloroethylene-induced color-vision loss.
44     Arch Environ Health 53:196-198.
45
46     Goh, VH; Chia, SE; Ong, CN. (1998) Effects of chronic exposure to low doses of trichloroethylene on steroid
47     hormone and insulin levels in normal men. Environ Health Perspect 106:41-44.
48
49     Goldman, JA. (1996) Connective tissue disease in people exposed to organic chemical solvents: systemic sclerosis
50     (scleroderma) in dry cleaning plant and aircraft industry workers.  JCR: J Clin Rheumatol 2:185-190.
51
52     Goldsworthy, TL; Popp, JA. (1987) Chlorinated hydrocarbon-induced peroxisomal enzyme activity in relation to
53     species and organ carcinogenicity.  Toxicol Appl Pharmacol 88:225-233.
54
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-250       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Goldsworthy, TL; Lyght, O; Burnett, VL; et al. (1988) Potential role of alpha-2 mu-globulin, protein droplet
  2     accumulation, and cell replication in the renal carcinogenicity of rats exposed to trichloroethylene,
  3     perchloroethylene, and pentachloroethane.  Toxicol Appl Pharmacol 96:367-379.
  4
  5     Goodman, DG; Ward, JM; Squire, RA; et al. (1979) Neoplastic and non-neoplastic lesions in aging F344 rats.
  6     Toxicol Appl Pharmacol 48:237-248.
  7
  8     Graham, MJ; Wilson, SA; Winham, MA; et al. (1994) Lack of peroxisome proliferation in marmoset liver following
  9     treatment with ciprofibrate for 3 years. Fundam Appl Toxicol 22:58-64.
10
11     Grasl-Kraupp, B; Huber, W; Schulte-Hermann, R. (1993) Are peroxisome proliferators tumour promoters in rat
12     liver? In: Gibson,  G and Lake, B , eds. Perixosomes: biology and importance in toxicology and medicine. London:
13     Taylor and Francis; pp. 667-693.
14
15     Green, T; Odum, J. Nash,  JA; et al. (1990) Perchloroethylene-induced rate kidney tumors: an investigation of the
16     mechanisms involved and their relevance to human. Toxicol Appl Pharmacol 103:77-89.
17
18     Green, SM; Khan, MF; Kaphalia, BS; et al. (2001) Immunohistochemical localization of trichloroacylated protein
19     adducts in tetrachloroethene-treated mice. J Toxicol Environ Health A 63:145-157.
20
21     Griffin, J M; Blossom, S J; Jackson, S K; Gilbert, K M; Pumford, N R  (2000a) Trichloroethylene accelerates an
22     autoimmune response by Thl T cell activation in MRL +/+ mice. Immunopharmacology 46:123-137.
23
24     Griffin, JM; Gilbert, KM;  Pumford, NR. (2000b) Inhibition of CYP2E1 reverses CD4+ T-cell alterations in
25     trichloroethylene-treated MRL+/+ mice. Toxicol Sci 54:384-389.
26
27     Griffin, JM; Gilbert, KM;  Lamps, LW; et al. (2000c) CD4(+) T-cell activation and induction of autoimmune
28     hepatitis following trichloroethylene treatment in MRL+/+ mice. Toxicol Sci 57:345-352.
29
30     Grobin, AC; Matthews, DB; Devaud, LL; et al. (1998) The role of GABA(A) receptors in the acute and chronic
31     effects of ethanol. Psychopharmacology (Berl) 139:2-19.
32
33     Gualandi, G. (1987) Use of alpha- and beta-tubulin mutants for the study of spontaneous and induced chromosomal
34     mis-distribution inAspergillus nidulans. Mutat. Res. 178: 33-41.
35
36     Hake, CL; Stewart, RD. (1977) Human exposure to tetrachloroethylene: inhalation and skin contact. Environ Health
37     Perspect 21:231-238.
38
39     Hakkola, J; Pasanen, M; Hukkanen, J; et al. (1996a) Expression of xenobiotic-metabolizing cytochrome  P450 forms
40     in human full-term placenta. Biochem Pharmacol 51:403 -411.
41
42     Hakkola, J; Raunio, H; Purkunen, R; et al. (1996b) Detection of cytochrome P450 gene expression in human
43     placenta in first trimester of pregnancy. Biochem Pharmacol  52:379-83.
44
45     Hakkola, J; Pelkonen, O; Pasanen, M; et al. (1998) Xenobiotic-metabolizing cytochrome P450 enzymes  in the
46     human feto- placenta! unit: role in intrauterine toxicity. Crit Rev Toxicol 28:35-72.
47
48     Haider, CA; Warne, TM; Hatoum, NS. (1984) Renal toxicity of gasoline and related petroleum naphthas in male rats.
49     In: Mehlman, MA; Hemstreet, GP;  Thorpe, JJ; et al., eds. Advances in modern environmental toxicology, vol. VII.
50     Renal effects of petroleum hydrocarbons. Princeton, NJ:  Princeton Scientific.
51
52     Hanefeld, M; Kemmer, C; Kadner,  E. (1983) Relationship between morphological changes and lipid-lowering action
53     of p-chlorphenoxyisobutyric acid (CPIB) on hepatic mitochondria and peroxisomes in man. Atherosclerosis 46:239-
54     246.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                       4-251       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Hanioka, N; Jinno, H; Toyo'oka, T; et al. (1995) Induction of rat liver drug-metabolizing enzymes by
  2    tetrachloroethylene. Arch Environ Contam Toxicol 28:273-280.
  3
  4    Hansen, J; Raaschou-Nielsen, O; Christensen, JM; et al. (2001) Cancer incidence among Danish workers exposed to
  5    trichloroethylene. JOccup Environ Med 43:133-139.
  6
  7    Harap, KR. (1975) Deviant metabolic patterns in malignant disease. In: Ambrose, EJ; Roe, FJC; eds. Biology of
  8    cancer, 2nd ed. London: Cox and Wyman.
  9
10    Hardell, L; Eriksson, M; Lenner, P; et al. (1981) Malignant lymphoma and exposure to chemicals, especially organic
11    solvents, chlorophenols and phenoxy acids: a case-control study.  Br J Cancer 43:169-176.
12
13    Hardell, L; Bengtsson, NO; Jonsson, U; et al. (1984) Aetiological aspects on primary liver cancer with special
14    regard to alcohol, organic solvents and acute intermittent porphyriaBan epidemiological investigation. Br J Cancer
15    50:3896397.
16
17    Hardin, BD; Bond, GP; Sikov, MR; et al. (1981)  Testing of selected workplace chemicals for teratogenic potential.
18    Scand J Work Environ Health 7 Suppl 4:66-75.
19
20    Hardwick, JP; Song, BJ; Huberman, E; et al. (1987) Isolation, complementary DNA sequence, and regulation of rat
21    hepatic lauric acid omega-hydroxylase (cytochrome P-450 LA omega): Identification of a new cytochrome P-450
22    gene family. J Biol Chem 262:801-810.
23
24    Harrington, JM; Whitby, H; Gray, CN; et al. (1989) Renal disease and occupational exposure to organic solvents: a
25    case referent approach. Br J Ind Med 46:643-650.
26
27    Harrington-Brock, K; Doerr, CL; Moore, MM. (1998) Mutagenicity of three disinfection by-products: di- and
28    trichloroacetic acid and chloral hydrate in L5178Y/TK +/- (-)3.7.2C mouse lymphoma cells.  Mutat Res 413:265-
29    276.
30
31    Harris, RA; Mihic, SJ; Dildy-Mayfield, JE; et al.  (1995) Actions of anesthetics on ligand-gated ion channels: role of
32    receptor subunit composition.  FASEB J 9:1454-1462.
33
34    Harris, NL; Jaffe, ES; Diebold, J; et al. (2000a) The World Health Organization classification of hematological
35    malignancies report of the Clinical Advisory Committee Meeting, Airlie House, VA, November 1997.  Mod Pathol
36    13:193-207.
37
38    Harris, NL; Jaffe, ES; Diebold, J; et al. (2000b) Lymphoma classification-from controversy  to consensus: the
39    R.E.A.L. and WHO Classification of lymphoid neoplasms. AnnOncol 11 Suppl 1:3-10.
40
41    Hartmann, A; Speit, G. (1995) Genotoxic effects  of chemicals in the  single cell gel (SCO) test with human blood
42    cells in relation to the induction of sister-chromatid exchanges (SCE). Mutat Res 346:49-56.
43
44    Hasmall, S; Orphanides, G; James,  N; et al. (2002) Down regulation of lactoferrin by PPAR-a ligands: role in
45    perturbabion of hepatocyte proliferation and apoptosis. Toxicol Sci 68:304-313.
46
47    Haworth, S; Lawlor, T; Mortelmans, K; et al. (1983) Salmonella mutagenicity test results for 250 chemicals.
4 8    Environ Mutagen 5 Suppl 1:1-142.
49
50    Hayes, JR; Condie, LW, Jr.; Borzelleca, JF. (1986)  The subchronic toxicity of tetrachloroethylene
51    (perchloroethylene) administered in the drinking  water of rats.  Fundam Appl Toxicol 7:119-125.
52
53    Heineman, EF; Cocco, P; Gomez, MR; et al. (1994) Occupational exposure to chlorinated aliphatic hydrocarbons
54    and risk of astrocytic brain cancer.  Am J Ind Med 26:155-169.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-252       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Hemminki, K; Franssila, E; Vainio, H (1980) Spontaneous abortions among female chemical workers in Finland.
  2     Int Arch Occup Environ Health 45:123-126.
  3
  4     Hernberg, S; Korkala, ML; Asikainen, U; et al. (1984) Primary liver cancer and exposure to solvents. Int Arch
  5     Occup Environ Health 54:147-153.
  6
  7     Hernberg, S; Kauppinen, T; Riala, R; et al. (1988) Increased risk for primary liver cancer among women exposed to
  8     solvents. ScandJ Work Environ Health 14:356-365.
  9
10     Herren-Freund, SL; Pereira, MA; Khoury, MD; et al. (1987) The carcinogenicity of trichloroethylene and its
11     metabolites, trichloroacetic acid and dichloroacetic acid, in mouse liver.  Toxicol Appl Pharmacol 90:183-189.
12
13     Herrinton, LJ. (1998) Epidemiology of the revised European-American lymphoma classification subtypes.
14     Epidemiol Rev 20:187-203.
15
16     Heuvel, JPV. (1999) Peroxisome proliferator-activated receptors (PPARS) and carcinogenesis.  Toxicol Sci 47:1-8.
17
18     Hill, AB. (1965) The environment and disease: association or causation? Proc R Soc Med 58:295-300.
19
20     Hinnen, U; Schmid-Grendelmeier, P; Muller, E; et al. (1995) Exposure to solvents in scleroderma: disseminated
21     circumscribed scleroderma (morphea) in a painter exposed to perchloroethylene. Schweiz Med Wochenschr
22     125:2433-2437.
23
24     Hinson, JA; Roberts, DW. (1992) Role of covalent and noncovalent interactions in cell toxicity: effects on proteins.
2 5     Annu Rev Pharmacol Toxicol 3 2:471 -510.
26
27     Hinton, RH; Mitchell, DE; Mann, A; et al. (1986) Effect of phthalic acid esters on the liver and thyroid Environ
28     Health Perspect 70:195-210.
29
30     Holson, RR; Pearce, B. (1992) Principles and pitfalls in the analysis of prenatal treatment effects in multiparous
31     species. Neurotoxicol Teratol 14:221-228.
32
33     Honma, T; Hasegawa, H; Sato, M;  et al.  (1980a) Changes of free amino acid content in rat brain after exposure to
3 4     trichloroethylene and tetrachloroethylene. Ind Health 18:1-7.
35
36     Honma, T; Sudo, A; Miyagawa, M; et al. (1980b) Effects of exposure to trichloroethylene and tetrachloroethylene
37     on the contents of acetylcholine, dopamine, norepinephrine and serotonin in rat brain. IndHealth 18:171-178.
38
39     Houten, L, Sonnesso, G. (1980) Occupational exposure and cancer of the liver.  Arch Environ Health 35:51-53.
40
41     HSIA (Halogenated Solvents Industry Alliance). (2004) Comment entitled AEvaluation of EPA'sneurotoxicity of
42     tetrachloroethylene.@ ORD-2003-0014-0005. Available online at
43     http://docket.epa.gov/edkpub/do/EDKStaffItemDetailView?objectId=090007d48025686c.
44
45     Hudnell, HK; Schreiber, JS. (2004) Residential tetrachloroethylene exposure: response. Environ Health Perspect
46     112(15):A864.
47
48     Hudnell, HK; Boyes, WK; Otto, DA; et al. (1996a) Battery of neurobehavioral tests recommended to ATSDR:
49     solvent-induced deficits in microelectronic workers. Toxicol Ind Health 12:235-243.
50
51     Hudnell, HK; Otto, DA; House, DE. (1996b) The  influence of vision on computerized neurobehavioral test scores: a
52     proposal for improving test protocols. Neurotoxicol Teratol 18:391-400.
53
54     Hudnell, HK; House, D; Schmid, J; et al. (2001) Human visual function in the North Carolina clinical study on
55     possible estuary-associated syndrome. J Toxicol Environ Health A 62:575-594.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         4-253       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Huff, J. (1995) Mechanisms, chemical carcinogenesis, and risk assessment: cell proliferation and cancer. Am J Ind
  2    Med 27:293-300.
  3
  4    Huff, J. (1996) Response: alpha-2-mu-Globulin nephropathy, posed mechanisms, and white ravens.  Environ Health
  5    Perspect 104:1264-1267.
  6
  7    Huss, JM; Kelly, DP. (2005) Mitochondria! energy metabolism in heart failure: a question of balance.  J Clin Invest
  8    115:547-555.
  9
10    IARC (International Agency for Research on Cancer). (1995) Tetrachloroethylene. In: IARC monographs on the
11    evaluation of carcinogenic risks to humans; vol. 63: dry cleaning, some chlorinated solvents and other. Lyon,
12    France.
13
14    Ikeda, M; Imanura, T. (1973) Biological half-life of trichloroethylene and tetrachloroethylene in human subjects. Int
15    Arch Arbeitsmed. 31(3):209-224.
16
17    Ikeda, M; Otsuji, H. (1972) A comparative study of the excretion of Fujiwara reaction-positive substances in urine
18    of humans and rodents given trichloro- or tetrachloro-derivatives of ethane and ethylene. Br J Ind Med 29:99-104
19
20    ILSI (International Life Sciences Institute). (1992) Similarities and differences between children and adults:
21    implications for risk assessment. New York: ILSI Press.
22
23    Infante-Rivard, C; Siemiatycki, J; Lakhani, R; et al. (2005) Maternal exposure to occupational solvents and
24    childhood leukemia.  Environ Health Perspect 113(6):787-92.
25
26    IOM (Institute of Medicine). (2002) Gulf war and health.  Vol. 2: Insecticides and solvents.  Washington, DC:
27    National Academies Press.
28
29    Ishmael, J and Dugard, P H  (2006) A review of perchloroethylene and rat mononuclear cell leukemia. Regul
30    Toxicol Pharmacol 45:178-184.
31
32    Issemann, I; Green, S. (1990) Activation of a member of the steroid hormone receptor superfamily by peroxisome
33    proliferators. Nature 347:645-650.
34
35    Issemann, I; Prince, RA; Tugwood, JD; et al. (1993) The peroxisome proliferator-activated receptor:retinoid X
3 6    receptor heterodimer is activated by fatty acids and fibrate hypolipidaemic drugs.  J Mol Endocrinol 11:37-47.
37
3 8    Ito, YO ; Yamanoshita, N ; Asaeda,  Y; et al. (2007). Di(2-ethylhexyl)phthalate induces hepatic tumorigenesis
39    through a peroxisome proliferator-activated receptor -a-independent pathway.  J Occup Health 49(3): 172-82.
40
41    Jalouli, M; Carlsson, L; Ameen, C; et al. (2003) Sex difference in hepatic peroxisome proliferator-activated receptor
42    -a expression: influence of pituitary and gonadal hormones.  Endocrinology  144:101-109.
43
44    Jang, JY; Droz, PO. (1997) Ethnic differences in biological monitoring of several organic solvents. II. A simulation
45    study with a physiologically based pharmacokinetic model.  Int Arch Occup Environ Health 70:41-50.
46
47    Ji, J; Hemminki, K. (2005a) Occupation and upper aerodigestive tract cancers: a follow-up study in Sweden. J
48    Occup Environ Med 47:785-795.
49
50    Ji, J; Hemminki, K. (2005b) Occurrences of leukemia subtypes by socioeconomic and occupational group. J Occup
51    Environ Med 47:1131-1140.
52
53    Ji, J; Hemminki, K. (2006) Socioeconomic/occupational risk factors for lymphproliferative diseases in Sweden.
54    Ann Epidemiol 16:370-376.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-254      DRAFT-DO NOT CITE OR QUOTE

-------
  1     Jirtle, RL; Meyer, SA. (1991) Liver tumor promotion: effect of phenobarbital on EOF and protein kinase C signal
  2     transduction and transforming growth factor-beta 1 expression. Dig Dis Sci 36:659-668.
  3
  4     JISA (Japan Industrial Safety Association). (1993) Carcinogenicity study of tetrachloroethylene by inhalation in rats
  5     and mice. Data No. 3-1. Kanagawa, Japan. Available from: IRIS Information Desk, U.S. Environmental Protection
  6     Agency, Washington, DC.
  7
  8     Johnson, PD; Dawson, BV; Goldberg, SJ. (1998) Cardiac teratogenicity of trichloroethylene metabolites.  J Am Coll
  9     Cardiol 32:540-545.
10
11     Jonker, D; Woutersen, RA; Feron, VJ. (1996) Toxicity of mixtures of nephrotoxicants with similar or dissimilar
12     mode of action. Food ChemToxicol 34:1075-1082.
13
14     Juntunen, J. (1986) Occupational toxicology of trichloroethylene with special reference to neurotoxicity. Dev
15     Toxicol Environ Sci 12:189-200.
16
17     Kaerlev, L; Teglbjaerg, PS; Sabroe, S; et al. (2000) Occupation and small bowel adenocarcinoma: a European case-
18     control study.  Occup Environ Med 57:760-766.
19
20     Kafer, E. (1985) Tests which distinguish induced aneuploidy and crossing-over from secondary segregation in
21     Aspergillus treated with chloral hydrate or x-rays. Mutat. Res. 167: 9-34.
22
23     Karlsson, JE; Rosengren, LE; Kjellstrand, P; et al. (1987) Effects of low-dose inhalation of three chlorinated
24     aliphatic organic solvents on deoxyribonucleic acid in gerbil brain. Scand J Work Environ Health 13:453-458.
25
26     Kato, I; Koenig, KL; Watanabe-Meserve, H; et al. (2005) Personal and  occupational exposure to organic solvents
27     and risk of non-Hodgkin's  lymphoma (NHL) in women (United States). Cancer Causes and Control 16:1215-1224.
28
29     Kaufman AS; McLean JE;  Reynolds  CR.  (1988) Sex, race, residence, region, and education differences on the  11
30     WAIS-R subtests. J ClinPsychol 44:231-248.
31
32     Kerbey, AL; Randle, PJ; Cooper, RH; et al. (1976) Regulation of pyruvate dehydrogenase in rat heart: mechanism of
3 3     regulation of proportions of dephosphorylated and phosphorylated enzyme by oxidation of fatty acids and ketone
34     bodies and of effects of diabetes: role of coenzyme A, acetyl-coenzyme A and reduced and oxidized nicotinamide-
35     adenine dinucleotide. BiochemJ 154:327-348.
36
37     Kernan, GJ; Ji, BT; Dosemeci, M; et al. (1999) Occupational risk factors for pancreatic cancer: a case-control study
3 8     based on death certificate from 24 U.S. states.  Am J Ind Med. 36(2):260-70.
39
40     Kjellstrand, P; Holmquist, B; Kanje, M; et al. (1984) Perchloroethylene: effects on body and organ weights and
41     plasma butyrylcholinesterase activity in mice.  Acta Pharmacol Toxicol (Copenh) 54:414-424.
42
43     Kjellstrand, P; Holmquist, B; Jonsson, I; et al. (1985) Effects of organic solvents on motor activity in mice.
44     Toxicology 35:35-46.
45
46     Klaunig, J; Ruch, R; DeAngelo, A; et al. (1988) Inhibition of mouse hepatocyte intercellular communication by
47     phthalate esters. Cancer Lett 43:65-71.
48
49     Klaunig, JE; Babich, MA; Baetcke, KP; et al. (2003) PPAR-a agonist-induced rodent tumors: modes of action and
50     human relevance. Crit Rev Toxicol 33(6):655-780.
51
52     Kobayashi, S; Hutcheon, DE; Regan, J. (1982) Cardiopulmonary toxicity of tetrachloroethylene. J Toxicol Environ
53     Health 10:23-30.
54
55     Kok, RJ; Haas, M; Moolenaar, F; et al. (1998) Drug delivery to the kidneys and the bladder with the low molecular
56     weight protein lysozyme. RenFail 20:211-217.

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         4-255      DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2     Koppel, C; Arndt, I; Arendt, U; et al. (1985) Acute tetrachloroethylene poisoning-blood elimination kinetics during
  3     hyperventilation therapy. JToxicol ClinToxicol23:103-115.
  4
  5     Kraupp-Grasl, B; Huber, W; Putz, B; et al. (1990) Tumor promotion by the peroxisome proliferator nafenopin
  6     involving a specific subtype of altered foci in rat liver. Cancer Res 50:3701-3708.
  7
  8     Kraupp-Grasl, B; Huber, W; Taper, H;  et al. (1991) Increased susceptibility of aged rats to hepatocarcinogenesis by
  9     the peroxisome proliferator nafenopin and the possible involvement of altered liver foci occurring spontaneously.
10     Cancer Res 51:666-671.
11
12     Kriebel, D; Zeja, A; Eisen, EA; et al. (2004) Quantitative evaluation of the effects of uncontrolled confounding by
13     alcohol and tobacco in occupational cancer studies. Int J Epidemiol 33:1040-1045.
14
15     Kumar, P; Prasad, AK; Dutta, KK. (2000) Steroidogenic alterations in testes and sera of rats exposed to
16     trichloroethylene (TCE)by inhalation.  HumExp Toxicol 19:117-121.
17
18     Kurata, Y; Kidachi, F; Yokoyama, M; et al. (1998) Subchronic toxicity of di(2-ethylhexyl)phthalate in common
19     marmosets: Lack of hepatic peroxisome proliferation, testicular atrophy or pancreatic acinar cell hyperplasia.
20     Toxicol Sci 42:49-56.
21
22     Kurien, B T; Hensley, K; Bachmann, M;  Scofield, R H (2006) Oxidatively modified autoantigens in autoimmune
23     diseases. Free Radic Biol Med 41:549-556.
24
25     Kylin, B; Sumegi, I; Yllner, S. (1965) Hepatotoxicity of inhaled trichloroethylene and tetrachloroethylene: long-
26     term exposure.  Acta Pharmacol Toxicol (Kbh) 22:379-385.
27
28     Kyrklund, T; Haglid, K. (1991) Brain lipid composition in guinea pigs after intrauterine exposure to
29     perchloroethylene. Pharmacol Toxicol 68:146-148.
30
31     Kyrklund, T; Ailing, C; Kjellstrand, P;  et al. (1984) Chronic effects of perchloroethylene on the composition of lipid
32     and acyl groups in cerebral cortex and hippocampus of the gerbil. Toxicol Lett 22:343-349.
33
34     Kyrklund, T; Kjellstrand, P; Haglid, KG.  (1987) Lipid composition and fatty acid pattern of the gerbil brain after
35     exposure to perchloroethylene. Arch Toxicol 60:397-400.
36
37     Kyrklund, T; Kjellstrand, P; Haglid, KG.  (1988) Effects of exposure to Freon 11, 1,1,1-trichloroethane or
3 8     perchloroethylene on the lipid and fatty-acid composition of rat cerebral cortex.  Scand J Work Environ Health
39     14:91-94.
40
41     Kyrklund, T; Kjellstrand, P; Haglid, KG.  (1990) Long-term exposure of rats to perchloroethylene, with and without
42     a post-exposure solvent-free recovery period: effects on brain lipids. Toxicol Lett 52:279-285.
43
44     Kyyronen, P; Taskinen, H; Lindbohm, ML; et al. (1989) Spontaneous abortions and congenital malformations
45     among women  exposed to tetrachloroethylene in dry cleaning. J Epidemiol Community Health 43:346-351.
46
47     Lacey, JV, Jr.; Garabrant, DH; Laing, TJ; et al. (1999) Petroleum distillate solvents as risk factors for
48     undifferentiated connective tissue disease (UCTD). Am J Epidemiol 149:761-770.
49
50     Lagakos, SW; Wessen, B; Zelen, M. (1986) Analysis of contaminated well water and health effect in Woburn,
51     Massachusetts.  Journal of the American Statistical Association 81:583-596.
52
53     Liang, JC and Pacchierotti, F. (1988). Cytogenetic investigation of chemically-induced aneuploidy in mouse
54     spermatocytes.  Mutat. Res. 201: 325-335.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-256      DRAFT-DO NOT CITE OR QUOTE

-------
  1     Lake, BG; Evans, JG; Cunninghame, ME; et al. (1993) Comparison of the hepatic effects of nafenopin and
  2     WY14643 on peroxisome proliferation and cell replication in the rat and Syrian hamster. Environ Health Perspect
  3     5:241-247.
  4
  5     Lane, S E; Watts, R A; Bentham, G; Innes, N J; Scott, D G (2003) Are environmental factors important in primary
  6     systemic vasculitis? A case-control study. Arthritis Rheum 48:814-823.
  7
  8     Lapinskas, PJ; JC, Gorton. (1997) Peroxisome proliferator-activated receptor alpha: central mediator of peroxisome
  9     proliferatortoxicity. CUT Activities. Vol 17 No 1.  1-9. Research Triangle Park, NC.
10
11     Lapsley, M; Akers, K; Norden, AG. (1998) Sensitive assays for urinary retinol-binding protein and beta-2-
12     glycoprotein-1 based on commercially available standards. AnnCrinBiochem35 (Pt 1):115-119.
13
14     Lash, LH; Fisher, JW; Lipscomb, JC; et al. (2000a) Metabolism of Trichloroethylene.  Environ Health Perspect 108:
15     (2):177B200.
16
17     Lash, LH; Parker, JC; Siegel-Scott, C. (2000b) Modes of Action of trichloroethylene for kidney tumorigenesis.
18     Environ Health Perspect 108:(2):225-240.
19
20     Lash, LH; Parker, JC. (2001) Hepatic and renal toxicities associated with perchloroethylene.  Pharmacol Rev
21     53:177-208.
22
23     Lash, LH; Qian, W; Putt, DA;  et al. (2002) Renal toxicity of perchloroethylene and s-(l,2,2-
24     trichlorovinyl)glutathione in rats and mice: sex- and species-dependent differences. Toxicol App Pharmacol
25     179:163-171.
26
27     Laslo-Baker, D; Barrera, M; Knittel-Keren, D; et al. (2004) Child neurodevelopmental outcome and maternal
28     occupational exposure to solvents. Arch Pediatr Adolesc Med 158(10):956-61.
29
30     Laughter, AR; Dunn, CS, Swanson CL; et al. (2004) Role of the peroxisome proliferator-activated receptor a
31     (PPARa) in responses to trichloroethylene and metabolite, trichloroacetate and dichloroacetate in mouse liver.
32     Toxicology 203:83-98.
33
34     Lauwerys, R; Herbrand, J; Buchet, J P; et al. (1983) Health surveillance of workers exposed to tetrachloroethylene
35     in dry- cleaning shops. Int Arch Occup Environ Health 52:69-77.
36
37     Lee, SS; Pineau, T; Drago, J; et al. (1995) Targeted disruption of the alpha isoform of the peroxisome proliferator-
3 8     activated receptor gene  in mice results in abolishment of the pleiotropic effects of peroxisome proliferators.  Mol
39     Cell Biol 15:3012-3022.
40
41     Lehmann, I; Rehwagen, M; Diez, U; etal. (2001) Enhanced in vivo IgE production and T cell polarization toward
42     the type 2 phenotype in association with indoor exposure to VOC: results of the LARS study. Int J Hyg Environ
43     Health 204:211-221.
44
45     Lehmann, I; Thoelke, A; Rehwagen, M; et al. (2002) The influence of maternal exposure to volatile organic
46     compounds on the cytokine secretion profile of neonatal T cells. Environ Toxicol 17:203-210.
47
48     Letkiewicz, F; Johnston P; Macaluso  C; et al. (1982)  Occurrence of tetrachloroethylene in drinking water, food, and
49     air. Prepared by JRB Associates JRB Project No. 2-613-03-852-29, for U.S. Environmental Protection Agency
50     contract 86-01-6388, task 29.
51
52     Lezak, MD. (1995) Neuropsychological assessment, 3rd ed. New York: Oxford University Press.
53     Lin, RS; Kessler, II. (1981) A multifactorial model for pancreatic cancer in man. Epidemiologic evidence. JAMA
54     245:147-152.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-257       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Lindbohm, ML; Taskinen, H; Sallmen, M; et al. (1990) Spontaneous abortions among women exposed to organic
  2    solvents. Am JIndMed 17:449-463.
  3
  4    Lock, EA; Sani, Y; Moore, RB; et al. (1996) Bone marrow and renal injury associated with haloalkene cysteine
  5    conjugates in calves. Arch Toxicol 70:607-619.
  6
  7    Loizou, GD. (2001) The application of physiologically based pharmacokinetic modeling in the analysis of
  8    occupational exposure to perchloroethylene. Toxicol Lett 124:59-69.
  9
10    Lowengart, RA; Peters,  JM; Cicioni, C; et al. (1987) Childhood leukemia and parents' occupational and home
11    exposures. J Natl Cancer Inst 79:39-46.
12
13    Lumpkin, MH; Bruckner, JV; Cambell, JL; et al. (2003) Plasma binding of trichloroacetic acid in mice, rats, and
14    humans under cancer bioassay and environmental exposure conditions.  Drug Metab Disp 31 (10):1203-1207.
15
16    Lundberg, I; Alfredsson, L; Plato, N; Sverdrup, B; Klareskog, L; Kleinau, S (1994) Occupation, occupational
17    exposure to chemicals and rheumatological disease. A register based cohort study.  Scand J Rheumatol 23:305-310.
18
19    Lybarger, JA; Lichtveld, MY; Amler, RW. (1999) Biomedical testing of the kidney forpersons exposed to
20    hazardous substances in the environment. RenFail 21:263-274.
21
22    Lynge,E. (1994) Danish Cancer Registry as a resource for occupational research. JOccupMed 36:1169-1173.
23
24    Lynge, E; Thygesen, L.  (1990) Primary liver cancer among women in laundry and dry-cleaning work in Denmark.
25    Scand J Work Environ Health 16:108-112.
26
27    Lynge, E; Carstensen, B; Andersen, O. (1995) Primary liver cancer and renal cell carcinoma in laundry and dry-
28    cleaning workers in Denmark.  Scand J Work Environ Health 21:293-295.
29
30    Lynge, E; Anttila, A; Hemminki, K. (1997) Organic solvents and cancer. Cancer Causes Control 8:406-419.
31
32    MA DPH (Massachusetts Department of Public Health) (1997) Woburn childhood leukemia follow-up study. Final
33    eport. Available online at
34    http://www.mass.gov/Eeohhs2/docs/dph/environmental/investigations/woburn childhood leukemia follow.pdf.
35
36    MacNaughton, MG; Uddin, DE. (1984) Toxicology of mixed distillate and high-energy synthetic fuels. Mehlman,
37    MA; Hemstreet, GP; Thorpe, JJ; et al.,  eds. Advances in modern environmental toxicology. Vol.7. Renal effects of
38    petroleum hydrocarbons. Princeton, NJ: Princeton Scientific Publishers, Inc.; pp. 121-132.
39
40    Maitre, A; Hours, M; Bonneterre, V; et al. (2004) Systemic sclerosis and occupational risk factors: role of solvents
41    and cleaning products.  JRheumatol 31:2395-2401.
42
43    Mallin, K. (1990) Investigation of a bladder cancer cluster in northwestern Illinois. Am J Epidemiol 132:896-106.
44
45    Malone, KE; Koepsell, TD; Dating, JR; et al. (1989) Chronic lymphcytic leukemia in relation to chemical exposures.
46    Am J Epidemiol  130:1152-1158.
47
48    Maloney, EK; Waxman, DJ.  (1999) Trans-activation of PPAR-alpha and PPARgamma by structurally diverse
49    environmental chemicals.  Toxicol Appl Pharmacol 161:209-218.
50
51    Maltoni, C; Cotti, G. (1986) Results of long-term carcinogenicity bioassay s of tetrachloroethylene on Sprague-
52    Dawley rats administered by ingestion. Acta Oncologica 7(1): 11-26.
53
54    Mandel, JS; McLaughlin, JK; Schlehofer, B; et al. (1995) International renal-cell cancer study. IV.  Occupation.  Int
55    J Cancer 61:601-605.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        4-258      DRAFT-DO NOT CITE OR QUOTE

-------
  1    Manzo, L; Artigas, F; Martinez, E; et al. (1996) Biochemical markers of neurtoxicity: a review of mechanistic
  2    studies and applications. Hum Exper Toxicol 15(suppl l):S20-235.
  3
  4    Marcus, SL; Miyata, KS; Zhang, B; et al. (1993) Diverse peroxisome proliferator-activated receptors bind to the
  5    peroxisome proliferator-responsive elements of the rat hydratase/dehydrogenase and fatty acyl-CoA oxidase genes
  6    but differentially induce expression.  ProcNatlAcad Sci USA 90:5723-5727.
  7
  8    Maronpot, RR; Fox, T; Malarkey, DE; et al. (1995) Mutations in the ras proto-oncogene: clues to etiology and
  9    molecular pathogenesis of mouse liver tumors. Toxicology 101:125-156.
10
11    Marsman, DS. (1991) Peroxisome proliferatorhepatocarcinogenesis in rats.  Thesis. University of North Carolina,
12    Chapel Hill, NC.
13
14    Marsman, DS; Popp, JA. (1994) Biological potential of basophilic hepatocellular foci and hepatic adenoma induced
15    by the peroxisome proliferator, Wy-14,643. Carcinogenesis 15:111-117.
16
17    Marsman, DS; Cattley, RC; Conway, JG; et al. (1988) Relationship of hepatic peroxisome proliferation and
18    replicative DNA synthesis to the hepatocarcinogenicity of the peroxisome proliferators di(2-ethylhexyl)phthalate
19    and [4-chloro-6-(2,3-xylidino)-2- pyrimidinylthio]acetic acid (Wy-14,643) in rats.  Cancer Res 48:6739-6744.
20
21    Marsman, DS; Goldsworthy, TL; Popp, JA. (1992) Contrasting hepatocytic peroxisome proliferation, lipofuscin
22    accumulation and cell turnover for the hepatocarcinogens Wy-14,643 and clofibric acid. Carcinogenesis  13:1011-
23    1017.
24
25    Mattsson, JL; Albee, RR; Yano, BL; et al. (1998) Neurotoxicologic examination of rats exposed to 1,1,2,2-
26    tetrachloroethylene (perchloroethylene) vapor for 13 weeks. Neurotoxicol Teratol 20:83-98.
27
28    Mayes,  MD. (1999) Epidemiologic Studies of Environmental Agents and Systemic Autoimmune Diseases.  Environ
29    Health Perspect 107:743-748.
30
31    McCredie, M; Stewart, JH. (1993) Risk factors for kidney cancer in New South Wales. IV. Occupation.  Br J Ind
32    Med 50:349-354.
33
34    McDermott, M J; Mazor, K A; Shost, S J; Narang, R S; Aldous, K M; Storm, J E (2005) Tetrachloroethylene (PCE,
3 5    Perc) levels in residential dry cleaner buildings in diverse communities in New York City.  Environ Health Perspect
36    113:1336-1343.
37
38    McDonald, AD; Armstrong, B; Cherry, NM; et al. (1986) Spontaneous abortion and occupation. J Occup Med
39    28:1232-1238.
40
41    McDonald, AD; McDonald, JC; Armstrong, B; et al. (1987) Occupation and pregnancy outcome. Br J Ind Med
42    44:521-526.
43
44    McKinney, LL; Weakley, FB; Eldridge, AC; et al. (1957) S-(dichlorovinyl) L-cysteine: an agent causing fatal
45    aplastic anemia in calves.  J AM Chem Soc 51:16946.
46
47    McKone, TE; Daniels, JL (1991) Estimating human exposure through multiple pathways from air, water, and soil.
48    Regul Toxicol Pharmacol 13:36-61.
49
50    McLaughlin, JK; Blot, WJ. (1997) A critical review of epidemiology studies of trichloroethylene and
51    perchloroethylene and risk of renal-cell cancer. Int Arch Occup Environ Health 70:222-231.
52
53    McLaughlin, JK; Malker, HS; Stone, BJ; et al.  (1987) Occupational risks for renal cancer in Sweden. Br J Ind Med
54    44:119-123.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-259       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Meckler, LC; Phelps, DK. (1966) Liver disease secondary to tetrachloroethylene exposure. A case report. JAMA
  2    197:662-663.
  3
  4    Meek, ME; Bucher, JR; Cohen, SM; et al. (2003) A framework for human relevance analysis of information on
  5    carcinogenic modes of action. Crit Rev Toxicol 33(6):591-653.
  6
  7    Meijer, J; Afzelius, BA. (1989) Effects of clofibrate treatment and of starvation on peroxisomes, mitochondria, and
  8    lipid droplets in mouse hepatocytes: a morphometric study. J Ultrastruct Mol Struct Res 102:87-94.
  9
10    Mellemgaard, A; Engholm, G; McLaughlin, JK; et al. (1994)  Occupational risk factors for renal-cell carcinoma in
11    Denmark. ScandJ Work Environ Health 20:160-165.
12
13    Melnick, RL. (2001) Is peroxisome proliferation an obligatory precursor step in the carcinogenicity of di(2-
14    ethylhexyl)phthalate (DEHP)? Environ Health Perspect 109:437-442.
15
16    Melnick, RL. (2002) The IARC evaluation of di02-ethylhexl)phthalate (DEHP): a flawed decision based on an
17    untested hypothesis. Int J Occup Environ Health 8:284-286.
18
19    Melnick, RL; Kohn, MC; Huff, J. (1997) Weight of evidence  versus weight of speculation to evaluate the alpha2u-
20    globulin hypothesis. Environ Health Perspect 105:904-906.
21
22    Mergler, D;  Blain, L. (1987) Assessing color vision loss among solvent-exposed workers.  Am J Ind Med 12:195-
23    203.
24
25    Mergler, D;  Blain, L; Lagace, JP. (1987) Solvent related colour vision loss: an indicator of neural damage? Int Arch
26    Occup Environ Health 59:313-321
27
28    Mergler, D;  Blain, L; Lemaire, J; et al. (1988a) Colour vision impairment and alcohol consumption. Neurotoxicol
29    Teratol 10:255-260.
30
31    Mergler, D;  Belanger, S; De Grosbois, S; et al. (1988b) Chroma! focus of acquired chromatic discrimination loss
32    and solvent exposure among printshop workers. Toxicology 49:341-348.
33
34    Mergler, D;  Huel, G; Bowler, R; et al. (1991) Visual dysfunction among former microelectronics assembly workers.
35    Arch Environ Health 46:326-334.
36
37    Mergler, D;  Huel, G; Belanger, S; et al.  (1996) Surveillance of early neurotoxic dysfunction. Neurotoxicol 17:803-
38    812.
39
40    Meskar, A; Plee-Gautier, E; Amet, Y; et al. (2001) Alcohol-xenobiotic interactions: role of cytochrome P4502E1.
41    Pathol Biol (Paris) 49(9):696-702.
42
43    Mester, B; Neiters, A; Deeg, E; et al. (2006) Occupation and malignant lymphoma: a population based case control
44    study in Germany. Occup Environ Med 63:17-26.
45
46    Meyer, K; Lee, JS; Dyck, PA; et al. (2003) Molecular profiling of hepatocellular carcinomas developing
47    spontaneouwly in acyl-CoA oxidase deficient mice: comparison with liver tumors induced in wild-type mice by a
48    peroxisome proliferator and a genotoxic carcinogen. Carcinogenesis 24:975-984.
49
50    Mihic, SJ. (1999) Acute effects of ethanol on GAB AA and glycine receptor function. Neurochemlnt 35:115-123.
51
52    Miligi, L; Seniori Constantini, A; Crosignani, P; et al. (1999)  Occupational, environmental, and life-style factors
53    associated with the risk of hematolymphopoietic malignancies in women.  Am J Ind Med 36:60-69.
54
55    Miligi, L; Seniori Costantini, A; Benvenuti, A; et al. (2006) Occupational exposure to solvents and the risk of
56    lymphomas. Epidemiology 17:552-561.

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                         4-260      DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2    Monster, AC; Boersma, G; Steenweg, H. (1979) Kinetics of tetrachloroethylene in volunteers; influence of exposure
  3    concentration and work load. Int Arch Occup Environ Health 42:303-309.
  4
  5    Moody, DE; Reddy, JK; Lake, BG; et al. (1991) Peroxisome proliferation and nongenotoxic carcinogenesis:
  6    commentary on a symposium. Fundam Appl Toxicol 16:233-248.
  7
  8    Moore, MM; Harrington-Brock, K. (2000) Mutagenicity of trichloroethylene and Its metabolites: implications for
  9    the risk assessment of trichloroethylene. Environ Health Perspect 108 Suppl 2:215-223.
10
11    Morrow, LA; Ryan, CM; Hodgson, MJ; et al. (1990) Alterations in cognitive and psychological functioning after
12    organicsolvent exposure. J Occup Med 32:444- 450.
13
14    Morton, W; Marjanovic, D. (1984) Leukemia incidence by occupation in the Portland-Vancouver metropolitan area.
15    Am JInd Med 6:185-205.
16
17    Morton, W and Marjanovic, D  (1984) Leukemia incidence by occupation in the Portland-Vancouver metropolitan
18    area. Am J Ind Med 6:185-205.
19
20    Moser, VC; Cheek, BM; MacPhail, RC. (1995) A multidisciplinary approach to lexicological screening: III.
21    Neurobehavioral toxicity. J Toxicol Environ Health 45:173-210.
22
23    Moser, VC; Phillips, PM; McDaniel, KL; et al. (1999) Behavioral evaluation of the neurotoxicity produced by
24    dichloroacetic acid in rats. Neurotoxicol Teratol 21:719-731.
25
26    Moslen, MT; Reynolds, ES; Szabo, S. (1977).Enhancement of the metabolism and hepatotoxicity of
27    trichloroethylene and perchloroethylene. Biochem Pharmacol 26(5):369-75.
28
29    Motohashi, Y; Miyazaki, Y; Takano,  T. (1993) Assessment of behavioral effects of tetrachloroethylene using a set
30    of time-series analyses. Neurotoxicol Teratol 15:3-10.
31
32    Mundt, KA; Birk, T; Burch, MT. (2003) Critical review of the epidemiological literature on occupational exposure
33    to perchloroethylene and cancer. Int Arch Occup Environ Health 76:473-491.
34
35    Mueser, KT; McGurk, SR. (2004) Schizophrenia. Lancet 363:2063-2072.
36
37    Murakami, K; Horikawa, K. (1995) The induction of micronuclei in mice hepatocytes and reticulocytes by
38    tetrachloroethylene. Chemosphere 31:3733-3739.
39
40    Mutti, A; Smargiassi, A. (1998) Selective vulnerability of dopaminergic systems to industrial chemicals: risk
41    assessment of related neuroendocrine changes.  Toxicol Ind Health 14:311-23.
42
43    Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-cleaners.
44    Lancet 340:189-193.
45
46    Muttray, A; Wolff, U; Jung, D; et al. (1997) Blue-yellow deficiency in workers exposed to low concentrations of
47    organic solvents. Int Arch Occup Environ Health 70:407-412.
48
49    Mazzullo, MS; Grilli, G; Lattanzi, G; et al. (1987).  Evidence of DNA binding activity of perchloroethylene. Res
50    Commun Chem Pathol Pharmacol 58(2): 215-35.
51
52    Nagano, K; Nishizawa, T; Yamamoto, S; et al. (1998) Inhalation carcinogenesis studies of six halogenated
53    hydrocarbons in rats and mice. In: Chiyotani, K; Hosoda, Y; Aizawa, Y, eds. Advances in the prevention of
54    occupational respiratory diseases: Proceedings of the 9th international conference  on occupational respiratory
55    diseases, Kyoto, Japan, 13-16, October 1997. Amsterdam: Elsevier; pp. 741-746.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        4-261       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Nakai, JS; Stathopulos, PB; Campbell, GL; et al. (1999) Penetration of chloroform, trichloroethylene and
  2     tetrachloroethylene through human skin. J Toxicol Environ Health A 58:157-170.
  3
  4     Nakajima, T; Kamijo, Y; Usuda, N; et al. (2000) Sex-dependent regulation of hepatic peroxisome proliferation in
  5     mice by trichloroethylene via peroxisome proliferator-activated receptor alpha (PPAR-alpha). Carcinogenesis
  6     21:677-682.
  7
  8     Nakatsuka, H; Watanabe, T; Takeuchi, Y; et al. (1992) Absence of blue-yellow color vision loss among workers
  9     exposed to toluene or tetrachloroethylene, mostly at levels below occupational exposure limits.  Int Arch Occup
10     Environ Health 64:113-117.
11
12     Narotsky, MG; Kavlock, RJ. (1995) A multidisciplinary approach to toxicological screening: II.  Developmental
13     toxicity. J Toxicol Environ Health 45:145-171.
14
15     NCI (National Cancer Institute). (1977) Bioassay of tetrachloroethylene for possible carcinogenicity. DHEW Pub.
16     (NIH)77-813.
17
18     Nelson, BK; Taylor, BJ;  Setzer, JV; et al. (1980) Behavioral teratology of perchloroethylene in rats. J Environ
19     Pathol Toxicol 3:233-250.
20
21     Nelson, MA; Bull, PJ. (1988) Induction of strand breaks in DNA by trichloroethylene and metabolites in rat and
22     mouse liver in vivo.  Toxicol Appl Pharmacol 94:45-54.
23
24     Nelson, MA; Lansing, AJ; Sanchez, IM; et al. (1989) Dichloroacetic acid and trichloroacetic acid-induced DNA
25     strand breaks are independent of peroxisome proliferation. Toxicology 58:239-248.
26
27     Nelson, DE; Emont, SL;  Brackbill, RM; et al. (1994) Cigarette smoking prevalence by occupation in the United
28     States. A comparison between 1978 to  1980 and 1987 to  1990. J Occup Med 36:516-525.
29
3 0     Nemali, MR; Usuda, N; Reddy, MK; et al. (1988) Comparison of constitutive and inducible levels of expression of
31     peroxisomal beta-oxidation and catalase genes in liver and extrahepatic tissues of rat.  Cancer Res 48:5316-5324.
32
33     Newman, TB; Hulley, SB. (1996) Carcinogenicity of lipid-lowering drugs. JAMA 275:55-660.
34
3 5     Nietert, PJ; Sutherland, SE; Silver, RM; et al. (1998) Is occupational organic solvent exposure a risk factor for
36     scleroderma? [published erratum appears in Arthritis Rheum 1998 Aug;41(8):1512]. Arthritis Rheum
37     41:111161118.
38
39     Nomiyama, K; Liu, SJ; Nomiyama,  H. (1992) Critical levels of blood and urinary cadmium, urinary beta 2-
40     microglobulin and retinol-binding protein for monitoring cadmium health effects. IARC Sci Publ 325-340.
41
42     Nowell, PC. (1991) Genetic instability and tumor development. Basic Life Sci 57:221-228.
43
44     NRC (National Research Council). (1993) Pesticides in the diets of infants and children. Washington, DC: National
45     Academy Press.
46
47     NRC (National Research Council). (1995) Biologic markers in urinary toxicology. Washington, DC: National
48     Academy Press.
49
50     NRC (National Research Council). (2005) Assessing the human health risks of trichloroethylene: Key scientific
51     issues.
52
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-262       DRAFT-DO NOT CITE OR QUOTE

-------
  1    NTP (National Toxicology Program). (1983) Toxicology bioassay of pentachloroethane (CAS No.76-01-7) in
 2    F344/N rats and B6C3F1 mice (gavage study). NIH Publication No. 83-1788, NTP TR 232.
 3
 4    NTP (National Toxicology Program). (1986a) Toxicology and carcinogenesis studies of tetrachloroethylene
 5    (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3F1 mice. 311.  Available from: National
 6    Toxicology Program, National Institutes of Health, Public Health Service, U.S. Department of Health and Human
 7    Services.
 8
 9    NTP (National Toxicology Program). (1986b) Toxicology and carcinogenesis studies of isophorone (CAS No. 78-
10    59-1) in F344/N rats and B6C3F1 mice (gavage studies). NIH Publication No. 86-2547.  NTPTR291.
11
12    NTP (National Toxicology Program). (1987) Toxicology and carcinogenesis studies of dimethyl methylphosphonate
13    (CAS No. 756-79-6) in F344/N rats and B6C3F1 mice (gavage studies). Nffl Publication No. 87-2579, NTP TR
14    323.
15
16    NTP (National Toxicology Program). (1988) Toxicology and carcinogenesis studies of d-limonene  (CAS No.
17    598927-5) in F344/N rats and B6C3F1 mice (gavage studies). NIH Publication No. 88-2902, NTP  TR 347.
18
19    NTP (National Toxicology Program). (1989) Toxicology and carcinogenesis studies of hexachloroethane (CAS No.
20    67-72-1) in F344/N rats (gavage studies).  NTP TR No. 361, NIH Publication No. 89-2816. Available from:
21    National Toxicology Program, National Institutes of Health, Public Health Service, U.S. Department of Health and
22    Human Services.
23
24    NTP (National Toxicology Program). (1990a) Toxicology and carcinogenesis studies of d-limonene in F344/N rats
25    and B6C3F1 mice (gavage studies). NTP TR 347.  Washington, DC, U.S. Department of Health and Human
26    Services.
27
28    NTP (National Toxicology Program). (1990b) Carcinogenesis studies of trichloroethylene (without
29    epiochlorohydrin) in F344/N rats and B6C3F1 mice. NIH Publication No. 90-1799, NTP TR 243.
30
31    NTP (National Toxicology Program). (2000a) Toxicology and carcinogenesis studies of chloral hydrate (CAS NO.
32    302-17-0) in B6C3F1 mice (gavage study). NTP TR 502, Nffl Publication No. 00-4436.  Available from: National
33    Toxicology Program, National Institutes of Health, Public Health Service, U.S. Department of Health and Human
34    Services.
35
36    NTP (National Toxicology Program). (2000b) Toxicology and carcinogenesis studies of chloral hydrate (ad libitum
37    and feed restricted) (CAS NO. 302-17-0) in male B6C3F1 mice (gavage study).  NTP TR 503, Nffl Publication No.
38    00-4437. Available from: National Toxicology Program, National Institutes of Health, Public Health Service, U.S.
3 9    Department of Health and Human Services.
40
41    Nutrition Foundation. (1983) The relevance of mouse liver hepatoma to human carcinogenic risk: a report of the
42    International Expert Advisory Committee to the Nutrition Foundation. Washington, DC.
43
44    NYS DOH (New York State Department of Health). (1997) Tetrachloroethylene ambient air criteria document.
45    Final Report. Albany, NY.
46
47    NYS DOH (New York State Department of Health) (2000) Evaluation of residential exposure to tetrachloromethane
48    biomarkers of dose and neurological tests. Final deliverable from NYS DOH to U.S. Environmental Protection
49    Agency under Cooperative Agreement #CR 824400-01.
50
51    NYS DOH (New York State Department of Health) (2004) Comments on ANeurotoxicity of Tetrachloroethylene
52    (perchloroethylene)-Discussion Paper. @ EPA/600/P-03/005A.
53
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        4-263       DRAFT-DO NOT CITE OR QUOTE

-------
  1    NYS DOH (New York State Department of Health) (2005a) Improving human risk assessment for
  2    tetrachloroethylene by using biomarkers and neurobehavioral testing. U.S. EPA Star Grant #R827445. Grant
  3    #R827446. Available online at
  4    http://cfpub.epa.gov/ncer abstracts/index.cfm/fuseaction/displav.abstractDetail/abstract/977/reprort/O.
  5
  6    NYS DOH (New York State Department of Health)  (2005b) Pumpkin patch day care center investigation. Final
  7    Report. New York State Department of Health, Bureau of Toxic Substance Assessment, Division of Environmental
  8    Health Assessment, Center for Environmental Health
  9
10    NYS DOH (New York State Department of Health)  (2005c) Pumpkin patch day care center follow-up evaluation.
11    Final Report. New York State Department of Health, Bureau of Toxic Substance Assessment, Division of
12    Environmental Health Assessment, Center for Environmental Health.
13
14    NYS OAG (New York State Office of Attorney General). (2004a) Comments from Judy Schreiber, Ph.D. for EPA
15    Docket ORD-2003-0014. March 1, 2004.
16
17    NYS OAG (New York State Office of Attorney General). (2004b) Comment focusing on APeer review workshop of
18    the neurotoxicity of tetrachloroethylene discussion paper: submitted by Judith S. Schreiber, Ph.D. Chief Scientist,
19    Albany Environmental Protection Bureau. ORD-2003-0014-0007.  Available online at
20    http://docket.epa.gov/edkpub/do/EDKStaffItemDetailView?objectId=090007d48038f97c.
21
22    Odum, J; Green, T; Foster,  JR; et al. (1988) The role of trichloracetic acid and peroxisome proliferation in the
23    differences in carcinogenicity of perchloroethylene in the mouse and rat. Toxicol Appl Pharmacol 92:103-112.
24
25    O=Flaherty, EJ; Scott, W; Schreiner, C; et al. (1992) A physiologically based kinetic model of rat and mouse
26    gestation: disposition of a weak acid.  Toxicol Appl Pharmacol 112:245-256.
27
28    Olsen, J; Hemminki, K; Ahlborg, G; et al. (1990) Low birthweight, congenital malformations, and spontaneous
29    abortions among dry-cleaning workers in Scandinavia. Scand J Work Environ Health 16:163-168.
30
31    Olson, MJ; Garg, BD; Murty, CV; et al. (1987) Accumulation of alpha 2u-globulin in the renal proximal tubules of
32    male rats exposed to unleaded gasoline. Toxicol Appl Pharmacol 90:43-51.
33
34    Onofrj, M; Thomas, A; Paci, C; et al. (1998) Optic neuritis with residual tunnel vision in perchloroethylene toxicity.
35    J Toxicol Clin Toxicol 36:603-607.
36
37    Opdam, JJ. (1989) Intra and interindividual variability in the kinetics of a poorly and highly metabolising solvent.
38    BrJIndMed46:831-845.
39
40    Opdam, JJ; Smolders, JF. (1986) Alveolar sampling and fast kinetics of tetrachloroethylene in man. I. Alveolar
41    sampling. Br J Ind Med 43:814-824.
42
43    Pahler, A; Parker, J; Dekant, W. (1999) Dose-dependent protein adduct formation in kidney, liver, and blood of rats
44    and in human blood after perchloroethene inhalation. Toxicol Sci 48:5-13.
45
46    Palmer, CN; Hsu, MH; Griffin,  KJ; et al. (1998) Peroxisome proliferator activated receptor-alpha expression in
47    human liver. Mol Pharmacol 53:14-22.
48
49    Parent, ME; Siemiatycki, J; Fritschi, L. (2000) Workplace exposures and oesophageal cancer. Occup Environ Med
50    57:3256334.
51
52    Park, JH; Spengler, JD; Yoon, DW; et al. (1998) Measurement of air exchange rate of stationary vehicles and
53    estimation of in-vehicle exposure. J Expo Anal Environ Epidemiol 8:65-78.
54
55    Partanen, T; Heikkila, P; Hernberg, S; et al. (1991) Renal cell cancer and occupational exposure to chemical agents.
56    Scand J Work Environ Health 17:231-239.

                   This document is  a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        4-264      DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2     Pastino, GM; Yap, WY; Carroquino, M. (2000) Human variability and susceptibility to trichloroethylene [In Process
  3     Citation]. Environ Health Perspect 108 Suppl 2:201-214.
  4
  5     Paulu, C; Aschengrau, A; Ozonoff, D. (1999) Tetrachloroethylene-contaminated drinking water in Massachusetts
  6     and the risk of colon-rectum, lung, and other cancers. Environ Health Perspect 107:265-271.
  7
  8     Paulu, C; Aschengrau, A; Ozonoff D. (2002) Exploring associations between residential location and breast cancer
  9     incidence in a case-control study. Environ Health Perspect 110:471-478.
10
11     Pegg, DG; Zempel, JA; Braun, WH; et al. (1979) Disposition of tetrachloro(14C)ethylene following oral and
12     inhalation exposure in rats.  Toxicol Appl Pharmacol 51:465-474.
13
14     Pereira, MA. (1996) Carcinogenic activity of dichloroacetic acid and trichloroacetic acid in the liver of female
15     B6C3F1 mice. Fundam Appl Toxicol 31:192-199.
16
17     Pereira, MA; Wang, W; Kramer, PM; et al. (2004a) Prevention by methionine of dichloroacetic acid-induced liver
18     cancer and DNA hypomethylation in mice. Toxicol Sci 77:243-248.
19
20     Pereira, MA; Tao, L; Wang, W; et al. (2004b)  Modulation by celecoxib and difluoromethylornithine of the
21     methylation of DNA and the estrogen receptor-{-a} gene in rat colon tumors.  Carcinogenesis 25:1917-1923.
22
23     Perrin, MC; Opler, MG; Harlap, S, et al. (2007) Tetrachloroethylene  exposure and risk of schizophrenia: Offspring
24     of dry cleaners in a populaton birth cohort, preliminary findings.  Schizophr Res. 90(l-3):251-4.
25
26     Perrone, CE; Shao, L; Williams, GM. (1998) Effect of rodent hepatocarcinogenic peroxisome proliferators on fatty
27     acyl-CoA oxidase, DNA synthesis, and apoptosis in cultured human and rat hepatocytes. Toxicol. Appl Pharmacol.
28     150:277-286.
29
30     Pesch, B; Haerting, J; Ranft, U; et al. (2000a) Occupational risk factors for renal cell carcinoma: agent-specific
31     results from a case-control study in Germany. MURC Study Group.  Multicenter urothelial and renal cancer study.
32     IntJEpidemiol 29:1014-1024.
33
34     Pesch, B; Haerting, J; Ranft, U; et al. (2000b) Occupational risk factors for urothelial carcinoma: agent-specific
3 5     results from a case-control study in Germany. MURC Study Group. Multicenter Urothelial and Renal Cancer. Int J
36     Epidemiol 29:238-247.
37
3 8     Peters, JM; Vanden Heuvel, JP. (2002) Peroxisome, peroxisome proliferators and peroxisome-proliferator activated
39     receptors (PPARs). In: Vanden Heuvel, JP; Perdew,  GH; Mattes, WB; et al.; eds. Cellular and molecular toxicology.
40     Amsterdam: Elsevier Science, pp 133-158.
41
42     Peters, JM; Cattley, RC; Gonzalez, FJ. (1997a) Role of PPARalpha in the mechanism of action of the nongenotoxic
43     carcinogen and peroxisome proliferator Wy-14,643.  Carcinogenesis  18:2029-2033.
44
45     Peters, JM; Hennuyer, N; Staels, B; et al. (1997b). Alterations in lipoprotein metabolism in peroxisome proliferator-
46     activated receptor-alpha-deficient mice. J Biol Chem 272:27303-27312.
47
48     Peters, J; Rusyn, I; Rose, M; et al. (2000) Peroxisome proliferator-activated receptor alpha is restricted to hepatic
49     parenchymal cells, not Kupffer cells: implications for the mechanism of action of peroxisome proliferators in
50     hepatocarcinogenesis Carcinogenesis 21:823-826.
51
52     Phillips, RD; Moran, EJ; Dodd, DE, et al. (1987) A 14-week vapor inhalation toxicity study of methyl isobutyl
53     ketone. Fundam Appl Toxicol 9:380-388.
54
55     Poet, TS; Weitz, KK; Gies, RA; et al. (2002) PBPK modeling of the percutaneous absorption of perchloroethylene
56     from a soil matrix in rats and humans. Toxicol Sci 67:17-31.

                   This document is a draft for review purposes only and  does not constitute Agency policy
       06/06/08                                         4-265       DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2    Pohl, HR; Roney, N; Wilbur, S; et al. (2003) Six interaction profiles for simple mixtures. Chemosphere 53:183-197.
  3
  4    Pohlabeln, H; Boffetta, P; Ahrens, W; et al. (2000) Occupational risks for lung cancer among nonsmokers.
  5    Epidemiology 11:532-538.
  6
  7    Popp, JA. (1984) Mouse liver neoplasia. Washington, DC: Hemisphere.
  8
  9    Popp, W; Muller, G; Baltes-Schmitz, B; et al. (1992) Concentrations of tetrachloroethylene in blood and
10    trichloroacetic acid in urine in workers and neighbors of drycleaning shops. Int Arch Occup Environ Health
11    63(6):393-395.
12
13    Potter, CL; Chang, LW; DeAngelo, AB; et al. (1996) Effects of four trihalomethanes on DNA strand breaks, renal
14    hyaline droplet formation and serum testosterone in male F-344 rats. Cancer Lett 106:235-242.
15
16    Price, RG; Taylor, SA; Olivers, I; et al. (1996) Development and validation of new screening tests for nephrotoxic
17    effects. HumExpToxicol 15 Suppl 1:S10-S19.
18
19    Rachootin, P; Olsen, J. (1983) The risk of infertility and delayed conception associated with exposures in the Danish
20    workplace. J Occup Med 25:394-402.
21
22    Rampy, LW; Quast, JF; Balmer, MF; et al. (1978) Results of a long-term inhalation toxicity study:
23    perchloroethylene in rats.  Toxicology Research Laboratory, Health and Environmental Research, Dow Chemical
24    Company, Midland, MI.
25
26    Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
27    groundwater for a bathing and showering determination.  Risk Anal 13:37-49.
28
29    Rao, MS; Reddy, JK. (1991) An overview of peroxisome proliferator-induced hepatocarcinogenesis.  Environ
30    Health Perspect 93:205-209.
31
32    Rea, WJ; Ross, GH; Johnson, AR; et al. (1991) Chemical sensitivity in physicians. Bol Asoc Med P R 83:383-388.
33
34    Reddy, JK; Lalwani, ND. (1983) Carcinogenesis by hepatic peroxisome proliferators:  evaluation of the risk of
3 5    hypolipidemic drugs and industrial plasticizers to humans.  Crit Rev Toxicol 12:1-58.
36
37    Reddy, JK; Rao, MS. (1986) Peroxisome proliferators and cancer: mechanisms and implications.  Trends in
38    Pharmacol Sci 7:438-443.
39
40    Reddy, JK; Azarnoff, DL; Hignite, CE. (1980) Hypolipidaemic hepatic peroxisome proliferators form a novel class
41    of chemical carcinogens. Nature 283:397-398.
42
43    Reddy, JK; Lalwani, ND; Qureshi, SA; et al. (1984) Induction of hepatic peroxisome proliferation in nonrodent
44    species, including primates. Am JPathol 114:171-183.
45
46    Reddy, JK; Reddy, MK; Usman, MI; et al. (1986) Comparison of hepatic peroxisome proliferative effect and its
47    implication for hepatocarcinogenicity of phthalate esters, di(2- ethylhexyl) phthalate, and di(2-ethylhexyl) adipate
48    with a hypolipidemic drug. Environ Health Perspect 65:317-327.
49
50    Regan, D. (1989) Human brain electrophysiology. New York: Elsevier Science.
51
52    Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
53    for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
54    306.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                       4-266       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Renwick, AG. (1998) Toxicokinetics in infants and children in relation to the ADI and TDI. Food Additives and
  2    Contaminants 15 Suppl: 17-35.
  3
  4    Richmond, RE; Carter, JH; Carter, HW; et al. (1995) Immunohistochemical analysis of dichloroacetic acid (DCA)-
  5    induced hepatocarcinogenesis in male Fischer (F344) rats. Cancer Lett 92:67-76.
  6
  7    Rijhsinghani, KS; Abrahams, C; Swerdlow, MA; et al. (1986) Induction of neoplastic  lesions in the livers of C57BL
  8    x C3HF1 mice by chloral hydrate. Cancer Detect Prev 9:279-288.
  9
10    Ris, H. 1949. The anaphase movement of chromosomes in the spermatocytes of the grasshopper. Biol Bull. 96:90-
11    106.
12
13    Roberts, RA; Soames, AR; Gill, JH; et al. (1995) Non-genotoxic hepatocarcinogens stimulate DNA synthesis and
14    their withdrawal induces apoptosis, but in different hepatocyte populations. Carcinogenesis 16:1693-1698.
15
16    Roberts, RA; James, NH; Hasmall, SC; et al. (2000) Apoptosis and proliferation in nongenotoxic carcinogenesis:
17    species differences and role of PPAR alpha. Toxicol Lett 112-113:49-57.
18
19    Rose, ML;  Rivera, CA; Bradford, BU; et al. (1999) Kupffer cell oxidant production is central to the mechanism of
20    peroxisome proliferators. Carcinogenesis 20:27-33.
21
22    Rosengren, LE; Kjellstrand, P; Haglid, KG. (1986) Tetrachloroethylene: levels of DNA and S-100 in the gerbil CNS
23    after chronic exposure. Neurobehav Toxicol Teratol 8:201-206.
24
25    Rothman, KJ; Greenland, S. (1998) Modern epidemiology.  New York: Lippincott, Williams, and Wilkins.
26
27    Rowe, VK; McCollister, DD; Spencer, HC; et al. (1952) Vapor toxicity of tetrachloroethylene for laboratory animals
28    and human subjects. Arch Ind Hyg Occup Med 5:556-579.
29
30    Roy, AK; McMinn, DM; Biswas, NM. (1975). Estrogenic inhibition of the hepatic synthesis of a2u-globurininthe
31    rat. Endocrinology 97:1501-1508.
32
33    Ruder, AM; Ward, EM; Brown, DP. (1994) Cancer mortality in female and male dry-cleaning workers. J Occup
34    Med 36:867-874.
35
36    Ruder, AM; Ward, EM; Brown, DP. (2001) Mortality in dry-cleaning workers: an update.  Am  J Ind Med 39:121-
ST    132.
38
39    Russo, A; Pacchierotti, F; Metalli, P. (1984) Nondisjunction induced in mouse spermatogenesis by chloral hydrate, a
40    metabolite of trichloroethylene. Environ. Mutagen. 6: 695-703.
41
42    Rusyn, I; Rose, ML; Bojes, HK; et al. (2000a) Novel role of oxidants in the molecular mechanism of action of
43    peroxisome proliferators. Antioxid Redox Signal 2:607-621.
44
45    Rusyn, I; Denissenko, MF; Wong, V; et al. (2000b) Expression of base excision repair enzymes in rat and mouse
46    liver is induced by peroxisome proliferators and is dependent upon carcinogenic potency. Carcinogenesis 21:2141-
47    2145.
48
49    Rusyn, I; Kadiiska, MB; Dikalova, A; et al. (2001) Phthalates rapidly increase production of reactive oxygen species
50    in vivo: Role of Kupffer cells. Mol. Pharmacol 59:744-750.
51
52    Saillenfait, AM; Langonne, I; Sabate, JP. (1995) Developmental toxicity of trichloroethylene, tetrachloroethylene
53    and four of their metabolites in rat whole embryo culture. Arch Toxicol 70:71-82.
54
55    Saland, G. (1967) Accidental exposure to perchloroethylene. NY State J Med 67:2359-2361.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        4-267       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Sallmen, M; Lindbohm, ML; Kyyronen, P; et al. (1995) Reduced fertility among women exposed to organic
  2     solvents. Am JIndMed 27:699-713.
  3
  4     Salmon, AG; Kizer, KW; Zeise, L; et al. (1995) Potential carcinogenicity of chloral hydrateCa review.  J Toxicol
  5     Clin Toxicol 33:115-121.
  6
  7     Sapone, A; Peters, JM; Sakai, S; et al. (2000) The human peroxisome proliferator-activated receptor alpha gene:
  8     Identification and functional characterization of two natural allelic variants. Pharmacogenetics 10:321-333.
  9
10     Sarangapani, R; Gentry, PR; Covington, TR; et al. (2003) Evaluation of the potential impact of age- and gender-
11     specific lung morphology and ventilation rate on the dosimetry of vapors.  Inhal Toxicol 15:987-1016.
12
13     Savolainen, H; Pfaffli, P; Tengen, M; Vainio, H. (1977a) Biochemical and behavioural effects of inhalation
14     exposure to tetrachloroethylene and dichloromethane. J Neuropathol Exp Neurol 36:941-949.
15
16     Savolainen, H; Pfaffli, P; Tengen, M; Vainio, H. (1977b) Trichloroethylene and 1,1,1-trichloroethane: effects on
17     brain and liver after five days intermittent inhalation. Arch Toxicol 38:229-237.
18     Sawyers, CL; Denny, CT; Witte, ON. (1991) Leukemia and the disruption of normal hematopoiesis. Cell 64:337-
19     350.
20
21     Schatten, H; Chakrabarti, A. (1998) Centrosome structure and function is altered by chloral hydrate and diazepam
22     during the first reproductive cell cycles in sea urchin eggs. Eur J Cell Biol 75:9-20.
23
24     Schattenberg, JM; Galle, PR; Schuchmann, M. (2006) Apoptosis in liver disease. Liver Int 26:904-911.
25
26     Scherr, PA; Hutchison, GB; Neiman, RS. (1992) Non-Hodgkin's lymphoma and occupational exposure. Cancer Res
27     52:5503s-5509s.
28
29     Schlehofer, B; Heuer, C; Blettner, M; et al. (1995) Occupation, smoking and demographic factors, and renal cell
30     carcinoma in Germany. Int JEpidemiol 24:51-57.
31
32     Schlesselman, J. (1982) Case-control studies. Design, conduct, analysis.  New York: Oxford University Press.
33
34     Schlichting, LM; Wright, PF; Stacey, NH. (1992) Effects of tetrachloroethylene on hepatic and splenic
3 5     lymphocytotoxic activities in rodents.  Toxicol Ind Health 8:255-266.
36
37     Schreiber, JS; House, S; Prohonic, E; et al. (1993) An investigation of indoor air contamination in residences above
38     dry cleaners. Risk Anal 13:335-344.
39
40     Schreiber, JS. (1993) Predicted infant exposure to tetrachloroethene in human breastmilk.  Risk Anal 13:515-524.
41     Schreiber. JS. (1997) Transport of organic chemicals to breast milk: Tetrachloroethene case study. In Kacew S,
42     Lambert G (eds): Environmental Toxicology and Pharmacology of Human Development. Washington, DC, Taylor
43     and Francis.
44
45     Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
46     tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect 110:655-664.
47
48     Schultz, R; Yan, W; Toppari, J; et al. (1999) Expression of peroxisome proliferator-activated receptor alpha
49     messenger ribonucleic acid and protein in human and rat testis. Endocrinology 140:2968-2975.
50
51     Schultze, MO; Klubes, P; Perman, V; et al. (1959) Blood dyscrasia in calves induced by S-(dichlorovinyl)-L-
52     cysteine. Blood 14:1015-1025.
53
54     Schumann, AM; Quast, JF; Wantanabe, PG. (1980) The pharmacokinetics and macromolecular interactions of
55     perchloroethylene in mice and rats as related to oncogenicity. Toxicol Appl Pharmacol 55:207-219.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         4-268      DRAFT-DO NOT  CITE OR  QUOTE

-------
  1     Schwetz, BA; Leong, KJ; Gehring, PJ. (1975) The effect of maternally inhaled trichloroethylene, perchloroethylene,
  2     methyl chloroform, and methylene chloride on embryonal and fetal development in mice and rats.  Toxicol Appl
  3     Pharmacol 32:84-96.
  4
  5     Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
  6     11:579-583.
  7
  8     Seliger, B. (2005) Strategies of tumor immune evasion.  BioDrugs 19:347-354.
  9
10     Seniori Costantini, A; Miligi, L; Kreibel, D; et al. (2001) A multicenter case-control study in Italy on
11     Hematolymphopoietic neoplasms and occupation. Epidemiology 12:78-87.
12
13     Seppalainen, AM; Antti-Poika, M. (1983) Time course of electrophysiological findings for patients with solvent
14     poisoning: adescriptive study. ScandJ Work Environ Health 9:15-24.
15
16     Shafer, TJ; Bushnell, PJ; Benignus, VA; et al. (2005) Perturbation of voltage-sensitive Ca2+ channel function by
17     volatile organic solvents. 129. J Pharmacol Exp Ther 315:1109-1118.
18
19     Sharanjeet-Kaur; Mursyid, A; Kamaruddin, A; et al. (2004) Effect of petroleum derivatives and solvents on colour
20     perception.  ClinExp Optom 87:339-343.
21
22     Sharpe, CR; Rochon, JE; Adam, JM; et al. (1989) Case-control study of hydrocarbon exposures in patients with
23     renal cell carcinoma. CMAJ 140:1309-1318.
24
25     Shawn, SH; Robins, JM. (1986) Comment. J Am StatAssoc 84:604-609.
26
27     Sheldon, L; Handy, R; Hartwell, W; et al. (1985) Human exposure assessment to environmental chemicals: nursing
28     mothers study. Final report. Available from: Research Triangle Institute, Research Triangle Park, NC.
29
30     Sher, T; Yi, HF; McBride, OW; et al. (1993) CDNA cloning, chromosomal mapping, and functional characterization
31     of the human peroxisome proliferator activated receptor. Biochemistry 32:5598-5604.
32
33     Shu, XO;  Steward, P; Wen, WO; et al. (1999) Parental occupational exposure to hydrocarbons and risk of acute
34     lymphocytic leukemia in offspring. Cancer Epidemiol Biomarkers Prev 8(9):783-791.
35
36     Silverman, DT; Levin, LI; Hoover, RN; et al. (1989a) Occupational risks of bladder cancer in the United States: I.
37     Whitemen.  JNatl Cancer Inst 81:1472-1480.
38
39     Silverman, DT; Levin, LI; Hoover, RN. (1989b) Occupational risks of bladder cancer in the United States: II
40     Nonwhitemen. JNatl Cancer Inst 81:1480-1483.
41
42     Simpson ,AE; Brammar, WJ; Pratten, MK; et al. (1995) Translactational induction of CYP4A expression in 10.5-
43     day neonatal rats by the hypolipidemic drug clofibrate. Biochem Pharmacol. 50:2021-2032.
44
45     Simpson, AE; Brammar, WJ; Pratten, MK; et al. (1996) Placenta! transfer of the hypolipidemic drug, clofibrate,
46     induces CYP4A expression in 18.5-day fetal rats. Drug Metab Disp 24:547-554.
47
48     Singh, M; Sinha, U. (1976).  Chloral hydrate induced haploidization mAspergillus nidulans. Experientia
49     32:114-115.
50
51     Singh, M; Sinha, U. (1979).  Mitotic haploidization and growth ofAspergillus nidulans  on media containing chloral
52     hydrate.  JCytol Genet 14:1-4.
53
54     Skowron, J; Miranowicz-Dzierzawska, K; Zapor, L; et al. (2001) Interactions of some organic solvents:
55     hydrocarbons and chloroalkene. Int J Occup Saf Ergon 7(l):35-47.
56

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-269      DRAFT-DO NOT  CITE OR QUOTE

-------
  1     Smith, EM; Miller, ER; Woolson, RF; et al. (1985) Bladder cancer risk among laundry workers, dry cleaners, and
  2     others in chemically-related occupations.  JOccupMed 27:295-297.
  3
  4     Smith, MK; Randall, JL; Read, EJ; et al. (1989) Teratogenic activity of trichloroacetic acid in the rat. Teratology
  5     40:445-451.
  6
  7     Sokol, L; Loughran, TP. (2006) Large granular lymphocyte leukemia. The Oncologist 11:263-273.
  8
  9     Solet, D; Robins, TG. (1991) Renal function in dry cleaning workers exposed to perchloroethylene.  Am J Ind Med
10     20:601-614.
11
12     Solet, D; Robins, TG; Sampaio, C. (1990) Perchloroethylene exposure assessment among dry cleaning workers. Am
13     Ind HygAssoc 151:566-574.
14
15     Solleveld, HA; Haseman, JK; McConnell, EE. (1984) National history of body weight gain, survival, and neoplasia
16     in the F344 rat. J Natl Cancer Inst 72:929-940.
17
18     Sonnenfeld, N; Hertz-Picciotto, I; Kaye, WE. (2001) Tetrachloroethylene in drinking water and birth outcomes at
19     the US Marine Corps base at Camp Lejeune, NC. Am J Epidemiol 154:902-908.
20
21     Sora, S; Agostini-Carbone, ML. (1987) Chloral hydrate, methylmercury hydroxide and ethidium bromide affect
22     chromosomal segregation during meiosis of Saccharomyces cervisiae. Mutat Res. 190: 13-17.
23
24     Sparrow, GP. (1977) A connective tissue disorder similar to vinyl chloride disease in a patient exposed to
25     perchlorethylene.  ClinExpDermatol2:17-22.
26
27     Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
28     of subjects chronically exposed to organic solvents (II)]. G Ital Med Lav Ergon 19:85-88.
29
30     Spirtas, R; Stewart, PA; Lee, JS; et al. (1991) Retrospective cohort mortality study of workers at an aircraft
31     maintenance facility. I. Epidemiological results.  Br J Ind Med 48:515-530.
32
3 3     Stauber, AJ; Bull, RJ. (1997) Differences in phenotype and cell replicative behavior of hepatic tumors induced by
34     dichloroacetate (DCA) and trichloroacetate (TCA).  Toxicol Appl Pharmacol 144:235-246.
35
36     Stauber, A; Bull, RJ; Thrall, BD. (1998) Dichloroacetate and trichloroacetate promote clonal expansion of
37     anchorage-independent hepatocytes in vivo and in vitro.  Toxicol Appl Pharmacol 260:287-294.
38
39     Stemhagen, A; Slade, J; Altaian, R; et al. (1983) Occupational risk factors and liver cancer. A retrospective case-
40     control study of primary liver cancer in New Jersey. Am J Epidemiol 117:443-454.
41
42     Sterchele, PF;  Sun, H; Peterson, RE; et al. (1996) Regulation of peroxisome proliferator-activated receptor-alpha
43     mRNA in rat liver. Arch Biochem Biophys 326:281-289.
44
45     Stevenson, DE; McClain, RM; Popp, JA; et al. (1990) Mouse liver carcinogenesis: mechanisms and species
46     comparisons. New York: Wiley-Liss.
47
48     Stewart, RD; Dodd, HC. (1964) Absorption of carbon tetrachloride, trichloroethylene, tetrachloroethylene,
49     methylene chloride and 1,1,1-trichloroethane through the human skin. J Ind Hyg 25:429-446.
50
51     Stewart, RD; Erley, DS; Schaeffer, AW; et al. (1961) Accidental vapor exposure to anesthetic concentrations of a
52     solvent containing tetrachloroethylene. Indus Med Sug 30:327-330.
53
54     Stewart, RD; Baretta, ED; Dodd, HC; et al. (1970) Experimental human exposure to tetrachloroethylene. Arch
55     Environ Health 20:225-229.
56

                    This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-270       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Stewart, RO; Hake, CL; Wu, A; et al. (1977) Effects of perchloroethylene drag interactions on behavior and
  2     neurological function. DHEW (NIOSH) Publ. No. 77-191. Department of Health, Education, and Welfare.
  3     Washington, DC.
  4
  5     Stonard, MD; Phillips, PG; Foster, JR; et al. (1986a) Alpha 2u-globulin: measurement in rat kidney and relationship
  6     to hyaline droplets. Clin Chim Acta 160:197-203.
  7
  8     Storer, RD; McKelvey, TW; Kraynak, AR; et al. (1996) Revalidation of the in vitro alkaline elution/rat hepatocyte
  9     assay for DNA damage: improved criteria for assessment of cytotoxicity and genotoxicity and results for 81
10     compounds. Mutat Res 368:59-101.
11
12     Storm, JE; Mazor, KA. (2004) Update of residential tetrachloroethylene exposure and decreases in visual contrast
13     sensitivity. Environ Health Perspect 112(15):A862-863.
14
15     Story, DL; Meierhenry, EF; Tyson, CA; et al. (1986) Differences in rat liver enzyme-altered foci produced by
16     chlorinated aliphatics and phenobarbital. Toxicol Ind Health 2:351-362.
17
18     Stott, WT. (1988). Chemically induced proliferation of peroxisomes: Implications to risk assessment. Regul Toxicol
19     Pharmacol 8:125-159.
20
21     Stromberg, PC. (1985) Large granular lymphocyte leukemia in F344 rats: model for human T gamma lymphoma,
22     malignant histiocytosis, and T-cell chronic lymphocytic leukemia. Am J Pathol 119:517-519.
23
24     Styles, JA; Wyatt, I; Coutts, C. (1991) Trichloroacetic acid: studies on uptake and effects on hepatic DNA and liver
25     growth in mouse.  Carcinogenesis 12:1715-1719.
26
27     Suarez, L; Weiss, NS; Martin, J. (1989) Primary liver cancer death and occupation in Texas. Am J Ind Med 15:167-
28     175.
29
30     Subramoniam, A; Goel, SK; Pandya, KP; et al. (1989) Influence of trichloroethylene treatment on phosphoinositides
31     in rat brain.  Toxicol Lett 49:55-60.
32
33     Sverdrap, B; Kallberg, H; Bengtsson, C; et al. (2005) Association between occupational exposure to mineral oil and
34     rheumatoid arthritis: results from the Swedish EIRA case-control study. Arthritis Res Ther 7:R1296-R1303.
35
36     Swenberg, JA; Lehman-McKeeman, LD. (1999) A 2 Urinary-globulin-associated nephropathy as a mechanism of
37     renal tubule cell carcinogenesis in male rats. In: Capen, C; Dybing, E; Rice, J; et al.; eds. Species differences in
3 8     thyroid, kidney and urinary bladder carcinogenesis. Lyon: IARC Scientific Publications No 147; pp 95-118.
39
40     Swenberg, JA; Short, B; Borghoff, S; et al. (1989) The comparative pathobiology of alpha 2u-globulin nephropathy.
41     Toxicol Appl Pharmacol 97:35-46.
42
43     Szakmary, E; Ungvary, G; Tatrai, E. (1997) The offspring-damaging effect of tetrachloroethylene in rats, mice, and
44     rabbits. Central Eur J Occup Environ Med 3 (1): 31 -3 9.
45
46     Tang, ML; Kemp, AS; Thorburn, J; et al. (1994) Reduced interferon-gamma secretion in neonates and subsequent
47     atopy. Lancet 344:983-985.
48
49     Tanios, MA; El, GH; Rosenberg, B J; et al. (2004) Can we still miss tetrachloroethylene-induced lung disease? The
5 0     emperor returns in new clothes.  Respiration 71:642-645.
51
52     Tao, L; Kramer, PM; Ge, R; et al. (1998) Effect of dichloroacetic acid and trichloroacetic acid on DNA methylation
53     in liver and tumors of female B6C3F1 mice. Toxicol Sci 43:139-144.
54
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-271       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Tao, L; Yang, S; Xie, M; et al. (2000) Hypomethylation and overexpression of c-jun and c-myc protooncogenes and
  2     increased DNA methyltransferase activity in dichloroacetic and trichloroacetic acid-promoted mouse liver tumors.
  3     Cancer Lett 158:185-193.
  4
  5     Taskinen, H; Anttila, A; Lindbohm, ML; et al. (1989) Spontaneous abortions and congenital malformations among
  6     the wives of men occupationally exposed to organic solvents.  Scand J Work Environ Health 15:345-352.
  7
  8     Tateishi, T; Nakura, H; Asoh, M; et al. (1997) A comparison of hepatic cytochrome P450 protein expression
  9     between infancy and postinfancy. Life Sciences 61(21):2567-74.
10
11     Teschke, K; Morgan, MS; Checkoway, H; et al. (1997) Surveillance of nasal and bladder cancer to locate sources of
12     exposure to occupational carcinogens. Occup Environ Med 54:443 -451.
13
14     Thangada, S; Alvares, K; Mangino, M; et al. (1989) An in vitro demonstration of peroxisome proliferation and
15     increase in peroxisomal beta-oxidation system mRNAs in cultured rat hepatocytes treated with ciprofibrate. FEES
16     Lett 250:205-210.
17
18     Thomas, KW; Pellizzari, ED; Perritt, RL; et al. (1991) Effect of dry-cleaned clothes on tetrachloroethylene levels in
19     indoor air, personal air, and breath for residents of several New Jersey homes. J Expo Anal Environ Epidemiol
20     1:475-490.
21
22     Till, C; Koren, G; Rovet, JF.  (2001a) Prenatal exposure to organic solvents and child neurobehavioral performance.
23     Neurotoxicol Teratol 23(3):235-245.
24
25     Till, C; Westall, CA; Rovet, JF; et  al. (200 Ib) Effects of maternal occupational exposure to organic solvents on
26     offspring visual functioning:  a prospective controlled study. Teratology 64(3):134-141.
27
28     Till C; Rovet JF; Koren G; et al. (2003) Assessment of visual functions following prenatal exposure to organic
29     solvents. Neurotoxicology 24(4-5):725-31.
30
31     Till, C; Westall, CA; Koren,  G; et al. (2005) Vision abnormalities in young children exposed prenatally to organic
32     solvents. Neurotoxicology 26:599-613.
33
34     Tinston, DJ. (1994) Perchloroethylene: A multigeneration inhalation study in the rat. CTL/P/4097. Available from:
35     IRIS Information Desk, U.S.  Environmental Protection Agency.
36
37     Tomatis, L; Aitio, A; Wilbourn, J;  et al. (1989) Human carcinogens so far identified. Jpn J Cancer Res 80:795-807.
38
39     Tontonoz, P; Hu, E; Spiegelman, DM. (1994) Stimulation of adipogenesis in fibroblasts by PPARgamma2, alipid-
40     activated transcription factor. Cell 79:1147-1156.
41
42     Topping, MD; Forster, HW; Dolman, C; et al. (1986) Measurement of urinary retinol-binding protein by enzyme-
43     linked immunosorbent assay, and its application to detection of tubular proteinuria. ClinChem 32:1863-1866.
44
45     Toraason, M; Clark, J; Dankovic, D; et al. (1999) Oxidative stress and DNA damage in Fischer rats following acute
46     exposure to trichloroethylene or perchloroethylene [in process citation]. Toxicology 138:43-53.
47
48     Toraason, M; Butler, MA; Ruder, A; et al. (2003) Effect of perchloroethylene, smoking, and race on oxidative DNA
49     damage in female dry cleaners. Mutat Res 539:9-18.
50
51     Travier N;  Gridley, G; De Roos, AJ; et al. (2002) Cancer incidence of dry cleaning, laundry, and ironing workers in
52     Sweden. Scand J Work Environ Health 28:341-348.
53
54     Trevisan, A; Cristofori, P; Fanelli,  G. (1999) Glutamine synthetase activity in rat urine as sensitive marker to detect
55     S3 segment-specific injury of proximal tubule induced by xenobiotics. Arch Toxicol 73:255-262.
56

                    This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-272       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Trevisan, A; Macca, I; Rui, F; et al. (2000) Kidney and liver biomarkers in female dry-cleaning workers exposed to
  2    perchloroethylene.  Biomarkers 5:399-409.
  3
  4    Tugwood, JD; Issemann, I; Anderson, RG; et al. (1992) The mouse peroxisome proliferator activated receptor
  5    recognizes a response element in the 5* flanking region of the rat acyl CoA oxidase gene. EMBO 11:433-439.
  6
  7    Tugwood, JD; Aldridge, TC; Lambe, KG; et al. (1996) Peroxisome proliferator-activated receptors: structures and
  8    function. Ann NY Acad Sci 804:252-265.
  9
10    U.S. EPA (Environmental Protection Agency). (1980) Ambient water quality criteria for tetrachloroethylene. Office
11    of Water, Washington, DC; EPA 440/5-80-073.  Available from: National Technical Information Service,
12    Springfield, VA.
13
14    U.S. EPA (Environmental Protection Agency). (1985a) Health assessment document for tetrachloroethylene
15    (perchloroethylene). Office of Health and Environmental Assessment, Office of Research and Development,
16    Washington, DC; EPA/600/8-82/005F. Available from: National Technical Information Service, Springfield, VA.
17    PB-85-249696/AS.
18
19    U.S. EPA (Environmental Protection Agency). (1985b) Drinking water criteria document for tetrachloroethylene.
20    Available from: National Technical Information Service, Springfield, VA; PB86-118114.
21
22    U.S. EPA (Environmental Protection Agency), (1985c) Chemical carcinogens: review of the science and its
23    associated principles. Office of Science and Technology Policy. Federal Register 50:10372-10442.
24
25    U.S. EPA (Environmental Protection Agency). (1986a) Addendum to the health assessment document for
26    tetrachloroethylene (perchloroethylene). Prepared by the Office of Health and Environmental Assessment,
27    Washington, DC; EPA/600/8-82/005FA. Available from: National Technical Informtion Service, VA; PB-86-
28    174489/AS.
29
30    U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for carcinogen risk assessment. Federal Register
31    51(185):34014-34025.
32
33    U.S. EPA (Environmental Protection Agency). (1987a) Evaluation of the carcinogenicity of unleaded gasoline.
34    Office of Health and Environmental Assessment, Office of Research and Development, Washington, DC;
35    EPA/600/8-87/001. Available from: National Technical Information Service, Springfield, VA; PB-87-186151/AS.
36
37    U.S. EPA (Environmental Protection Agency). (1987b) Health assessment document for beryllium.  Office of
3 8    Health and Environmental Assessment, Office of Research and Development, Washington, DC; EP A/600/8-84/026F.
39    (Pages 7-84.) Available from: National Technical Information Service, Springfield, VA.
40
41    U.S. EPA (Environmental Protection Agency). (1991a) Response to issues and the data submissions on the
42    carcinogenicity  of tetrachloroethylene (perchloroethylene).  Office of Health and Environmental Assessment,
43    Washington, DC; EP A/600/6-91/002F. Available from: National Technical Information Service, Springfield, VA.
44
45    U.S. EPA (Environmental Protection Agency). (1991b) Alpha-2u-globulin: association with chemically induced
46    renal toxicity and neoplasia in the male rat.  Risk Assessment Forum, Office of Research and Development,
47    Washington, DC; EPA/625/3-91/0IF.  Available from: National Technical Information Service, Springfield, VA.
48
49    U.S. EPA (Environmental Protection Agency). (1997a) Chemical and radiation leukemogenesis in humans and
5 0    rodents and the value of rodent models for assessing risks of lymphohematopoietic cancers. Office of Research and
51    Development, Washington, DC; EPA/600/R-97/090. Available from: National Technical Information Service,
52    Springfield, VA; PB97-208185.
53
54    U.S. EPA (Environmental Protection Agency). (1997) Exposure factors handbook. Vol  1. National Center for
55    Environmental Assessment, Office of Research and Development, Washingotn, DC; EPA/600/P-95/002Fa.
56    Available online at http://www.epa.gov/ncea.

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                       4-273       DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2    U.S. EPA (Environmental Protection Agency). (2000a) Toxicological review of chloral hydrate.  National Center for
  3    Environmental Assessment, Washington, DC; EPA/635/R-00/006.  Available online at http://www.epa.gov/iris.
  4
  5    U.S. EPA (Environmental Protection Agency). (2000b) Toxicological review of vinyl chloride.  Integrated Risk
  6    Information System (IRIS). National Center for Environmental Assessment, Washington, DC; EPA/635/R-00/004.
  7    Available online at http://www.epa.gov/iris.
  8
  9    U.S. EPA (Environmental Protection Agency). (200la) Trichloroethylene health risk assessment: synthesis and
10    characterization [external review draft]. National Center for Environmental Assessment, Office of Research and
11    Development, Washington, DC; EPA/600/P-01/002A. Available online at
12    http://oaspub.epa.gov/eims/eimscomm.getfile7p  download id=4580.
13
14    U.S. EPA (Environmental Protection Agency). (200Ib) Sources, emission and exposure for trichloroethylene (TCE)
15    and related chemicals. National Center for Environmental Assessment, Washington, DC; EPA/600/R-00/099.
16
17    U.S. EPA (Environmental Protection Agency). (2002) Toxicological review of vinyl chloride. National Center for
18    Environmental Assessment, Washington, DC; EPA/635/R-00/004.  Available online at http://www.epa.gov/iris.
19
20    U.S. EPA (Environmental Protection Agency). (2002) Child-specific exposure factors handbook. Interim report.
21    EPA-600-P-00-002B.  U.S. Environmental Protection Agency, Washington, DC.
22
23    U.S. EPA (Environmental Protection Agency). (2003a) IRIS summary of toxicological effects of dichloroacetic acid.
24    National Center for Environmental Assessment, Washington, DC. Available online at http://www.epa.gov/iris.
25
26    U.S. EPA. (Environmental Protection Agency). (2003b) Neurotoxicity of tetrachloroethylene (perchloroethylene):
27    discussion paper.  National Center for Environmental Assessment, Washington, DC; EPA/600/P-03/005 A.
28    Available online at http://www.epa.gov/ncea.
29
30    U.S. EPA (Environmental Protection Agency). (2003c) Framework for cumulative risk assessment. Risk
31    Assessment Forum, Washington, DC; EPA/630/P-02/001F. Available online at http://www.epa.gov/ncea/raf.
32
33    U.S. EPA (Environmental Protection Agency). (2003d) Proposed science policy: PPAR-aagonist-mediated
34    hepatocarcinogenesis in rodents and relevance to human health risk assessment (available online at
35    http://www.epa.gov/scipolv/sap/meetings/2003/december9/peroxisomeproliferatorsciencepolicvpaper.pdf) and
36    minutes of FIFRA Scientific Advisory Panel Meeting, Arlington, VA (available online at
37    http://www.epa.gov/oscpmont/sap/2003/december9/meetingminutes.pdf).
38
39    U.S. EPA (Environmental Protection Agency). (2004) Summary report of the peer review workshop on the
40    neurotoxicity of tetrachloroethylene (perchloroethylene) discussion paper. National Center for Environmental
41    Assessment, Washington, DC; EPA/600/R-04/041. Available online at http://www.epa.gove/ncea.
42
43    U.S. EPA (Environmental Protection Agency). (2005a) Trichloroethylene issue paper 3: role of peroxisome
44    proliferator-activated receptor agonism and cell signaling in trichloroethylene toxicity. National Center for
45    Environmental Assessment, Washington, DC; EPA/600/R-05/024.  Available online at
46    http://oaspub.epa.gov/eims/eimscomm.getfile7p  download id=438646.
47
48    U.S. EPA (Environmental Protection Agency). (2005b) Guidelines for carcinogen risk assessment. Federal Register
49    70(66)17765-18717.  Available online at http://www.epa.gov/cancerguidelines.
50
51    U.S. EPA (Environmental Protection Agency). (2005c) Supplemental guidance for assessing cancer susceptibility
52    from early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
53    online at http://www.epa.gov/cancerguidelines.
54
55    U.S. EPA (Environmental Protection Agency). (2005d) Human testing; proposed plan and description of review
56    process. Federal Register 70(25):6661-6667.

                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                       4-274        DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2     U.S. EPA (Environmental Protection Agency). (2006) Aging and toxic response: issues relevant to risk assessment.
  3     EPA/600/P-03/004A. Available online at http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=156648.
  4
  5     UAREP (Universities Associated for Research and Education in Pathology). (1983) Hydrocarbon toxicity; acute,
  6     subchronic and chronic effects in relation to unleaded gasoline exposure of rodents with comments on the
  7     significance to human health. Contract No. PS-6 UAREP (504-2) with the American Petroleum Institute. Bethesda,
  8     MD.
  9
10     Uetrecht, J; Zahid, N; Rubin, R. (1988) Metabolism of procainamide to a hydroxylamine by human neutrophils and
11     mo no nuclear leukocytes.  Chem Res Toxicol 1:74-78.
12
13     Ulm, K; Henschler, D; Vamvakas, S. (1996) Occupational exposure to perchloroethylene. Cancer Causes Control
14     7:284-286.
15
16     Umezu, T; Yonemoto, J; Soma, Y; et al.  (1997) Behavioral effects of trichloroethylene and tetrachloroethylene in
17     mice. Pharmacol Biochem Behav 58:665-671.
18
19     Valles, EG; Laughter, AR; Dunn, CS; et al. (2003) Role of the peroxisome proliferator-activated receptor alpha in
20     responses to diisononyl phthalate. Toxicology 191:211-225.
21
22     Valic, E; Waldhor, T; Konnaris, C; et al. (1997) Acquired dyschromatopsia in combined exposure to solvents and
23     alcohol. Int Arch Occup Environ Health 70:403-406.
24
25     Vamvakas, S; Dekant, W; Berthold, K; et al. (1987) Enzymatic transformation of mercapturic acids derived from
26     halogenated alkenes to reactive and mutagenic intermediates. Biochem Pharmacol 36:2741-2748.
27
28     Vamvakas, S; Dekant, W; Henscher, D. (1989a) Genotoxicity of haloalkene and haloalkane glutathione S-
29     conjugates in porcine kidney cells.  Toxicol In Vitro 3:151-156.
30
31     Vamvakas, S; Herkenhoff, M; Dekant, W;  et al. (1989b) Mutagenicity of tetrachloroethene in the  Ames test--
32     metabolic activation by conjugation with glutathione.  J Biochem Toxicol 4:21-27.
33
34     Vamvakas, S; Dekant, W; Henschler, D.  (1989c) Assessment of unscheduled DNA synthesis in a  cultured line of
3 5     renal epithelial cells exposed to cysteine  S-conjugates of haloalkenes and haloalkanes. Mutat Res 222:329-335.
36
37     Vamvakas, S; Kochling, A; Berthold, K; et al. (1989d) Cytotoxicity of cysteine S-conjugates: structure-activity
38     relationships. Chem Biol Interact 71:79-90.
39
40     Vartiainen, T; Pukkala, E; Strandman, T; et al. (1993) Population exposure to tri- and tetrachloroethylene and cancer
41     risk: two cases of drinking water pollution. Chemosphere 27:1171-1181.
42
43     Vaughan, TL; Stewart, PA; Davis, S; et al.  (1997) Work in dry cleaning and the incidence of cancer of the oral
44     cavity, larynx, and oesophagus. Occup Environ Med 54:692-695.
45
46     Verplanke, AJ; Leummens, MH; Herber, RF. (1999) Occupational exposure to tetrachloroethene and its effects on
47     the kidneys. J Occup Environ Med 41:11-16.
48
49     Vieira, I; Sonnier, M; Cresteil, T. (1996) Developmental expression of CYP2E1 in the human liver:
50     hypermethylation control of gene expression during the neonatal period. Eur J Biochem 238(2):476-83.
51
52     Vieira, V; Aschengrau, A; Ozonoff, D. (2005) Impact of tetrachloroethylene-contaminated drinking water on the
53     risk of breast cancer: Using a dose model to assess exposure in a case-control study. Environ Health 4:3.
54
55     Vingrys, AJ; King-Smith, PE. (1988) A quantitative scoring technique for panel tests of color vision.  Invest
56     Ophthalmol Vis Sci 29:50-53.

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-275       DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2     Vohl, MC; Lepage, P; Gaudet, D; et al. (2000) Molecular scanning of the human PPAR-a gene. Association of the
  3     1162v mutation with hyperapobetalipoproteinemia. J Lipid Res 41:945-952.
  4
  5     Volkel, W; Friedewald, M; Lederer, E; et al. (1998) Biotransformation of perchloroethene: dose-dependent
  6     excretion of trichloroacetic acid, dichloroacetic acid,  and N-acetyl-S- (trichlorovinyl)-L-cysteine in rats and humans
  7     after inhalation.  Toxicol Appl Pharmacol 153:20-27.
  8
  9     Vyskocil, A; Emminger, S; Tejral, J; et al. (1990) Study on kidney function in female workers exposed to
10     perchloroethylene. Hum Exp Toxicol 9:377-380.
11
12     WHO, IPCS Tetrachloroethene (Cicads 68, 2006). World Health Organization, International Programme on
13     Chemical Saftey. Available online at http://www.inchem.org/documents/cicads/cicads/cicad68.htm.
14
15     Walgren, JE; Kurtz, DT; McMillan, JM. (2000a) The effect of the trichloroethylene metabolites trichloroacetate and
16     dichloroacetate on peroxisome proliferation and DNA synthesis in cultured human hepatocytes. Cell Biol Toxicol
17     16:257-273.
18
19     Walgren, JE; Kurtz, DT; McMillan, JM. (2000b) Expression of PPAR(-a) in human hepatocytes and activation by
20     trichloroacetate and dichloroacetate. Res  Commun Mol Pathol Pharmacol  108:116-132.
21
22     Walgren, JL; Jollow, DJ; McMillan, JM. (2004) Induction of peroxisome proliferation in cultured hepatocytes by a
23     series of halogenated acetates. Toxicol 197(3)189-197.
24
25     Walker, JT; Burnett, CA; Lalich, NR;  et al. (1997) Cancer mortality among laundry and dry cleaning workers. Am J
26     IndMed 32:614-619.
27
28     Walles, SA. (1986) Induction of single-strand breaks in DNA of mice by trichloroethylene and tetrachloroethylene.
29     Toxicol Lett 31:31-35.
30
31     Wang, S; Karlsson, JE; Kyrklund, T; et al. (1993) Perchloroethylene-induced reduction in glial and neuronal cell
32     marker proteins in rat brain. Pharmacol Toxicol 72:273-278.
33
34     Wang, JL; Chen, WL; Tsai, SY; et al.  (2001) An in vitro model for evaluation of vaporous toxicity of
3 5     trichloroethylene and tetrachloroethylene to CHO-K1 cells.  Chem Biol Interact 137:139-154.
36
37     Ward, JM; Griesemer, RA; Weisburger, EK. (1979) The mouse liver tumor as an endpoint in carcinogenesis tests.
38     Toxicol Appl Pharmacol 51:389-397.
39
40     Warner, JA; Miles, EA; Jones, AC; et al. (1994) Is deficiency of interferon gamma production by allergen triggered
41     cord blood cells apredictor of atopic eczema? 103. ClinExp Allergy 24:423-430.
42
43     Warren, DA; Reigle, TG; Muralidhara, S; et al. (1996) Schedule-controlled operant behavior of rats following oral
44     administration of perchloroethylene: time course and relationship to blood and brain solvent levels. J Toxicol
45     Environ Health 47:345-362.
46
47     Wartenberg, D; Reyner, D; Scott, CS.  (2000) Trichloroethylene and cancer: epidemiologic evidence. Environ
48     Health Perspect 108 Suppl 2:161-176.
49
50     Watanabe, K; Fuji, H; Takahashi, T; et al. (2000) Constitutive regulation of cardiac fatty acid metabolism through
51     peroxisome proliferator-activated receptor -a associated with age-dependent cardiac toxicity. J Biol Chem
52     275:22293-22299.
53
54     Weber, G; Lea, MA. (1966) The molecular correlation concept of neoplasia.  Adv Enzyme Regul 4:115-145.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        4-276       DRAFT-DO NOT CITE OR QUOTE

-------
  1     Weisenburger, DD. (1992) Pathological classification of non-Hodgkin's lymphoma for epidemiological studies.
  2     Cancer Res 52:5456s-5462s.
  3
  4     Weiss, NS. (1995) Cancer in relation to occupational exposure to perchloroethylene. Cancer Causes Control
  5     6:257B66.
  6
  7     White, IN; Razvi, N; Gibbs, AH; et al. (2001) Neoantigen formation and clastogenic action of HCFC-123 and
  8     perchloroethylene in human MCL-5 cells. Toxicol Lett 124:129-138.
  9
10     Wick, MJ; Mihic, SJ; Ueno, S; et al. (1998) Mutations of gamma-aminobutyric acid and glycine receptors change
11     alcohol cutoff: evidence for an alcohol receptor? Proc Natl Acad Sci U S A 95:6504-6509.
12
13     Williams, RL; Creasy, RK; Cunningham, GC; et al. (1982) Fetal growth and perinatal viability in California. Obstet
14     Gynecol 59:624-632.
15
16     Windham, GC; Shusterman, D; Swan, S; et al. (1991) Exposure to organic solvents and adverse pregnancy outcome.
17     Am J of IndMed 20:241-259.
18
19     Yamakawa-Kobayashi, K; Ishiguro, H; Arinami, T; et al. (2002) A Val227Ala polymorphism in the peroxisome
20     proliferator activated receptor -a (PPAR-a) gene is associated with variations in serum lipid levels. J Med Genet
21     39:189-191.
22
23     Yamakura, T; Mihic, SJ; Harris, RA. (1999) Amino acid volume and hydropathy of a transmembrane site determine
24     glycine and anesthetic  sensitivity of glycine receptors. J Biol Chem 274:23006-23012.
25
26     Yang, Q.; Ito S; Gonzalez, FJ. (2007). Hepatocyte-restricted constitutive activation of PPAR-a induces
27     hepatoproliferationbut not hepatocarcinogenesis. Carcinogenesis 28(6): 1171-7.
28
29     Yoon, JS; Mason JM; Valencia R; et al. (1985) Chemical mutagenesis testing in Drosophila. IV. Results of 45
30     coded compounds tested for the National Toxicology Program. Environ Mutagen 7: 349-367.
31
32     Yount, EA; Felten, SY; O=Connor, BL; et al. (1982) Comparison of the metabolic and toxic effects of 2-
33     chloropropionate and dichloroacetate.  J Pharmacol Exp Ther 222:501-508.
34
3 5     Youssef, J; Badr, M. (1999) Biology of senescent liver peroxisomes: role in hepatocellular aging and disease.
36     Environ Health Perspect 107:791-797.
37
38     Yu, S; Cao, WQ; Kashireddy, P; et al. (2001) Human peroxisome proliferator-activated receptor a (PPARa) supports
39     the induction of peroxisome proliferation in PPARa deficient mouse  liver. J Biol Chem 276:42485-42491.
40
41     Zanelli, U; Puccini, P;  Acerbi, D; et al. (1996) Induction of peroxisomal beta-oxidation and P-450 4A-dependent
42     activities by pivalic and trichloroacetic acid in rat liver and kidney. Arch Toxicol 70:145-149.
43
44     Zhang, B; Marcus, SL; Sajjadi, FG; et al. (1992) Identification of a peroxisome proliferator-responsive element
45     upstream of the gene encoding rat peroxisomal enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase.
46     Proceedings of the National Academy of Sciences of the US A 89:7541 -7545.
47
48     Zhang, B; Marcus, S; Miyata, K; et al. (1993) Characterization of protein-DNA interactions within the peroxisome
49     proliferator-responsive element of the rat hydratase-dehydrogenase gene. J Biol Chem 268:12939-12945.
50
51     Zhou, YC; Waxman, DJ. (1998) Activation of peroxisome proliferator-activated receptors by chlorinated
52     hydrocarbons and endogenous steroids. Environ Health Perspect 106(suppl 4):983-988.
53
54     Zingg, JM: Jones, PA.  (1997) Genetic and epigenetic aspects of DNA methylation on genome expression, evolution,
55     mutation and carcinogenesis. Carcinogenesis 18:869-882.


                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         4-277       DRAFT-DO NOT CITE OR QUOTE

-------
 1                              5. DOSE-RESPONSE EVALUATION
 2
 O
 4    5.1.  INHALATION REFERENCE CONCENTRATION (RfC)
 5           Although the RfD is commonly presented first in the IRIS toxicological reviews, the RfC
 6    is presented in Section 5.1 and the RfD in Section 5.2 because the available data were primarily
 7    from inhalation exposure and pharmacokinetic modeling was available to carry out route-to-
 8    route extrapolation of the RfC to the oral route of exposure.
 9           The RfCl for tetrachloroethylene is derived through a process of (1) considering all
10    studies and selecting the adverse health effects that occur at the lowest exposure concentration,
11    (2) selecting the point of departure (POD)2 at which the adverse health effect either is not
12    observed or would occur at a relatively low prevalence (e.g., 10%), (3) deriving the POD in
13    terms of the human equivalent concentration (HEC), and (4) reducing this exposure
14    concentration by uncertainty factors (UFs) to account for uncertainties in the extrapolation from
15    the study conditions to an estimate of human environmental exposure.  This is EPA's first
16    attempt to define a tetrachloroethylene RfC for IRIS. Health assessments from other agencies,
17    more fully described in Appendix A, have included a criterion for noncarcinogenic effects
18    associated with inhalation exposure based on neurotoxic effects observed in human
19    epidemiologic studies (NYS DOH,  1997; ATSDR, 1997).
20
21    5.1.1. Choice of Principal Study and Critical Effect
22           The database of human and  animal studies on inhalation toxicity of tetrachloroethylene is
23    adequate to support derivation of inhalation reference values.  A number of targets of toxicity
24    from chronic exposure to tetrachloroethylene, include the nervous system, liver, kidney,
25    reproductive system, and developing fetus, with published reports in both animals and humans.
26    Greatest consideration is given to human data, if adequate, to develop an RfC.
27           Neurological  effects were judged to be associated with lower tetrachloroethylene
28    concentrations. This finding is in agreement with Rao and Brown (1993), who, using categorical
29    analysis methods, identified neurological effects as the most sensitive noncancer toxicity
              The RfC is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to
      the human population (including sensitive subgroups) that is likely to be without appreciable risk of deleterious
      effects during a lifetime. It can be derived from a NOAEL, a LOAEL, or a benchmark concentration, with
      uncertainty factors generally applied to reflect limitations of the data used. It is generally used in EPA's noncancer
      health assessments. The RfC, like the RfD for oral exposure, is based on the assumption that thresholds exist for
      certain toxic effects, such as liver pathology, but may not exist for other toxic effects, such as carcinogenicity.
             2
              The POD denotes a dose at the lower end of the observed dose-response curve where extrapolation to
      lower doses begins. For effects other than cancer, the POD is either a NOAEL, a LOAEL if no NOAEL can be
      identified, or a modeled point (e.g., a BMCL10 or an LED10) if the data are suitable for dose-response modeling.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-1         DRAFT-DO NOT CITE  OR QUOTE

-------
 1    endpoint. A number of studies assessing neurobehavioral effects in both humans and rodents are
 2    available for RfC analysis. The epidemiologic body of evidence is characterized by studies that
 3    used standardized neurobehavioral batteries.  In addition, some studies employed assessment of
 4    visual function (Cavalleri et al., 1994; Schreiber et al., 2002; Echeverria et al., 1994, 1995), a
 5    neurological outcome known to be sensitive to volatile organic compounds. Most
 6    epidemiological studies have examined occupational exposure to tetrachloroethylene. Two
 7    epidemiological studies examined residential exposure to tetrachloroethylene (Altmann et al.,
 8    1995; Schreiber et al., 2002).  Together, the epidemiologic evidence supports an inference of a
 9    broad range of cognitive, behavioral, and visual functional deficits following tetrachloroethylene
10    exposure (U.S. EPA, 2004).
11           The research in animal models on the effects of tetrachloroethylene on functional
12    neurological endpoints consists of screening studies (functional observation battery, motor
13    activity) or effects on sensory system function as assessed by evoked potential.  Some
14    consistency is seen in the animal models, with effects on motor activity and motor function
15    following exposure to tetrachloroethylene in either the adult or the developmental period,
16    changes in evoked potentials following acute and subchronic exposures, and replication of
17    observed alterations in brain DNA, RNA,  or protein levels and brain weight changes.
18           Of the studies discussed in Chapter 4, a number of epidemiologic studies of neurological
19    effects in either occupational workers or residential subjects with tetrachloroethylene  exposure
20    are considered for the principal study with which to identify the POD. No single epidemiologic
21    study stands  out as a superior candidate for identifying the POD, as all of the available studies
22    have limitations. However, some studies are considered more desirable as a principal or critical
23    study than other studies for the reasons below. The epidemiologic studies by Ferroni  et al.
24    (1992) and Spinatonda et al. (1997) have associated uncertainties related to incomplete reporting
25    and are methodologically weaker than other epidemiologic  studies, such as those of Seeber
26    (1989), Altmann et al. (1995), or Echeverria et al. (1994, 1995), that assessed neurobehavioral
27    functions. The studies by Echeverria et al. (1994, 1995) are not informative for supporting a
28    POD because the authors attribute observed visual  function effects to past higher
29    tetrachloroethylene concentrations and historical exposure data that are not available.  Table 5-1
30    identifies study characteristics and the rationale for the principal study considered in the RfC
31    analysis.
32           Seeber (1989) reports effects on visuo-spatial function, as does the residential  study by
33    Altmann et al. (1995).  Both studies are considered adequate for quantitative analysis, given the
34    numbers of study subjects (with Seeber et al., 1989, having the larger number of study subjects),
35    and their use of appropriate statistical methods, including methods to adjust for potentially

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-2         DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
        Table 5-1. Summary of rationale for principal study selection
       Consideration/
         approach
                  Type of data
                             Decision
      Quality of study
                 Animal
                 neurotoxicity
                 studies
Animal neurotoxicity studies are considered as supporting studies. An
RfC/RfD from human data, if available and of adequate quality, is
preferred for reducing interspecies extrapolation uncertainties.
      Quality of study
                 Human
                 neurotoxicity
                 studies
Both occupational and residential studies on tetrachloroethylene exposure
contain uncertainties regarding their use for quantitative analysis.  Some of
these epidemiology studies carry greater weight for quantitative analysis
than other studies. The occupational studies of Seeber (1989) and Cavelleri
et al. (1994), and the residential study of Altmann et al. (1995) are
considered to carry greater weight for quantitative analysis than Ferroni et
al. (1992), Echeverria et al. (1994, 1995), and Spinatonda et al. (1997).
      Measurement
      tool
                 Standardized
                 neurobehavioral
                 battery
Both occupational and residential epidemiology studies assessed
neurobehavioral function using a standardized neurobehavioral battery.
The battery has been widely administered to occupational population in
different setting with a reasonably high degree of validity. WHO and
ATSDR recommend these test methods to evaluate nervous system deficits
in adults and children.
      Endpoint
                 Deficits in
                 neurological
                 domains such as
                 attention, motor
                 function,
                 vigilance,or
                 visuo-spatial
                 function.
There is congruence of neurological effects observed in studies of both
residential and occupational populations. These domains are also sensitive
to acute tetrachloroethylene exposure in controlled human studies. The
consistency of observed effects between occupational and residential
populations and their persistence with lower tetrachloroethylene
concentration, as experienced by residential populations, provide a strong
rationale for a study of lower-level exposures as the basis for the RfC.
      Relevance of
      exposure
      scenario
                 Epidmiology
                 studies of
                 residential
                 populations
Tetrachloroethylene exposure to residential populations is of lower
concentration and of chronic duration compared to acute duration and
higher concentration exposure to occupational populations.  Additionally,
potential to tetrachloroethylene peak or intensity concentrations is more
common with occupational exposures. A study of residential exposure, if
adequate and of similar quality as an occupational epidemiology study, is
preferred for supporting the RfC because it better represents exposure
scenarios of interest to EPA.
 3
 4
 5
 6
 7
 8
 9
10
11
12
confounding factors.  However, in both studies, statistical analyses that adjusted for potential
confounders may have not have been fully complete due to the use of categorical variables.
        The report by Cavalleri et al. (1994) is consistent with the growing body of literature
indicating that chronic exposure to a variety of volatile organic solvents, including
tetrachloroethylene, toluene, styrene, and carbon disulfide, is associated with deficits in visual
perception measured either as deficits in  color vision or deficits in VCS (see Section 4.6.1);
visual perception is a sensitive test of neurological impairment.  The study authors reported
poorer performance on a test of color vision among dry cleaning operators.  A statistically
significant lower prevalence of tetrachloroethylene-exposed dry cleaners had perfect scores on a
             This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                        5-3          DRAFT-DO NOT CITE OR QUOTE

-------
 1    test of color perception, and their mean score was higher as compared with those of controls,
 2    indicating impaired color vision among dry cleaners. No effects were observed in laundry
 3    workers with exposure to a lower mean TWA concentration than that of dry cleaning operators.
 4    These effects of tetrachloroethylene-induced deficits in color vision are supported by other
 5    studies (Muttray et al., 1997; Gobba et al., 1998).
 6           Deficits in visual function were also reported in Schreiber et al. (2002), a study originally
 7    designed as a pilot for a study composed of a larger number of subjects, assessed
 8    tetrachloroethylene with air monitoring and markers in biological samples (biomarkers).  This
 9    pilot study was expanded after its inception to include control subjects and tests  of VCS (NYS
10    OAG, 2004). The findings in Schreiber et al. (2002), a first report of VCS  deficits, need
11    replication. For these reasons, although this study contributes to the weight of evidence for
12    hazard identification, it is less desirable than the residential study of Altmann et  al. (1995) as a
13    critical study for developing an RfC.
14           The study by Altmann et al. (1995) is chosen as the principal study  for a  number of
15    reasons. First, it adopted a standardized neurobehavioral battery to evaluate neurological effects
16    in residents. Tests in this battery have been widely administered to occupational populations in
17    different settings with a reasonably high degree of validity (Anger et al., 2000).  Additionally,
18    several public health organizations, such as the World Health Organization and the ATSDR in
19    the United States,  recommend these test methods to evaluate nervous system deficits in adults
20    and children (Anger et al., 1994, 2000, 2003; ATSDR, 1996; Amler et al., 1994). Second, there
21    is congruence of neurological effects observed in studies of both residential and  occupational
22    populations. As shown in Table 4-1 (Chapter 4), decrements in a number of neurological
23    domains such as attention, motor function, and vigilance reported by Altmann et al. (1994) are
24    also reported for occupationally exposed populations.  The consistency of these effects between
25    the two populations and their persistence with lower tetrachloroethylene concentration, as
26    experienced by residential populations, provide a strong rationale for  a study of lower-level
27    exposures as the basis for the RfC.  Last, a study of residential  exposures is preferred for
28    quantitative analysis because it better represents exposure scenarios of interest to EPA.
29    Table 5-2 identifies studies and outcomes considered for quantitative  analysis.
30           Table 5-1 summarizes chronic, subchronic or longer-term, and developmental toxicity
31    studies considered for derivation of the inhalation RfC and are a subset of the body of evidence
32    on tetrachloroethylene more fully described in Chapter 4.  These studies are considered
33    supportive of a POD and an RfC because they report effects associated with lower exposure
34    concentrations or are  studies with multiple experimental exposures, allowing exploration of
35    benchmark dose (BMD)  approaches.  For each study, Table 5-1 identifies the species; the

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-4         DRAFT-DO NOT CITE OR QUOTE

-------
o
Oi
Table 5-2.  Inhalation studies considered in the development of an RfC
Organ/ system
Liver
Kidney
CNS
Study
Brodkin et al.
(1995)
Gennari et al.
(1984)
Kjellstrand et al.
(1984)
NTP(1986)
JISA(1993)
Verplanke et al.
(1999)
Trevisan et al.
(2000)
Franchini et al.
(1983)
Mutti et al.
(1992)
NTP(1986)
Altmann et al.
(1995)
Schreiber et al.
(2002)
Species
Human
Human
Mouse
Mouse
Mouse
Human
Human
Human
Human
Rat
Human
Human
Duration/ exposure
route
20 years mean duration
12 years mean duration
Subchronic (4 weeks)
continuous
Chronic bioassay
(104 weeks)
Chronic
(104 weeks)
4 years duration
(geometric mean)
15 years duration
(geometric mean)
14 years mean duration
10 years duration
Chronic bioassay
(104 weeks)
10.6 years median
duration
5.8 years mean duration
NOAEL/LOAEI/
(ppm)
4Ji, 19.8
(overall mean, 16)
11.3 (mean)

0,2,37,75,150
0, 100, 200
0, 10, SO, 250
L2 (mean)
M
(mean)
Ifl
15 (median)
0, 200. 400
0.7 (mean)
0.2 (median)
0.1 (residents, median
and mean), maybe as
high as 0.4 (mean) and
0.3 (median
Effect
Hepatic
parenchymal
changes
Gamma glutamyl
transpeptidase
Increased liver
weight
Increased liver
degeneration,
necrosis
Increased
angiectasis
Retinol binding
protein
Glutamine
sythetase
Lysozyme
Urine and serum
markers of
nephrotoxicity
Increased
karyomegaly,
megalonuclear-
cytosis
Verbal function ,
cognitive
function,
vigilance, vision
Visual contrast
sensitivity
Human equivalent continuous
concentrations'" (ppm)
NOAEL/LOAEL
LOAEL: 2
LOAEL: 4
LOAEL: 9
LOAEL: 18
NOAEL: 2
LOAEL: 0.3
LOAEL: 3
LOAEL: 4
LOAEL: 5
LOAEL: 36
LOAEL: 0.7
LOAEL: 0.4
BMCLC
BMCL10: 0.5
NA
BMCLS: 0.6
BMCLS/P: 1.4-10
NA
BMCL10: 2
BMCL10/P: 4.3-23
NA
NA
NA

BMCL10: 2.2
NA
NA
oo S3'



   I


   I

   §
   ***.

   £3'
  1
   TO'
^ %
 I  Co

"" §
   I
H I
O
H
W

-------
Table 5-2. Inhalation studies considered in the development of an RfC (continued)
Organ/ system
CNS (cont.)
Study
Schreiber et al.
(2002)
Cavalleri et al.
(1994)
Spinatonda et al.
(1997)
Seeber(1989)
Ferroni et al.
(1992)
Echeverria et al.
(1995)
Kjellstrand et al.
(1984)
Rosengren et al.
(1986)
Mattsson et al.
(1998)
Wang et al.
(1993)
Species
Human
Human
Human
Human
Human
Human
Mouse
Gerbil
Rat
Rat
Duration/ exposure
route
4 years mean duration
8.8 years mean duration
Inhalation (no
information on
duration)
>10 years mean
duration
10.6 years mean
duration
15 (high-exposure
group) years mean
duration
Subchronic (4 weeks)
continuous
Subchronic (12 weeks,
with 16- week follow-
up) continuous
Subchronic (13 weeks)
6hrs/d, 5d/w
Subchronic (12 weeks)
continuous
NOAEL/LOAEI/
(ppm)
0.3 (daycare workers,
mean and median)
2
8 (median)
12,53
JJ
41 (operators)
0,9,22,75,150
0, 6Q, 300
0, 50, 200, 800
0, 300. 600
Effect
Visual contrast
sensitivity
Dyschromatopsia
Reaction time
Visuospatial
function,
information
processing speed
Reaction time,
continuous
performance
Endocrine:
prolactin
Visuospatial
function
Butyryl
cholinesterase
Brain: protein,
DNA
concentration
Flash-evoked
potential
Reduced brain
weight, DNA,
protein
Human equivalent continuous
concentrations'" (ppm)
NOAEL/LOAEL
LOAEL: 0.1
LOAEL: 3
LOAEL: 3
LOAEL: 4
LOAEL: 5
LOAEL: 4
LOAEL: 15
NOAEL: 9
LOAEL: 60
NOAEL: 36
NOAEL: 300
BMCLC
NA
NA
NA
NA
NA
NA
NA
NAe
NA
NA
NA

-------
      Table 5-2.  Inhalation studies considered in the development of an RfC (continued)
Organ/ system
Reproductive/
developmental
Study
Eskenazi et al.
(1991)
Nelson et al.
(1980)
Beliles et al.
(1980)
Tinston(1994)
Carney et al.
(2006)
Species
Human
Rat
Mouse
Rat
Rat
Duration/ exposure
route
9.5 years mean
duration, occupational
7 hrs/day on gestation
day 7-1 3 or 14-20
5 days exposure;
1, 4, and 10 week
follow-up
Developmental —
multigeneration; 6
hrs/day, 5 days/wk
Developmental — 6
hrs/day on GD 6-19
NOAEL/LQAELa
(ppm)
12
0,100,900
0, 100. 500
0,100,300,1,000
0, 65, 250, 600
Effect
Sperm quality
Decreased
weight gain in
offspring;
CNS: behavior,
brain
acetylcholine
Sperm quality
F2A pup deaths
by day 29;
FlandF2
generations:
CNS depression
Decreased fetal
and placental
weight and
incomplete
ossification of
thoracic
veretebral
centra
Human equivalent continuous
concentrations'" (ppm)
NOAEL/LOAEL
LOAEL: 0.4
NOAEL: 29
NOAEL: 21
NOAEL: 18
NOAEL: 16
BMCLC
NA
NA
NA
BMCLoi: 1.8
NA
Experimental/observational NOAEL is underlined, LOAEL is double-underlined.
Calculated using RfC methodology for a Category 3 gas, extrathoracic effects, and adjusted to equivalent continuous exposure; occupational exposures were
multiplied by 5/7(days) x 10/20 (nrVday, breathing rate).
BMCL10 is the lower bound on concentration associated with a 10% response over background. BMCLS is the lower bound on dose associated with a one
standard deviation change in the mean over the control response. This corresponds to an excess risk of approximately 10% for developing a level of the
endpoint above the 98th percentile (or below the 2nd percentile) of the control distribution for normally distributed effects. BMCLx/p is the estimation, based on
total metabolism, for either type of benchmark response (percent change or standard deviation).

-------
        Table 5-2. Inhalation studies considered in the development of an RfC (continued)

d Atmospheric monitoring indicated slightly higher exposure levels were experienced by subjects.  Schreiber et al. (2002) found mean tetrachloroethylene
  concentrations of 0.2 ppm (0.09 ppm, median) of four families living in apartments above active dry cleaning and two families living in an apartment building
  where dry cleaning had ceased 1 month earlier.  Ambient monitoring of these six apartments during a period of active dry cleaning indicated exposure to
  higher concentrations, mean = 0.4 ppm (median 0.2 ppm).
e Benchmark modeling not feasible; exposure-response relationship showed very little gradation among responses aside from apparently maximal response in
  mid- and high-dose groups relative to control and low-dose groups.

-------
 1    exposure duration; the ambient (experimental) concentration or, for epidemiologic studies, the
 2    mean concentration; the observed effect; and the exposure concentration, identified as the
 3    NOAEL or the LOAEL. Additionally, HECs for LOAELs or NOAELs are presented so as to
 4    better allow examination of effect levels across studies and species. The HECs in Table 5-1 are
 5    calculated using the RfC methodology for a category 3 gas, extrathoracic effects, and adjusted to
 6    equivalent continuous exposure (U.S. EPA, 1994).
 7
 8                        NOAEL * [HEcj = NOAEL * [ADJ] (ppm) x (Hb/g)A/Hb/g)H
 9    where:
10           NOAEL* [REG] = the NOAEL or analogous effect level such as the benchmark
11                           concentration (BMC)
12
13           NOAEL*[ADJ] = the NOAEL or analogous effect level adjusted for duration of
14                           experimental regimen; experimental exposure times duration (number of
15                           hrs exposed/24 hrs) times week (number of days of exposure/7 days)
16
17           (Hb/g)A/Hb/g)n = the ratio of the blood/gas (air) partition coefficient of the chemical  for the
18                           laboratory animal species to the human value.  The value of 1 is used for
19                           the ratio if (Hb/g)A > Hb/g)H
20
21    Response levels in Table 5-1 are presented as either a NOAEL, a LOAEL if the study did not
22    identify a NOAEL, or a modeled BMCL3 if the study results were suitable for modeling, using
23    BMD methodology (U.S. EPA, 2000). Five studies (Brodkin et al., 1985; Kjellstrand et al.,
24    1984; JISA, 1993; NTP, 1986; Tinston, 1994) reported toxicity in other organs besides the
25    nervous system and are presented for the comparative purposes. These studies had experimental
26    designs with multiple exposure concentrations and quantitative information that were sufficient
27    for quantitative analysis using BMD methodologies.
28           Ideally,  an examination using pharmacokinetically derived dose metrics is preferred
29    when well validated models are available.  An alternative procedure for deriving a POD using
30    pharmacokinetic modeling to estimate metabolite production for liver toxicity data sets is also
31    examined in this document.
             3 A BMCL10 is the lower 95% bound on the dose or concentration associated with a 10% extra risk
      compared to background. In general, the benchmark approach is a superior methodology to the LOAEL/NOAEL
      approach because it makes more complete use of exposure/ response data rather than being limited by the sample
      size of the study group that happened to be exposed at the LOAEL or NOAEL.  The BMD approach identifies doses
      that are not restricted to being one of the study exposure levels, another improvement over the LOAEL/NOAEL
      approach, particularly when doses are widely spaced.  BMDs correspond to specific response levels, such as 10%
      extra risk, facilitating comparisons across studies and endpoints. Because the BMD corresponds to an adverse effect
      level, it should be treated conceptually as a LOAEL.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-9         DRAFT-DO NOT CITE  OR QUOTE

-------
 1    5.1.2. Method of Analysis
 2           The present analysis defines a POD using the traditional NOAEL/LOAEL approach in
 3    addition to using BMD modeling where feasible. Further, PBPK modeling was used with
 4    suitable studies in animals in order to inform the process of extrapolating to human equivalent
 5    exposures.  The use of these alternative approaches has the potential to add information to the
 6    NOAEL/LOAEL approach.
 7           Altmann et al. (1995) reports a mean 8-hr TWA of 0.7 ppm (4.8 mg/m3) for residents
 8    exposed to tetrachloroethylene from living in close proximity to a dry cleaning establishment.
 9    This mean concentration is used as the POD for the RfC derivation. The POD is not adjusted for
10    exposure duration as is the general practice when using an occupational study. Instead, an
11    assumption was made that residents were continuously exposed. In other words, no further
12    adjustments to the estimated exposure level to approximate continuous exposure levels were
13    considered necessary due to the lack of information concerning the duration and length of
14    exposure of the study population.
15           The application of BMD methodologies offer advantages over traditional
16    LOAEL/NOAEL approaches, however,  exposure in Altmann et al. (1995) is reported as a group
17    mean concentration and does not allow fitting of BMD models. Data sufficient for BMD
18    modeling generally came from animal experiments, with the exception of one human study.  This
19    was the Brodkin et  al. (1985) study, which found an increasing incidence of hepatic  parenchymal
20    changes in laundry  workers with increasing exposure to perchloroethylene. There was no
21    concurrent control group, so substitution of a background level was necessary (Hartwell et al.,
22    1985).  BMD modeling of the two reported groups plus the substituted control group yielded a
23    BMCLio of 0.5 ppm (see Table 5-1). In addition to the lack of a control group, another
24    limitation of the result modeled from the Brodkin et al. (1985) study is that hepatic parenchymal
25    changes appear to be a less severe endpoint, as all of the participants had normal liver function
26    measurements, and its  relationship to frank liver disease is not known.  Despite these
27    uncertainties, the result provides support for the POD derived from the Altmann et al. (1995)
28    study.  Furthermore, Eskenazi et al. (1991), who observed effect on sperm quality at a similar
29    mean exposure concentration as that of Altmann et al.  (1995) and Schreiber et al. (2002), support
30    the POD of the critical study. Table 5A-1 in the appendix provides  details of the BMD
31    modeling.
32           The animal  studies suitable for BMD modeling addressed liver and kidney toxicity and
33    pup death. For liver toxicity in mice (increased liver weights [Kjellstrand et al., 1984]), a human
34    equivalent BMCLS4 of 0.6 ppm was estimated using administered concentration, and a BMCLs/p
             BMCLS = Lower bound on dose (concentration) associated with a one standard deviation change in the
      mean over the control response. This corresponds to an excess risk of approximately 10% for the proportion of
      individuals above the 98th percentile (or below the 2nd percentile) of the control distribution for normally distributed
      effects (U.S. EPA, 2000).
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-10       DRAFT-DO NOT CITE OR QUOTE

-------
 1    of 1.4-10 ppm (9.6-69 mg/m3) was derived using pharmacokinetic models, assuming that total
 2    metabolism was the relevant dosimeter for animals and humans (see Tables 5A-2 and 5A-3 for
 3    modeling details).  For liver toxicity in rats (increased angiectasis [JISA, 1993]), BMD modeling
 4    yielded a BMCLio of 3.7 ppm (25 mg/m3) using administered concentration and a BMCLio/p of
 5    4.3-23 ppm (29-156 mg/m3) using pharmacokinetic modeling (see Table 5A-6). PBPK models
 6    are described in Section 3.5 and are more fully considered in the cancer dose-response analysis
 7    (see Section 5.4.4.2).  Karyomegaly  (kidney) was observed in male rats in the chronic study by
 8    the NTP (1986). BMD modeling yielded a BMCL75 of 29 ppm (near the lowest exposure tested)
 9    and a BMCLio of 2.2 ppm (15 mg/m3). Because the lowest exposure was associated with a
10    relatively high response, estimation of the BMCLio  is somewhat tenuous, and modeling using
11    total metabolism was not pursued. Last, BMD modeling of pup deaths through Day 29 in the
12    F2A generation of a multigeneration study (Tinston, 1994) yielded a BMCLoi of 1.8 ppm (12
13    mg/m3). These BMD analyses are more fully presented in Tables 5A-2 through 5A-9 and are
14    provided in support of the choice of the Altmann et  al. (1995) study as the most relevant data
15    source for developing the RfC.
16
17    5.1.3.  Reference Concentration (RfC) Derivation, Including Application of Uncertainty
18           Factors
                                                3
19           The NOAEL of 0.7 ppm (4.8 mg/nr) from Altmann et al. (1995) is the POD, as described
20    above. The POD is reduced by the following UFs.5
             BMCLx/p = Lower bound on dose (concentration), where X denotes the benchmark response (either in
      percent or one standard deviation), based on dose metric of total metabolism as estimated by a pharmacokinetic
      model. This subscript distinguishes these BMCLs from those based on administered exposure.
              RfCs apply to lifetime human environmental exposure, including exposures of sensitive subgroups.
      Differences between study conditions and conditions of human environmental exposure may make a dose that
      appears safe in an experiment not safe in the environment. UFs account for differences between study conditions
      and conditions of human environmental exposure. These include the following:
        1.  Variation from average humans to sensitive humans: RfCs apply to the human population, including sensitive
           subgroups, but studies rarely target  sensitive humans. Sensitive humans could be adversely affected at doses
           lower than those that affect a general study population; consequently, general population NOAELs are reduced
           to cover sensitive humans.
        2.  Uncertainty in extrapolation from animals to humans: If an RfC is developed from animal studies, the animal
           NOAEL is reduced to reflect pharmacokinetic and pharmacodynamic factors that may make humans more
           sensitive than animals.
        3.  Uncertainty in extrapolating from subchronic NOAELs to chronic NOAELs:  RfCs apply to lifetime exposure,
           but sometimes the best data come from shorter studies. Lifetime exposure can have effects that do not appear
           in a shorter study; consequently, a safe dose for lifetime exposure can be less than the safe dose for a shorter
           period. If an RfC is developed from less-than-lifetime studies, the less-than-lifetime NOAEL is adjusted to
           estimate a lifetime NOAEL.
        4.  Uncertainty in extrapolation from LOAELs to NOAELs:  RfCs estimate a dose with appreciable risks, but
           sometimes adverse effects are observed at all study doses. If an RfC is developed from a dose where there are
           adverse effects, that dose is adjusted to estimate a NOAEL.
        5.  Other factors to reflect professional assessment of scientific uncertainties not explicitly treated  above,
           including completeness of the overall database, minimal sample size, or poor exposure characterization.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-11        DRAFT-DO NOT CITE OR QUOTE

-------
 1       1. Human variation: The UF of 10 is applied for human variation. Although human
 2          residential data were used as the basis for the POD (Altmann et al., 1995), the overall
 3          database does not support the use of a value other than the default  10-fold UFn. The
 4          rationale for this determination is based on several considerations.  First, Altmann et al.
 5          (1995) excluded subjects with disorders such as hypertension, neurological or
 6          endocrinological  diseases (e.g., diabetes), impaired vision, or impairment of joints.
 7          Hence, subjects in Altmann et al.  (1995) are considered to be a select population,
 8          analogous to an occupational population and subject to selection bias known as the
 9          "healthy worker effect."  The use of these  exclusion criteria and the small numbers of
10          subjects in Altmann et al.  (1995; « = 37, 14 exposed and 23  control subjects) suggest that
11          the range of human variation in a larger and more diverse population is not likely
12          represented by this study.  Second, no information is presented in Altmann et al. (1995)
13          with which to examine variation between subjects.  Third, the sparse data available on
14          tetrachloroethylene indicate the presence of pharmacokinetic variation in the human
15          population. One report described variation in tetrachloroethylene blood concentrations
16          among nine subjects exposed acutely to tetrachloroethylene and observed a twofold
17          spread in the ratio of alveolar air concentrations to atmospheric concentrations (Opdam,
18          1989).
19
20          Gentry et al. (2003), Clewell et al. (2004), and Pelekis et al. (2001) present
21          pharmacokinetic modeling simulations of pharmacokinetic variation between adults and
22          children in tetrachloroethylene parent and  its metabolites. As the authors themselves
23          indicated, validation of these results for various life stages and further refinement of the
24          parameters in the model are necessary before the results of such an analysis can be
25          considered for use in risk assessment.  Further investigation of variability in the
26          parameters used in the Clewell et al. (2004) analysis is also needed before their results
27          can be used to address pharmacokinetic uncertainty for age and gender variability.
28
29          Given an adequate database, or after adjustment in the assessment  for deficiencies in the
30          database, a reference value incorporating the default 10-fold factor for human variation is
31          believed to adequately address likely susceptibilities in children. A thorough evaluation
32          of the animal and human hazard data for tetrachloroethylene identified the developing
33          fetus and the young child as susceptible life stages (populations).  As described in Section
34          4.9, data-derived noncancer outcomes of concern in children for perinatal exposure are
35          (1) spontaneous abortions, (2) childhood mortality, and (3) neurological impairment.
36          This assessment contains a database uncertainty factor addressing,  in part, limitations in
37          life stage-related data in human and rodent studies.
38
39          Section 4.9 also describes differential  opportunities for exposure to children. However,
40          susceptibilities of this nature are not addressed in the determination of the UF for human
41          variation, but rather are used to establish a context for the hazard and dose-response
42          evaluation and are further addressed in the exposure assessment and subsequent risk
43          characterization.
44
45       2. Animal-to-human uncertainty. This UF is used when the POD is supported by an animal
46          study. When the POD is supported  by a human study, this UF is not needed.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-12        DRAFT-DO NOT CITE OR QUOTE

-------
 1       3.  Subchronic-to-chronic uncertainty. A factor to address the potential for more severe
 2           toxicity from chronic or lifetime exposure to tetrachloroethylene is not used in this
 3           assessment. The epidemiologic studies, except for Schreiber et al. (2002), are all of
 4           median duration of exposures of more than 15% of a 70-year lifespan.  There are no data
 5           to suggest that continuing exposure to tetrachloroethylene can increase the severity of
 6           effects; duration-response trends are not generally evident in the human studies.
 7
 8       4.  LOAEL-to-NOAEL uncertainty. The default value of 10 is applied for use of a LOAEL
 9           because of the lack of a NOAEL in Altmann et al. (1995).
10
11       5.  Database uncertainty. A threefold database UF has been applied to address the lack of
12           data to adequately characterize the hazard and dose-response in the human population.
13           The rationale for this database UF is based on several considerations. There is human
14           evidence of neurotoxicity following tetrachloroethylene exposure, with both visual
15           system dysfunction and cognitive performance deficits. However, these studies have
16           limitations, and in particular lack adequate data to address childhood or other life stage
17           susceptibility. There is also a lack of animal studies (including in developing animals)
18           designed to clearly investigate these neurotoxicity findings and define and characterize
19           the exposure-response relationship.

20           A broad range of animal  toxicology data are available for the hazard assessment of
21           tetrachloroethylene, as described throughout this document. Included in these studies are
22           short-term and long-term bioassays in rats and mice (see Chapter 4 and Table 4-2);
23           neurotoxicology studies in rats, mice, and gerbils (see Tables 4-6 and 4-7); prenatal
24           developmental toxicity studies in rats, mice, and rabbits and a two-generation
25           reproduction study in rats (see Table 4-10); and numerous supporting genotoxicity and
26           metabolism studies. Nevertheless, critical data gaps have been identified. Data from
27           acute studies in animals (Warren et al., 1996;  Umeza et al., 1997) suggest that cognitive
28           function is affected by exposure to tetrachloroethylene. These studies do not address the
29           exposure-response relationship for subchronic and chronic tetrachloroethylene exposures
30           on cognitive functional deficits observed in humans (e.g., Seeber, 1989; Echeverria et al.,
31           1994;  and Altmann et al., 1995).
32
33           Even more importantly, there is a lack of cognitive testing in both developmentally
34           exposed animals and adult animals following exposures to  tetrachloroethylene that are
35           longer than acute durations of exposure. For another critical outcome,  visual function,
36           there has been a limited evaluation of visual function in rodents, with the exception of the
37           evoked potential studies by Mattsson  et al. (1998). Visual  system dysfunction and
38           processing of visual spatial information are sensitive endpoints in human studies. The
39           exposure-response relationship of these functional deficits  could be evaluated more
40           definitively with studies using homologous methods that examine retinal and visual
41           function in experimental  animals.  These types of studies could help elucidate whether
42           there are both peripheral  and central effects of tetrachloroethylene exposure on visual
43           perception, and they could be used as an animal model to better define the exposure-
44           response relationships.
45
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-13       DRAFT-DO NOT CITE OR QUOTE

-------
 1           Additionally, the database of human epidemiological studies includes studies of liver and
 2           kidney toxicity (see Tables 4-1 and 4-3), neurotoxicity (see Tables 4-4 and 4-5), and
 3           developmental/reproductive effects (see Table 4-8).  The reference value is established on
 4           the basis of a sensitive neurological effect in adult humans (Altmann et al., 1995), and
 5           some characterization of the response of  children to tetrachloroethylene exposure was
 6           found in limited data for a similar neurological (visual system) parameter (Schreiber et
 7           al., 2002).
 8
 9           Although the toxicological database is considered adequate for establishing a reference
10           value, some uncertainties remain. In both the Altmann et al. (1995) and the Schreiber et
11           al. (2002) studies, there was a lack of robust sample  size and an inadequate dose-response
12           characterization for potentially susceptible human populations following
13           tetrachloroethylene exposures. Although the Altmann et al.  (1995) study (with a LOAEL
14           of 0.7 ppm for healthy adult subjects) was used in setting the reference value (based on a
15           number of considerations that are summarized in Section 5.2.1), the Schreiber et al.
16           (2002) study, using an alternative visually based testing paradigm, identified adverse
17           visual effects at 0.4 ppm (see Section 4.6.1.2.11). Additionally, in a postnatal
18           neurotoxicity study in mice (Fredriksson et al., 1993), persistent neurological effects (i.e.,
19           increased locomotion and decreased rearing behavior at 60 days of age, measured 43 days
20           after exposure ceased) were observed at an oral dose of 5 ppm, with no NOAEL,
21           although this study did not conform to traditional toxicity testing guidelines (see Section
22           4.6.2.2). The possibility exists that if adequate, robust, dose-response data based on the
23           most appropriate neurophysiological  and  cognitive tests were available, the exposure
24           eliciting an adverse response (and hence the POD for the reference value) could be lower
25           than that established on the basis of deficits in visuo-spatial and cognitive function
26           following tetrachloroethylene exposure in healthy adults (Altmann et al., 1995).
27
28           A total UF of 300 was applied to this effect level:  10 for human variation, 10 for
29    consideration of LOAEL to NOAEL, and 3 for database uncertainties.
30
31                     RfC  = NOAEL I UF
32                           =0.7ppm(4.8mg/m3)/300
33                           = 0.002 ppm (0.02 mg/m3, rounded to 1 significant digit}
34
35    where:
36           0.7 ppm  = LOAEL for neurological effects in residents exposed to tetrachloroethylene
37                      (Altmann etal., 1995)
38
39           300      = composite UF chosen to account for extrapolation from a LOAEL to a
40                     NOAEL in humans, intra-individual variability in humans, and uncertainties
41                      in the database
42
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-14       DRAFT-DO NOT CITE OR QUOTE

-------
 1    5.1.4. Supporting Studies
 2          PODs and reference values (RfVs) that could be derived from supporting neurotoxicity
 3    studies identified in Table 5-2 (see Section 5.1.1) are presented below in Figure 5-1 to allow a
 4    comparison with the critical study. Not all studies of neurotoxic effects identified in Table 5-2
 5    are presented in Figure 5-1; however, these studies are a sample of human and animal data sets
 6    for some of the more sensitive measures of neurotoxic endpoints. Vision or visual function
 7    effects are observed in the human and rodent studies and one study in gerbils reports changes in
 8    brain chemistry (Schreiber et al., 2002; Altmann et al., 1995; Cavalerri  et al., 1994; Seeber,
 9    1989; Mattsson et al., 1998;  Rosengren et al.,  1986). Effect magnitudes could be identified for
10    three studies and ranged from a 5% change in color vision index to roughly a 15% decrement in
11    several tests on a neurobehavioral evaluation battery (summarized in Figure 5-1).  In the absence
12    of tetrachloroethylene data to inform uncertainty factors, the analysis uses the default values as
13    discussed in Section 5.1.3:  a factor of 10 to extrapolate from a LOAEL to a NOAEL; a factor of
14    10 for human variation; and a factor of 3 for database deficiencies. For the rodent studies of
15    Mattson et al. (1998) and Rosengren et al.  (1986), PODs represent human equivalent
16    concentrations for a category 3 gas adopting EPA's RfC methodology (U.S. EPA, 1994) and an
17    uncertainty factor of 3 addresses uncertainties associated with extrapolating from animal data to
18    an average human.  Three studies are of subchronic exposure duration,  and extrapolation to
19    chronic exposure duration is achieved using a factor of 10 for the studies  of Mattsson et al.
20    (1998) and Schreiber et al. (2002; daycare employees).  A subchronic to chronic factor of 3,
21    rather than 10, was applied for Rosengren  et al. (1986) in light of the large overall uncertainty for
22    this study associated with extrapolating from a LOAEL to NOAEL, from animal to humans, for
23    human variation, and for database deficiencies; the total uncertainty factor was 3,000.
24          One study in Figure 5-1 used a neurological test for visual contrast sensitivity and  yields
25    LOAELs (PODs) of 0.1 and 0.4 ppm for day care workers or apartment dwellers adjacent to dry
26    cleaners, respectively (Schreiber et al., 2002).  An RfV developed from this study  would be
27    lower than that of Altmann et al. (1995) but, as more fully discussed in Section 5.1.1, this  study
28    was not  considered as the critical study due to its design as a pilot for a larger study conducted
29    by the NYS DOH (see Sections 5.2.1 and 5.2.2). LOAELs are higher than that of Altmann et al.
30    (1995) for the occupational studies of Cavalleri et al. (1994) and Seeber (1989) and reflect higher
31    tetrachloroethylene concentrations in the occupational setting when compared to residential
32    concentrations.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-15        DRAFT-DO NOT CITE OR QUOTE

-------
                               Inhalation Neurotoxicity RfVs
              Human     Human     Human     Human     Human
                                                                           Gerbil
                                                                                      -- 100
Q.
c
o
O
   0.001
  0.0001
 0.00001
   0.01
c
01
u
c
o
O
                                                                                      -- 0.01
                                                                                      -- 0.001
KEY
  +   Point of departure
      (POD)
I    I  UF, extrapolation from a
      lowest-observed-
      adverse-effect level
      (LOAEL) to a no-
      observed-adverse-effect
      level (NOAEL)
      UF, animal-to-human
      extrapolation
      UF, human variation
      UF, extrapolation from
      subchronic exposure
      duration to chronic
      exposure duration
 H  UF, database
      deficiencies
  O  Reference value (RfV)
                Visual      Visual   Vision, cognitive Vision, 5% Visual function,   Visual       Brain
               contrast     contrast     function,    change in  ~10% change  function     chemistry
               (day care   (apartment  vigilance," ~15%   color      in neuro-    (Mattsson    (Rosengren
              employees)   residents)   decrement   confusion   behavioral   etal., 1998)   etal.,1986)
               (Schreiber   (Schreiber    (Altmann      index       tests
              etal.,2002)   etal.,2002)   etal., 1995)   (Cavalleri (Seeber, 1989)
                                            etal., 1994)
     Principal study.
   Figure 5-1.  Array of PODs and reference values for a subset of neurotoxic effects of studies in Table 5-2.

-------
 1          PODs and inhalation organ-specific RfVs are presented for selected studies in Figure 5-2
 2    to give perspective on the RfC derived from the adverse neurotoxic effects in Altmann et al.
 3    (1995) and to provide information on other systemic effects associated with tetrachloroethylene
 4    exposure. Toxicity to the liver, kidney, developing fetus, and reproductive organs are observed
 5    at higher mean or median tetrachloroethylene concentrations than Altmann et al. (1995). The
 6    POD for a given study is the tetrachloroethylene concentration associated with the LOAEL,
 7    NOAEL, or lower bound on a benchmark concentration (BMCL).  Benchmark concentration
 8    models are fit to JISA (1993) and Tinston (1994; see Appendix 5, Tables 5A-5 and 5A-7).
 9    Furthermore, a pharmacokinetic model of Bois et al. (1996) and scaling of the body weight to the
10    3/4th power is adopted for liver weight changes in JISA (1993) to obtain human equivalent
11    concentrations, a treatment consistent with the cancer dose-response analysis of liver tumors and
12    more fully described in Section 5.4.3.1. PODs developed from pharmacokinetic models of Rao
13    and Brown (1993) and Reitz et al. (1996), reflecting uncertainties associated with
14    tetrachloroethylene metabolism, are up to an order of magnitude higher than that of Bois et al.
15    (1996). The POD for JISA (1993) using the Bois et al. (1996) pharmacokinetic model is shown
16    on Figure 5-2.
17          Organ-specific inhalation RfVs for endpoints besides neurotoxicity are developed by a
18    procedure similar to that for neurotoxicity and was carried out for Altmann et al. (1995). Default
19    values in the absence of data-informed adjustment factors are adopted to account for
20    uncertainties in the analysis. With human studies, the default values as discussed in
21    Section 5.1.3 are a factor of 10 to extrapolate from a LOAEL in Altmann et al. (1995) and Mutti
22    et al. (1992) to a NOAEL;  a factor of 10 for human variation; and a factor of 3 for database
23    deficiencies. Another uncertainty factor is adopted to account for uncertainty in extrapolating
24    laboratory animal data to the case of average healthy humans (U.S. EPA, 1994). Typically, this
25    factor addresses residual uncertainties not associated with default dosimetric adjustment like the
26    human equivalent concentration. The POD for Tinston (1994) is a human equivalent
27    concentration for a category 3 gas and an uncertainty factor of 3 addresses uncertainty associated
28    with extrapolating animal data to the average healthy human case.  In the case of liver
29    angiectasis, cross-species scaling of total rate of metabolism using body weight to the 3/4th
30    power was used for describing toxicological equivalence because of the extensive rationale
31    supporting it (U.S. EPA, 1992). The methodology achieves a human equivalent concentration
32    that is expected to approximate AUC or ppm equivalence across species for a category 3 gas. An
33    animal to human uncertainty factor of 3 addresses non-pharmacokinetic uncertainties such as
34    pharmacodynamics as suggested in the RfC  framework (U.S. EPA, 1994).
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-17       DRAFT-DO NOT CITE OR QUOTE

-------
                     Inhalation Organ-Specific RfVs
    100
                      Human
                                     Human
                                                                     Mouse
Q.
a.
c
o
«
I
o
c
o
o
   0.001
  0.0001
    0.01
                                                                                KEY
                                                                           Point of departure
                                                                           (POD)
                                                                           UF, extrapolation from a
                                                                           lowest-observed-
                                                                           adverse-effect level
                                                                           (LOAEL) to a no-
                                                                           observed-adverse-effect
                                                                           level (NOAEL)
                                                                           UF, animal-to-human
                                                                           extrapolation
                                                                           UF, human variation
                                                                           UF, database
                                                                           deficiencies
                                                                           Reference value (RfV)
                   CMS, -15%
                 decrement in verbal
                 function, cognitive
                 function, vigilance,
                   and vision3
                  (Altmann et al.,
                      1995)
Kidney, increase in   Developmental/
  mean urinary    Reproductive, 1%
 concentration of
  several renal
   enzymes
    (Multti
  etal., 1992)
increase in fetal
death, BMCLoi
   as POD
   (Tinston
 etal., 1994)
  Liver, 10%
 increase in liver
  angiectasis,
BMCLio/p as POD
  (JISA, 1993)
 aPrincipal study.
  Figure 5-2. Organ-specific reference values for inhalation exposure to tetrachloroethylene.

-------
 1    5.1.5. Previous Inhalation Assessment
 2          There is no previous EPA RfC assessment for tetrachloroethylene with which to compare
 3    and contrast the RfC developed in this assessment.  Other assessments identified in Appendix A
 4    have derived a noncancer reference value from the human evidence.  California's drinking water
 5    assessment on tetrachloroethylene (Cal EPA, 2001) derived a public health goal (PHG)6 for
 6    noncancer effects from a geometric mean in Altmann et al. (1995) and two occupational studies:
 7    Spinatonda et al. (1997) and Ferroni et al. (1992). The most recent assessment of
 8    tetrachloroethylene by the NYS DOH (1997) used a slightly different set of neurotoxicity  studies
 9    than did California and presented reference criteria for neurotoxicity  derived from Cavalleri et al.
10    (1994), Altmann et al. (1995), and  Seeber (1989). NYS DOH (1997) considered the reference
11    criterion of Seeber (1989) as best providing a sufficient margin of exposure over the air levels of
12    tetrachloroethylene associated with CNS effects.
13          ATSDR (1997), on the other hand, based its chronic MRL on Ferroni et al. (1992).
14    ATSDR (1997) considered Altmann et al. (1995) to provide a NOAEL, a conclusion inconsistent
15    with the assessments by New York State and California.  ATSDR, however, noted that the
16    Altmann et al. (1995) study suggested a need to characterize neurotoxic  effects in populations
17    exposed to very low levels of tetrachloroethylene. A second report (Schreiber et al., 2002) of
18    visual functional deficits in two populations exposed to tetrachloroethylene at lower ambient
19    concentrations  that were similar to those of Altmann et al. (1995) has become available since the
20    publication of the ATSDR toxicological profile.
21          A difference between this and previous assessments is in the previous assessments'
22    treatment of human variation, particularly the residential study by Altmann et al. (1995). A
23    choice other than the UF of 10 has been adopted in the assessments by California and New York
24    State. A presumption underpinning this choice is that the residential  population studied by
25    Altmann et al. (1995) is more reflective of the general adult population than of an occupational
26    population and any accompanying  selection bias that is often associated  with a healthier worker
27    population. Although there is some merit in this opinion, observations in Altmann et al. (1995)
28    are of a German population: 14 adults with exposure to tetrachloroethylene. These individuals
29    likely do not represent the full range of human variation found in a large and ethnically  diverse
30    population such as the United States population.  Furthermore, as noted,  the Altmann et al.
31    (1995) study excluded subjects with disorders such as hypertension, neurological or
32    endocrinological diseases (e.g., diabetes), impaired vision,  or impairment of joints; hence, these
33    subjects can be considered as having an overall good health status, and individuals whose
             PHG is conceptually similar to an RfD.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-19        DRAFT-DO NOT CITE OR QUOTE

-------
 1    diseases may have increased their susceptibility to tetrachloroethylene effects were not included
 2    in this study.
 O
 4    5.2. ORAL REFERENCE DOSE (RfD)
 5           Ideally, the studies of greatest duration of exposure and conducted via the oral route of
 6    exposure have the most confidence for derivation of an RfD.7 An earlier assessment of
 7    tetrachloroethylene oral noncancer toxicity by EPA, for example, identified liver toxicity in
 8    Buben and O'Flaherty (1985) as the critical effect for developing an RfD (U.S. EPA, 1988).
 9    However, the application of pharmacokinetic models for a route-to-route extrapolation of the
10    inhalation studies expands the oral database.  Cal EPA (2001), for example, carried out a route-
11    to-route extrapolation of the human inhalation studies of neurotoxic effects to develop a PHG for
12    oral tetrachloroethylene exposure, based on a route-to-route extrapolation of inhalation
13    neurotoxicity studies.
14
15    5.2.1. Choice of Principal Study and Critical Effects
16           Toxicity to several targets, including the liver, kidney, nervous system, and reproductive
17    system and to the developing fetus is seen in rodents with oral tetrachloroethylene exposure.
18    Effects have been observed at these targets in acute studies (28 days or less),  longer
19    term/sub chronic studies (90 days), or chronic studies (1 year or more). At higher doses (above
20    approximately 1,000 mg/kg-day), targets of oral tetrachloroethylene toxicity include the liver,
21    kidney, nervous system, lymphatic system, reproductive system, and developing fetus (ATSDR,
22    1997).  There are few studies at lower doses  as compared to the number of studies of inhalation
23    exposure. Several targets of toxicity from oral exposure are similar to targets observed with
24    inhalation exposures, i.e., liver and kidney.
25           No epidemiologic studies of oral exposure were suitable for quantitative analysis,
26    although these studies did provide information for hazard identification.  Four studies of
27    subchronic oral exposure in mice or rats (Buben and O'Flaherty, 1985; Jonker et al., 1996;
28    Berman et al., 1995;  Hayes et al.,  1986) are available, as is a developmental study in mice of oral
29    exposure to tetrachloroethylene (Fredriksson et al., 1993).  As discussed  above, EPA previously
30    developed an RfD from Buben and O'Flaherty (1985). A significant effect on liver weight was
31    seen in this study.
              The RfD is expressed in units of milligrams per kilogram body weight per day (mg/kg-day). In general,
      the RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
      population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious noncancer
      effects during a lifetime.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-20        DRAFT-DO NOT CITE OR QUOTE

-------
 1          The CNS is a sensitive target for tetrachloroethylene inhalation toxicity, as discussed in
 2    Section 5.1.1. This assessment has attempted to expand the database for derivation of an RfD
 3    using relevant inhalation data and route-to-route extrapolation with the aid of a PBPK model (see
 4    Section 3.5) to the POD of Altmann et al. (1995). The nervous system is an expected target with
 5    lower oral tetrachloroethylene exposures, in view of the fact that other organ systems such as the
 6    liver the kidney  are also common targets associated with both inhalation and either oral routes of
 7    subchronic or chronic exposure.  The similarity of effects in these organ systems with either oral
 8    or inhalation exposure to tetrachloroethylene supports the use of route extrapolation to compare
 9    PODs for oral and inhalation exposure. For these reasons, the inhalation study in humans by
10    Altmann et al. (1995) is chosen as the principal study for supporting the RfD.
11
12    5.2.2.  Methods of Analysis, Including Models
13          The present analysis defines a POD using the traditional NOAEL/LOAEL approach in
14    addition to using BMD modeling where feasible. This assessment has attempted to expand the
15    database for derivation of an RfD using relevant inhalation data and route-to-route extrapolation
16    with the aid of a PBPK model  (see Section 3.5). Several factors support the use of route-to-route
17    extrapolation for tetrachloroethylene. Tetrachloroethylene has been shown to be rapidly and
18    well absorbed by both the oral and inhalation routes of exposure (ATSDR, 1997).  Additionally,
19    the metabolic pathways and kinetics of excretion with oral exposure are similar to those of
20    inhalation exposure (ATSDR,  1997). Furthermore, the data for oral administration indicate a
21    pattern of effects similar to that of inhalation exposure, including effects on the liver and  kidney.
22    PBPK modeling was also used with suitable studies in animals in order to inform the process of
23    extrapolating to  HECs.  The use of these alternative approaches has the potential to add
24    information to the NOAEL/LOAEL approach.
25          PBPK modeling was used to derive the oral dose that would result in  the same
26    tetrachloroethylene in blood AUC as that following a continuous inhalation exposure of 0.7 ppm,
27    the LOAEL from the inhalation study by Altmann et al. (1995). A hypothetical drinking  water
28    scenario of 9 equal drinking water incidents, spaced 2 hrs apart allowing for an 8-hr sleep period,
29    was judged to be a reasonable baseline. Use of this scenario generated an estimated total oral
30    ingestion of 1.1  mg/kg-day of tetrachloroethylene, leading to the same steady-state blood
31    tetrachloroethylene AUC as a continuous inhalation exposure of 0.7 ppm using the PBPK model
32    of Rao and Brown (1993).
33          A route-to-route extrapolation based on a venous blood dose metric is more robust than
34    one based on another dose metric such as the amount of metabolized tetrachloroethylene, and it
35    provides a strong rationale for using blood AUC as a dose metric for extrapolating between

                 This  document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-21        DRAFT-DO NOT CITE OR QUOTE

-------
 1    exposure routes. Venous blood concentration is well-validated in the Rao and Brown (1993)
 2    model; hence, little model uncertainty is associated with its estimation (see the pharmacokinetic
 3    discussion in Chapter 3). Furthermore, as noted in Section 5.1.1, the use of blood
 4    tetrachloroethylene provides some attempt to account for breathing rates and to adjust for kinetic
 5    nonlinearities related to tetrachloroethylene absorption, and it is assumed to better reflect
 6    tetrachloroethylene pharmacokinetics than use of default methodologies. The chemical species
 7    responsible for tetrachloroethylene-induced neurotoxic effects has not been demonstrated, but
 8    blood tetrachloroethylene is presumed to be one step in the MOA pathway and is used as a
 9    marker for the dose metric associated with neurologic  effects.
10          The route-to-route extrapolation starts with the estimation of the average venous blood
11    tetrachloroethylene AUC resulting from continuous inhalation exposure at the LOAEL of 0.7
12    ppm. The venous blood tetrachloroethylene AUC at steady state resulting from continuous
13    exposure to 0.7 ppm tetrachloroethylene is estimated to be 68.3 mg/min/L, according to the Rao
14    and Brown (1993) model.  This model does not address pharmacokinetic variation in the  human
15    population.  An analogous curve corresponding to a drinking water scenario is provided in
16    Figures 5-3 and 3-10.  The drinking water scenario models a  subject consuming water every
17    2 hrs except during sleep, which was assumed to be for 8 hrs. An assumption of the amount of
18    water consumed is also not necessary, because blood concentrations of tetrachloroethylene are
19    solely dependent on the amount of compound ingested during each drinking episode. The curve
20    depicted in Figures 5-3 and 3-10 yields the same AUC as does continuous inhalation exposure to
21    0.7 ppm and corresponds to ingestion of 76 mg/day. Therefore, the extrapolation of an
22    inhalation exposure of 0.7 ppm (4.8 mg/m3) using the PBPK  model of Rao and Brown (1993)
23    yields the same blood concentration of tetrachloroethylene as does ingestion of 1.1 mg/kg-day.
24          Table 5-3 summarizes the results of animal studies of oral exposures that represent the
25    lower end of the dose-response curve. The doses shown in Table 5-3 are expressed in human
26    equivalent terms—using mg/kg3/4-day scaling—to enable interspecies comparisons (U.S. EPA,
27    1992). For liver effects, an alternative procedure of using the pharmacokinetic model of total
28    metabolism as the dose metric for extrapolating between species was carried out. Potential
29    PODs are presented as either a NOAEL or a modeled LEDX when the study results were suitable
30    for modeling.  Among the four studies identified in Table 5-2, a significant effect on liver weight
31    is seen in the study by Buben and O'Flaherty (1985), with a NOAEL at a duration-adjusted
32    human equivalent dose of 2 mg/kg-day.  The modeled human equivalent BMDLs was 5 mg/kg-
33    day in terms of administered exposure and 3.4-32 mg/kg/day using the available
34    pharmacokinetic models. As discussed above, EPA previously developed an RfD from Buben
35    and O'Flaherty (1985).

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-22       DRAFT-DO NOT CITE OR  QUOTE

-------
     0.10 n
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
                                        10             15

                                        Time (hours)
                                                                            20
       Figure 5-3. Time course of venous blood concentration in humans as
       predicted by the Bois et al. (1996), Rao and Brown (1993), and Reitz et al.
       (1996) PBPK models for ingested tetrachloroethylene.  A total of 76 mg of
       tetrachloroethylene was orally delivered via drinking water in nine bolus doses
       spaced 2 hrs apart for a duration of 18 hrs, followed by 8 hrs of no dosing. The
       dashed line indicates the steady-state blood concentration level due to inhaled
       tetrachloroethylene of 0.7 ppm exposure concentration that results in the same
       area under the curve as above the curve, integrated over a 24-hr period. The
       alveolar ventilation rate was 9.3 L/min (total inspiratory rate 13.9 L/min) and the
       ventilation-to-perfusion ratio was equal to 1.3. The three models result in nearly
       equal concentrations at this exposure concentration.
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   5-23        DRAFT-DO NOT CITE OR QUOTE

-------
 1    5.2.3. Reference Dose (RfD) Derivation, Including Application of Uncertainty Factors
 2          The POD of 1.1 mg/kg-day obtained from route-to-route extrapolation from Altmann et
 3    al. (1995) is similar to the POD of 2 mg/kg-day observed for liver toxicity in mice orally exposed
 4    to tetrachloroethylene (see Table 5-4). Several studies have identified similar PODs for adverse
 5    liver effects, indicating some degree of confidence in Altmann et al. (1995) as the basis for the
 6    POD and for calculating an RfD.  Studies of inhalation exposure show that humans respond to
 7    tetrachloroethylene's neurotoxic effects at lower concentrations than do rodents, indicating
 8    humans as a potentially more sensitive species for these effects. Humans are also expected to be
 9    sensitive to neurotoxicity from oral tetrachloroethylene exposure, and there are no data to
10    indicate otherwise. Hence, the LOAEL of 1.1 mg/kg-day derived from Altmann et al. (1995)
11    represents the POD for the RfD for oral exposure to tetrachloroethylene.
12          To address differences between study conditions and conditions of lifetime human
13    environmental exposure, the POD is reduced by UFs that consider specific areas of uncertainty.
14    The following areas of uncertainty were evaluated for this RfD:
15          1.  Human variation. Because the critical study is the same for this RfD derivation as for
16              the RfC, the 10-fold factor applied to the POD for the RfC is used as a default in this
17              assessment for human variation.
18
19          2.  Animal-to-human uncertainty. Because the critical study is in humans, this factor is
20              not needed. For animal data, a factor of 3 would be used for extrapolation of
21              pharmacodynamics, because body weight scaling was used for calculating the human
22              equivalent doses (see Table 5-4).
23
24          3.  Subchronic-to-chronic uncertainty. As with the RfC derivation, described in Section
25              5.1.3, the POD is based on a study of 10.6-year median duration of exposure, or 15%
26              of a 70-year lifespan. There are no data to suggest that continuing exposure to
27              tetrachloroethylene can increase the severity of effects; exposure-response trends are
28              not evident in the human studies, and there is no systematic trend in the relation
29              between the LOAEL and treatment duration for neurotoxicity in animals. A factor to
30              address the potential for more  severe toxicity from lifetime exposure to
31              tetrachloroethylene is not used in this assessment.
32
            4.  LOAEL-to-NOAEL uncertainty.  LOAELs are observed in this human study,  and, as
                in the RfC derivation, a 10-fold factor is adopted in this assessment to approach the
                range where a negligible response could be expected.
33
34          5.  Database uncertainty.  A factor of 3 is used for the database uncertainties for the
35              same reasons as those for the derivation of the inhalation RfC.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-24        DRAFT-DO NOT CITE OR QUOTE

-------
        Table 5-3.  Oral studies considered in analysis of the oral RfD
Organ/
system
Liver
Liver,
kidney
Liver
Central
nervous
system
Whole
animal
Study
Buben and
O'Flaherty (1985)
Jonker et al.
(1996)
Herman et al.
(1995)
Fredriksson et al.
(1993)
Hayes et al.
(1986)
Species
Mouse
(40 g)
Rat
Rat
Mouse
Rat
Duration/exposure
route
Subchronic (6
weeks)/oral gavage
Subchronic (4
weeks)/oral gavage
14 days/oral gavage
Post-natal days 10-
16/oral, gavage
Subchronic (91
days)/oral drinking
water
Dose/exposure
(NOAEL/LOAEU)"
(mg/kg-day)
0, 20, 100. 200, 500,
1,000, 1,500, 2,000
0, 600. 2,400
0, 50, 150, 500, 1.500.
5,000
0, 5, 320
0, 0 (vehicle control),
14, 400, 1,400
Effect
Liver weight,
triglycerides
Liver weight, enzyme
levels; kidney weight,
kidney enzyme levels
Liver weight, ALT
Day 60: increased
locomotion, decreased
rearing
Body weightd
Human equivalent doses b
(mg/kg-day)
NOAEL/
LOAEL
NOAEL: 2
LOAEL: 133
NOAEL: 77
LOAEL: 0.5C
NOAEL: 4
BMDLC
BMDLS: 5.0
BMDLS/P: 3.4-32
NA
NA
NA
NA
a NOAELs are underlined once, LOAELs are double-underlined.
b Human equivalent doses calculated using RfD methodology (Barnes and Dourson, 1988) and scaled to the ratio of body weight to the 0.75 power, i.e.,
  multiplied by [(animal body weight in kg)/human body weight (70kg)]025.  Also adjusted for daily exposure by multiplying by (5 days)/(7 days) where
  relevant.
 0 BMDL is the lower bound on dose associated with a one standard deviation change in the mean over the control response. This corresponds to an excess risk
  of approximately 10% for developing a level of the endpoint above the 98th percentile (or below the 2nd percentile) of the control distribution for normally
  distributed effects.  BMDLS/P is the BMDLS using pharmacokinetic modeling for relating metabolites in the experimental animals to the responses, and for
  reflecting extrapolation to humans (see Appendix 5A for details).
 0 For Fredriksson et al. (1993), a regression was fit to the data with body weight estimated as 8g, the body weight at day 13, the midpoint of 10-16 days. The
  human equivalent NOAEL  is 5  x (0.008/70)°25 = 0.5  mg/kg-day.
 d Female rats (LOAEL was 400 mg/kg-day). Authors describe lower body weight gain as significant; however, statistical testing is not presented in the
  published paper (only that there was a statistical difference \p < 0.05] between treatment and control groups).

-------
        Table 5-4. Oral RfV:  point of departure and uncertainty factors


Study
Altmann et al.
(1995): humans
Buben and
O'Flaherty (1985):
mice, increased
liver/body weight
Fredriksson et al.
(1993)
Hayes etal. (1986)


Human
equivalent
concentration"
0.7 ppm
(LOAEL)
NA


NA

NA
Oral
human
equivalent
dose
(mg/kg-day)
l.lb
3.4-32c


0.5

4


Human
variation
10
10


10

10


Animal
to
human
1
3C


3

3


Subchronic
to
chronic
1
10


1

10


LOAEL
to
NOAEL
10
1


10

1


Database
o
J
o
J


3

o
J


Composite
uncertainty
factor
300
1000


1000

1000


RfV
(mg/kg-day)
4 x 10'3
3 x 1(T3 to
3 x 10'2

5 x 1Q-4

4 x lO'3
a Ambient concentration is assumed to represent continuous exposure in the residential studies.
b Equivalent oral exposure from application of PBPK model of Rao and Brown (1993), on the basis of equivalent AUC of blood tetrachloroethylene for humans.
0 See Table 5A-9 for the dose-response modeling summary and extrapolation to human equivalent exposure. These human equivalent doses lie in the nonlinear
  portion of the exposure versus rate of metabolism relationship.  However, for the purpose of comparing alternate reference doses, the calculation presented
  here is equivalent to applying the uncertainty factors prior to conversion from human equivalent metabolized dose to human equivalent administered dose.

-------
 1          A total UF of 300 was applied to this effect level:  10 for human variation, 10 for
 2    consideration of LOAEL to NOAEL, and 3 for uncertainties in the database.
 O
 4          RfD  =NOAELIUF
 5               = 1.1 mg/kg-day / 300
 6               = 0.0037 mg/kg-day (rounded to 0.004 mg/kg-day)
 7
 8    where:
 9          1.1 mg/kg-day   = the oral exposure POD equivalent of 0.7 ppm (4.8 mg/m3) continuous
10                           inhalation exposure LOAEL for neurologic effects in residents
11                           exposed tetrachloroethylene (Altmann et al., 1995) estimated via
12                           PBPK modeling
13
14          300            = composite UF chosen to account for extrapolation from a LOAEL to a
15                           NOAEL in humans, intra-individual variability in humans, and
16                           uncertainties in the database
17
18          In summary, an RfD for tetrachloroethylene was developed through a route-to-route
19    extrapolation from the POD in Altmann et al. (1995), which reports neurological toxicity in
20    residents exposed to tetrachloroethylene.  The oral exposure POD equivalent to the 0.7 ppm (4.8
21    mg/m3) continuous inhalation exposure LOAEL was estimated via PBPK modeling to be 1.1
22    mg/kg-day.  A composite UF of 300 was obtained by multiplying  factors of 10 for average to
23    sensitive human variation, 10 for starting from an effect level instead of a NOAEL,  and 3 for
24    database uncertainties. Dividing the POD by a composite UF of 300 yields an RfD  of 4 xlO"3
25    mg/kg-day.  This RfD is equivalent to a drinking water concentration of 0.12 mg/L, assuming a
26    body weight of 70 kg and a daily water consumption of 2 L.
27
28    5.2.4.  Supporting Studies
29          PODs and oral RfVs from studies in Table 5-3 are arrayed in Figure 5-4 to give a
30    perspective on the oral RfD supported by Altmann et al. (1995). Effects include liver weight
31    changes (POD [BMDLs/p] of 3.4 mg/kg-day from Buben and O'Flaherty [1985]), body weight
32    changes (POD [NOAEL] of 4 mg/kg-day from Hayes et al. [1986]), and developmental
33    neurotoxicity (POD [LOAEL] of 0.5 mg/kg-day from Fredriksson et al. [1993]). These PODs
34    span a range which includes the oral-equivalent POD (LOAEL) of 1.1 mg/kg-day of Altmann et
35    al. (1995), converted using a  pharmacokinetic model for route-to-route extrapolation (Rao and
36    Brown [1993]). The POD from Buben and O'Flaherty (1985) is a standard deviation change in
37    mean liver-to-body weight over control using benchmark dose models and  using
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-27       DRAFT-DO NOT CITE OR QUOTE

-------
                           Oral Organ-Specific RfVs
    100
                    Human
                                   Mouse
                                                  Mouse
                                                                  Rat
                                                                                         KEY
                                                                                           +   Point of departure
                                                                                               (POD)
                                                                                         I    I UF, extrapolation from a
                                                                                               lowest-observed-
                                                                                               adverse-effect level
                                                                                               (LOAEL) to a no-
                                                                                               observed-adverse-effect
                                                                                               level (NOAEL)
                                                                                               UF, animal-to-human
                                                                                               extrapolation

                                                                                               UF, human variation
                                                                                         I    I UF, extrapolation from
                                                                                               subchronic exposure
                                                                                               duration to chronic
                                                                                               exposure duration
                                                                                         ^H UF, database
                                                                                               deficiencies

                                                                                           O  Reference value (RfV)
 0.0001
aPrincipal study.
CMS, -15% decrement     Neuro-
  in verbal function,     developmental
  cognitive function,     (Fredriksson
 vigilance, and vision3    et al., 1993)
 (Altmann et al., 1995)
                                                Liver, one S.D.
                                              change in the mean
                                              liver weight over the
                                               control response
                                                 (Buben and
                                               O'Flaherty, 1985)
 Systemic
  toxicity
(Hayes et al.,
   1986)
  Figure 5-4. Oral organ-specific reference values for exposure to tetrachloroethylene.

-------
 1    pharmacokinetic models for relating metabolites in the experimental animals to the responses,
 2    and for reflecting extrapolation to humans (see Appendix 5A and Table 5A-8 for details).
 3    Default uncertainty factor values in the absence of data-informed adjustment factors are adopted
 4    to account for uncertainties in the analysis (see Table 5-3).  A factor of 10 is used to extrapolate
 5    from a LOAEL in Altmann et al. (1995), and Fredriksson et al. (1993) to a NOAEL because an
 6    effect occurred at the lowest dose studied; this factor is 1.0 for the Buben and O'Flaherty (1985)
 7    and Hayes et al. (1986) studies because there was no effect observed at the lowest dose studied.
 8    In all studies a factor of 10 is used for human variation and a factor of 3 for database
 9    deficiencies. For rodent studies, an uncertainty factor of 3 is adopted to account for uncertainty
10    in extrapolating laboratory animal data to the case of average healthy humans (U.S. EPA, 1994).
11    Typically, this factor addresses residual uncertainties not associated with default dosimetric
12    adjustment like the human equivalent concentration. Scaling experimental doses in the rodent
13    studies to the 3/4th power achieves a human equivalent dose that is considered lexicologically
14    equivalent. For Buben and O'Flaherty (1985), specifically, body weight scaling of the
15    pharmacokinetically derived total rate of metabolism similarly produces a human equivalent
16    dose that is assumed to be lexicologically equivalent. The uncertainty factor of 3 is used to
17    account for residual uncertainty such as pharmacodynamic processes.  For studies of subchronic
18    exposure duration, a factor of 10 accounts for uncertainty associated with extrapolating to
19    chronic exposure duration.
20          RfVs in the oral rodent studies of weight changes in Buben and O'Flaherty (1985) and
21    Hayes et al. (1986) were 3 x 10"3 and 4 x  10"3 mg/kg-day, respectively. The RfD of Altmann et
22    al. (1995) at the higher end of the range and with a total uncertainty factor (300) less than total
23    uncertainty factor of 1000 was applied to the two rodent studies. A developmental neurotoxicity
24    study in animals with persistent effects on motor activity yielded a NOAEL of 5 mg/kg-day and
25    a RfV of 5 x 10"4 mg/kg-day (Fredriksson et al., 1993).  However, this study was not considered
26    as the principal study for chronic exposure due to uncertainties associated with study design and
27    its level of confidence (see Section 4.6.2.2).
28
29    5.2.5. Previous Oral Assessment
30          EPA previously suggested an RfD of 1  x 10'2 mg/kg-day (U.S. EPA, 1988), which was
31    supported by an adjusted NOAEL of 14 mg/kg-day in Buben and O'Flaherty (1985), and a
32    composite UF of 1,000 (10 for extrapolation from  the rat to humans, 10 for human variation,  and
33    10 for extrapolating to chronic exposure conditions). A human study is now available, using
34    route-to-route extrapolation, and is preferred to animal data.  The composite UF of 300 in the
35    current analysis is smaller than that used in the previous analysis, reflecting fewer uncertainties

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-29        DRAFT-DO NOT CITE OR QUOTE

-------
 1    associated with using human data. More recently, Cal EPA (2001) developed a PHG for oral
 2    exposure to tetrachloroethylene from the studies by Altmann et al. (1995), Spinatonda et al.
 3    (1997), and Ferroni et al. (1992) and conversion factors for breathing and absorption rates for an
 4    extrapolation from the inhalation to the oral exposure route. The PHG, calculated by taking the
 5    geometric mean of these three studies, of 0.032 mg/kg-day is higher than the current RfD (U.S.
 6    EPA, 1987) and the RfD developed in this assessment.  On the other hand, ATSDR (1997) did
 7    not develop subchronic- and chronic-duration oral MRLs, although an acute MRL was
 8    developed from the study by Fredriksson et al. (1993).  ATSDR (1997) noted neurological
 9    effects as the principal effect of tetrachloroethylene in humans and the scarcity of data in animals
10    from subchronic and chronic studies on this endpoint.
11          One difference between this assessment and the health assessment from California (Cal
12    EPA, 2001) are  choices  of studies and UFs. The Schreiber et al. (2002) and NYSDOH (2005)
13    studies, which support the findings of Altmann et al.  (1995) of neurotoxic effects in residentially
14    exposed populations, were not available at the time of the California assessment. Another
15    difference between this assessment and previous assessments is their treatment of human
16    variation. A choice other than the default of 10 was adopted in the California assessment, based
17    on a presumption that the residential population studied by Altmann et al. (1995) is more
18    reflective of the general adult population than of an occupational population and any
19    accompanying selection bias that is often associated with a healthier worker population.
20          Although there is some merit in this opinion,  the observations in Altmann et al. (1995)
21    are of a German population of 14 adults with exposure to tetrachloroethylene.  These individuals
22    do not likely represent the full range of human variation found in a large and ethnically diverse
23    population such as the UNITED STATES population. The study excluded subjects with
24    disorders such as hypertension, neurological or endocrinological diseases (e.g., diabetes),
25    impaired vision, or impairment of joints; therefore, the subjects in Altmann et al. (1995) likely
26    had overall good health  status. Individuals with diseases that may have increased their
27    susceptibility to tetrachloroethylene  effects were not  included in the study.
28
29    5.3. UNCERTAINTIES IN INHALATION REFERENCE CONCENTRATION (RFC)
30         AND ORAL REFERENCE DOSE (RfD)
31          Risk assessments need to portray associated uncertainty. The following discussion
32    identifies uncertainties associated with the RfC or RfD for tetrachloroethylene.  As presented
33    earlier in this chapter (see Sections 5.1.2, 5.1.3, 5.2.2, and 5.2.3), the uncertainty factor approach,
34    following EPA practices and RfC and RfD guidance  (U.S. EPA, 1993, 1994), was applied to a
35    POD, a LOAEL from an epidemiologic study of neurobehavioral effects. Factors accounting for
36    uncertainties associated with a number of steps in the analyses were adopted to account for
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-30       DRAFT-DO NOT CITE OR QUOTE

-------
 1    extrapolating the POD, the starting point in the analysis, to a no-adverse-effect concentration or
 2    dose (LOAEL to NOAEL) given insufficient data in the principal study for benchmark dose
 3    modeling, to a diverse population of varying susceptibilities, and to account for database
 4    deficiencies. These extrapolations are carried out with default approaches instead of factors
 5    derived from data on tetrachloroethylene given the paucity of experimental tetrachloroethylene
 6    data to inform individual steps.  As further explained below, limited information is available on
 7    human variation in blood tetrachloroethylene concentration and can provide qualitative
 8    information on uncertainties associated with human variation.  Evaluation of a
 9    tetrachloroethylene exposure dose or concentration likely to be without an appreciable risk of
10    chronic adverse health effects over a lifetime and associated uncertainties relies on chemical -
11    specific data to describe dose-response curves, on the breadth of the database for evaluating
12    toxicity in a number of organs, and on characteristics of these data.
13           A broad range of animal toxicology and human epidemiologic data is available for the
14    hazard assessment of tetrachloroethylene, as described throughout the previous section
15    (Chapter 4). Included in these studies are short-term and long-term bioassays in rats and mice
16    (see Table 4-2, Chapter 4); neurotoxicology studies in humans, rats, mice, and gerbils (see
17    Tables 4-4, 4-5, 4-6,  and 4-7); prenatal developmental toxicity studies in rats, mice, and rabbits
18    and a two-generation reproduction study in rats (see Table 4-10); and numerous supporting
19    genotoxicity and metabolism studies.  Toxicity associated with inhalation exposure to
20    tetrachloroethylene is observed in the liver, kidney, central nervous system, reproductive organs,
21    and the developing fetus (see Chapter 4, Table 5-1, and Figure 5-2). Liver, kidney, and
22    neurodevelopmental  effects are observed with oral exposure (see Chapter 4, Table 5-2, and
23    Figure 5-4). Nevertheless, critical data gaps have been identified and uncertainties associated
24    with data deficiencies are more fully discussed below.
25           Neurotoxicity appears to be a sensitive organ system as previously identified in more
26    limited analyses of Rao and Brown (1993) and Guth et al. (1997).  The neurotoxic effects
27    observed in a residential population (Altmann et al., 1995) are similar to those observed in
28    occupational populations exposed at higher mean tetrachloroethylene concentration (Seeber,
29    1989, Echeverria et al., 1994).  Schreiber et  al.  (2002) observed visual effects (visual contrast
30    sensitivity) among residents co-located near dry cleaning establishments; however, this study
31    was a pilot for a larger study.  The larger study (NYS DOH, 2005) has become available as a
32    final report and appears supportive of this pilot study
33
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-31        DRAFT-DO NOT CITE OR QUOTE

-------
 1    5.3.1. Point of Departure
 2          A POD based on a LOAEL or NOAEL is, in part, a reflection of the particular exposure
 3    concentration or dose at which a study was conducted. It lacks characterization of the dose-
 4    response curve and for this reason is less informative than a POD defined as a BMC or a BMD
 5    obtained from benchmark dose-response modeling. With respect to neurotoxicity of
 6    tetrachloroethylene, benchmark dose-response models are fit to five data sets (Buben and
 7    O'Flaherty, 1985; JISA, 1993; NTP,  1986; Brodkin et al.,  1995; Tinston,  1994) with sufficient
 8    information. The choice of benchmark dose model does not lead to significant uncertainty in
 9    estimating the POD since benchmark effect levels were within the range of experimental data.
10    Parameter uncertainty can be assessed through confidence intervals and probabilistic analysis.
11    Each description of parameter uncertainty assumes that the underlying model and associated
12    assumptions are valid.  Uncertainty in the animal dose-response data can be assessed through the
13    ratio of BMCs to their BMCLs.  These generally do not exceed a factor of two at the POD
14    identified in Tables 5-1 and 5-2.
15          Effects in the CNS and in other organ systems (liver, kidney, reproductive, and
16    developmental) in occupational populations and in animals are observed at higher average
17    tetrachloroethylene concentrations than the Altmann et al. (1995) residential study.  As more
18    fully discussed in Section  5.1, uncertainties in other studies of neurotoxicity and of other organ
19    systems differ from those of Altmann et al. (1995).  For both occupational and residential
20    populations, studies do not describe a NOAEL and human variation is not well characterized in
21    study subjects. Uncertainties associated with the occupational studies include the following:  (1)
22    potential for neurobehavioral effects at lower exposures and (2) exposure pattern differences
23    between occupational and residential studies with peaks characterizing occupational exposures.
24    For animal studies,  uncertainties are associated with extrapolating high concentration exposure
25    typically of subchronic duration to genetically inbred rodents to infer a concentration of
26    tetrachloroethylene that is likely  to be without an appreciable risk of adverse health effects over a
27    lifetime to a diverse human population.
28
29    5.3.2. Extrapolation from Laboratory Animal Studies to Humans
30          Extrapolating from animals to humans embodies further issues and uncertainties. First,
31    the effect and its magnitude associated with the concentration at the POD in rodents is
32    extrapolated to human response.  Pharmacokinetic models are useful to examine species
33    differences in pharmacokinetic processing.  This was possible for liver toxicity where limited
34    MO A information suggests metabolism as important to toxicity.  The ranges of BMCLs
35    presented for liver effects  (a 10-fold range of estimates of tetrachloroethylene metabolism)
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-32        DRAFT-DO NOT CITE OR QUOTE

-------
 1    demonstrates the uncertainty in tetrachloroethylene pharmacokinetic models. The discrepancies
 2    between the models and with experimental data may point to large uncertainties in the
 3    parameters used in these models. Because the accuracy of the models has been evaluated only
 4    against blood and breath concentrations of the parent compound, their reliability for predicting
 5    total metabolites is an unknown.  The use of all three of these pharmacokinetic models to provide
 6    a range of risk estimates is intended to capture some of this uncertainty.
 7
 8    5.3.3. Human Variation
 9           Heterogeneity among humans is another uncertainty associated with extrapolating doses
10    from animals to humans. Uncertainty related to human variation needs consideration, also, in
11    extrapolating dose from a subset or smaller sized population, say of one sex or a narrow range of
12    life stages typical of occupational epidemiologic studies, to a larger, more diverse population.
13           In the absence of tetrachloroethylene-specific data on human variation, a factor of 10 was
14    used to account for uncertainty associated with human variation. Human variation may be larger
15    or smaller; however, tetrachloroethylene-specific data to examine the potential magnitude of
16    over- or under-estimation are few. The pharmacokinetic model of Clewell et al. (2004) of mean
17    physiological parameters to explore age-dependent pharmacokinetic differences is suggestive of
18    a 2-fold variation in blood tetrachloroethylene levels (see Chapters 3 and 5). Bois et al. (1996),
19    revised by Chiu and Bois (2006), have examined uncertainty and variation in a
20    tetrachloroethylene pharmacokinetic model describing the amount of tetrachloroethylene
21    metabolism. This analysis suggests large uncertainty is associated with estimating the quantity
22    of tetrachloroethylene metabolism in humans.
23
24    5.3.4. Database Uncertainties
25           Critical data gaps have been identified: uncertainties associated with database
26    deficiencies on developmental, immunological,  and neurotoxic effects.  Most notably, data
27    characterizing dose-response relationships and chronic visual-spatial functional deficits and the
28    cognitive effects of tetrachloroethylene exposure under controlled laboratory conditions are
29    lacking. Several halogenated organic solvents have been linked with altered immune system
30    function in both animals and humans (e.g., toluene, TCE).  Additional data from inhalation, oral,
31    and dermal exposures, at different durations, are needed to assess the potential immunotoxicity
32    of tetrachloroethylene.  This lack of data combined with the concern that other structurally
33    related solvents have been associated with immunotoxicity contributes to uncertainty in the
34    database for tetrachloroethylene.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-33       DRAFT-DO NOT CITE OR QUOTE

-------
 1          Data from acute studies in animals (Warren et al., 1996; Umeza et al., 1997) suggest that
 2    cognitive function is affected by exposure to tetrachloroethylene. These studies do not address
 3    the exposure-response relationship for subchronic and chronic tetrachloroethylene exposures on
 4    cognitive functional deficits observed in humans (e.g., Seeber, 1989; Echeverria et al., 1994; and
 5    Altmann et al., 1995). Even more importantly, there is a lack of cognitive testing in both
 6    developmentally exposed animals and adult animals following exposures to tetrachloroethylene
 7    that are longer than acute durations of exposure.  Visual system dysfunction and processing of
 8    visual spatial information are sensitive endpoints in human studies.  The exposure-response
 9    relationship of these functional deficits could be evaluated more definitively with studies using
10    homologous methods that examine retinal and visual function in experimental animals.
11    However, there has been a limited evaluation of visual function in rodents, with the exception of
12    the evoked potential studies by Mattsson et al. (1998). These types of studies could help
13    determine whether there are both peripheral and central effects of tetrachloroethylene exposure
14    on visual perception, and they could be used as an animal model to better define the exposure-
15    response relationships.
16          Subjects in the epidemiologic studies comprise adults, and some characterization of the
17    response of children to tetrachloroethylene exposure was found in limited data for a similar
18    neurological (visual system) parameter (Schreiber et al., 2002) and in a larger number of subjects
19    (NYS DOH, 2005) using other visually based testing paradigms.  Additionally, in a postnatal
20    neurotoxicity study in mice (Fredriksson et al., 1993), persistent neurological  effects (i.e.,
21    increased locomotion and decreased rearing behavior at 60 days of age, measured 43 days after
22    exposure ceased) were observed at an oral dose of 5 ppm, with no NOAEL, although this study
23    did not conform to traditional toxicity testing guidelines (see Section 4.6.2.2). There is
24    uncertainty that if adequate, robust, dose-response data based on the most appropriate
25    neurophysiological and cognitive tests were available, the exposure eliciting an adverse response
26    (and hence the POD for the reference value)  could be lower than that established on the basis of
27    deficits in visuo-spatial  and cognitive function following tetrachloroethylene exposure in healthy
28    adults (Altmann et al., 1995).
29
30    5.4. CANCER DOSE-RESPONSE ASSESSMENT
31          The following dose-response assessment was developed following the guidelines in the
32    Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).  As discussed in Section 4.10.1,
33    there is some indication from epidemiologic  investigations that a human cancer risk is associated
34    with exposure to tetrachloroethylene.  Sufficient human data linked with exposure
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-34       DRAFT-DO NOT CITE OR QUOTE

-------
 1    characterizations from these studies have not been available, but estimating cancer risk from
 2    these studies might be feasible in the future.
 3          Further, as detailed in Section 4.10.4., the available body of MO A information is not
 4    sufficient to derive quantitative, biologically based, or toxicodynamic models for low-dose
 5    extrapolation from animal data. Moreover, current literature does not identify a nonlinear MOA
 6    for tetrachloroethylene carcinogenicity.  Therefore, consistent with the 2005 cancer guidelines, a
 7    default low-dose linear model is indicated for use with the animal data to estimate human cancer
 8    risk.
 9          There is evidence that one or more tetrachloroethylene metabolites may be involved in
10    some of the carcinogenicity associated with tetrachloroethylene exposure (see Section 4.10.4).
11    PBPK models are available to estimate total metabolism in laboratory rodents and humans from
12    inhalation and oral exposure to tetrachloroethylene. The dose-response discussion below
13    describes where the PBPK models have been used to estimate human carcinogenic risk arising
14    from tetrachloroethylene exposure through their impact on high-dose to low-dose extrapolation
15    in animals, interspecies extrapolation, and route-to-route extrapolation.
16
17    5.4.1.  Choice of Study/Data with Rationale and Justification
18          As discussed in Chapter 4, there are several chronic studies in rats and mice: an oral
19    gavage study in mice and female rats by NCI (1977) and two inhalation studies in mice and rats
20    (NTP, 1986; JISA, 1993). These studiesestablished that the administration of
21    tetrachloroethylene, either by ingestion or by inhalation to sexually mature rats and mice, results
22    in increased incidence of tumors.  In at least two studies, several tumor sites showed statistically
23    significantly increased rates with increasing tetrachloroethylene administration:  MCL in male
24    and female rats and hepatocellular adenomas and carcinomas in male and female mice. Other
25    cancer dose response assessments of tetrachloroethylene have relied on the tumor data from the
26    NCI and NTP studies (see Appendix A).
27          This analysis considers all three bioassays but focuses primarily on the JISA (1993) study
28    results. First, the JISA (1993) study included lower exposures than did the two earlier bioassays
29    for both species tested, which makes it a stronger study for deriving dose-response relationships
30    for risk assessment purposes, insofar as all other aspects of these studies can be considered
31    comparable.  For mice, the lowest exposure concentration of 10 ppm was 10-fold lower and the
32    mid-dose of 50 ppm was 2-fold lower than the lower exposure concentration in the NTP (1986)
33    inhalation study (100 ppm).  For rats, the low-exposure concentration of 50 ppm was fourfold
34    lower than in the NTP study (200 ppm).  Second, no other dose response modeling appears to be
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-35        DRAFT-DO NOT CITE OR QUOTE

-------
 1    available for the JISA (1993) study, whereas the incidence of hepatocellular tumors and MCL in
 2    the NTP (1986) study have been extensively analyzed for previous assessments.
 O
 4    5.4.2.  Dose-Response Data
 5    5.4.2.1.  Liver Tumors in Mice
 6          All three bioassays showed increases in hepatocellular tumors in male and female mice.
 7    Table 5-5 summarizes these incidence patterns. Because hepatic adenomas and carcinomas are
 8    considered part of the same continuum of tumor development, and adenomas may be
 9    differentiated from carcinomas only on the basis of size, this analysis emphasizes the combined
10    incidence of these two tumor types. Historical data from the Japan Bioassay Research Center
11    (JBRC), where the JISA (1993) study was conducted, indicate that the control liver tumor
12    incidences in this study were fairly typical for this laboratory (see Table 5-6). Specifically, the
13    incidence in controls was 28% for males and 6% for females; the averages for the laboratory
14    were 23% and 2% and the upper bounds were 42% and 8%, respectively, for carcinomas.8
15          The results of the inhalation studies are reasonably consistent when adjusted for
16    background tumor incidence (see Figures 5-5a and 5-5b). Liver tumor incidence among male
17    mice in the JISA (1993) study did not follow a clearly monotonic pattern, with a higher response
18    in the low-dose group than seems consistent with the pattern in responses in the other dose
19    groups.  Taking into account this variability in the responses, however, the dose-response
20    patterns for the male and female mice in the NTP (1986) and JISA (1993) studies appear
21    reasonably concordant.
22          Several issues complicate comparisons of the NCI (1977) gavage study results with those
23    of the other chronic bioassays. First, dosing lasted 78 weeks rather than 104 weeks as in the
24    inhalation studies, so in making direct comparisons it might be expected that the observed tumor
25    incidence in the NCI (1977) study would underestimate the incidence associated with 104 weeks
26    of exposure. Second, this oral gavage study had a variable dosing schedule, with doses that were
27    increased by 100 mg/kg-day in the low-dose group and by 200 mg/kg-day in the high-dose group
28    after 11 weeks  of study. Consequently, association of a  constant level of exposure with the
29    observed effects must be inferred rather than measured.  In addition, surviving animals were
30    maintained without further exposure until final sacrifice  in week 90.
31          The NCI (1977) exposures were recalculated  on a basis consistent with other EPA
32    estimates of chronic toxicity, in which the cumulative exposure received over the full period of
            o
             Combined historical incidence of adenomas and carcinomas was not available. Presumably the incidence
      of carcinomas slightly underestimates the combined incidence of adenomas and carcinomas.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-36       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
              Table 5-5. Tumor incidence and estimated metabolized doses in mice
              exposed to tetrachloroethylene
Bioassay
Doses/Exposures
Administered
Continuous
Equivalent
Sex
Body
Weight3
(kg)
Total
Metabolism1"
(mg/kg-day)
Survival-
Adjusted
Tumor
Incidence0
(%)
Hepatocellular adenomas and carcinomas
NCI (1977)d
B6C3FJ mice
Gavage:
5 days/wk,
78wks
NTP (1986)
B6C3FJ mice
Inhalation:
6 hrs/day,
5 days/wk,
104 wks
JISA (1993)
Crj:BDFl mice
Inhalation:
6 hrs/day,
5 days/wk,
104 wks
Vehicle control
450 mg/kg-day
900 mg/kg-day
Vehicle control
300 mg/kg-day e
600 mg/kg-day
Oppm
100 ppm
200 ppm
0 ppm
100 ppm
200 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oppm
10 ppm
50 ppm
250 ppm
Oe
332 mg/kg-day
663 mg/kg-day
Oe
239 mg/kg-day
478 mg/kg-day
0
18 ppm
36 ppm
0
18 ppm
36 ppm
0
1.8 ppm
9.0 ppm
45 ppm
0
1.8 ppm
9.0 ppm
45 ppm
Male
Female
Male
Female
Male
Female
0.030
0.025
0.037
0.032
0.048
0.035
0
35
47
0
34
46
0
27
41
0
31
45
0
3.4
14
36
0
4
18
47
2/20 (10)
32/48 (67)
27/45 (60)
0/20 (0)
19/48 (40)
19/45 (42)
17/49 (35)
31/47 (70)
41/50 (82)
4/45 (9)
17/42 (40)
38/48 (79)
13/46 (28)
21/49 (43)
19/48 (40)
40/49 (82)
3/50 (6)
3/47 (6)
7/48 (15)
33/49 (67)
Malignant hemangiosarcomasf, liver or spleen
JISA (1993)
Crj:BDFl mice
6 hrs/day,
5 days/wk,
104 wks
Oppm
10 ppm
50 ppm
250 ppm
0
1.8 ppm
9.0 ppm
45 ppm
Male
0.048
0
3.4
14
36
2/46 (4)
1/49 (2)
6/48 (13)
9/49 (18)
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
      a Average body weight reached during adulthood.
      b As calculated using the Reitz et al. (1996) pharmacokinetic model for mice, using alveolar ventilation rate, at
       67% of total ventilation (see Section 3.5). Total metabolism was estimated from the simulated bioassay exposure
       pattern, i.e., the amount estimated to be metabolized following an increment of exposure (gavage dose or 6-hr
       inhalation exposure). Adjustment for continuous exposure followed by multiplying the exposure by (5 days/7
       days). Figure 3-9 illustrates the correspondence of total metabolism with administered exposure estimated by this
       model for mice weighing 0.025 kg: at 100 ppm, approximately 47 mg-equivalent (eq)/kg-day of metabolites are
       estimated to be produced. For the purposes of this assessment, this is assumed to be equivalent to 47 mg-eq/kg-
       day x 5/7 = 34 mg-eq/kg-day of metabolites resulting from continuous exposure.  Note that this level is higher
       than the 31 mg-eq/kg-day estimated for 0.032 kg female mice and the 27 mg-eq/kg-day for 0.037 kg male mice in
       the NTP (1986) study in this table, illustrating the dependence of the PBPK model on body size. This dependence
       is not tabulated or graphed in this document.

                  This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                        5-37         DRAFT-DO NOT CITE OR QUOTE

-------
  1            Table 5-5. Tumor incidence and estimated metabolized doses in mice
 2            exposed to tetrachloroethylene (continued)
 O

 4     °  Animals dying before the first appearance of the tumor of interest but no later than week 52 were omitted from
 5      the totals because these animals were presumed not to have adequate time on study to develop tumors.
 6     d  No adenomas were reported in this study. Because hepatic adenomas and carcinomas are considered part of the
 7      same continuum of tumor development, and adenomas have been distinguished from carcinomas only on the basis
 8      of size, the correspondence of this observation to the other studies is not clear.
 9     e  Gavage doses listed were increased after 11 weeks by 100 mg/kg-day in each low-dose group or by 200 mg/kg-
10      day in each high-dose group.  Animals surviving the 78-week exposure period were observed until the week 90
11      study termination.  Lifetime average daily (administered) doses (LADDs) were calculated as follows:
12
13            LADD (mg/kg-day) = Cumulative administered dose (mg/kg)/(total days on study)
14                              = {[(initial dose level x 11 weeks) + (increased dose level x 67 weeks)]/90 weeks}
15                              x (5 days/7 days)
16
17       Male mice received LADDs of 332 and 663 mg/kg-day, and female mice received 239 and 478 mg/kg-day.
18     f These tumors were reported as hemangioendotheliomas in the JISA (1993) report.  The term has been updated to
19      hemangiosarcoma. Note that these incidences do not match those tabulated in Table 12 of the JISA report
20      summary.  The incidences reported here represent a tabulation of malignant hemangioendotheliomas from the
21      individual animal data provided in the JISA report.
22
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                       5-38        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
            Table 5-6.  Historical control data of the Japan Bioassay Research Center,
            Crj/BDFl mouse, 104-week studies
Tumor types
Inhalation, feeding, and drinking
studies (19 studies)
Total incidence
(%)
Range
(%)
Inhalation studies only (9 studies)
Total incidence
(%)
Range
(%)
Male mice
Liver
hepatocellular adenoma
hepatocellular carcinoma
Spleen
hemangioma3
hemangiosarcoma3
165/947 (17.4)
215/947 (22.7)
17/946 (1.8)
30/946 (3.2)
4.0-34.0
2.0-42.0
0-10.0
0-8.0
92/448 (20.5)
105/448 (23.4)
8/448 (1.8)
12/448 (2.7)
10.0-30.6
10.0-36.7
0-8.0
0-6.0
Female mice
Liver
hepatocellular adenoma
hepatocellular carcinoma
Spleen
hemangioma3
hemangiosarcoma3
50/949 (5.3)
22/949 (2.3)
8/949 (0.9)
3/949 (0.3)
2.0-10.0
0-8.0
0-6.0
0-2.0
18/449 (4.0)
14/449 (3.1)
5/449 (1.1)
3/449 (0.7)
2.0-6.0
0-8.0
0-6.0
0-2.0
4
5
6
7
     3 The terms "hemangioendothelioma:  benign" and "hemangioendothelioma" in the original study have been
       changed to "hemangioma" and "hemangiosarcoma," respectively.

     Source: Attachment to letter dated September 5, 2001, from K. Nagano, Japan Bioassay Research Center, Japan
     Industrial Safety and Health Association, to R. McGaughy, U.S. EPA.  Available from hotline.iris@epa.gov.
                 This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                      5-39        DRAFT-DO NOT CITE OR QUOTE

-------
                          Male Mice
                                                              Female Mice
          t°
          0)0'
          to
          §-
          Q.
          W 02 -
          £
                   Administered concentration (ppm)
                                               o
                                               iO.8
                                               §.,
                                               2J0.6
                                               £
                                               _§°5
                                               0)0.4
                                               W>
                                               §«
                                               Q.
                                               0)0.2
                                               0)
                                                             Administered concentration (ppm)
                          Male Mice
                                                              Female Mice
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
          s,
          §».H
          3 '
          C
          |0.5
          0)0.4
          to
          £o.3j
                   Total metabolism (mg-equiv/kg/d)
                                               C
                                               00.7
                                               (A
                                               §0,
                                               Q.
                                               »0.2
                                               £..
                                                                                       d.
                                                              Total metabolism (mg-equiv/kg/d)
       Figure 5-5.  Mouse liver tumor responses (hepatocellular adenomas and
       carcinomas) for three chronic bioassays (Table 5-5), plotted against
       continuous equivalent concentration (ppm) and total tetrachloroethylene
       metabolism (mg-equivalents/kg-day), for male and female mice.
observation is prorated to obtain an average lifetime daily exposure.  Table 5-5 provides the
recalculated exposures.  It is not clear in this case, however, that a simple TWA over the period
of observation is the most suitable representation of tetrachloroethylene exposure in the NCI
(1977) study, due to the substantial changes in the dosing pattern, as noted above.  In addition,
mortality was significantly increased in both treated groups over that of controls, suggesting that
the maximum tolerated dose had been exceeded.  Note that although no adenomas were reported
in the NCI (1977) study, some of the reported carcinomas may have been adenomas. This factor
should be taken into account when comparing the incidence of carcinomas in the NCI (1977)
study with the combination of adenomas and carcinomas in the inhalation studies (see
Table 5-5). Consequently, it was not feasible to compare this dose-response with that from the
inhalation studies on an administered mg/kg-day basis.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-40       DRAFT-DO NOT CITE OR QUOTE

-------
 1          In addition to hepatocellular adenomas and carcinomas, the JISA (1993) study
 2    demonstrated increased hemangiomas and hemangiosarcomas in the liver and spleen of mid- and
 3    high-dose male mice (see Table 5-5; Cochran-Armitage trend test, p = 0.004). The incidence in
 4    control and low-dose male mice was similar to the JBRC historical control incidence for spleen
 5    only (3.2%, range 0-8%; see Table 5-6).  This finding was not replicated in the NCI (1977) or
 6    NTP (1977) studies (tumors noted in the NTP male mice  livers:  controls, 3/49; low dose, 2/49;
 7    high dose, 2/50; tumors noted in the NTP historical controls for  all sites: 4.4%, range 2-8%
 8    (http://ntp.niehs.nih.gov/ntp/research/database searches/historical controls/path/m  inhar.txt).
 9    Both the JBRC and NTP historical controls showed similar background levels of hemangiomas
10    and hemangiosarcomas, although the JBRC data included only the spleen, whereas the NTP data
11    included all sites.
12          Because tetrachloroethylene's metabolites have been implicated in its liver toxicity (see
13    Section 4.10.4.1), and because a pharmacokinetic model was available to estimate metabolism
14    levels  in mice (based on the work of Reitz et al., 1996;  see Section 3.5), the hepatocellular tumor
15    responses in the three chronic bioassays were compared in terms of total metabolism of
16    tetrachloroethylene (see Figures 5-5c and 5-5d).  Here it can be seen that the hepatocellular
17    tumor dose-response in the gavage study appears to be  quite similar to that of the inhalation
18    studies. Note further that, from an empirical  point of view, the dose-response patterns for the
19    inhalation  studies collectively appear to follow an approximately linear relationship, whether the
20    exposure measure is the administered concentration or total metabolism. In other words, these
21    data do not clearly suggest one dose metric over the other as being more closely associated with
22    the liver tumors or that some other dose metric would be  preferred for characterizing cancer
23    incidence in this range of exposure.
24
25    5.4.2.2. Mononuclear Cell Leukemia in Rats
26          The NTP (1986) and JISA (1993) studies demonstrated increased MCL incidences for
27    male and female rats (see Table 5-7). The NCI study did not demonstrate any MCL increases in
28    rats. However, the investigators considered this  study inconclusive because of low survival, so
29    the NCI study neither confirms nor refutes the findings of the NTP and JISA studies.
30          The responses in the NTP (1986) study were approximately twofold higher than for the
31    corresponding groups in the JISA (1993) study, including the control groups. Control groups for
32    both laboratories were consistent with their respective historical controls (see Table 5-8 for the
33    JISA historical controls). Like the hepatocellular tumor results in  mice (see Section 5.4.2.1), the
34    MCL results from the NTP and JISA studies  were plotted in terms of additional risk versus
35    administered concentration (see Figure  5-6).  Note that MCL risk has been considered previously

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-41        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
              Table 5-7. Incidence of mononuclear cell leukemia, kidney tumors, and
              brain gliomas in rats exposed to tetrachloroethylene by inhalation
Bioassay
Exposure concentration
(ppm)
Administered
Continuous
equivalent
Sex
Body
weight (kg)
Total
metabolism"
(mg/kg-day)
Su rvival-ad justed
tumor incidence1"
(%)
Mononuclear cell leukemia
NTP (1986)
F344/N rats
Inhalation
6 hrs/day,
5 days/wk,
104 wks
JISA (1993)
F344/DuCrj rats
Inhalation
6 hrs/day,
5 days/wk,
104 wks
0
200
400
0
200
400
0
50
200
600
0
50
200
600
0
36
72
0
36
72
0
9
36
108
0
9
36
108
Male
Female
Male
Female
0.44
0.32
0.45
0.3
0.0
3.6
5.0
0.0
5.1
6.9
0.0
1.4
3.6
6.1
0.0
2.0
5.1
8.7
28/50 (56)
37/48 (77)
37/50 (74)
18/50 (36)
30/50 (60)
29/50 (58)
11/50(22)
14/50 (28)
22/50 (44)
27/50 (54)
10/50 (20)
17/50 (34)
16/50 (32)
19/50 (38)
Kidney: tubular cell adenoma or adenocarcinoma
NTP (1986)
JISA (1993)
0
200
400
0
50
200
600
0
36
71
0
9
36
110
Male
Male
0.44
0.45
0.0
3.6
5.0
0.0
1.4
3.6
6.1
1/49 (2)
3/47 (6)
4/50 (8)
1/50 (2)
2/50 (4)
1/50 (2)
2/50 (4)
Brain gliomas
NTP (1986)
JISA (1993)
0
200
400
0
50
200
600
0
36
71
0
9
36
110
Male
Male
0.44
0.45
0.0
3.6
5.0
0.0
1.4
3.6
6.1
1/50 (2)
0/48 (0)
4/50 (8)
2/50 (4)
0/50 (0)
0/50 (0)
0/50 (0)
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
     a As calculated by the Reitz et al. (1996) pharmacokinetic model for rats using alveolar ventilation rate at 67% of total
       ventilation (see Section 3.5). Total metabolism was estimated from the simulated bioassay exposure pattern, that is,
       the amount estimated to be metabolized following an increment of exposure (6-hr inhalation exposure). Adjustment
       for continuous exposure followed by multiplying the exposure by (5 days/7 days).  Figure 3-9 illustrates the
       correspondence of total metabolism with administered exposure estimated by this model for rats weighing 0.3 kg:  at
       200 ppm, approximately 7.1 mg-equivalent (eq)/kg-day of metabolites are estimated to be produced. For the
       purposes of this assessment, this is assumed to be equivalent to 7.1 mg-eq/kg-day x 5/7 = 5.1 mg-eq/kg-day of
       metabolites resulting from continuous exposure (see metabolite levels above for the JISA female rats). Note that
       this level is higher than the 3.6 mg-eq/kg-day estimated for 0.45 kg rats in the JISA study, illustrating the
       dependence of the PBPK model on body size.  This dependence is not tabulated or graphed in this document.
      b Animals dying before the first appearance of the tumor of interest but no later than week 52 were omitted from the
       totals because these animals were presumed to have had inadequate time on study to develop these tumors.
     Sources: NTP (1986) and JISA (1993).
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        5-42         DRAFT-DO NOT CITE OR QUOTE

-------
1
2
            Table 5-8.  Historical control data of the Japan Bioassay Research Center,
            F344/DuCrj (Fischer) rat, 104-week studies
Tumor types
Inhalation, feeding, and drinking
studies
(23 studies)
Total incidence (%)
Range
(%)
Inhalation studies only
(11 studies)
Total incidence
(%)
Range
(%)
Male rats
Mononuclear cell leukemia
Kidney
Renal cell adenoma
Renal cell carcinoma
147/1149(12.8)
2/1149(0.2)
2/1149(0.2)
6.0-22.0
0-2.0
0-2.0
76/549 (13.8)
1/549 (0.2)
2/549 (0.4)
6.0-22.0
0-2.0
0-2.0
Female rats
Mononuclear cell leukemia
Kidney
Renal cell adenoma
Renal cell carcinoma
147/1048 (14.0)
1/1048(0.1)
0/1048 (0.0)
2.0-26.0
0-2.0
NA
68/448 (15.2)
1/448 (0.2)
0/448 (0.0)
8.0-20.0
0-2.0
NA
4
5
     Source: Attachment to letter from K. Nagano to R. McGaughy 9/5/01. Available from IRIS Information Desk.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   5-43        DRAFT-DO NOT CITE OR QUOTE

-------
                     Male Rats
                                                              Female Rats
 2
 3
 4
 5
 6
 7
 8
 9
10
1 1
12
13
14
15
16
17
18
19
0.4 -i

^*
O
-go.3
o
o
to
3
c
I02
0)
to
c
Q.0.1
to
a:

n n s

-A- NTP, 1986
-e-JISA, 1993
^__^°

^^_^-~-"~~~~~^
^^_^-—-~~"'^

%C^^
/y~~~~~~~~-t.
//

//
//
//
r
/ a.
«,,,,,,,,,,,
0.4 -i

^*
O
"C 0.3
O
o
to
3
C
I02
0)
to
c
o
Q0.1
to

-------
 1          For female rats (see Figures 5-6b and d), the responses adjusted for control rates do not
 2    show as much concordance between studies as those for the male rats, with the JISA study still
 3    showing lower adjusted responses than the NTP study at comparable exposures. The low-dose
 4    females in the JISA study had a higher response relative to the pattern of responses in the other
 5    three groups and a higher response than would be expected from the dose-response pattern in the
 6    NTP study. The dose-response relationship for the JISA study appears to be slightly more linear
 7    for total metabolism than for administered concentration; however, both studies suggest some
 8    degree of saturation of effects in the available range of the dose metrics considered.  Although
 9    F344 rats were used in both studies, it is possible that there could be some differences
10    attributable to the specific lines of animals used at each laboratory and laboratory-specific
11    procedures.
12
13    5.4.2.3. Other Tumor Sites in Male Rats
14          Other elevated tumor incidences—brain gliomas and kidney tubule adenomas and
15    adenocarcinomas—were observed in male F344/N rats in the NTP study but not in the JISA
16    study (again, there were no corresponding data available for the NCI male rats). Table 5-7
17    summarizes the incidence data from both laboratories for these sites.  Brain gliomas in rats in the
18    NTP inhalation study were elevated.  In males, the incidences were 1/50, 0/48, and 4/50 in the
19    control, 200 ppm, and 400 ppm tetrachloroethylene groups, respectively.  This was a statistically
20    significant dose-related trend by  the life table test (p = 0.039) but not by the incidental tumor
21    trend test.  A similar trend was seen in female rats (1/50, 0/50, 2/50), but this trend was not
22    statistically significant.  Brain gliomas are rare tumors in NTP rat bioassays;  in male rats the
23    historical control incidence is 2/247 (0.8%) in the laboratory where this study was conducted,
24    and 4/1971 (0.2%) in the overall program.  Because these tumors had not been observed in
25    previous NTP studies of tetrachloroethylene, trichloroethylene, or pentachloroethane, and
26    because they appeared in the untreated groups, the NTP investigators concluded that they were
27    not related to tetrachloroethylene exposure.
28          On the other hand, the previous study of tetrachloroethylene was the NCI (1977) gavage
29    study, not an inhalation study; route-to-route differences are not implausible.  Although the other
30    solvents mentioned have been associated with similar adverse effects, there are other differences
31    among them. Also, the brain tumors in the high-dose males started occurring earlier (weeks 88,
32    96, 102, and 103) than in the control group (99 weeks), and in the high-dose  females they
33    occurred even earlier (75 and 78  weeks in the high-dose group vs. 104 weeks in the control
34    group). Therefore, although the  association between tetrachloroethylene exposure and brain
35    gliomas in rats in this data set is  not strong, it is still suggestive, especially considering that the

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-45       DRAFT-DO NOT CITE OR QUOTE

-------
 1    nervous system is a target of tetrachloroethylene exposure in humans and animals (see Sections
 2    4.5.3 and 5.1.1).
 3           In the JISA study, brain gliomas were observed only in male control rats. Historical
 4    brain tumor incidence data for this laboratory were not available, however.  Although F344 rats
 5    were used in both studies, it is possible that there could be some differences between laboratories
 6    attributable to the specific lines of animals used and laboratory-specific procedures. Given the
 7    low overall incidences relative to the other tumor sites, these data were not modeled.
 8           Kidney tubule cell adenomas and adenocarcinomas (see Table 5-7) were elevated in the
 9    exposed male rats, but they were not statistically significantly  elevated.  Statistical significance is
10    a secondary consideration in determining the biological significance of these tumors because
11    they are considered to be uncommon in NTP studies of rats. The investigators noted that these
12    tumors were observed among historical controls at about 0.2% in 1968 untreated control rats.
13    Further support for considering the relevance of this site comes from the evidence relating
14    trichloroethylene and rat kidney tumors (U.S. EPA, 2001).
15           There was no apparent trend in the incidence of kidney tubule cell adenomas and
16    adenocarcinomas among JISA male rats.  The incidence in all  groups was consistent with JISA
17    historical control data (see Table 5-8).
18
19    5.4.3.  Estimation of Dose Metrics for Dose-Response Modeling
20           The sequence of steps in estimating human equivalent  risks is illustrated in Figure 5-7,
21    with estimation of the dose inputs to the dose-response modeling being the first step.
22    Considerations for estimating continuous exposure levels equivalent to the intermittent animal
23    bioassay exposures differ according to whether administered exposures or metabolized doses
24    were used as the measure of dose, and they are discussed following the selection of each dose
25    metric.
26
27    5.4.3.1.  Dose Metric for Hepatocellular Carcinogenicity
28           There are several possible rationales to consider for the appropriate dose metric for
29    tetrachloroethylene-induced liver toxicity and carcinogenicity. The specific chemical species
30    responsible for adverse effects on the liver would be the preferred choice. As discussed in
31    Chapter 3 and Section 4.10.4.1, several metabolites associated with P450 metabolism occurring
32    in the liver have been identified. TCA, which is associated with liver toxicity when administered
33    directly, is considered a key product of this P450 oxidation pathway. However, the MOA for
34    tetrachloroethylene-induced liver toxicity and carcinogenicity  is not clear.  Further, TCA does
35    not appear to explain the liver carcinogenicity observed with tetrachloroethylene, because a
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-46       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
                 Administered dose/
                 concentration in bio assay
                      Animal PBPK model
                                                     Human equivalent
                                                     unit risk (per ppm) or
                                                     slope factor
                                                     (p er m g/kg B W- day)
                 Total metabolism
                 in anim als, m g- eq/kg B W- day
                                                                BMR - POD
                       Continuous exposure
                       adjustment for 5/7 days
                 Equivalent continuous total
                 metabolism
                 in anim als, m g- eq/kg B W- day
                                                     Human equivalent point
                                                     of departure, in
                                                     continuous environmental
                                                     exp osure level, in ppm
                                                     (inhalation) or
                                                     m g/k g B W- day (oral)
                      Dose-response model,
                      with, response data
                                                              Human PBPK model
                 Point of departure, in
                 m g- eq/kg B W- day of total
                 metabolism
                                       Bw  scaling
Human equivalent point
of departure,
in m g- e q/k g B W-day of
total m etabolism
       Figure 5-7.  Sequence of steps for extrapolating from tetrachloroethylene
       bioassays in animals to human-equivalent exposures expected to be
       associated with comparable cancer risk.
comparison of hepatocellular tumor incidence associated with direct TCA exposure appears to
underpredict the hepatocellular carcinoma incidence in the NTP and JISA studies when
characterized in terms of equivalent TCA exposures (see Appendix 4A in Chapter 4).
       Additional metabolites could play a role.  For instance, reactive intermediates such as
tetrachloroethylene oxide or trichloroacetyl chloride are hypothesized to be precursors to TCA,
and as reactive  compounds they would be likely candidates. However, their involvement in liver
toxicity remains unknown, and they have not been confirmed in the tetrachloroethylene
metabolic pathways.  Consequently, although it appears plausible that at least another compound
besides TCA contributes to tetrachloroethylene-induced hepatocarcinogenicity, none has been
identified nor can the amounts be estimated.
       Because of the uncertainty over which metabolite species are involved in causing liver
toxicity—and to what degree they are involved—total metabolism was considered the most
appropriate dose metric. Use of this dose metric does not require assuming that all of the
metabolites are responsible for tetrachloroethylene's liver carcinogenicity, however, only that the
rate of total metabolite production is proportional to the actual target dose in the target tissue,  at
least at very low exposures.  That is, if there is a constant relationship  between the surrogate dose
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   5-47       DRAFT-DO NOT CITE OR QUOTE

-------
 1    measure and the relevant measure of the actual carcinogen, at least at very low exposures, then
 2    dose-response modeling of each measure (if both were available) would yield the same cancer
 3    risk value in terms of environmental exposure. For instance, the concentration of the actual liver
 4    carcinogen(s) could be in equilibrium with the circulating blood, or it could be primarily
 5    generated and active in the liver; as long as proportionality of the surrogate dose measure is a
 6    reasonable assumption, cancer risk values can be estimated and interpreted accordingly.
 7    Consequently, the daily production rate of all metabolites of tetrachloroethylene corresponding
 8    to the bioassay exposure  patterns, as estimated using the PBPK model of Reitz et al. (1996), was
 9    used for the dose-response modeling of mouse liver tumors.
10          The second adjustment made prior to dose-response modeling was to  characterize the
11    intermittent bioassay exposures in terms of equivalent continuous exposure.  Because the
12    pharmacokinetic model cannot generate an AUC for total metabolism, due to the lack of
13    clearances for the individual metabolites, the daily production rate of all metabolites for five
14    days was averaged over seven days by multiplying it by 5/7 (0.71), under the assumption that
15    concentration multiplied  by time maintains a constant effect (C x t = k\ is likely to hold for very
16    low tetrachloroethylene exposures. The metabolism rates reported in Tables  5-5 and 5-7 reflect
17    this adjustment.
18
19    5.4.3.2.  Dose Metric for Rat Leukemias and Kidney Tumors
20          Experimental evidence suggests that a  GST-pathway metabolite (TCVC) is more likely to
21    be associated with the kidney tumors and possibly the leukemias than the P450 pathway (see
22    Sections 4.9.4.2 and 4.9.4.3). However, the measurements of glutathione-dependent metabolism
23    are from in vitro studies or they are measures of urinary excretion products and are, therefore,
24    not representative of the toxic species in vivo.  Consequently, insufficient data exist to
25    incorporate the GST-derived metabolites explicitly in the PBPK models.
26          Given the approximately linear dose-response relationship observed between leukemias
27    and total metabolism for  male rats (see Figure 5-6c), it appears plausible that the carcinogen
28    responsible for the leukemias may be approximately proportional to total metabolism.  The
29    situation is somewhat less clear for female rats due to the nonmonotonic dose-response patterns,
30    although the degree of saturation was less pronounced when the dose-response relationship was
31    considered in terms of total metabolism (Figure 5-6d).  Accordingly, total metabolism was
32    considered a better surrogate than administered concentration for the proximate carcinogen.
33    Because kidney tumors are associated with the same GST-metabolite as the leukemias—
34    somewhat more definitively than the leukemias—total metabolite production was also
35    considered as a dose metric for estimating the  male rat kidney dose-response  relationship.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-48       DRAFT-DO NOT CITE OR QUOTE

-------
 1          Adjustment for continuous exposure was the same as for the liver tumors (see Section
 2    5.4.3.1). That is, the only continuous exposure adjustment to the total metabolite dose metric
 3    that was needed was to average the metabolic rate for five days over seven days (by multiplying
 4    each by 5/7).
 5
 6    5.4.3.3. Dose Metric for Sites Not Addressed by Physiologically Based Pharmacokinetic
 1           (PBPK) Modeling
 8          For tumor sites for which the MOA is not clear, such as for male rat brain tumors (see
 9    Section 4.10.4), the administered concentration of tetrachloroethylene was used as the dose
10    metric. Use of this dose metric should provide plausible results as long as the concentration of
11    the proximate carcinogen(s) is proportional to administered concentration—at least at low
12    concentrations.  Because there is uncertainty in the identification of the carcinogenic agent for all
13    the sites, the dose-response relationships for all tumor sites were also estimated using this default
14    dose metric, for comparison purposes.
15          Allowance for extrapolation to continuous exposures was made before dose-response
16    modeling.  In all cases, administered inhalation concentrations (in ppm) were adjusted for
17    continuous exposure by averaging the five 6-hr daily exposures over the full week.  That is,
18    administered concentrations were multiplied by 6 hrs/24 hrs x 5 days/7 days (0.179) to yield
19    equivalent continuous concentrations. Tables 5-5 and 5-7 provide these adjusted concentrations.
20
21    5.4.4.  Extrapolation Methods
22          Extrapolation of tetrachloroethylene cancer risks observed in animal bioassays to humans
23    with continuous environmental exposure involved a number of methods, including dose-response
24    modeling in the range of observation, interspecies extrapolation, extrapolation to low exposures,
25    and route-to-route extrapolation. Section 5.4.4.1  and Figure 5-7 summarize the methods used to
26    extrapolate from the experimental data to humans.
27
28    5.4.4.1. Dose-Response Models and Extrapolation to Low Doses
29          As  discussed in Section 4.10.3, the available body of MOA information is not sufficient
30    to derive biologically based quantitative models for low-dose extrapolation.  No key events in
31    the tumor development process for tetrachloroethylene have been identified that would
32    determine the overall dynamics of such a model, nor are there experimental data specific to
33    tetrachloroethylene describing any of the underlying toxicodynamic processes, such as cell
34    replication rates.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-49        DRAFT-DO NOT CITE OR QUOTE

-------
 1          The multistage model (and the multistage-Weibull) has been used by EPA in the vast
 2    majority of quantitative cancer assessments because it has some parallelism to the multistage
 3    carcinogenic process and it fits a broad array of dose-response patterns. Occasionally the
 4    multistage model does not fit the available data, in which case an alternate model should be
 5    considered. The related multistage-Weibull model has been the preferred model when individual
 6    data are available for time-to-tumor modeling, which considers more of the observed response
 7    than does the simpler dichotomous response model. Use of this decision scheme has contributed
 8    to greater consistency among cancer risk assessments.
 9          Consequently, the multistage model was the primary tool considered for fitting the dose-
10    response data and is given by:
11
12                     P(d) =1 - exp(- q0-qixd-q2x£f-...-q6xd6)
13
14    where:
15          d   = exposure level and
16          q;  = parameters estimated in fitting the model
17
18    The multistage model in HMDS (Benchmark Dose Software, version 1.3.2; U.S. EPA, 2000) was
19    used for all multistage model fits.
20          Two tumor sites with statistically significantly decreased time to tumor were noted:  brain
21    gliomas in NTP male rats and MCL in the NTP female rats, especially  for the most severe stage
22    of leukemia observed (Stage 3).  The multistage-Weibull model, given by the following
23    equation, was also used to evaluate the importance of decreased time to tumor and intercurrent
24    mortality in interpreting these responses.
25
26                     P(d,t) = 1- exp[(- q0- qi  x d - q2 x d2 -.  . . - q6 x d5) x f]
27
28    where:
29          d     = exposure level
30          t     = time to observation of the tumor
31          q;, z  = parameters estimated in fitting the model
32
33    Note that when the time to observation of the tumor is not a significant contributor to the dose-
34    response relationship, the parameter z is estimated to have a value of 1, and the model reduces to
35    the simpler multistage model described just before the multistage-Weibull.  Tox_Risk (K.S.

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-50        DRAFT-DO NOT CITE OR QUOTE

-------
 1    Crump Group, Inc., ICF Kaiser International, Ruston, LA) was used for all multistage-Weibull
 2    model fits.
 3          Following dose-response modeling in the range of observation, the cancer risk values for
 4    extrapolation to low doses were derived from the lower bound on the dose/concentration
 5    associated with a level of risk from the low end of the observed range, usually 10% extra risk,
 6    consistent with the 2005 cancer assessment guidelines. Extra risk has been used consistently
 7    throughout EPA risk assessments and is given by:
 8
 9                 Extra risk = [P(d) -P(0)] / [1 -P(0)]
10
11    where:
12          P(d) = estimated response at dose d and
13          P(0) = estimated response in the control group
1 A
14
15    The slope factor (risk per mg/kg-day for oral exposure) and risk per unit concentration (risk per
16    mg/L for drinking water exposure, or per |ig/m3 for inhalation exposure) are estimated by
17    dividing the risk level by its associated POD:
18
19          Risk/(unit of exposure) = Extra risk/(lower confidence bound on associated exposure)
20
21    5.4.4.2.  Extrapolation to Human Equivalent Environmental Exposure
22          For extrapolation of risk to humans, this assessment used two approaches that were
23    dependent on the relevant dose metric: the EPA RfC methodology (U.S. EPA, 1994), which
24    applies when chemical-specific pharmacokinetic data are lacking, and EPA's cross-species
25    scaling methodology (U.S. EPA,  1992), which applies to exposures characterized in mg/kg-day,
26    whether parent chemical or metabolite.
27
28    5.4.4.2.1. Metabolized dose. Because of the availability of PBPK models to estimate a plausible
29    dose metric that addresses the differential metabolism of tetrachloroethylene between laboratory
30    rodents and humans, extrapolation to human equivalent environmental exposure entailed two
31    steps.  First, consistent with the 2005 cancer guidelines (U.S. EPA, 2005a), EPA's methodology
32    for cross-species scaling (U.S. EPA, 1992) was considered when toxicological equivalence for
33    the relevant tumor sites was addressed.  Then, the human equivalent exposure in terms of
34    metabolized dose was estimated via the human PBPK models.  These  considerations are further
35    described below.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  5-51        DRAFT-DO NOT CITE OR QUOTE

-------
 1          EPA's cross-species scaling methodology was used for describing toxicological
 2    equivalence because of the extensive rationale supporting it (U.S. EPA, 1992).  Briefly, the
 3    methodology maintains that, in the absence of adequate information to the contrary, toxicological
 4    equivalence across species is determined through equal average lifetime concentrations or AUCs
 5    of the carcinogen. The most typical application of this methodology is to oral exposures in
 6    mg/kg-day, with no pharmacokinetic or pharmacodynamic  information. In this circumstance, the
 7    correspondence of equal AUCs is equivalent to considering the exposures in terms of
 8    mg/kg3/4-day, and is achieved by multiplying animal exposures by (BWanimai/BWhuman)1/4. Note
 9    that this equivalence across species entails the cross-species correspondence of internal doses in
10    terms of AUCs or mg/kg3/4-day, which is implicit in the frequent default case, i.e., oral
11    carcinogens without chemical-specific pharmacokinetic data.  In other words, each time a
12    carcinogen is scaled from animals to humans on the basis of mg/kg3/4-day, an implicit
13    assumption is that internal doses are equipotent in terms of mg/kg3/4-day ("cross-species
14    scaling"), not mg/kg-day ("body-weight scaling").  Accordingly, when pharmacokinetic data are
15    available that relate administered concentration to the overall metabolized dose of the
16    carcinogen, this methodology is still applicable; internal doses, as a fraction of administered
17    dose, should still  tend to produce equivalent effects when considered in terms of AUCs or
18    mg/kg3/4-day because metabolites are also subject to scale-affected clearance processes.  In other
19    words, the scaling may be thought of as applied to the administered dose adjusted by the fraction
20    metabolized. There is a wide body of empirical evidence that overall metabolic rates associated
21    with enzymatic processes scale with body weight to the 3/4 power (U.S. EPA, 1992).
22    Furthermore, because in this assessment the scaling is applied to an internal dose (namely, the
23    overall metabolic rate), it is applicable regardless of the route of exposure.
24          EPA has experience applying cross-species  scaling methodology in a number of
25    carcinogen assessments that have relied on pharmacokinetic modeling to characterize risks from
26    inhalation  exposure. Further, the vast majority of EPA carcinogen assessments have relied on
27    this method—all  oral slope factor estimates developed from animal bioassay data and all cancer
28    risk values developed from bioassay data and relying on PBPK models. Specific assessments
29    relying on  PBPK models include the previous tetrachloroethylene assessment, dichloromethane,
30    vinyl chloride, and trichloroethylene. In all cases, a scientific rationale was provided for the
31    cross-species scaling approach taken.
32          The previous tetrachloroethylene assessment (U.S. EPA,  1986) also used cross-species
33    scaling of total rate of metabolism  for the liver tumors and leukemias.  The dichloromethane
34    assessment used cross-species scaling (BW2/3) of the daily amount of inhaled dichloromethane
35    metabolized by a GST pathway (U.S. EPA, 1987, 1995).  The vinyl  chloride inhalation risk per

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-52       DRAFT-DO NOT CITE OR QUOTE

-------
 1    unit concentration involved a reactive metabolite whose AUC was judged to be proportional to
 2    the metabolite's tissue concentration (U.S. EPA, 2000); that is, AUCs (and responses) for this
 3    metabolite would tend to be equal for doses in terms of mg/kg-day rather than mg/kg3/4-day.9
 4    Most recently, EPA's trichloroethylene assessment (external draft) used AUCs of metabolites
 5    produced in the liver following inhalation exposure as being predictive  of human risk using the
 6    liver tumors observed in mice; for kidney tumors, the human equivalent risk was estimated using
 7    BW3/4 scaling of daily production of thiol in the kidney (U.S. EPA, 2001).
 8           Following the cross-species scaling methodology, metabolized tetrachloroethylene was
 9    scaled using mg/kg3/4-day in order to estimate equivalent toxic effects in humans (U.S. EPA,
10    1992).  This determination followed consideration of the reactivity of the dose metric and the
11    ability to estimate AUCs for the dose metric.  The involvement of reactive metabolites through
12    which all other metabolites may follow has been hypothesized; however, body-weight scaling
13    was not considered pertinent for tetrachloroethylene because the possible reactive metabolites
14    have not been confirmed and because the majority of the metabolites formed is accounted for by
15    TCA, a stable metabolite. Concerning estimation of AUCs, the PBPK models for
16    tetrachloroethylene provide the rate of overall metabolism in units of mg-equivalents/kg-day,
17    which is a rate or flux.  The models do not describe the kinetics of the overall metabolism and
18    therefore cannot provide  an AUC. As discussed in Section 3, this is because the clearances of all
19    but one of tetrachloroethylene's metabolites are unknown, and many of the metabolites
20    themselves have not been identified.  The metabolite whose clearance has been estimated is
21    TCA. While TCA is the  predominant metabolite, it is not clear that TCA is responsible for all
22    the observed toxicity (see Appendix 4A). For animals, the study-specific body weights were
23    used (see Tables 5-5  and  5-7), and for humans the default of 70 kg was  used.
24           It might appear that the use of such a procedure constitutes a "double counting" of
25    allometric scaling. This is not the case as is evident from the following explanation. The AUC
26    of the circulating stable metabolite (if available) leads to an equivalent average tissue
27    concentration of the metabolite X, Cx, for both species. This average concentration, when
28    applied over the lifetime  of a species, leads to equivalent risk across species. For simplicity,
29    consider a one-compartment model.  At steady-state, the production of X will be equal to the
30    clearance of X, so that
              Available human incidence data were judged to be concordant with this interpretation of the animal data
      for vinyl chloride; consequently, no cross-species scaling factor was considered necessary. The cross-species
      scaling methodology (U.S. EPA, 1992) points out that, in general, body-weight scaling for reactive metabolites
      entails assuming that the metabolite is removed from its target by spontaneous action, never leaves the tissue in
      which it is formed, does not form lexicologically active macromolecular adducts, and that there are no species
      differences in persistence. That is, body-weight scaling of a reactive metabolite would not be expected to result in
      cross-species lexicological equivalence in all cases.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-53       DRAFT-DO NOT CITE OR QUOTE

-------
 1                          Rmet=VdxBWxCxxkcl,
 2
 3    where:
 4          Rmet   = rate of production of X
 5          Vd     = fractional volume of distribution
 6          BW   = body weight (converted to liters)
 7          Cx     = concentration of X and
 8          kci     = clearance of X in units of I/time
 9
10    Then, for the concentration Cx equivalent in both species:
11
12                     CX = [Rmet/BW X kcl xVd]H = [Rmet/BW X kcl X Vd\A
13
14    where H and A refer to human and animal.  It is safe to assume that Vd is the same across
15    species.  Then, [RmetIBW x kci\H = [RmetIBW x &C/]A. Now, kd (with units of I/time) is known to
16    scale according to BW1'4 (U.S. EPA, 2005a).  Thus, the AUC approach leads to
17
18                     Rmet (H/8 WH   = Rmet (A)IB WA
19
20    This is the  scaling approach used in this assessment due to lack of data to pursue an AUC
21    approach explicitly.
22          In the last step of the extrapolation to human equivalent PODs, the PODs in  terms of
23    metabolized dose were extrapolated to environmental inhalation and oral exposures using
24    pharmacokinetic modeling. As discussed in Section 3.5, three human PBPK models were
25    considered, owing to insufficient data to distinguish between these models at low environmental
26    concentrations, especially concerning validation of total metabolite levels.  These models
27    represent the work of Reitz et al. (1996), Rao and Brown (1993), and Bois et al. (1996), as
28    adapted by EPA (see Section 3.5 for more details). Because use of the human PBPK models
29    indicated that the correspondence between total metabolism and administered concentration was
30    linear below 0.1-1 ppm (see Figure 3-10), conversion factors (slopes) derived from  Figure 3-10
31    were applied to estimate the human equivalent PODs in terms of administered concentration;
32    e.g., human equivalent POD (ppm) = human equivalent POD (mg-eq/kg-day) x conversion
33    factor ([mg-eq/kg-day]/ppm). See footnotes d-f in Table 5-9 and footnotes b-d in Table 5-11 for
34    the inhalation and oral conversion factors.
35

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-54       DRAFT-DO NOT CITE OR QUOTE

-------
 1    5.4.4.2.2. Administered concentration as dose metric. For those sites for which
 2    pharmacokinetic-adjusted doses were not available or not otherwise relevant, EPA's default RfC
 3    methodology was used (U.S. EPA, 1994). Tetrachloroethylene is considered a Category 3 gas
 4    because it is water soluble and perfusion limited, and it has systemic (extra-respiratory) effects.
 5    Because the ratio of blood/air partition coefficients for the experimental animal species relative
 6    to humans is greater than or equal to 1 (for F344 rats,  18.9/10.3 = 1.8; for B6C3Fi mice,
 7    17.5/10.3 = 1.7), a default value of 1 was used for this ratio (U.S. EPA, 1994). Consequently,
 8    when administered inhalation concentrations were used as the dose metric, the concentrations
 9    were considered equipotent across species for extrapolating risk to humans.  Therefore, no
10    further extrapolation was necessary with the resulting PODs in the units of human equivalent
11    environmental exposure levels.
12
13    5.4.5.  Cancer Risk Values
14           Human cancer risk was assessed using six different sex-species animal tumor data sets
15    and three different human PBPK models of total metabolism.  The results of the dose-response
16    modeling using the data from the inhalation animal studies are discussed below, followed by
17    route-to-route extrapolation for estimating human cancer risk via oral exposure to
18    tetrachloroethylene.  Finally, a discussion of quantitative and qualitative uncertainties underlying
19    the risk estimation process is provided.
20
21    5.4.5.1. Dose-Response Modeling Results
22           The dose-response modeling relying on total metabolism as the dosimeter is illustrated in
23    Figures 5-8a through 5-13a, and it is summarized in Table 5-9. The estimation of risk per unit
24    concentration associated with each tumor site is summarized in Tables 5-9 (identification of
25    PODs) and  5-10 (conversions of PODs to risk per unit concentration).  The dose-response
26    modeling relying on administered concentration is illustrated in Figures 5-8b through 5-13b and
27    summarized in Table 5-11. Site-specific modeling results  and conversions to human equivalent
28    risk values are discussed below.  In all cases, linear extrapolation from the PODs was carried out
29    because of the lack of information supporting another extrapolation approach (U.S. EPA, 2005a).
30
31    5.4.5.1.1. Mouse tumors.
32    5.4.5.1.1.1. Hepatocellular tumors, male mice. The  dose-response modeling results from the
33    hepatocellular adenomas or carcinomas in male mice of the JISA bioassay using total
34    metabolism (via PBPK modeling) led to human equivalent PODs (BMCLios) ranging from 1.8
35    ppm (Bois model) to 18 ppm (Rao and Brown model) tetrachloroethylene in air (see Table 5-9
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-55        DRAFT-DO NOT CITE OR QUOTE

-------
                   0.9

                   O.E!

                   0.7
                T3
                'If
                ts
                I
                •I  0.5
                fO
                £  0.4

                   0.3

                   0.2
                                    Multistage Model with 0.95 Confidence Level
                         Multistage
                            BMDL
                               BMD
                                        10     15      20
                                                 dose
                                               25
30
35
                  17:0705/222002
 2
 3
 4
 5
 6
 7
 8
 9
10
Figure 5-8a.  Incidence of hepatocellular adenomas and carcinomas in male
mice (JISA, 1993) corresponding to total tetrachloroethylene metabolism
(mg-eq/kg-day) and multistage model fit showing BMC and BMCL at 10%
extra risk. Data from Table 5-5.
                   17" 10 05^22072
Figure 5-8b.  Incidence of hepatocellular adenomas and carcinomas in male
mice (JISA, 1993) corresponding to human equivalent continuous
tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
and BMCL at 10% extra risk. Data from Table 5-5
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  5-56       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
                            0         10

                   144:06272002
                                                        30
40
51
                                                 dose
            Figure 5-9a. Incidence of hepatocellular adenomas and carcinomas in female
            mice (JISA, 1993) corresponding to total tetrachloroethylene metabolism
            (mg-eq/kg-day) and multistage model fit showing BMC and BMCL at 10%
            extra risk.  Data from Table 5-5.
                                   Multistage Model with 0.95 Confidence Level
 6
 1
 8
 9
10
                                                                    40
                                                 dose
                14:5206/272002
           Figure 5-9b. Incidence of hepatocellular adenomas and carcinomas in female
           mice (JISA, 1993) corresponding to human equivalent continuous
           tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
           and BMCL at 5% extra risk. Data from Table 5-5.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  5-57       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
                                             dose
           17:2005/222002
       Figure 5-10a. Incidence of malignant hemangiosarcomas in male mice (JISA,
       1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
       and multistage model fit showing BMC and BMCL at 10% extra risk.  Data
       from Table 5-5.

                               Multistage Model with 0.95 Confidence Level
                  0.3
                 0.25
             S    0.2
             u
             Qj
             I
             O
             ~o
             ¥
                 0.15
                  0.1
                 0.05
                        Multistage
                             BMDL
                                   BMD
                                    10
                                           20
30
40
                                                 dose
               13:4305/252005
       Figure 5-10b. Incidence of malignant hemangiosarcomas in male mice
       (JISA, 1993) corresponding to human equivalent continuous
       tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
       and BMCL at 5% extra risk. Data from Table 5-5.

           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  5-58       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
6
7
8
9
              •c
              Qj

              g
              o
                 O.G


                 0.5


                 0.4


                 0.3


                 0.2


                 0.1
                        Multistage
                            BMDL   BMD
                                 1
                                                  3
                                                 dose
                17:2705/222002
           Figure 5-lla. Incidence of mononuclear cell leukemia in male rats (JISA,
           1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
           and multistage model fit showing BMC and BMCL at 10% extra risk. Data
           from Table 5-7.

                                   Multistage Model with 0.95 Confidence Level
                0.6
                0.5
                0.4
                0.3
                0.2
                0.1
                       Multistage
                       .BMDL  BMP
                                 20
                                           40
60
SO
100
120
                                                 dose
               17:3805/222002
           Figure 5-1 Ib. Incidence of mononuclear cell leukemia in male rats (JISA,
           1993) corresponding to human equivalent continuous tetrachloroethylene
           exposure (ppm) and multistage model fit showing BMC and BMCL at 5%
           extra risk. Data from Table 5-7.
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   5-59        DRAFT-DO NOT CITE OR QUOTE

-------
                   0.5

                  0.45

               ,   0.4
               m
               I  0,35

                   0.3

                  0.25

                   0.2

                  0.15

                   0.1
                          Multistage
                                  BMDL
                                                    dose
1
2
3
4
5
6
 7
 8
 9
10
                 17:4205/222002
            Figure 5-12a. Incidence of mononuclear cell leukemia in female rats (JISA,
            1993) corresponding to total tetrachloroethylene metabolism (mg-eq/kg-day)
            and multistage model fit showing BMC and BMCL at 10% extra risk. Data
            from Table 5-7.
                                    Multistage Model with 0.95 Confidence Level
              -a
              Oi
              T>
                 0.5

                0.45

                 0.4

                0.35

                 0.3

                0.25

                 0.2

                0.15

                 0.1
                         Multistage
                            BMDL
BMD
                          0
                                  20
    40
GO
80
100
120
                                                  dose
               17:4405/222002
           Figure 5-12b. Incidence of mononuclear cell leukemia in female rats (JISA,
           1993) corresponding to human equivalent continuous tetrachloroethylene
           exposure (ppm) and multistage model fit showing BMC and BMCL at 5%
           extra risk. Data from Table 5-7.
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   5-60       DRAFT-DO NOT CITE OR QUOTE

-------
                                  Multistage Model with 0.95 Confidence Level
I
2
3
4
5
               0.15 -
           III
           t:
           CD
           o
           iLJ
           £
               0.05 -
             09:47 03/09 2005
           Figure 5-13a. Incidence of kidney adenomas and adenocarcinomas in male
           rats (NTP, 1986) corresponding to total tetrachloroethylene metabolism (mg-
           eq/kg-day) and multistage model fit showing BMC and BMCL at 5% extra
           risk. Data from Table 5-7.
           m
           Tj
           IZ
           o
          tj
           £5
              0.15
               0.1
              0.05
                              10
                                      20
30
40
50
60
70
                                                dose
6          Figure 5-13b. Incidence of kidney adenomas and adenocarcinomas in male
7          rats (NTP, 1986) corresponding to human equivalent continuous
8          tetrachloroethylene exposure (ppm) and multistage model fit showing BMC
9          and BMCL at 5% extra risk. Data from Table 5-7.

               This document is a draft for review purposes only and does not constitute Agency policy
    06/06/08                                  5-61        DRAFT-DO NOT CITE OR QUOTE

-------
        Table 5-9. Dose-response modeling summary for tumor sites using total tetrachloroethylene metabolites as the
        dosimeter; tumor incidence data from JISA (1993) and NTP (1986)
Tumor type
Hepatocellular
adenomas or
carcinomas
(JISA, 1993)
Hemangiosarcomas,
spleen or liver
(JISA, 1993)
Mononuclear cell
leukemia, rats
(JISA, 1993)
Kidney tumors
(NTP, 1986)
Group
Male
mice
Female
mice
Male
mice
Male rats
Female
rats
Male rats
Modeling summary"
MLE coefficients
qo = 0.34
qj = 0.0049
q3 = 2.3 x ID'5
qo = 0.055
q2 = 4.6 x ID'4
q0 = 0.032
qj= 0.0051
qo = 0.21
Q! = 0.078
q2 = 0.0020
qo = 0.24
qj = 0.026
q0 = 0.020
Q!= 0.013
/7-value
0.15
0.75
0.48
0.68
0.48
1
POD in terms of metabolized dose (mg-
eq/kg-day)
Bioassay
estimate1"
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC10
BMCL10
BMC05
BMCL05
12
3.6
15
9.8
21
12
1.3
0.81
4.1
2.0
4.1
1.9
Human
equivalent0
1.9
0.58
2.2
1.5
3.4
1.9
0.37
0.23
1.1
0.51
1.2
0.53
Human equivalent POD, in terms of
environmental exposure (ppm)
Rao and
Brown
(1993)d
59
18
68
44
100
59
11
7
32
16
35
16
Reitz et al.
(1996)e
14
4.2
16
11
24
14
2.6
1.6
7.5
3.7
8.2
3.8
Bois et al.
(1996)f
5.9
1.8
6.8
4.4
10
5.9
1.1
0.7
3.2
1.6
3.5
1.6
aModel: multistage model, extra risk. Coefficients estimated in terms of mg-equivalents/kg-day, as estimated for the experimental animals, and adjusted to
  estimate equivalent continuous exposure.
bBioassay estimates illustrated in Figures 5-8a through 5-13a.
°Human equivalent points of departure were derived by dividing the bioassay estimate by [70 kg/ animal body weight (kg)]0 25. Animal body weights provided
 in Tables 5-5 and 5-7.
dAt exposures below 1 ppm, approximately 0.033 (mg-eq/kg-day)/ppm inhaled tetrachloroethylene was estimated to be metabolized at steady state (see
  Figure 3-10).
eAt exposures below 1 ppm, approximately 0.14 (mg-eq/kg-day)/ppm inhaled tetrachloroethylene was estimated to be metabolized at steady state (see
Figure 3-10).
fAt exposures below 0.1 ppm, approximately 0.33 (mg-eq/kg-day)/ppm inhaled tetrachloroethylene was estimated to be metabolized at steady state (see
Figure 3-10).

-------
1
2
3
4
5
Table 5-10. Human equivalent risk per unit concentration, in terms of
continuous environmental exposure, derived using total tetrachloroethylene
metabolites as the dosimeter; tumor incidence data from JISA (1993) and
NTP (1986)
Tumor type
Hepatocellular
adenomas or
carcinomas
(JISA, 1993)
Hemangiosarcomas,
spleen or liver
(JISA, 1993)
Mononuclear cell
leukemia, rats
(JISA, 1993)
Kidney tumors
(NTP, 1986)
Group
Male mice
Female mice
Male mice
Male rats
Female rats
Male rats
Concentrations above which these risks per
unit concentration should not be used due
to nonlinearity of metabolism and dose-
response
Human equivalent risk per unit concentration", continuous
environmental exposure, (ppm) *
Rao and Brown
(1993)
5.7 x ID'3
2.3 x lO'3
1.7 x 10'3
1.4 x ID'2
6.4 x ID'3
3.1 x 10'3
1 ppm
Reitz et al.
(1996)
2.4 x ID'2
9.6 x lO'3
7.2 x 10'3
6.1 x ID'2
2.7 x ID'2
1.3 x lO'3
0.1 ppm
Bois et al.
(1996)
5.7 x ID'2
2.3 x lO'2
1.7 x lO'2
1.4 x lO'1
6.4 x ID'2
3.1 x 10'2
1 ppm
       aRisk per unit concentration calculated by dividing the risk level by the lower bound on its risk-specific
         environmental concentration. See Table 5-9 for the risk levels and risk-specific concentrations.
                  This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                    5-63        DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
       Table 5-11. Dose-response summary and cancer risk estimates using
       continuous equivalent administered tetrachloroethylene levels as dosimeter,
       from NTP (1986) and JISA (1993)
Tumor type
Hepatocellular
adenomas or
carcinomas, male
mice
Hepatocellular
adenomas or
carcinomas, female
mice
Hemangiosarcomas,
spleen or liver,
male mice
Mononuclear cell
leukemia, male rats
Mononuclear cell
leukemia, female rats
Kidney tubular cell
adenoma or
adenocarcinoma,
male rats
Source
JISA
(1993)
JISA
(1993)
JISA
(1993)
JISA
(1993)
JISA
(1993)
NTP
(1986)
Modeling summary"
MLE dose
coefficients
q0 = 0.34
qj= 0.013
q3 = 7.8 x ID'6
qo = 0.056
q1 = 0.0076
q2 = 3.6 x lO'4
qo = 0.041
qj = 0.0041
q0 = 0.25
qj= 0.0051
qo = 0.26
qj= 0.0017
qo = 0.022
q1 = 9.6 x ID'4
/7-value
0.17
0.83
0.27
0.51
0.34
0.75
MLE PODb
(ppm)
BMC10 = 8.1
BMC05 = 5.4
BMC05=12
BMC05 = 10
BMC05 = 29
BMC05 = 53
Lower bound
on POD
(ppm)
BMCL10 = 2.8
BMCL05 = 2.1
BMCLoi = 6.9
BMCL05 = 6.4
BMCL05 = 13
BMCL05 = 24
Risk per unit
concentration
(ppm)la'c'd
3.6 x ID'2
2.4 x ID'2
7.2 x 10'3
7.8 x ID'3
3.8 x ID'3
2.1 x ID'3
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
aUsing dose coefficients in terms of administered ppm of tetrachloroethylene adjusted to equivalent continuous
  exposure, consistent with RfC methodology (U.S. EPA, 1994), and the multistage model, extra risk:
  P(d) = 1 - exp( - q0 - qi x d - q2 x d2 x . . . q6 x d6).  See Tables 5-5 and 5-7 for input data.
kpOD results (MLEs and lower bounds) illustrated in Figures 5-8b through 5-13b.
Consistent with 2005 cancer guidelines; risk per unit concentration calculated by dividing the appropriate risk level
  by its risk-specific total metabolite level.
dRisks per unit concentration, which are approximations for low-dose extrapolation, should not be used with
  exposures greater than the POD from which they were derived without considering the curvature of the dose-
  response function (at left).
Dose-response modeling of the male mouse liver tumor data using administered exposure fit the
data points as well as when using total metabolism, with the control and lowest exposure groups
again having the poorest fit. This dose-response modeling led to a human equivalent POD
(BMCLio) of 2.8 ppm tetrachloroethylene in air (see Table 5-11 and Figure 5-8b). The
corresponding central tendency estimate was approximately threefold higher, at 8.1 ppm. Linear
                    This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                     5-64        DRAFT-DO NOT CITE OR QUOTE

-------
                                                                                           _2
 1      extrapolation from this POD led to a human equivalent risk per unit concentration of 3.6 x  10
 2      per ppm, about twofold lower than the upper end of the range obtained using total metabolism.
 3            Hepatocellular tumors, female mice.  The dose-response modeling of the hepatocellular
 4      adenomas or carcinomas in female mice from the JISA bioassay and the consideration of total
 5      metabolism (via the PBPK models) led to human equivalent PODs (BMCLios) ranging from 4.4
 6      ppm (Bois et al. [1996] model) to 44 ppm (Rao and Brown [1993] model) tetrachloroethylene in
 7      air (see Table 5-9; Figure 5-9a). The corresponding central tendency estimates are
 8      approximately 1.5-fold higher,  at 6.8-68 ppm.
 9            Linear extrapolation from the PODs above for hepatocellular tumors in female mice was
10      carried out because of the lack of information supporting another extrapolation approach. This
11      led to risks per unit concentration that were approximately 2.5-fold lower than those for the male
12      mice, at 2.3 x  10"3 per ppm (Rao and Brown [1993] model 0.1/44 = 0.0023) to 2.3 x 10"2 per ppm
13      (Bois et al. [1996] model; 0.1/0.4.4 = 0.023; see Table 5-9 and Figure 5-9a).
14            The dose-response modeling results from these same tumor data—but using administered
15      inhalation exposure as the dose metric (without PBPK modeling)—led to a human equivalent
16      POD (BMCL05) of 2.1 ppm (see Table 5-11 and Figure 5-9b).  Note that, because the range of
17      experimental data extended below 10% extra risk, the risk per unit concentration was based on
18      5% extra risk.  The corresponding central tendency estimate is approximately 2.5-fold higher, at
19      5.4 ppm. Linear extrapolation from this POD led to a risk per unit concentration of 2.4 x 10"2 per
20      ppm (0.05/2.1  = 0.024), which is virtually identical to the upper end of the risk per unit
21      concentration range obtained using total metabolism as the dose metric.
22            The dose-response relationship in terms of administered exposure (Figure 5-9b) appears
23      somewhat more linear than when expressed in terms of metabolized dose (Figure 5-9a), but the
24      PODs  relying on the PBPK models have relatively narrower confidence intervals. However,  the
25      confidence intervals associated with both dose  metrics are fairly typical of adequate dose-
26      response fits, and neither dose metric is clearly better on a purely empirical basis.
27
28      5.4.5.1.1.2. Hemangiosarcomas.  Hemangiosarcomas of the liver and spleen were also observed
29      in the JISA male mice. Because these tumors differ etiologically from the hepatocellular
30      adenomas and carcinomas, they were modeled  separately. Dose-response modeling using total
31      metabolism led to human equivalent PODs (BMCLios) ranging from 5.9  ppm (Bois et al. [1996]
32      model) to 59 ppm (Rao and Brown [1993] model) tetrachloroethylene in  air (see Table 5-9;
33      Figure 5-10a). The corresponding central tendency estimates are approximately  1.7-fold higher,
34      at 10-100 ppm.
                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   5-65        DRAFT-DO NOT CITE OR QUOTE

-------
 1             Linear extrapolation from the PODs above for hemangiosarcomas in male mice led to
                                                                         -3
 2      human equivalent risk per unit concentration ranging from 1.7 x 10" per ppm (Rao and Brown
                                                 -v-2
 3      [1993] model; 0.1/59 = 0.0017) to 1.7 x 10'z per ppm (Bois et al.  [1996] model; 0.1/5.9 = 0.017)
 4      tetrachloroethylene in air (see Table 5-10), approximately 3.5-fold less than the corresponding
 5      risks per unit concentration for the other male mouse liver tumors. These results raise some
 6      concern that total cancer risk based on the male mice data may be underestimated slightly by
 7      considering only the hepatocellular adenomas and carcinomas.  An analysis combining the risks
 8      from these two sites indicates an overall risk from the male mice data of 3.4 x 10"2 per
 9      mg-equiv/kg-day, about 20% higher than  the risk per unit concentration estimated for
10      hepatocellular adenomas and carcinomas  alone.10
11             Dose-response modeling using human equivalent continuous administered concentration
12      led to a human equivalent POD (BMCLos) of 6.9 ppm tetrachloroethylene in air (see Table 5-11
13      and Figure 5-10b). The corresponding central tendency estimate is approximately 1.7-fold
14      higher,  at 12 ppm.  However, although this fit was technically adequate (p > 0.1), the model  did
15      not fit as well as the model using total metabolism in the region of the low and middle exposures;
16      the dose-response relationship is essentially a straight line between the high-dose group and  the
17      control  group.
               10In order to gain some understanding of the total risk from multiple tumor sites in male mice, a sum of
        risks across tumor sites was considered. This combined risk does not constitute double-counting if it can be
        assumed that the hepatic adenomas and carcinomas were mechanistically independent from the hemangiosarcomas.
        If there is some dependence between the tumor types, then the combined risk would tend to be an overestimate of
        the total risk.
               A statistically appropriate approach was used to sum the maximum likelihood estimates (MLE) of unit
        potency across these tumor sites for male mice in the JISA study, assuming independence of the tumor sites.
        Specifically, an estimate of the 95% upper bound on the summed unit risk, corresponding to the region of 10~4 extra
        risk in the two dose-response curves, where the slopes were reasonably constant and stable. Assuming a normal
        distribution for the individual risk estimates, the variance of the risk estimate for each tumor site can be derived
        from its 95% upper confidence limit (UCL) according to the formula

                                   95% UCL = MLE + 1.645 x standard error (MLE)                        (1)

        after solving for the standard error

                                   standard error (MLE) = (95% UCL - MLE)/1.645                        (2)

        where 1.645 is the z-statistic corresponding to a one-sided 95% confidence interval. Then the result is squared the
        result to obtain the variance of each MLE.
               The variances of the MLEs for the two tumor sites were summed to obtain the variance of the sum of the
        MLEs. Then the standard error of the summed risk was obtained by taking the square root of the variance. The
        95% UCL on the sum of the MLEs was then calculated using equation (1) above.
               The resulting 95% UCL on the summed unit risk  was 3.4 x 10~2 per mg-equiv/kg-day, about 20% higher
        than the unit risk estimated at the POD at 10% for hepatocellular adenomas and carcinomas alone. That is, at 3.6
        mg-equiv/kg-day (see Table 5-8), the extra risk for hepatocellular adenomas and carcinomas in male mice is 2.8
        x 10"2 per mg-equiv/kg-day (0.1/3.6 mg-equiv/kg-day).
                    This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                      5-66        DRAFT-DO NOT CITE OR  QUOTE

-------
 1            Linear extrapolation from the POD based on administered concentration led to a human
 2      equivalent risk per unit concentration of 7.2 x 10"3 per ppm (0.01/1.4 = 0.0074)
 3      tetrachloroethylene in air (see Table 5-11), approximately twofold less than the risk per unit
 4      concentration obtained for male mouse hepatocellular tumors and administered concentration.
 5
 6      5.4.5.1.2. Rat leukemias.
 1      5.4.5.1.2.1.  Male rats. Table 5-9 provides the dose-response model coefficients for the curve fit
 8      of the male rat MCL data shown in Figure 5-1 la.  The dose-response modeling using total
 9      metabolism (via the PBPK models) led to human equivalent PODs (BMCLios) ranging from 0.7
10      ppm (Bois et al.  [1996] model)  to 7 ppm (Rao and Brown [1993] model) tetrachloroethylene in
11      air. The corresponding central tendency estimates are approximately 1.5-fold higher, at 1.1-11
12      ppm.
13            Linear extrapolation from the PODs above for MCL in male rats led to human equivalent
14      risks per unit concentration ranging from 1.4 x 10"2 per ppm (Rao and Brown [1993] model;
15      0.1/7 = 0.014) to 1.4 x 10'1 per ppm (Bois et al. [1996] model; 0.1/0.7 = 0.14)
16      tetrachloroethylene in air (see Table 5-10).
17            The dose-response modeling results from these same tumor data but using administered
18      inhalation exposure as the dose metric (without PBPK modeling) led to a human equivalent POD
19      (BMCLos) of 6.4 ppm (see Table 5-11 and Figure 5-1 Ib). The corresponding central tendency
20      estimate is approximately 1.5-fold higher, at 10 ppm. Linear extrapolation from this POD led to
21      a risk per unit concentration of 7.8 x 10"3 per ppm (0.05/6.4 = 0.0073). Although the model fit in
22      terms of administered concentration was technically adequate (p = 0.51), the model
23      overestimated the control response and underestimated the mid-dose group response, leading to a
24      risk per unit concentration 2- to 20-fold lower than those obtained using total metabolism as the
25      dose metric.
26
27      5.4.5.1.2.2.  Female rats. It was noted earlier (see Section 5.4.2.2) that the dose-response pattern
28      of MCL for female rats in the JISA study also was not monotonic.  In this case, the "best" model
29      fit in terms of both dose metrics (see Figures 5-12a and 5-12b) provided adequate fits overall
30      (p = 0.48 and 0.27), but fit the control and low-exposure group responses least well.  The model
31      fit in terms of metabolized dose provided a better fit of the control response, although both dose
32      metrics lead to approximately the same estimated response for the low dose, at -27%, compared
33      with the 34% observed. Although both models would appear to underestimate extra risk in this
34      region of the dose-response for female rat leukemias, it is not clear in this particular set of
                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   5-67       DRAFT-DO NOT CITE OR QUOTE

-------
 1      responses that the fit to the low dose should be emphasized over fitting as many of the responses
 2      as possible.
 3            The dose-response modeling of MCL in female rats using total metabolism (via the
 4      PBPK models) led to human equivalent PODs (BMCLi0s) ranging from 1.6 ppm (Bois et al.
 5      [1996] model) to 16 ppm (Rao and Brown [1993] model) tetrachloroethylene in air.  The
 6      corresponding central tendency estimates are twofold higher, at 3.2-32 ppm.
 7            Linear extrapolation from the PODs above for MCL in female rats led to human
 8      equivalent risks per unit concentration ranging from 6.4 x 10"3 per ppm (Rao  and Brown [1993]
 9      model; 0.1/7 = 0.014) to 6.4 x 10"2 per ppm (Bois et al. [1996] model; 0.1/0.7 = 0.14)
10      tetrachloroethylene in air (see Table 5-10).  The dose-response modeling using administered
11      inhalation exposure as the dose metric led to a human equivalent POD (BMCLio) of 13 ppm (see
12      Table 5-11 and Figure 5-12b). The corresponding central tendency estimate is approximately
13      twofold higher, at 29 ppm. Linear extrapolation from this POD led to a risk per unit
14      concentration of 3.8 x 10"3 per ppm (0.1/13 = 0.0038), about 1.6- to 16-fold lower than those
15      obtained using total metabolism as the dose metric. Although the risks per unit concentration for
16      female rats were about twofold lower than those for the male rats, this relationship among the
17      dose metrics is very similar to that seen with male rat MCL.
18            There was an indication of accelerated occurrence of leukemias in female rats in the NTP
19      study,  but the addition of time-to-tumor in the multistage model did not significantly affect the
20      estimate from that study. There was no similar observation of earlier leukemia incidence with
21      increasing exposure in the JISA  study.
22
23      5.4.5.1.3. Rat kidney tumors. Table 5-9 provides the dose-response model coefficients for the
24      curve fit of the male rat kidney adenocarcinomas and carcinomas seen in the NTP study (see
25      Figure 5-13a). The dose-response modeling using total metabolism (via the PBPK models) led
26      to human equivalent PODs (BMCL0ss) ranging from 1.6 ppm (Bois et al. [1996] model) to 16
27      ppm (Rao and Brown [1993] model) tetrachloroethylene in air. Note that, because the 10% extra
28      risk response level fell above the range of experimental data, the POD was based on 5% extra
29      risk which the available data did span. The corresponding central tendency estimates are
30      approximately twofold higher than their lower bounds, at 3.5-35 ppm.
31            Linear extrapolation from the PODs above for kidney tumors in male  rats led to human
32      equivalent risks per unit concentration ranging from 3.1 x 10"3 per ppm (Rao  and Brown [1993]
33      model; 0.05/1.6 = 0.0031) to 3.1 x 10'2 per ppm (Bois et al. [1996] model; 0.05/16 = 0.031)
34      tetrachloroethylene in air (see Table 5-10).  These risks per unit concentration were the lowest of
35      those estimated for all sites.

                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   5-68       DRAFT-DO NOT CITE OR QUOTE

-------
 1            The dose-response modeling results from these same tumor data but using administered
 2      inhalation exposure as the dose metric (without PBPK modeling) led to a human equivalent POD
 3      (BMCLos) of 24 ppm (see Table 5-11, Figure 5-13b).  The corresponding central tendency
 4      estimate is approximately twofold higher, at 53 ppm. Linear extrapolation from this POD led to
 5      a risk per unit concentration of 2.1 x 10"3 per ppm (0.05/24 = 0.0021).
 6
 7      5.4.5.1.4. Summary and discussion of site-specific dose response modeling.  Dose-response
 8      modeling of the candidate data sets presented no particular difficulties.  As noted in the
 9      preceding descriptions of modeling results, lower bounds on the central tendency estimates
10      (maximum likelihood estimate [MLEs]) of the PODs tended to be within twofold of the central
11      estimates. The only exception was for male mice, with a threefold difference between the MLEs
12      and their lower bounds.
13            The slopes of the dose-response curves at the PODs were estimated and found to be
14      within 1.6-fold of the corresponding risks per unit concentration in all cases, reflecting the
15      mostly low-dose linear dose-response relationships estimated within the lower region of the
16      observed data ranges. Because of the similarity of the slopes to the risks per unit concentration
17      and the apparent lack of potential for sublinear dose response behavior in the range of exposure
18      below the experimental  data, these slopes are not shown.
19            Figure 5-14 shows the relative magnitudes of the risks per unit concentration associated
20      with each tumor site.  It is interesting to note that the risks per unit concentration estimated using
21      administered concentration are not consistently the lowest or highest risk values among the
22      different estimates for each tumor site.
23            For example, the risk per unit concentration estimated from female mouse hepatocellular
24      tumors using administered concentration (2.4 x 10"2 per ppm) is approximately equal to the
25      upper end of the range estimated using metabolized tetrachloroethylene (2.3 x 10"2 per ppm).
26      Similarly for the male mice, the risk per unit concentration using administered concentration is
27      about twofold lower than the upper end of the range using metabolized dose (3.6 x 10"2 per ppm
28      vs. 5.7 x 10"2 per ppm, respectively).  In contrast, the risks per unit concentration for MCL
29      estimated using administered concentration (7.8 x 10"3 per ppm, males;  3.8 x 10"3 per ppm,
30      females) are about twofold lower than the lower end of the range estimated using metabolized
31      tetrachloroethylene (1.4 x 10"2 per ppm, males;  6.4 x 10"3 per ppm, females).  Some of this
32      variation is attributable to the differing shapes of the dose-response curves for the two different
33      dose metrics for each site and the variability in the bioassay responses.  Overall, the interleaving
34      of the results from the two types of dosimetric, administered concentration and PBPK-estimated
35      metabolism, underscores some uncertainty in identifying the appropriate dosimetric(s).

                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  5-69       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2      5.4.5.1.5. Concordance of animal and human risk estimates. Although sufficient human data
 3      linked with exposure characterizations are not available to derive cancer risk values, an analysis
 4      by van Wijngaarden and Hertz-Picciotto (2004) provides a limited perspective on the human
 5      cancer risk values estimated from animal bioassays. Van Wijngaarden and Hertz-Picciotto
 6      (2004) demonstrated a simple methodology using epidemiologic data for four chemical
 7      exposures including tetrachloroethylene. For tetrachloroethylene specifically, a linear dose-
 8      response model was fit to laryngeal cancer observations in the upper airway cancer case-control
 9      study of Vaughan et al. (1997).  Van Wijngaarden and Hertz-Picciotto (2004) presented both an
10      EDoi and LEDoi (effective dose for a 1% additional lifetime risk over background and the lower
11      confidence interval on this dose, called the TD1 and LCL1 in their paper) for humans exposed
12      for 45 years, 240 days/year, a standard occupational exposure scenario.  The ED0i was 228.40
13      mg/day and LEDoi was 60.16 mg/day. In order to compare these results with those derived from
14      the JISA (1993) study, we assumed a continuous lifetime exposure (70 years, 365 days/year,  and
15      20 mVday breathing rate), resulting in an equivalent ED0i of 4.8 mg/m3 and LED0i of 1.3 mg/m3.
16      Using the continuous lifetime equivalent LEDoi as the POD and a low-dose linear approach,  a
17      risk per unit concentration based upon Vaughan et al. (1997) is 8 x 10"6 per |ig/m3 (or,  0.01/1.3
18      x io3 |ig/m3). This estimate falls in the lower end of the range of cancer risk estimates  from
19      male and female rat MCL tumors in JISA (1993). A cancer risk estimate from human data using
20      the EDoi as the POD is 2 x 10'6 per |ig/m3 (or, 0.01/4.8 x IO3  |ig/m3).
21            While the analysis of van Wijngaarden and Hertz-Picciotto (2004) can provide some
22      insight on the rodent-based tetrachloroethylene cancer risk estimate, it is still quite limited due to
23      the possible biases in Vaughan et al. (1997) and other factors. While individual bias in Vaughan
24      et al. (1997) may influence observed risk estimates from this study in either a positive
25      (overestimate) or null (underestimate) direction, the overall direction of all bias is likely toward
26      the null. First, Vaughan et al. (1997) do not have exposure information on individual cases and
27      controls and make an assumption that case and controls are exposed to tetrachloroethylene
28      concentrations as described by industrial hygiene surveys in dry cleaning establishments. For
29      this reason, bias related to exposure misclassification is likely great in this study.  Second, as is
30      common to many population case-control studies, exposure prevalence to tetrachloroethylene is
31      low. Only 5 of 235 laryngeal cancer cases were identified as having exposure to
32      tetrachloroethylene, and 4 of these  5 cases as more likely than not as being exposed. Low
33      exposure prevalence may lead to reduced study power lead and imprecise estimates of the
34      relative risk (OR) that are not statistically significant.  Last, epidemiologic  evidence is available
35      to suggest an association between esophageal cancer and tetrachloroethylene, a site also

                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   5-70        DRAFT-DO NOT CITE OR QUOTE

-------
 1      examined by Vaughan et al. (1997).  The OR between esophageal cancer and tetrachloroethylene
 2      exposure was larger than that for laryngeal cancer, OR =11.9 (95% CI = 1.1-124). Cancer risk
 3      estimates based on esophageal cancer observations would lead to a higher estimate than that
 4      identified by van Wijngaarden and Hertz-Picciotto (2004). Also, Vaughan et al. (1997) only
 5      considered respiratory cancers; given the epidemiological results discussed earlier these results
 6      may represent an underestimate of total risk. An examination of site concordance with animal
 7      observations, additionally, is not possible because the rat is a poor model for laryngeal cancer.
 8      Odds ratios in Vaughan et al. (1997) are adjusted for a number of possible confounders such as
 9      age, sex, education, study period, alcohol consumption, and cigarette smoking, and the use of
10      adjusted odds ratios is a strength of the van Wijngaarden and Hertz-Pi cci otto (2004) analysis.
11
12      5.4.5.2. Recommended Inhalation Unit Risk
13            Human inhalation cancer risk has been assessed using several different gender-species
14      animal tumor data sets and three different human PBPK models  of total metabolism rate.  These
15      results have been discussed above and are summarized in Figure 5-14.
16            In choosing which species-sex combination is most relevant for extrapolating to humans,
17      the MOA information does not provide a clear rationale. Although target organ concordance is
18      not a prerequisite for evaluating the implications of animal study results  for humans (U. S. EPA,
19      2005a), it is notable that the leukemias (in both sexes of rats) support the observation of
20      lymphopoietic cancers in individuals employed as dry cleaners and degreasers, and the liver
21      tumors (in both sexes of mice) support the observation of liver tumors in dry cleaners (see
22      Section 4.10.1.1.2).
23            The male rat leukemia data provide the most sensitive response of the four
24      species-sex combinations in the JISA study for deriving a unit risk, defined as the
25      plausible upper-bound excess lifetime cancer risk estimated to result from continuous
26      exposure to tetrachloroethylene per unit of concentration. From  Table 5-10, the
27      recommended unit risk value range is 1.4  x 10~2 to 1.4 x 10"1 per ppm, or 2 x 10~6 to
28      2 x 10~5 per ug/m3. This range reflects uncertainty in the choice of pharmacokinetic
29      model.
30            Comparison with previous EPA assessment:  EPA (U.S. EPA, 1986,  1991) reported an
31      overall unit risk  of 5.8 x 10"7 per |ig/m3 (3.9  x 10"3 per ppm), which was a geometric mean of six
32      risks per unit concentration from the 1986 NTP study: male and female  rat leukemias, male and
33      female mouse liver carcinomas, and male and female mouse liver adenomas and carcinomas.
34      The highest risk per unit concentration in that range was 9.5 x 10"7 per |ig/m3, corresponding to
35      the leukemias in male rats from the NTP study (using total metabolism as the dosimeter).

                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  5-71        DRAFT-DO NOT CITE OR QUOTE

-------
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
                               O-I---A----
                         O---A	O-
                                      ..-.{V}	[7]   MCL, male rats
                                        -Q---O---B
                  A----
          A--O
                                •O	H>
                             -Q
                                    a
MCL, female rats


Hepatocellular tumors, male mice


Hepatocellular tumors, female mice

Hemangiosarcomas, male mice

             I
Kidney tumors,(male rats
            0.001               0.01                0.1                 1
                     Human equivalent tetrachloroethylene unit risks (per ppm)
                                                                                  10
       Figure 5-14.  Comparison of inhalation risks per unit concentration for
       tetrachloroethylene derived from rodent bioassays using four different dose
       metrics—continuous equivalent inhalation concentration (0), Bois et al. (1996)
       PBPK model (n), Reitz et al. (1996) PBPK model (o), and Rao and Brown (1993)
       PBPK model (A).  See Table 5-9 for PBPK model-derived estimates and Table 5-
       10 for estimates relying on administered tetrachloroethylene.
       This analysis supports a unit risk 14-fold higher than in EPA's 1991 assessment.  This
difference is attributable to number of considerations.  A comparison of the results from the two
bioassays, using the Reitz et al. (1996) model to characterize internal dose for both data sets but
not extrapolating to humans, indicates that the JISA study leads to risks per unit concentration
that are approximately twofold lower than those from the NTP study (not shown), if all else can
be considered equal. The remaining differences between the human equivalent inhalation risks
per unit concentration are attributable to differences in the particular PBPK models used, the
change in cross-species scaling factor from BW2/3 to BW3/4 (U.S. EPA, 1992),  and use of the
most sensitive response rather than a (geometric) mean of the significant tumor responses.
Concerning the latter decision, use of a mean response treats the observations as if all are equal
likely alternatives (in the case of geometric means, the highest responses are disproportionately
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   5-72       DRAFT-DO NOT  CITE OR QUOTE

-------
 1      discounted relative to the lower responses). Use of the most sensitive response acknowledges
 2      the weight contributed by all of the observed responses as independent indicators of human risk,
 3      and provides a plausible upper bound on potential human risk.
 4
 5      5.4.5.3. Recommended Oral Slope Factor
 6            The oral slope factor was developed from inhalation data because the only available oral
 7      bioassay was less relevant for extrapolating to lifetime risk in humans, for several reasons. First,
 8      the study was conducted by gavage at relatively high doses.  Human exposures are more likely
 9      not to occur in boluses, and high doses are associated at least with saturable metabolism
10      processes which may involve a different profile of toxicological processes than those prevalent at
11      more likely environmental exposure levels. Also, the animals were dosed for only
12      approximately 75% of the more usual 2-year period (NCI, 1977), making the oral study less
13      useful for estimating lifetime risk. Route-to-route extrapolation from the inhalation PODs
14      developed from the JISA study (see Table 5-9) was carried out using the human pharmacokinetic
15      models described in Section 3.5. Table 5-12 summarizes the resulting slope factors. Because the
16      oral slope factors are linear conversions of the inhalation risks per unit concentration, no figure
17      analogous to Figure 5-14 is provided; such a figure would be identical to Figure 5-14 with the
18      exception that the x-axis would reflect mg/kg-day units rather than ppm units.
19            The same arguments that led to selecting the range based on male rat leukemias
20      for the inhalation unit risk apply to the oral slope factor. In order to  account for the
21      uncertainty contributed by the human PBPK models, the oral slope factor is given by the
22      range 1 x 10"2 to 1 x 10"1 per mg/kg-day.  This range is equivalent to drinking water risks
23      per unit concentration of 4 x 10"7 to 4 x 10"6 per |ig/L of tetrachloroethylene in water
24      (assuming 70 kg body weight and a daily water consumption of 2 L/day). The
25      recommended slope factor range is 1  x 10~2 to 1 x  10"1 per mg/kg-day. This range
26      reflects uncertainty in the choice of pharmacokinetic model.
27            Comparison with previous EPA assessment:  EPA (U.S. EPA, 1985) reported a slope
28      factor of 5.1 x 10"2 per mg/kg-day, based on the liver tumor incidence in female mice in the NCI
29      (1977) oral  gavage study, total metabolized dose, and BW2/3 cross-species scaling.  This value
30      falls near the center of the range developed in the current assessment.
31
32      5.4.5.4.  Quantitative Adjustment for Sensitive Populations
33            Although a mutagenic MO A would indicate increased early-life susceptibility, there are
34      no data exploring whether there is differential sensitivity to tetrachloroethylene carcinogenicity
35      across life stages.  This lack of understanding about potential differences in metabolism and
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                    5-73       DRAFT-DO NOT CITE OR QUOTE

-------
  1
  2
  3
  4
      Table 5-12.  Summary of tetrachloroethylene oral slope factors, estimated from
      dose-response modeling of inhalation-exposed animals and by extrapolation to oral
      exposure using pharmacokinetic models
Tumor type
Hepatocellular adenomas
or carcinomas, mice
Hemangiosarcomas, male
mice
Mononuclear cell
leukemia, rats
Kidney tumors
Group
Male
mice
Female
Mice
Male
mice
Male rats
Female
rats
Male rats
(NTP)
POD (and extra
risk), in terms of
human equivalent
metabolized dose
(mg-eq/kg-day) a
0.58
1.5
1.9
0.23
0.51
0.53
(10%)
(10%)
(10%)
(10%)
(10%)
(5%)
Concentrations above which these risks per unit concentration
should not be used due to nonlinearity of metabolism and dose-
response.
Oral slope factor (mg/kg-day) * b
Rao and
Brown
(1993)
model c
5.6 x 10'3
2.2 x lO'3
1.7 x 10'3
1.4 x 1Q-2
6.3 x 1Q-3
1.5 x lO'3
1 mg/kg-day
Reitz et al.
(1996)
model d
2.4 x 10'2
9.4 x lO'3
7.0 x 10'3
5.9 x 1Q-2
2.7 x 1Q-2
6.2 x lO'3
1 mg/kg-day
Bois et al.
(1996) model6
5.3 x 10'2
9.4 x lO'3
1.6 x 10'2
1.4 x 1Q-1
6.3 x 1Q-2
1.5 x lO'2
1 mg/kg-day
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
See Table 5-9 for derivation of human equivalent metabolite estimates.
Points of departure in the previous column were converted to human equivalent oral doses using the
pharmacokinetic models detailed below (intermediate calculation not shown), then converted to risks per unit
concentration by dividing extra risk by the corresponding risk-specific oral doses.
At exposures below about 1 mg/kg-day (divided between nine equally spaced doses during waking hours),
approximately 0.033 (mg-eq/kg-day)/(mg/kg-day) of ingested tetrachloroethylene was estimated to be
metabolized at steady-state, using the Rao and Brown (1993) model modified for oral exposure (see Section 3.5
and Figure 3-13)  The conversion factor was calculated by dividing 0.01 mg-eq/kg-day (the total metabolite
production in 24 hrs) by the total oral dose of tetrachloroethylene estimated to produce that level of metabolites,
21mg/70kg.
At exposures below about 1 mg/kg-day (divided between nine equally spaced doses during waking hours),
approximately 0.14 (mg-eq/kg-day)/(mg/kg-day) of ingested tetrachloroethylene was estimated to be metabolized
at steady-state, assuming  that the proportional relationship observed between the Rao and Brown (1993) model
and the Reitz et al. (1996) model for the inhalation route holds for oral exposure (see Section 3.5 and Figure 3-
13).  The conversion factor was calculated by dividing 0.01 mg-eq/kg-day (the total metabolite production in 24
hrs) by the total oral dose of tetrachloroethylene estimated to produce that level of metabolites, 5.1 mg/70 kg.
At exposures below about 0.1 mg/kg-day (divided between nine equally spaced doses during waking hours),
approximately 0.31 (mg-eq/kg-day)/(mg/kg-day) of ingested tetrachloroethylene was estimated to be metabolized
at steady-state, assuming  that the proportional relationship observed between the Rao and Brown (1993) model
and the Bois et al. (1996) model for the inhalation route holds for oral exposure (see Section 3.5 and Figure 3-13).
The conversion factor was calculated by dividing 0.01 mg-eq/kg-day (the total metabolite production in 24 hrs) by
the total oral dose of tetrachloroethylene estimated to produce that level of metabolites, 2.25 mg/70 kg.
                      This document is a draft for review purposes only and does not constitute Agency policy
         06/06/08                                        5-74        DRAFT-DO NOT CITE OR QUOTE

-------
 1      susceptibility across exposed human populations thus represents a source of uncertainty.
 2      Nevertheless, the existing data do support the possibility of a heterogenous response that may
 3      function additively to ongoing or background exposures, diseases, and biological processes. As
 4      noted in Section 4.9.5, there is some evidence that certain subpopulations may be more
 5      susceptible to exposure to tetrachloroethylene.  These subpopulations include early and later life
 6      stages and groups defined by health and nutrition status, gender, race/ethnicity, genetics,  and
 7      multiple exposures and cumulative risk. As discussed below, these considerations strengthen the
 8      scientific support for the choice of a linear non-threshold extrapolation approach. However,
 9      because the MOA for tetrachloroethylene has not been established, it is not appropriate to derive
10      age-adjustment factors for early life exposures, as discussed in Supplemental Guidance for
11      Assessing Susceptibility from Early-Life Exposure to Carcinogens (U.S. EPA, 2005b).
12
13      5.4.6. Discussion of Uncertainties in Cancer Risk Values
14             A number of uncertainties underlie the cancer unit risk for tetrachloroethylene. These are
15      discussed in the following paragraphs.  Specifically addressed is the impact on the assessment of
16      issues such as the use of models and extrapolation approaches, the reasonable alternatives and
17      the choices made and the data gaps identified.  In addition, the use of assumptions, particularly
18      those underlying the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) is explained
19      and the decision concerning the preferred approach is given  and justified. Several of the
20      uncertainties with the largest impact cannot be  considered quantitatively.  Thus an overall
21      integrated quantitative uncertainty analysis is not presented.  Section 5.4.6.1 and Table 5-13
22      summarize principal uncertainties.
23
24      5.4.6.1. Sources of Uncertainty
25      5.4.6.1.1. Human population variability.  The extent of inter-individual variability in
26      tetrachloroethylene metabolism has not been characterized.  As noted above, several enzymes of
27      the oxidative and GSH metabolism, notably CYP2E1, CYP3A4, GSTZ, GSTA, GSTM, and
28      GSTT, show genetic polymorphisms with the potential for variation in production of specific
29      metabolites.  Tetrachloroethylene metabolism has been shown to increase by inducers of
30      CYP450 enzymes such as toluene, phenobarbital, and pregnenolone-16 alpha-carbonitrile,
31      whereas CYP enzyme inhibitors such as SKF 525A, metyrapone, and carbon monoxide have
32      been shown to decrease tetrachloroethylene metabolism. Additionally,  chronic exposure to
33      tetrachloroethylene has been shown to cause  self-induction of metabolism. Human population
34      variability has also been discussed in Chapter 3.
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  5-75        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
       Table 5-13. Summary of uncertainties in tetrachloroethylene cancer unit
       risk estimate
Consideration/
Approach
Human
population
variability in
metabolism and
response/
sensitive
subpopulations
Low-dose
extrapolation
procedure





Dose metric




Species /gender
combination
(see Table 5-14)









PBPK model





Cross-species
scaling





Impact on unit risk
Low-dose risk f to
an unknown extent





Departure from
EPA's Guidelines for
Carcinogen Risk
Assessment POD
paradigm, if justified,
could | or t risk per
unit concentration an
unknown extent
Alternatives could t
or J, risk per unit
concentration by an
unknown extent

Human risk could J,
or t, depending on
relative sensitivity









10-fold range in risk
per unit
concentration among
the three available
models

Alternatives could J,
or t risk per unit
concentration (e.g.,
3.5-fold| [scaling by
BW) or t 2-fold
[scaling by BW2/3])

Decision
Considered
qualitatively





Multistage
model to
determine POD,
linear low-dose
extrapolation
from POD
(default
approach)
Considered
total
metabolism and
administered
concentration
Male rat MCL











All are
considered




BW3/4 (default
approach)





Justification
No data to support range of human
variability/sensitivity, including whether children are
more sensitive. Mutagenic MOA (if established)
would indicate increased early-life susceptibility.



Available MOA data do not inform selection of
dose-response model but do not support non-
linearity (mutagenicity is plausible contributor and
cannot be ruled out); male rat MCL data are linear in
observed range; linear approach in absence of clear
support for an alternative is generally supported by
scientific deliberations supporting EPA's Guidelines
for Carcinogen Risk Assessment.
Experimental evidence supports a role for
metabolism in toxicity, but actual responsible
metabolites are not clearly identified.


MCL is the largest response and is reproducible
across studies, despite high background response
rate. There are no MOA data to guide extrapolation
approach for any choice. It was assumed that
humans are as sensitive as the most sensitive rodent
gender/species tested; true correspondence is
unknown. The carcinogenic response occurs across
species. Generally, direct site concordance is not
assumed; consistent with this view, some human
tumor types are not found in rodents (i.e., cervical,
esophageal cancer deaths) and rat and mouse tumor
types also differ.
There is no scientific basis for choosing among
pharmacokinetic results for estimating total
metabolism of tetrachloroethylene given limitations
in available data. The highest value provides a
reasonable upper estimate of potential human cancer
risk.
There are no data to support alternatives. Because
the dose metric was not an AUC, BW3/4 scaling was
used to calculate equivalent cumulative exposures
for estimating equivalent human risks.


           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                   5-76        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
             Table 5-13.  Summary of uncertainties in tetrachloroethylene cancer unit
             risk estimate (continued)
Consideration/
Approach
Bioassay
Statistical
uncertainty at
POD
Impact on unit risk
t risk per unit
concentration 2-fold
if NTP study used
J, risk per unit
concentration 1.6-
foldif ECio used
rather than LECio
Decision
JISA study
LEC (default
approach for
calculating
plausible upper
bound)
Justification
JISA study used the lowest experimental exposures
(reduces extrapolation uncertainty)
Limited size of bioassay results in sampling
variability; lower bound is 95% confidence interval
on concentration.
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
     LECIO = 95% lower confidence limits on the air concentrations associated with a 10% extra risk of cancer incidence

            A separate issue is that the human variability in response to tetrachloroethylene is also
     poorly understood. The effect of metabolic variation, including potential implications for
     differential toxicity, has not been well studied.
            Although a mutagenic MOA would indicate increased early-life susceptibility, there are
     no data exploring whether there is differential sensitivity to tetrachloroethylene carcinogenicity
     across life stages.  This lack of understanding about potential differences in metabolism and
     susceptibility across exposed human populations thus represents a source of uncertainty.
     Nevertheless, the existing data do support the possibility of a heterogenous response that may
     function additively to ongoing or background exposures, diseases, and biological processes.  As
     noted in Chapter 4 (see Section 4.9.5), there is some evidence that certain subpopulations may be
     more susceptible to exposure to tetrachloroethylene. These subpopulations include early and
     later life stages and groups defined by health and nutrition status, gender, race/ethnicity,
     genetics, and multiple exposures and cumulative risk.  As discussed in the section on low-dose
     extrapolation below, these considerations strengthen the scientific support for the choice of a
     linear non-threshold extrapolation approach.

     5.4.6.1.2.  Choice of low-dose extrapolation approach. The MOA is a key consideration in
     clarifying how risks should be estimated for low-dose exposure. MOA data are lacking or
     limited for all candidate cancer endpoints for tetrachloroethylene (i.e., rat MCL and kidney
     tumors, mouse hepatocellular tumors and hemangiosarcomas). When the MOA cannot be
     clearly defined, EPA uses a linear approach to estimate low-exposure risk, based on the
     following broad and long-term  scientific assumptions, which supported the development of the
     EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   5-77       DRAFT-DO NOT CITE OR QUOTE

-------
 1       •   A chemical's carcinogenic effects may act additively to ongoing biological processes,
 2           given that diverse human populations are already exposed to other agents and have
 3           substantial background incidence of various tumors.
 4
 5       •   A broadening of the dose-response curve in the human population (less rapid fall-off with
 6           dose) and, accordingly, a greater potential for risks from low-dose exposures (see Zeise et
 7           al., 1987; Lutz et al., 2005) would result for two reasons. First, even if there is a
 8           threshold concentration at the cellular level, that threshold is likely to be different among
 9           different individuals.  Secondly, greater variability in response to exposures in the
10           heterogeneous human population would be anticipated than in controlled laboratory
11           species and conditions (due to, e.g., genetic variability, disease states, age).
12
13       •   The use of linear extrapolation provides consistency across assessments as well as
14           plausible upper-bound risk estimates that are believed to be health-protective (U.S. EPA,
15           2005a).
16
17           The extent to which the overall uncertainty in low-dose risk estimation could be reduced
18    if the MOA for tetrachloroethylene were known with a high degree of confidence is of interest,
19    but clear data on the MOA of tetrachloroethylene is not available, and even if it were,
20    incorporation of MOA into dose-response modeling might not be straightforward and might not
21    significantly reduce the uncertainty about low-dose extrapolation.  This is because the MOA as
22    well as other factors, especially human response variability, are determinants of the  dose-
23    response function in humans.
24           This chemical assessment also evaluates the extent to which a collection of mathematical
25    functions, fit to one of the tetrachloroethylene bioassay data sets and extrapolated down to low
26    doses, could inform uncertainty. There is not sufficient information regarding the MOA to
27    support a chemical-specific inference about dose-response behavior at low dose for
28    tetrachloroethylene. Thus, it is of interest to observe how different functions fit to the tumor data
29    may diverge when extrapolated downward. Much previous experience has supported a general
30    mathematical  property that different curves, though fitting observed experimental data well,
31    often diverge widely when extrapolated to doses well outside the observed range.  Indeed,  the
32    inability of curve-fitting procedures to provide useful compound-specific information about low
33    dose risks has been a principal motivation for the "model free" approach of straight  line
34    extrapolation from  a POD within the observed range of the data (Krewski and van Ryzin, 1981;
35    NRC, 1983).
36           Calculations here considered four alternative functional forms frequently used for
37    noncancer dose-response assessment in the observable range of the experimental data
38    (multistage, Weibull, log-logistic, and log-probit). These can accommodate a wide variety of
39    dose-response shapes, including threshold-like behavior.  These models were fit to the MCL data
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-78        DRAFT-DO NOT CITE OR QUOTE

-------
 1    in male rats using the EPA BMDS program without restricting the shape parameters of the latter
 2    three models (see Appendix 5B for more details). Parameters describing the risk distribution
 3    (mean, median and 95% upper and lower confidence bounds) were estimated by a bootstrap
 4    procedure because these parameters are not all readily available in the current BMDS software.
 5    In comparing these distributions from the bootstrap procedure, the mean (as a measure of central
 6    tendency) and the 95% upper and lower bounds calculated from the bootstrap procedure are of
 7    interest.  The resulting risk distributions were compared at two exposure levels—at a generalized
 8    POD and at an environmental level approximately 105-fold lower than the POD.
 9          At POD C, corresponding to a risk of approximately 0.1 using the mean estimate from
10    the multistage model, the bootstrap procedure yielded similar risk distributions for the four
11    models (see Figure 5-15a). The means and corresponding confidence bounds agree within an
12    order of magnitude and the spreads in the distributions (the distance between the upper and lower
13    confidence bounds) are within two orders of magnitude. Note that the probability calculations
14    are in terms of metabolized dose in the male rat and do not directly characterize human risks.
15          EPA also examined the bootstrap results from those same models at a dose that is lower
16    than the POD C by a factor of 105 (although EPA's actual low dose risk estimates are developed
17    using a linear extrapolation from a POD to the origin rather than using low-dose estimates from a
18    model).  Figure 5-15 illustrates these results.  In the region of extrapolated concentrations
19    (C x 10"5), the mean risks  of the latter three models (Weibull, log-logistic and log-probit) are
20    about one to three orders of magnitude higher than the mean of the  multistage model risks. The
21    spreads of all the models are quite broad, with a six order of magnitude 95% confidence interval
22    for the multistage and much greater spreads for the other three models.  The upper bounds of risk
23    for the other three models  are higher than that for multistage model, within about three orders of
24    magnitude, and their lower bounds of risk are much lower than that of multistage, by nine or
25    more orders of magnitude. With such large spreads in confidence intervals, the extrapolated
26    models in effect provide little information about low-dose risks.  The extrapolation of the
27    multistage model does result in estimates reasonably close to the low-dose estimates from the
28    model-independent straight line extrapolation from the POD, in that the mean and upper-bound
29    risks at the lower concentration are both within 10% of the estimates resulting from applying
30    linear extrapolation to the  results at the higher concentration.
31          This comparison of risk distributions has several limitations. First, the selected models
32    do not represent all possible models one might fit, serving primarily to illustrate a range  of
33    possibilities. That is, other models could be selected to yield more  extreme results, both higher
34    and lower than those shown here. Further, the results apply only to the prediction of MCL in
35    male rats. For  reasons discussed above concerning expected additivity to background processes

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-79        DRAFT-DO NOT CITE OR QUOTE

-------
                      •c
                               a. Predicted risks of MCL in male rats, corresponding
                                 to 1.5 mg-eq/kg-day total metabolism
                                   Multistage
Weibull           Log-logistic

   Dose-response model
Log-probit
                      •
                       §, 1.E-03 -
                       § 1.E-04

                      1
                       O1 1.E-05 -

                       Q.

                      £ 1.E-06 -
                                b. Predicted risks of MCL for male rats, corresponding
                                  to 1.5E-5 mg-eq/kg-day total metabolism
                                       LB=1E-12
                                                         LB=0
                                                                           LB=1E-21
                                                                                         Me iian=8E-26
                                                                                         LB: 1E-152
                                   Multistage
Weibull           Log-logistic

   Dose-response model
Log-probit
 1
 2             Figure 5-15. Illustration of sensitivity to model selection for low-dose extrapolation. The risk
 3             distributions associated with four dose-response models which adequately fit the
 4             tetrachloroethylene dose-response data for MCL in male rats (JISA, 1993) were compared.  The
 5             mean (0) and median (n) risks for each model are indicated with symbols, and the 5th and 95th
 6             percentiles are indicated by bars. Risks are in terms of metabolized tetrachloroethylene in male
 7             rats. Figure a shows the comparison at a generalized POD, selected as the mean exposure estimate
 8             from the multistage model corresponding to a risk of approximately 0.1—that is, 1.5 mg-eq/kg-
 9             day, equivalent to about 50 ppm as administered in the bioassay. Figure b compares risk
10             distributions at an exposure corresponding to  an environmental concentration of
11             tetrachloroethylene-approximately  105-fold lower than the POD, or 1.5 x 10~5 mg-eq/kg-day,
12             which is equivalent to about 50 x 10"5 ppm if  administered as in the bioassay.  Note that three
13             lower bounds (Weibull, log-logistic, and log-probit) and one median (log-probit) could not be
14             plotted on the graph. See Appendix 5B for more details.
                    This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                           5-80         DRAFT-DO NOT  CITE OR QUOTE

-------
 1    in humans and greater heterogeneity of human populations, more linear fits are expected to better
 2    capture the anticipated response in a human population.  The low-dose extrapolation to humans
 3    from threshold-like models (i.e., log-logistic and log-probit) carries a relatively greater degree of
 4    uncertainty than extrapolation from the multistage and Weibull fits.  These calculations illustrate
 5    the expected finding that alternative functional forms fit to the tetrachloroethylene tumor data
 6    yield a wide range of numerical values for probability of response when extrapolated down to
 7    low dose and are uninformative of the actual risk.
 8          Given the current state of scientific knowledge about tetrachloroethylene carcinogenicity,
 9    the straight line based risk estimates presented above form the preferred recommendation for
10    estimating a plausible upper-bound estimate of potential human risks from tetrachloroethylene.
11    This approach is  supported by both general scientific considerations, including those  supporting
12    the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), as well as chemical-specific
13    findings. The former include the scientific principles articulated above (the expectation that a
14    chemical functions additively to background exposures, diseases, and processes, that  variability
15    within the human population would broaden the dose-response curve and eliminate individual
16    thresholds if present, and that the approach provides consistency across assessments facilitating
17    direct comparison of the derived risk values). The latter includes evidence that, within the dose
18    range of the cancer bioassays,  the observable tumor response data are consistent with a linear
19    model and do not suggest occurrence of a threshold, and that variability in the human response
20    across the population is expected (see Human population variability, above).
21
22    5.4.6.1.3. Dose metric. Tetrachloroethylene is metabolized to several intermediates with
23    carcinogenic potential.  Although much data exist for TCA, several analyses indicate that TCA
24    alone is  not able to explain the toxicity associated with tetrachloroethylene exposure; therefore,
25    at least one other toxic agent appears to be involved.  Whether total metabolism, either as a
26    measure of a precursor or intermediate or as a surrogate directly proportional to the toxic
27    agent(s), is an adequate indicator of potential risk is unclear.  Use of administered dose (without
28    use of a  PBPK model) yields risk estimates intermediate between those based on the higher and
29    lower PBPK models. Consequently, a role for the parent compound has not been ruled out, nor
30    is it clear that the toxic agent(s) are not proportional to administered concentration.
31
32    5.4.6.1.4.  Choice of species/gender. The factors influencing the choice of rodent tumor data set
33    for human risk characterization are summarized in Table 5-14.  The carcinogenic response
34    occurs in rodents as well as in humans.  There is no information on tetrachloroethylene to
35    indicate  that the observed rodent tumors are not relevant to humans,  and there are no  non-rodent

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-81       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
       Table 5-14.  Summary of considerations for each rodent tumor type

Magnitude of response
Specificity of response
Replication of findings
in multiple bioassays
Other considerations
contributing to
uncertainty in rodent
data
Relevance to humans:
a) qualitative
(biologic) site
concordance
b) occurrence in
human studies
c) confidence in
MOA
Overall considerations
for choice of tumor
type
Mononuclear cell
leukemia
+++
Rats, both genders
4/4 study datasets
High background
rate
Yes
Yes, but exact
match of tumor
classification is not
found/may not be
possible
No data
Rodent response of
highest magnitude,
reproducible; no
MOA data
Liver
++
Mice, both
genders
6/6 study datasets
High background
rate in male mice
Yes
Yes, but
association is
weak
PPAR-a
activation may
contribute, but is
not sole MOA
Rodent response
of considerable
magnitude,
reproducible,
some MOA data
Kidney
++
Male rats only
1/2 study datasets
Rare tumor
(unlikely to be due
to chance, but low
incidence)
Yes
Yes, but
association is weak
Multiple MO As
may play a role
Rare tumor in
rodents and
humans; MOA
data are strongest
Hemangiosarcoma
+
Male mice only
1/3 study datasets
Rare tumor (unlikely
to be due to chance,
but low incidence)
Yes
No, but tumor type is
rare
No data
Rare tumor in rodents
and humans; no
MOA data
 4
 5
 6
 7
 8
 9
10
11
12
13
14
cancer bioassay data. Further, no tetrachloroethylene data exist to guide quantitative adjustment
for differences in sensitivity among rodents and humans. Human-rodent site concordance
generally is not assumed, e.g., due to potential differences in pharmacokinetics, DNA repair,
other protective systems across species and tissues (U.S. EPA, 2005a).  In keeping with this
view, certain tumors associated with tetrachloroethylene exposure in human mortality studies
(e.g., cervix  and esophagus) were not observed in rodents; cancer of the lymphoid system was
associated with tetrachloroethylene exposure in humans, with some evidence for an association
with bladder, kidney, and lung cancer.  In addition, rat and mouse tumor types also differ from
each other. Finally, conclusive MOA data are lacking for the observed rodent and human
tumors.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   5-82       DRAFT-DO NOT CITE OR QUOTE

-------
 1          MCL is the cancer response of highest magnitude, and it is reproducible in two bioassays
 2    and both genders. Although MCL has a high and variable incidence in unexposed F344 rats, a
 3    biologically and statistically significant increase over background was observed. The qualitative
 4    similarities among MCL to certain lymphoid cancers, and the implications regarding human
 5    relevance, are addressed in Section 4.8.2.4.1.2; also addressed is that elevated lymphoma
 6    mortality has been associated with tetrachloroethylene exposure in humans.  The MOA for MCL
 7    remains unexplored.
 8          Occasionally, if the multistage model does not adequately fit a data set, an alternate
 9    model can be used to determine the POD. In the case of female rat MCL data, the best-fitting
10    model (Weibull) allowed for a plateau and yielded estimates of risk per unit concentration 10-
11    fold higher than those from the multistage model fit of the male rat MCL data. While the female
12    rat MCL data suggest a plateau (also apparent for female rat MCL data from the NTP bioassay),
13    the multistage model fit was technically adequate (p = 0.48).
14          The mouse liver tumor is a robust finding in several studies, including in both sexes. As
15    is the case with MCL, the background for this tumor type is high especially in males. A
16    biologically and statistically significant increase over background was observed in males and
17    females.  There is evidence that activation of the PPAR-a receptor by the tetrachloroethylene
18    metabolite TCA contributes in part to the induction  of mouse liver tumors. However, it is not the
19    only operative MOA involved in hepatocellular tumorigenesis. Thus, the MOA remains
20    unresolved.
21          Two tumor types were observed in only one bioassay. Kidney tumors rarely occur in
22    unexposed rodents and were significantly elevated with tetrachloroethylene exposure in the male
23    rat NTP bioassay. The MOA is better understood for kidney tumors than for the other sites.
24    Hemangiosarcoma is another rare tumor associated with tetrachloroethylene exposure in the
25    male mouse JISA study.  There are no MOA data for hemangiosarcomas.
26
27    5.4.6.1.5. Physiologically basedpharmacokinetic (PBPK) model.  Toxicokinetic models are
28    used in this assessment for deriving dose metrics to  support dose-response analyses. The
29    evidence  suggests that by-products of tetrachloroethylene metabolism are responsible for liver
30    and kidney toxicity  and for carcinogenicity. Inhaled concentration of the parent compound is,
31    therefore, not an appropriate dosimeter for these effects, and pharmacokinetic modeling of daily
32    overall metabolized dose is expected to be an improvement in spite of the many attendant
33    uncertainties in the modeling. Of the available toxicokinetic models on tetrachloroethylene, the
34    assessment considers three recently developed models that describe parent tetrachloroethylene
35    and overall metabolism of the parent compound in humans. These models do not describe the

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-83       DRAFT-DO NOT CITE OR QUOTE

-------
 1    kinetics and transformation of total metabolic products or any individual metabolite. All three
 2    models provide reasonably good predictions of exhaled breath and blood tetrachloroethylene
 3    concentrations, so there is no particular basis for preferring one model over another. A 10-fold
 4    difference is shown in model predictions of the rate of metabolism in humans, a reflection of
 5    model differences in the values for the metabolic parameters. Because the accuracy of the
 6    models has been evaluated only against blood and breath concentrations of the parent
 7    compound—quantities that are insensitive to these parameters—the reliability of these models
 8    for predicting the rate of total metabolism in humans is unknown. Data on total metabolite levels
 9    are not available in humans, and the use of available urinary and blood TCA data is problematic.
10    The overall  difference in risk estimates using these three models is approximately 10-fold.
11
12    5.4.6.1.6. Cross-species scaling. An adjustment for cross-species scaling (BW3/4) was applied
13    to address toxicological equivalence of internal doses between each rodent species and humans,
14    consistent with the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a); the
15    approach is  detailed in Section 5.4.4.2.1.  It is assumed that, without data to the contrary, equal
16    risks result from equivalent constant exposures.  While the true correspondence of equipotent
17    tetrachloroethylene exposures across species is unknown, the use of BW3/4 scaling is expected
18    neither to over- or underestimate human risk (U.S. EPA, 1992).
19
20    5.4.6.1.7. Choice ofbioassay. The JISA inhalation bioassay provides data on the lowest
21    experimental exposures, and its use, therefore, reduces extrapolation uncertainty slightly. For
22    mice, the lowest-exposure concentration of 10 ppm was 10-fold lower than the lowest-exposure
23    concentration in the NTP inhalation study (NTP, 1986). For rats, the low-exposure concentration
24    of 50 ppm was fourfold lower than in the NTP study.  Although the JISA and NTP inhalation
25    bioassays used similar rodent strains, it is possible that differences in the animals used (in
26    addition to other unidentified factors) may have contributed to the twofold higher incidence of
27    hepatocellular tumors and MCL in the NTP study. The estimated risks for these sites are
28    consequently twofold lower than in previous EPA assessments which relied on the NTP bioassay
29    (U.S. EPA,  1991).
30
31    5.4.6.1.8. Statistical uncertainty at the Point of Departure (POD). Parameter uncertainty
32    within the chosen model reflects the limited sample size of the cancer bioassay.  For the
33    multistage model applied to this data set, there is a reasonably small degree of uncertainty  at the
34    10% extra risk level (the POD for linear low-dose extrapolation).
35

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-84        DRAFT-DO NOT CITE OR QUOTE

-------
 1    5.4.6.2.  Summary and Conclusions
 2          The uncertainties presented in Table 5-13 have a varied impact on risk estimates. Some
 3    suggest risks could be higher than was estimated, while others would decrease risk estimates or
 4    have an impact of an uncertain direction.  Several uncertainties are quantitatively
 5    characterized—the range of uncertainty in the PBPK models considered, together with the
 6    statistical uncertainty in the multistage modeling estimate, for the significantly increased rodent
 7    tumors.  Sensitivity to model selection is quantitatively explored in Figure 5-15, with a focus on
 8    thresholded, non-linear alternatives, illustrating the expected finding that such alternatives yield
 9    a wide range of estimates that are uninformative of the actual risk. Alternatives that would yield
10    higher risk estimates (e.g., supralinear models), which are equally scientifically valid, are not
11    presented. In addition, the results apply only to the prediction of MCL in male rats,  not in
12    humans. Due to limitations in the data, particularly regarding the MO A and relative human
13    sensitivity and variability, the quantitative impact of other uncertainties of potentially equal or
14    greater impact has not been explored. As a result, an integrated quantitative analysis that
15    considers all of these factors independently was not undertaken.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-85       DRAFT-DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
                                   APPENDIX 5A:
                       BENCHMARK DOSE MODEL RESULTS
                                  ABBREVIATIONS

AIC     = Akaike Information Criterion (see, e.g., U.S. EPA, 2000)

BMCXX   = Effective concentration at a specified level of extra risk; e.g., BMCio is the
            concentration corresponding to 10% extra risk, BMC0s corresponds to 5% extra risk

BMCs   = Effective concentration corresponding to a one standard deviation difference in the
            mean response from the control mean response (for continuous data).  This is
            approximately equivalent to 10% of the responses at the effective concentration
            being more extreme than 98% of the controls if the adverse response is an increase
            relative to the controls, or 10% of the responses being more extreme than 2% of the
            controls if the adverse response is a decrease relative to controls.

BMCLxx = Lower 95% confidence bound on the estimated BMCXX

BMCLs  = Lower 95% confidence bound on the estimated BMCs

NA      = Not applicable

POD     = Point of departure
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  5-86       DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
             Table 5A-1. Benchmark modeling summary:  hepatic parenchymal changes
             in humans with occupational exposure to tetrachloroethylene, data from
             Brodkin et al. (1995)
8-hr TWA exposure
(ppm)
0.0007
4.6
19.8
Quantal models
Multistage13
Gamma, Weibull,
Logistic
Probit
Probit, log-transformed dose
Equivalent continuous
exposure"
(ppm)
0.0003
1.6
7.1
Goodness of fit
/7-value
0.71
0.71
0.63
0.62
0.53
Number examined
26
9
18
AIC
72.4
72.4
72.5
72.5
72.7
Maximum
X2 residual
near POD
0.2
0.3
0.4
0.4
0.6
Number (%) with
parenchymal changes in
hepatic ultrasonograph
10 (38)
5(55)
13 (72)
BMC10
(ppm)
0.9
0.9
1.2
3.5
1.7
BMCL10
(ppm)
0.5
0.5
0.7
2.2
0.8
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
     a Exposures adjusted by 10/20 (m3/day) x 5/7 (days) to estimate equivalent continuous exposure levels.
      Measurements were taken from personal samplers for a subset of the individuals in the two higher-exposure
      groups. The background level of 0.0007 ppm is the high end of a range from Hartwell et al. (1985).
     b Multistage model selected as best fitting-model. Models had similar fits, multistage had lowest AIC and closest fit
      near the ECi0. Multistage model given by:

      P(d) = 1 - exp(- q0 -
      where:     d = continuous exposure level (ppm)
                 q0 = 0.40
                  i = o.n
                                            Multistage Model with 0.95 Confidence Level
                             O.B


                             0.7


                             0,



                             °£

                             0.4


                             0.3
                                  Multistage
                                   BMDL
                                         BMP
                                   0     1


                            13:4402/262002
                                                    3     4

                                                      dose
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      5-87        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
             Table 5A-2. Benchmark modeling summary: increased liver weight in
             female mice exposed to tetrachloroethylene, data from Kjellstrand et al.
             (1984)
Administered exposure,
24 hrs/day, 30 days
(ppm)
0
9
37
75
150
Continuous models"
Hillb
Power
Linear, Polynomial
Human equivalent
continuous exposure
(ppm)
0
9
37
75
150
Goodness of fit
p-value
0.56
0.001
0.003
N
42
11
10
10
10
AIC
510
524
522
Maximum
X2 residual
near POD
0.3
0.6
0.6
Female Mice
Liver weight (g)
mean V s.d.
108+13.5
142 + 25.6
210 + 24.2
241+41.4
230 + 31.1
BMCS
(ppm)
3.2
6.6
6.6
BMCLS
(ppm)
0.6
5.1
5.1
 5
 6
 7
 8
 9
10
11
12
13
14
     1 Nonconstant variance models fitted.
     3Hill model was the only adequately fitting model:

            P(d) = intercept + v * dosen/(kn + dose"),

            where:  intercept = 108.0
                        v= 193.6
                        n= 1.08
                        k=35.3
                                           Hill Model with 0.95 Confidence Level
                       150
                       100
                      1E:5402/D52002
                                                                           70
                                                                                 Fin
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      5-88        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
             Table 5A-3. Benchmark modeling summary: increased liver weight in
             female mice exposed to tetrachloroethylene, using response data from
             Kjellstrand et al. (1984) and human equivalent metabolized dose as dose
             metric
Administered exposure,
24 hr/day, 30 days
(ppm)
0
9
37
75
150
Continuous models'"
Hiir
Power0
Polynomial (linear)
Human equivalent
metabolized dose"
(mg-eq/kg-day)
0.0
3.2
12
19
28
Goodness of fit
p-value
0.58
0.78
0.03
N
42
11
10
10
10
AIC
510
515
513
Maximum
X2 residual
near POD
0.4
0.3
0.4
Female Mice
Liver weights (g)
mean V s.d.
108+13.5
142 + 25.6
210 + 24.2
241+41.4
230 + 31.1
BMCS
(mg-eq/kg-day)
1.1
0.7
1.8
BMCLs
(mg-eq/kg-day)
0.3
0.2
1.4
 6
 7
 8
 9
10
11
12
13
     a Metabolite levels were estimated using the Reitz et al. (1996) PBPK model, and adjusted to equivalent human
       doses using surface area scaling by multiplying by [0.03 kg/70 kg]025 = 0.144.
     b Nonconstant variance models fitted. Highest-dose group omitted due to poor fits for all models.
     0  Among the models with adequate fits (p > 0.1), the Hill and Power models had very similar BMCLss.  The
       average of these BMCLss was 0.3  mg-eq/kg-day. The Hill model fit is shown as a representative of the two fits.
       Using the three human pharmacokinetic models, the human equivalent inhalation exposures ranged from 1.4 to 10
       ppm (see Figure 3-9 and Chapter 3).
                                             Hill Model with 0.95 Confidence Level
                            150  -
                            100
                          09:43 02/DB 2002
                                                                              20
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      5-89        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
             Table 5A-4.  Benchmark modeling summary:  hepatic angiectasis in male
             mice exposed to tetrachloroethylene, data from JISA (1993)
Administered exposure,
6 hrs/day, 5 days/wk,
2 years (ppm)
0
10
50
250
Quantal models
Gamma b
Log-logistic b
Multistage b
Weibullb
Log-Probitb
Probit
Logistic
Human equivalent
continuous exposure"
(ppm)
0.0
1.8
8.9
45.0
Goodness of fit
p-value
0.57
0.78
0.66
0.59
1
0.02
0.02
N
50
50
50
50
AIC
161
161
160
161
161
166
167
Maximum
X2 residual
near POD
0.4
0.2
0.3
0.4
0
2.2
1.5
Incidence (%) of hepatic
angiectasis in male mice
1
3
12
30
BMC10
(ppm)
3.6
3.8
4.9
3.6
3.8
12.2
13.3
(2)
(6)
(24)
(60)
BMCL10
(ppm)
1.4
1.7
3.7
1.5
1.8
10
11
4
5
6
7
      a Exposure adjusted by 6/24 (hrs/day) x 5/7 (days/wk) to estimate equivalent continuous exposure levels.
      b All models except logistic and probit, had acceptable fits (p > 0.1), similar AICs, and BMCL10s within a factor of
       3 of each other. Average of these 5 BMCL10s is 2 ppm.  The log-probit model fit is shown as a representative fit.
                      0.8


                      0.7


                      0.6


                   "5 0.5
                    U
                    O)
                   g 0.4

                    o
                   '•3 0-3
                    CO

                   ^ 0.2


                      0.1


                        0
                                         Probit Model with 0.95 Confidence Level
                            Probit
                            BMDL
                                  BMD
                              0
                                        10
                                                                         40
                                                     dose
                     13:2502/082002
                 This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                       5-90         DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
              Table 5A-5. Benchmark modeling summary: hepatic angiectasis in male
              mice exposed to tetrachloroethylene, using data from JISA (1993) and
              human equivalent metabolized dose as dose metric
Administered exposure,
6 hr/day, 5 days/wk,
2 years
(ppm)
0
10
50
250
Quantal models
Gammab
Logisticb
Log-logistic13
Multistage13
Probitb
Log-Probitb
Weibullb
Human equivalent
metabolized dose"
(mg-eq/kg-day)
0.0
0.6
2.3
5.8
Goodness of fit
p-value
0.87
0.23
0.73
0.85
0.35
0.51
0.95
N
50
50
50
50
AIC
161
162
161
161
161
161
161
Maximum
X2 residual
near POD
0.1
1.3
0.2
0.1
1.1
0.4
0.04
Incidence (%) of hepatic
angiectasis in male mice
1
3
12
30
BMC10
(mg-eq/kg-
day)
1.2
2
1.2
1.1
1.8
1.2
1.2
(2)
(6)
(24)
(60)
BMCL10
(mg-eq/kg-
day)
0.6
1.6
0.6
0.6
1.5
0.6
0.6
 5
 6
 7
 8
 9
10
11
12
       Metabolite levels were estimated using the Reitz et al. (1996) PBPK model, were estimated adjusted to equivalent
       human doses using surface area scaling by multiplying by [0.048 kg/70 kg]0'25 = 0.162.
       All models achieved satisfactory fits (p > 0.1, similar AICs), except the logistic and probit models were least
       consistent with the data, having acceptable but relatively large %2 residuals near the BMCL10.  The BMCL10s from
       the remaining models were all 0.6 mg-eq/kg/day. The multistage model fit is shown as a representative fit.  Using
       the three human pharmacokinetic models with the POD of 0.6 mg-eq/kg-day, the human equivalent inhalation
       exposures ranged from 4.3-23 ppm.
                                            Multistage Model with 0.35 Confidence Level
                                   0       1


                            12:4702/082002
                                                        3

                                                       1	?
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                       5-91        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
             Table 5A-6. Benchmark modeling summary:  incidence of karyomegaly in
             male rats exposed to tetrachloroethylene, data from NTP (1986)
Administered exposure,
6 hrs/day, 5 days/wk,
2 years (ppm)
0
200
400
Quantal models
Gamma, Log-logistic, Log-
Probit, Weibull
Logistic
Multistage13
Probit
Human equivalent
administered dose
(ppm)a
0
36
71
Goodness of fit
p-value
N
49
49
50
AIC
Maximum
X2 residual
near POD
Incidence (and %) of
Karyomegaly in male rats
1
37
47
BMC10
(ppm)
(2)
(76)
(96)
BMCL10
(ppm)
NA
<0.01
0.98
<0.01
98.3
91
101
1.3
0.01
1.9
10.6
2.7
9.8
7.8
2.2
7.5
 4
 5
 6
 7
 8
 9
10
11
12
13
     a Exposure adjusted by 6/24 (hr/day) x 5/7 (days/wk) to estimate equivalent continuous exposure levels.
     b With only two nonzero exposure groups, options for fitting these data were limited. Among the models with two
       or less parameters to estimate, the multistage model was the only one to fit adequately  (p>0. 1).

       P(d) = 1 - exp(-q0 -
       where:        d = continuous exposure level (ppm)
                    qo = 0.02
                    qi = 0.04
                                          Multistage Model with 0.95 Confidence Level
                         0.8


                         O.B


                         0.4


                         0.2


                          0
                               Multistage
                              BMDLBMD
                                0     10


                       16:36 10/10 2001
                                            20     30     40     50

                                                     dose
                                                                       60
                                                                              70
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      5-92        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
              Table 5A-7.  Benchmark modeling summary:  deaths by Day 29 in offspring
              of rats exposed to tetrachloroethylene, data from Tinston (1994)
Administered exposure,
6 hrs/day, 5 days/wk, 11
weeks; every day during
gestation
(ppm)
0
100
300
1000
Nested models'"
Nested logistic
NCTRd
Rai and vanRyzind
Human equivalent
administered
exposure"
(ppm)
0
18
54
180
Goodness of fit
p-value
0.072
0.12
0.12
Number of F2A
Litters
23
22
21
19
AIC
656.9
656.9
656.9
Average per-litter percent
of deaths by Day 29
8.4
9.5
11.4
33.5
BMC0ic
(ppm)
24.4
23
23
BMCIV
(ppm)
2.1
1.8
1.8
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
     a Exposures adjusted to equivalent continuous exposures by multiplying by 6/24 (hrs/day) x 5/7 (days/wk).
     b All nested models fit best when including intralitter correlations. Litter size was not used as a litter-specific
       covariate.
     0 BMCLoi selected as relevant response level because of the severity of the response (pup death) and because the
       response level occurred within the range of the data set.
     d The NCTR and the Rai and van Ryzin model fits were identical for these data, whereas the nested Logistic did not
       provide an adequate fit.

             P(d)= l-exp[-(a + (3 xdp)],

             where:   d = continuous exposure level (ppm),
                     a =0.089
                     P = 4.38  x 10'5
                     p=1.73
                             0.4


                            0.35


                         -a   0.3
                         0

                         I  0.25
                         <

                         I   0.2
                         ro

                         "•  0.15


                             0.1


                            0.05
                                 NCTR
                                                NCTR Model with 0.95 Confidence Level
                                 BMDL      |BMD
                                                  50
                                                              100
                                                                            150
                                                           dose
                           15:1911/022004
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        5-93        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
             Table 5A-8.  Benchmark modeling summary:  liver to body weight ratio in
             rats exposed to tetrachloroethylene, data from Buben and O'Flaherty (1985)
Administered doses (5 days/wk,
6 weeks) (mg/kg-day)
0
20
100
200
500
1000
1500
2000


Continuous models'"
Hillc
Power
Polynomial, Linear

Goodness of fit
p-value
0.11
O.001
O.001

N
26
13
13
15
15
19
6
6


AIC
-24.1
28
24
Maximum
X2 residual
near POD
0.8
1
1
Liver weight/ body weight, mean
% + s.d.
5.21+0.46
5.51+0.40
5.97+0.40
6.45+0.46
7.35+0.62
7.89+0.70
8.10+0.66
9.00+0.27

BMDS
(mg/kg-day)
8.3
38
38

BMDLs
(mg/kg-day)
6.4
33
33
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
     a  Gavage doses were adjusted for daily exposure (x 5/7) and were adjusted to equivalent human doses using surface
       area scaling by multiplying by [0.048 kg/70 kg]025 = 0.162.
     b  Constant variance models used.
     0  Hill model was the only model to fit adequately.

            P(d) = intercept + v * dose7(kn + dose"),

            where:    intercept = 5.27
                           v = 4.18
                           n=l
                           k=58.2
                                              Hill Model with 0.95 Confidence Level
                          17:2201/302002
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      5-94        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
             Table 5A-9.  Benchmark modeling summary:  liver to body weight ratio in
             rats exposed to tetrachloroethylene, using data from Buben and O'Flaherty
             (1985) and human equivalent metabolized dose as dose metric
Administered doses
(5 days/wk, 6 wks)
(mg/kg-day)
0
20
100
200
500
1000
1500
2000
Continuous models'"
Hiir
Power"
Linear, Polynomial °
Continuous equivalent
metabolized dose"
(mg-eq/kg-day)
0.0
0.58
2.52
3.51
5.83
8.15
9.94
11.6
Goodness of
fit
/7-value
0.51
0.35
0.35
AIC
-30.9
-30.5
-32.5
N
26
13
13
15
15
19
6
6
Maximum
X2 residual
near POD
0.2
0.2
0.6
Liver weight/ body
weight, mean % + s.d.
5.21+0.46
5.51+0.40
5.97+0.40
6.45+0.46
7.35+0.62
7.89+0.70
8.10+0.66
9.00+0.27
BMDS
(mg-eq/kg-day)
8.8
7.9
7.9
BMDLS
(mg-eq/kg-day)
NA
6.5
6.5
 5
 6
 7
 8
 9
10
11
12
13
     '  Metabolites for mice were estimated using the Reitz et al. (1996) model and were adjusted for continuous daily
       exposure (x 5/7).
     ' Nonhomogeneous variance models.
     !  All continuous models fit reasonably well, except the Hill model could not provide a BMDL. The power model
       fit is shown as a representative of the other two fits. The BMDLS was converted to an equivalent human dose
       using surface area scaling by multiplying by [0.048 kg/70 kg]0'25 = 0.162, or 1.1 mg-eq/kg-day.  Using the three
       human pharmacokinetic models, the human equivalent oral exposures ranged from 3.4 to 32 mg/kg-day (see Table
       5-10 for conversion factors).
                                             Power Model with 0.95 Confidence Level
                                                                                  12
                          15:0307/11 2002
                  This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                      5-95         DRAFT-DO NOT CITE OR QUOTE

-------
 1                                       APPENDIX 5B:
 2             PROBABILITY DISTRIBUTIONS OF CANCER RISK ESTIMATES
 O
 4
 5          Given the importance of characterizing central tendency estimates of risk when feasible
 6    (U.S. EPA, 2005a), and the observation that MLEs resulting from typical dose-response analyses
 7    can be unstable, an analysis of the distributions underlying the range of site-specific
 8    tetrachloroethylene estimates of risk per unit concentration was undertaken. In addition, the
 9    distributions underlying estimates of risk per unit concentration based only mononuclear cell
10    leukemias was explored for three dose-response models frequently used for noncancer dose-
11    response assessment in the observable range of the experimental data: the Weibull, log-logistic,
12    and log-probit models.
13          The bootstrap analysis (Efron and Tibshirani, 1993) was used to characterize the
14    distributions of risk per unit concentration for the six tumor/sex types identified for
15    tetrachloroethylene. For each of the six data sets in Figure 5-14 (see Tables 5-4 and 5-6 for
16    group data), for each exposure group a simulated incidence level was generated using binomial
17    distribution with probability of success equal to the observed incidence.  This was repeated until
18    there were 10,000 simulated experiments for each tumor type. Then each simulated data set was
19    used to obtain estimates of BMDs using BMDS (U.S. EPA, 2000) in the same manner as for the
20    tetrachloroethylene assessment, including using the multistage model. The BMDs were
21    estimated at a benchmark response (BMR) of 10% extra risk for all sites except kidney tumors,
22    which were evaluated at 5% extra risk because 10% fell above the observed data. Distributions
23    of cancer slope values were obtained by calculating the distributions of the ratios BMR/BMDs.
                                                                            th      th
24    Upper and lower bounds on the linear extrapolation were determined by the 95  and 5
25    percentiles of the resulting distributions.
26          In the same manner as in the preceding paragraph, the bootstrap analysis was used to
27    characterize the distributions of risks per unit concentration resulting from fitting the male rat
28    leukemia data with the multistage, Weibull, log-logistic, and log-probit models. These models
29    were fit to the mononuclear cell leukemia data in male rats using the EPA BMDS program
30    without restricting the shape parameters of the latter three models.  Parameters describing the
31    risk distribution (mean, median and 95% upper and lower confidence bounds) were estimated by
32    a bootstrap procedure, because these parameters are not all readily available in the current
33    BMDS software. The resulting risk distributions were compared at two exposure levels—at a
34    generalized POD selected near the 10% response level estimated by the multistage model fit, at
35    1.5 mg-eq/kg-day, and at an environmental level  105-fold lower than the generalized POD, at 1.5
36    x 10"5 mg-eq/kg-day.  In comparing these distributions from the bootstrap procedure,  the mean
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-96       DRAFT-DO NOT CITE OR QUOTE

-------
 1    (as a measure of central tendency) and the 95% upper and lower bounds calculated from the
 2    bootstrap procedure are of interest.

 3    5.B.I. MULTISTAGE MODEL RESULTS
 4          Table 5B-1 compares cancer risk values calculated based on BMDS output (left half of
 5    table) with those calculated based on the bootstrap distribution of BMDs (right half of table).
 6    BMDS estimates the BMD as an MLE, and derives the 95% lower bound (BMDL) and the 95%
 7    upper bound (BMDU) on the BMD using the asymptotic distribution of the profile likelihood.
 8    Dividing the BMR by these values, one obtains estimates of the slopes of linear extrapolation to
 9    background responses from the BMDLs and BMDUs.
10          One can observe that there is generally a very good correspondence between asymptotic
11    (BMDS) results and re-sampling (bootstrap- based) results.  This is in agreement with analysis of
12    other models in BMDS (Moerbeek et al., 2004), but differs from the conclusions of Bailer and
13    Smith (1994). However, the latter paper's conclusions were based on 1,000 runs, and Moerbeek
14    et al. (2004) demonstrated that at least 2,000 runs are needed to stabilize confidence interval
15    estimates.  Additionally, risk estimates corresponding to the BMD were derived using the
16    average of the inverse distribution of BMDs.  While agreement with risk estimates calculated
17    using BMDS is generally good, for one data set (male mice liver tumors) the discrepancy is
18    noticeable, with the MLE (BMD) and bootstrap estimates differing by about 50% (8.07 x io"3 vs.
19    1.16x10"). The difference is due to asymmetry of the distribution of BMDs, so that the MLE
20    may be different from the average of the distribution in such  situations.  The estimate based on
21    the bootstrap average is therefore a preferred estimate of central tendency in such a case.
22
23    5.B.2. RESULTS USING ALTERNATE MODELS
24          At  1.5 mg-eq/kg-day, corresponding to a risk of approximately 0.1 using the mean
25    estimate of risk from the multistage model, the bootstrap procedure yielded similar risk
26    distributions for the four models (see Table 5B-2 and Figure  5B-la). The means and
27    corresponding confidence bounds agree within an order of magnitude and the spreads in the
28    distributions (the distance between the upper and lower confidence bounds) are within two
29    orders of magnitude. Note that the probability calculations are in terms of metabolized dose in
30    the male rat, and do not directly characterize human risks.
31          EPA also examined the bootstrap results from those same models at a dose IO5 -fold
32    lower (although EPA's actual low dose risk estimates are developed using a linear extrapolation
33    from a POD to the origin rather than using low-dose estimates from a model). These results are
34    shown in Table 5B-2 and Figure 5B-lb. In the region of extrapolated concentrations (1.5 x 10
-5
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   5-97       DRAFT-DO NOT CITE OR QUOTE

-------
 1   mg-eq/kg-day), the mean risks of the latter three models (Weibull, log-logistic and log-probit)
 2   are about one to three orders of magnitude higher than the mean of the multistage model risks.
 3   The spreads of all the models are quite broad, with a six order of magnitude 95% confidence
 4   interval for the multistage and much greater spreads for the other three models. The upper
 5   bounds of risk for the other three models are higher than that for multistage model, within about
 6   three orders of magnitude, and their lower bounds of risk are much lower than that of multistage,
 7   by nine or more orders of magnitude. With such large spreads in confidence intervals, the
 8   extrapolated models in effect provide little information about low-dose risks. The extrapolation
 9   of the multistage model does result in estimates reasonably close to the low-dose estimates from
10   the model-independent straight line extrapolation from the POD, in that the mean and upper-
11   bound risks at the lower concentration are both within 10% of the estimates resulting from
12   applying linear extrapolation to the results at the higher concentration.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                   5-98        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
             Table 5B-1. Comparison of BMDS and bootstrap-based cancer risk values
             derived using the multistage model to fit rodent bioassay data for tumor
             types associated with tetrachloroethylene exposure, and using total
             metabolites as the dose metric
Tumor
type
Male rat
kidney
tumorsb
Male mice
liver
hemangio-
sarcomas
Male mice
liver
tumors
Female
mice liver
tumors
Male rat
leukemias
Female rat
leukemias
Model
ordera
2
1
o
3
i
3
3
o
5
2
3
1
BMR
0.05
0.05
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
Slopes, (mg-eq/kg-day) , of linear
extrapolation using BMDS
estimates of:
BMD
1.22 x 10"2
1.22 x 10"2
4.82 x 10"3
4.82 x 10"3
8.07 x 10"3
5.33 x 10"3
7.64 x 10"2
7.64 x 10"2
2.46 x 10"2
2.46 x 10"2
BMDL
2.63 x 10"2
2.63 x 10"2
8.04 x 10"3
8.04 x 10"3
2.77 x 10"2
1.04 x 10"2
1.24 x 10"1
1.24 x 10"1
4.93 x 10"2
4.93 x 10"2
BMDU
NDC
ND
2.06 x 10"3
2.06 x 10"3
5.46 x 10"3
4.14 x 10"3
2.67 x 10"2
2.93 x 10"2
ND
ND
Bootstrap-based slopes,
-i
(mg-eq/kg-day)
Mean
1.20 x 10"2
1.24 x 10"2
4.61 x 10"3
4.88 x 10"3
1.16 x 10"2
6.02 x 10"3
6.75 x 10"2
6.89 x 10"2
2.36 x 10"2
2.49 x 10"2
Upper
Bound
2.32 x 10"2
2.52x 10"2
7.66 x 10"3
8.01 x 10"3
2.77 x 10"2
9.69 x 10"3
1.16 x 10"1
1.16 x 10"1
4.81 x 10"2
5.03 x 10"2
Lower
Bound
ND
ND
2.26 x 10"3
1.97 x 10"3
5.76 x 10"3
4.30 x 10"3
2.92 x 10"2
3.30 x 10"2
ND
ND
 6
 7
 8
 9
10
     a Order of polynomial in multistage model. Different order models were compared where various order models fit
      equally well.
     b Data set had 2 non-zero dose groups, all others had three. See Tables 5-4 and 5-6 for data and study details.
     0 ND = could not be determined.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                    5-99        DRAFT-DO NOT CITE OR QUOTE

-------
1
2
3
4
5
       Table 5B-2. Comparison of BMDS and bootstrap-based cancer risk values
       derived using several dose-response models to fit rodent bioassay data for
       mononuclear  cell leukemia incidence in male rats associated with
       tetrachloroethylene exposure, and using total metabolites as the dose metric
Model
Multistage
Log-probit
Log-logistic
Weibull
Dose,
mg-eq/kg-day
1.5
1.5x 10'5
1.5
1.5 x ID'5
1.5
1.5 x 10'5
1.5
1.5 x 10'5
MLE of risk
1.143 x ID'1
1.168 x ID'6
9.717 x ID'2
4 x ID'25
1.018 x 10'1
2.553 x 10'8
1.064 x 10'1
2.41 x 10'7
Bootstrap distribution of extra risk
Mean
9.442 x 1Q-2
9.172 x ID'7
1.037 x 1Q-1
8.172xlO-4
1.079 x 10'1
1.078 x 10'3
1.124 x 10'1
1.339 x 10'3
Median
9.853 x 1Q-2
9.933 x 1Q-7
9.544 x 1Q-2
1 x 1Q-26
9.937 x 10'2
1.401 x 10'8
1.052 x 10'1
1.643 x 10'7
S^percentile
1.163 x 1Q-2
4.546 x 1Q-13
8.366 x 1Q-4
2 x 10-170
4.320 x 10'3
2 x 10'21
5.594 x 10'3
0
95th percentile
1.672 x 1Q-1
1.819 x 1Q-6
2.426 x 1Q-1
1.008 x 1Q-5
2.429 x 10'1
7.776 x 10'4
2.452 x 10'1
1.893 x 10'3
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                 5-100      DRAFT-DO NOT CITE OR QUOTE

-------
      _>.
      '55
      0}
      a
                          Risks for 4 models for dose=1.5; log scale
          IO
          O
           I
• Multistage (M)
• Weibull (W)
n Logprobif (P)
• Loglogistic (L)

L 5th percentile 1 1
U 95th percentile M
— ~&r. . j\
LiVI UIVI
LW UW
LL LIL
l l l l
-15 -10 -5 0
                                          LOG of Risk
                        Risks for 4 models for dose=1.5E-5; log scale
      0)
      Q
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
to
o
          o
           I
• Multistage (M)
• Weibull (W)
n Logprobit (P)
• Loglogistic (L)
/
f\ L 5th percentile
1 1 U 95th percentile
I
I I
/' i
\ _^___^. 	 I 	 - -
	 —ttfr 	 uw
LW* UW
LL* UL
1 1 1 1
-15 -10 -5 0
                                LOG of Risk
                   * points are outside of the range of the graph

  Figure 5B-1.  Illustration of sensitivity to model selection for low-dose
  extrapolation. The risk distributions associated with four dose-response models
  which adequately fit the tetrachloroethylene dose-response data for MCL in male
  rats (NTP, 1986) were compared. Risks are in terms of metabolized
  tetrachloroethylene in male rats. Figure 5-2 shows the comparison at a
  generalized POD, selected as the mean exposure estimate, from the multistage
  model corresponding to a risk of approximately 0.1.  Figure 5-3 compares risk
  distributions at an exposure corresponding to environmental concentrations of
  tetrachloroethylene, approximately 105-fold lower than the POD.
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  5-101      DRAFT-DO NOT CITE OR QUOTE

-------
  1                                    REFERENCES FOR CHAPTER 5
  2
  O

  4     Anger, WK; Letz, R; Chrislip, DW; et al. (1994) Neurobehavioral test methods for environmental health studies of
  5     adults. Neurotoxicol Teratol 16:489-497.
  6
  7     Anger, WK; Liang, YX; Nell, V; et al. (2000) Lessons learned-15 years of the WHO-NCTB:  a review.
  8     Neurotoxicology 21:837-846.
  9
10     Anger, WK. (2003) Neurobehavioural tests and systems to assess neurotoxic exposure in the workplace and
11     community. Occup Environ Med 60:531-538.
12
13     Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
14     low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
15
16     Amler, RW; Lybarger, JA: Anger, WK; et al. (2004) Adoption of an adult environmental neurobehavioral test
17     battery. Neurotoxicol Teratol 16:525-530.
18
19     ATSDR (Agency for Toxic Substances and Disease Registry). (1996) Pediatric Environmental Neurobehavioral Test
20     Battery. Amler, RW; Gibertini, M; eds. Atlanta, GA: U.S. Department of Health and Human Services, Public
21     Health Service, Agency for Toxic Substances and Disease Registry.
22
23     ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological profile for tetrachloroethylene.
24     Springfield, VA.
25
26     Bailer, AJ: Smith, RJ. (1994) Estimating Upper Confidence Limits for Extra Risk in Quantal Multistage Models.
27     Risk Analysis 14(6): 1001-1010.
28
29     Barnes, DG; Dourson, M. (1988) Reference dose (RfD): description and use in health risk assessments. Regul
30     Toxicol Pharmacol 8:471-486.
31
32     Bellies, RP; Brusick, DJ; Mecler, FJ. (1980) Teratogenic-mutagenic risk of workplace contaminants:
33     trichloroethylene, perchloroethylene and carbon disulfide. Prepared by Litton Bionetics, Inc. NTIS Publication No.
34     PB-82 185-075, NIOSH Contract Report No. 201-77-0047. Available from: National Technical Information
3 5     Service, Springfield, VA.
36
37     Berman, E; Schlicht, M; Moser, VC; et al. (1995) A multidisciplinary approach to lexicological screening: I.
38     Systemic toxicity. J Toxicol Environ Health 45:127-143.
39
40     Bois, FY;  Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene.  Arch Toxicol
41     70:347-355.
42
43     Brodkin, CA; Daniell, W; Checkoway, H; et al. (1995) Hepatic ultrasonic changes in workers  exposed to
44     perchloroethylene.  Occup Environ Med 52:679-685.
45
46     Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
47     and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
48
49     Cal EPA (California Environmental Protection Agency).  (2001) Public health goal for tetrachloroethylene in
50     drinking water.  Prepared by the Office of Environmental Health Hazard Assessment. Available  online at
51     http://www.oehha.org/water/phg/pdf/PCEAug2001 .pdf.
52
53     CARD (California Air Resources Board). (1991) Proposed identification of perchloroethylene as a toxic air
54     contaminant. Oakland, CA.
55
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        5-102        DRAFT-DO NOT CITE OR QUOTE

-------
  1     CARD (California Air Resources Board). (1998) 1990-1996 Statewide perchloroethylene summary, ppb. Available
  2     online at http://www.arb.ca.gov/aqd/perc/pcstate.htm.
  3
  4     Cavalleri, A; Gobba, F; Paltrinieri, M; et al. (1994) Perchloroethylene exposure can induce colour vision loss.
  5     Neurosci Lett 179:162-166.
  6
  7     Chiu, WE; Bois, FY. (2006) Revisiting the population toxicokinetics of tetrachlorotethylene. Arch Toxicol 80:382-
  8     385.
  9
10     Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and gender-specific
11     pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79:381-393.
12
13     Echeverria, D; Heyer, N; Checkoway, H; et al. (1994) Behavioral investigation of occupational exposure to solvents:
14     Perchloroethylene among dry cleaners, and styrene among reinforced fiberglass laminators. Final Report. Prepared
15     for the Centers for Disease Control and Prevention under Grant No.  5 R01 OHo2719-03. Battelle Centers for Public
16     Health Research and Evaluation.
17
18     Echeverria, D; White, RF; Sampaio, C. (1995) A behavioral evaluation of PCE exposure in patients and dry
19     cleaners: a possible relationship between clinical and preclinical effects.  J Occup Environ Med 37:667-680.
20
21     Efron, B; Tibshirani, PJ. (1993) An Introduction to the Bootstrap. Chapman and Hall, San Francisco.
22
23     Eskenazi, B; Wyrobek, AJ; Fenster, L; et al. (1991) A study of the effect of perchloroethylene exposure on semen
24     quality in dry cleaning workers. Am J Ind Med 20:575-591.
25
26     Ferroni, C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
27     perchloroethylene.  Neurotoxicology 13:243-247.
28
29     Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
30     solvents. Int Arch Occup Environ Health 52:1-9.
31
32     Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
33     tetrachloroethylene  as neonates. Toxicol Lett 66:13-19.
34
3 5     Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
36     tetrachloroethylene.  Am J Ind Med 21:661-671.
37
38     Gentry, PR; Covington, TR; Clewell, HJ, III. (2003) Evaluation of the  potential impact of pharmacokinetic
39     differences on tissue dosimetry in offspring during preganancy and lactation. Regul Toxicol Pharmacol 38:1-16.
40
41     Gobba, F; Righi, E; Fantuzzi, G; et al. (1998) Two-year evolution of perchloroethylene-induced color-vision loss.
42     Arch Environ Health 53:196-198.
43
44     Guth, DJ; Carroll, RJ; Simpson, DG et al. (1997) Categorical regression analysis of acute exposure to
45     tetrachloroethylene. Risk Analysis 17:321-332.
46
47     Hartwell, TD; Crowder, JH; Sheldon, LA; et al. (1985) Levels of volatile organics in indoor air.  In: Proceedings of
48     the Air Control Pollution Association, 78th annual meeting, Vol. 3, 85-30.B., 2-12.
49
50     Hayes, JR; Condie,  LW, Jr.; Borzelleca, JF. (1986) The subchronic toxicity of tetrachloroethylene
51     (perchloroethylene) administered in the drinking water of rats.  Fundam Appl Toxicol 7:119-125.
52
53     JISA (Japan Industrial Safety Association). (1993) Carcinogenicity study of tetrachloroethylene by inhalation in rats
54     and mice. Data No.  3-1. Available from: EPA-IRIS  Information Desk.
55

                    This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         5-103       DRAFT-DO NOT  CITE OR QUOTE

-------
  1     Jonker, D; Woutersen, RA; Feron, VJ. (1996) Toxicity of mixtures of nephrotoxicants with similar or dissimilar
  2     mode of action. Food Chem Toxicol 34:1075-1082.
  3
  4     Kjellstrand, P; Holmquist, B; Kanje, M; et al. (1984) Perchloroethylene: effects on body and organ weights and
  5     plasma butyrylcholinesterase activity in mice.  Acta Pharmacol Toxicol (Copenh) 54:414-424.
  6
  7     Krewski, D; vanRyzin, J. (1981) Dose response models for quantal response toxicity data, in statistics and related
  8     topics. In:  Csorgo, M; Dawson, DA; Rao, INK; Saleh, ADMdE; eds. North-Holland Publishing Company.
  9
10     Lutz, WK;  Gay lor, DW; Conolly, RB; et al. (2005) Nonlinearity and thresholds in dose-response relationships for
11     carcinogenicity due to sampling variation, logarithmic dose scaling, or small differences in individual susceptibility.
12     Toxicol Appl Pharmacol 207:565-569.
13
14     Mattsson, JL; Albee, RR; Yano, BL; et al. (1998) Neurotoxicologic examination of rats exposed to 1,1,2,2-
15     tetrachloroethylene (perchloroethylene) vapor for 13 weeks. Neurotoxicol Teratol 20:83-98.
16
17     Moerbeek,  M; Piersma, AH; Slob, W. (2004) A Comparison of Three Methods for Calculating Confidence Intervals
18     for the Benchmark Dose. Risk Analysis 24(1):31-40.
19
20     Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-
21     cleaners. Lancet 340:189-193.
22
23     Muttray, A; Wolff, U; Jung, D; et al. (1997) Blue-yellow deficiency in workers exposed to low concentrations of
24     organic solvents. Int Arch Occup Environ Health 70:407-412.
25
26     NCI (National Cancer Institute). (1977) Bioassay of tetrachloroethylene for possible carcinogenicity.  DHEW Pub.
27     (NIH):77-813.
28
29     Nelson, BK; Taylor, BJ; Setzer, JV; et al.  (1980) Behavioral teratology of perchloroethylene in rats.  J Environ
30     Pathol Toxicol 3:233-250.
31
32     NRC (National Research Council). (1983) Risk assessment in the Federal government: managing the process.
3 3     Committee on the Institutional Means for  Assessment of Risks to Public Health, Commission on Life Sciences,
34     NRC. Washington, DC; National Academy Press
35
36     NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies fo tetrachloroethylene
37     (perchloroethylene) (CAS No. 127-18-4) inF344/N rats and B6C3F1 mice. NTP Technical Report 311. National
38     Institutes of Health, Public Health Service, U.S. Department of Health and Human Services, Washington, DC.
39
40     NYS DOH (New York State Department of Public  Health). (1997) Tetrachloroethene ambient air criteria document.
41     Final Report. Albany, NY.
42
43     NYS DOH (New York State Department of Health) (2005) Improving human risk assessment for
44     tetrachloroethylene by using biomarkers and neurobehavioral testing. U.S. EPA Star Grant #R827445. Grant
45     #R827446. Available online at
46     http://cfpub.epa.gov/ncer abstracts/index.cfm/fuseaction/displav.abstractDetail/abstract/977/report/O.
47
48     NYS OAG (New York State Office of Attorney General). (2004) Letter from Judy Schreiber, Ph.D., for EPA
49     Docket ORD-2003-0014. July 30, 2004.
50
51     Opdam, JJ. (1989) Intra and interindividual variability in the kinetics of a poorly and highly metabolising solvent. Br
52     JIndMed46:831-845.
53
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        5-104       DRAFT-DO NOT CITE OR QUOTE

-------
  1    Pelekis, M; Gephart, LA; Lerman, SE. (2001) Physiological-model-based derivation of the adult and child
  2    pharmacokinetic intraspecies uncertainty factors for volatile organic compounds. Regul Toxicol Pharmacol 33:12-
  3    20.
  4
  5    Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
  6    groundwater for a bathing and showering determination.  Risk Anal 13:37^9.
  7
  8    Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
  9    for physiologically based pharmacokinetic modeling in rats, mice, and humans.  Toxicol Appl Pharmacol 136:289-
10    306.
11
12    Rhomberg, LR. (2000) Dose-response analyses of the carcinogenic effects of trichloroethylene in experimental
13    animals. Environ Health Perspect 108(S2):343-358.
14
15    Rosengren, LE; Kjellstrand, P; Haglid, KG. (1986)  Tetrachloroethylene: levels of DNA and S-100 in the gerbil CNS
16    after chronic exposure. Neurobehav Toxicol Teratol 8:201-206.
17
18    Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
19    tetrachloroethylene and deficits in visual contrast sensitivity.  Environ Health Perspect 110:655-664.
20
21    Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
22    11:579-583.
23
24    Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
25    of subjects chronically exposed to organic solvents  (II)].  G Ital Med Lav Ergon 19:85-88.
26
27    Tinston, DJ. (1994) Perchloroethylene: A multigeneration inhalation study in the rat.  CTL/P/4097. Available from:
28    EPA IRIS Information Desk.
29
3 0    Trevisan, A; Macca, I; Rui, F; et al. (2000) Kidney  and liver biomarkers in female dry-cleaning workers exposed to
31    perchloroethylene.  Biomarkers 5:399-409.
32
33    Umezu, T; Yonemoto, J; Soma, Y; Miura, T. (1997) Behavioral effects of trichloroethylene and tetrachloroethylene
34    in mice. Pharmacol Biochem Behav 58:665-671.
35
36    U.S. EPA (Environmental Protection Agency). (1985) Health assessment document for tetrachloroethylene
3 7    (perchloroethylene). Office of Health and Environmental Assessment, Office of Research and Development,
38    Washington, DC; EPA/600/8-82/005F. Available from: National Technical Information Service, Springfield, VA;
39    PB-86-174489/AS.
40
41    U.S. EPA (Environmental Protection Agency). (1986) Addendum to the health assessment document for
42    tetrachloroethylene (perchloroethylene) [review draft]. Office of Health and Environmental Assessment, Office of
43    Research and Development, Washington, DC; EPA/600/8-82/005FA. Available from: National Technical
44    Information Service, Springfield, VA; PB-86-174489/AS.
45
46    U.S. EPA (Environmental Protection Agency). (1987) Technical analysis of new methods and data regarding
47    dichloromethane hazard assessments [review draft]. Office of Health and Environmental Assessment, Office of
48    Research and Development, Washington, DC; EPA/600/8-87/029A. Available from: National Technical
49    Information Service, Springfield, VA.
50
51    U.S. EPA (Environmental Protection Agency). (1988) IRIS summary of tetrachloroethylene RID.  Available online
52    at http://www.epa.gov/iris/subst/0106.htm.
53
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        5-105       DRAFT-DO NOT CITE OR QUOTE

-------
  1     U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
  2     carcinogenicity of tetrachloroethylene (perchloroethylene). Office of Health and Environmental Assessment,
  3     Washington, DC; EPA/600/6-91/002F.  Available from: National Technical Information Service, Springfield, VA.
  4
  5     U.S. EPA (Environmental Protection Agency). (1992) Draft report: a cross-species scaling factor for carcinogen risk
  6     assessment based on equivalence of mg/kg3/4/day. Federal Register 24152-24173.
  7
  8     U.S.EPA (1993) Reference Dose (RfD): Description and Use in Health Risk Assessments.
  9
10     U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
11     and application of inhalation dosimetry.  Office of Health and Environmental Assessment, Environmental Criteria
12     and Assessment Office, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical Information
13     Service,  Springfield, VA; PB2000-500023, and online at http://www.epa.gov/ncea.
14
15     U.S. EPA (Environmental Protection Agency). (1995) IRIS summary of dichloromethane. Available online at
16     http://www.epa.gov/iris/subst/0070.htm.
17
18     U.S. EPA (Environmental Protection Agency). (2000) Benchmark dose technical guidance document [external
19     review draft].  Risk Assessment Forum.  Washington, DC; EPA/630/R-00/001.  Available online at
20     http://www.epa.gov/ncea/raf.
21
22     U.S. EPA (Environmental Protection Agency). (2001) Trichloroethylene health risk assessment: synthesis and
23     characterization. National Center for Environmental Assessment, Office of Research and Development,
24     Washington, DC; EPA/600/P-01/002A. Available online at
25     http://oaspub.epa.gov/eims/eimscomm.getfile7p download id=4580.
26
27     U.S. EPA (Environmental Protection Agency). (2004) Summary report of the peer review workshop on the
28     neurotoxicity of tetrachloroethylene (perchloroethylene) discussion paper.  National Center for Environmental
29     Assessment, Washington, DC; EPA/600/R-04/041.  Available online at http://www.epa.gov/ncea.
30
31     U.S. EPA (Environmental Protection Agency). (2005a)  Guidelines for carcinogen risk assessment.  Federal Register
32     70(66)17765-17817. Available online at http://www.epa.gov/cancerguidelines.
33
34     U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
35     early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
36     online at http://www.epa.gov/cancerguidelines.
37
38     Verplanke, AJ; Leummens, MH; Herber, RF. (1999) Occupational exposure to tetrachloroethylene and its effects  on
39     the kidneys. J Occup Environ Med 41:11-16.
40
41     Wang, S; Karlsson, JE; Kyrklund, T; et al. (1993) Perchloroethylene-induced reduction in glial and neuronal cell
42     marker proteins in rat brain. Pharmacol Toxicol 72:273-278.
43
44     Warren, DA; Reigle, TG; Muralidhara, S; et al. (1996) Schedule-controlled operant behavior of rats following oral
45     administration of perchloroethylene: time course and relationship to blood and brain solvent levels. J Toxicol
46     Environ Health 47:345-362.
47
48     Zeise, L; Wilson, R; Crouch, E A. (1987) Dose-response relationships for carcinogens: a review. Environ Health
49     Perspect 73:259-306.
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        5-106       DRAFT-DO NOT CITE OR QUOTE

-------
 1                 6.  CHARACTERIZATION OF HAZARD AND DOSE-RESPONSE
 2
 3
 4      6.1. SUMMARY OF HUMAN HAZARD POTENTIAL
 5      6.1.1.  Exposure
 6            Tetrachloroethylene (CASRN 127-18-4) is a solvent used for cleaning clothes and for
 7      metal cleaning and degreasing.  It is a volatile liquid at room temperature.  The largest human
 8      exposure occurs indoors to workers in dry cleaning, laundry, and metal finishing facilities.
 9      Release of tetrachloroethylene into the air from these facilities also results in measurable outdoor
10      ambient air concentrations. Indoor residential exposure can also occur when dry cleaning
11      facilities are located within residential areas.  It has been detected in breast milk of women
12      exposed to tetrachloroethylene in ambient air in or near these facilities. Tetrachloroethylene can
13      enter water supplies, and it has been detected in drinking water. Exposure to airborne
14      tetrachloroethylene can  occur in homes via volatilization from tap water during showering as
15      well as from water ingestion in homes with contaminated drinking water (see Chapter 2 for more
16      information).
17
18      6.1.2.  Absorption, Metabolism, Distribution, and Excretion
19            Tetrachloroethylene rapidly enters body tissues after inhalation, ingestion, and dermal
20      exposure.  Tetrachloroethylene metabolism is considered to be well characterized in rodents but
21      not in humans. A significant portion of tetrachloroethylene inhaled by humans at ambient
22      concentrations is not metabolized (about 64% according to the pharmacokinetic model of Bois,
23      et al. 1996). The recovered metabolites in the urine represent only  a fraction of what is actually
24      metabolized (Bogen and McKone, 1988). Possible explanations for metabolites not reaching the
25      urine are (a) binding to plasma proteins, (b) biliary excretion, (c) enterohepatic circulation of
26      metabolites, (d) further metabolism of fat-sequestered parent compounds after the completion of
27      the studies, and (e) metabolism to currently unidentified metabolites. However, data to support
28      these hypotheses are sparse. The fraction of tetrachloroethylene metabolized appears to have a
29      strong dependence on the exposure concentration. The PBPK model by Bois et al. (1996)
30      predicts this fraction to be about 36% in humans at low environmental concentrations, whereas
31      the human data indicate a very small fraction would be metabolized at higher concentrations
32      (such as those corresponding to the laboratory animal bioassays; see Section 3.5 for more
33      details.)
34            There are two major routes of metabolism: (1) the predominant oxidative pathway,
35      which results in TCA and other urinary metabolites, as well as reactive intermediates and carbon
36      dioxide; and (2) the GSH conjugation pathway, which results in TCVG and TCVC that are
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  6-1       DRAFT - DO NOT CITE OR QUOTE

-------
 1      further processed to chemically reactive products that can bind to tissue macromolecules.
 2      Metabolism occurs predominately in the liver. Further, metabolism of the GSH metabolites,
 3      including activation by beta lyase, occurs in the kidney.  In addition, the CYP enzymes of the
 4      oxidative pathway as well as enzymes important to the GSH pathway are present in other tissues.
 5      Therefore, a potential exists for extrahepatic metabolism and formation of reactive metabolites at
 6      sites other than the liver and kidney (see Section 3.3.3.2).
 7             Many steps in the oxidative metabolism of tetrachloroethylene are well characterized in
 8      both animals and humans; however, not all proposed intermediates have been identified or
 9      detected. Although an initial epoxide metabolite has not been unequivocally demonstrated for
10      tetrachloroethylene, the epoxide intermediate is a reasonable1 proposal.  It has been chemically
11      synthesized, and it is metabolized to TCA when injected into rodents.  The tetrachloroethylene
12      epoxide intermediate is considered to be unstable and short-lived in vivo and is thought to
13      spontaneously rearrange and convert to other intermediate metabolites. Formation of
14      trichloroacetyl chloride directly from tetrachloroethylene via the mechanism of CYP-mediated
15      olefin oxidation without the obligatory formation of the epoxide intermediate has also been
16      postulated. The formation of trichloroacylated protein adducts in liver and kidney of rats, liver
17      of mice, and plasma of rats and humans following exposures to tetrachloroethylene provides
18      evidence of a metabolic intermediate that can react with tissue proteins.  TCA, a stable
19      metabolite, is believed to result primarily from the oxidation of tetrachloroethylene to the
20      potentially reactive trichloroacetyl choride.  TCA has been detected in the blood and urine of
21      humans and laboratory rodents, and excretion in urine is used as a biomarker for
22      tetrachloroethylene exposure (see Section 3.3.3 for more details).
23             Other steps in tetrachloroethylene oxidative metabolism are not as well characterized.
24      Both TCOH and TCOH-glucuronide have been detected in the urine of mice and humans
25      following tetrachloroethylene exposure; however, there is uncertainty about formation of TCOH
26      and its chloral hydrate precursor from tetrachloroethylene because not all studies have detected
27      TCOH as a urinary excretion product.  These different findings could be explained in several
28      ways:  (a)  TCOH could be an artifact of the analytical methodology, (b)  differences could exist
29      in analytical methodologies, (c) contamination could have occurred by unknown exposures to
30      another chemical, and (d) excretion of TCOH and its glucuronide might  be dependent on dose.
31      If the TCOH pathway exists in humans, the overall contribution to TCA from TCOH is expected
32      to be relatively small when compared with the amount of TCA resulting from trichloroacetyl
33      chloride.
               1 "Reasonable," as used is this chapter, is intended to imply the use of reasoned scientific judgment in data
        evaluation and decision-making, consistent with U.S. EPA practices and guidance (e.g., Guidelines for Carcinogen
        Risk Assessment, 2005a)
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-2       DRAFT - DO NOT CITE OR QUOTE

-------
 1            DC A is an intermediate metabolite that has been identified in the urine of rats, but not
 2      humans, exposed by inhalation to tetrachloroethylene. It is not clear whether DCA is a product
 3      of further metabolism of TC A or a product of another pathway originating with GSH
 4      conjugation, or both.  The amount of DCA produced from tetrachloroethylene oxidative
 5      metabolism could vary across species and is likely to be relatively small when compared with the
 6      amount of TCA produced. Data on mutational changes in tetrachloroethylene-induced liver
 7      tumors in mice also support a limited role for DCA in tetrachloroethylene toxicity.
 8            Quantitatively, the GSH conjugation pathway is relatively minor when compared with the
 9      P450 conjugation pathway. Urinary mercapturates comprise from 1% to as little as 0.03% of
10      total recovered urinary metabolites (Green et al., 1990; Birner et al., 1996); however, this urinary
11      excretion product does not reflect the amount of compound going through the GSH pathway, but,
12      rather, it reflects only the portion that is excreted.  The amount of the  mercapturate product
13      excreted in the urine also does not reflect the amount of the more important portion that is
14      converted to toxic by-products through further metabolism. TCVG is the conjugation product of
15      tetrachloroethylene, and its cleavage product, TCVC, reacts with a kidney enzyme, beta lyase, to
16      produce metabolites that are mutagenic in bacteria and are also cytotoxic.  These metabolites are
17      believed to contribute to tetrachloroethylene-induced kidney toxicity.  Uncertainty exists as to
18      the relative contribution of GSH metabolism to toxicity in humans as  compared with the rat due
19      to study differences in reported amounts of human tetrachloroethylene GSH metabolism as
20      measured by excreted mercapturate.  These differences may be due, in part, to different chemical
21      assay methodology or to problems resulting from the stability of the chemical product being
22      measured, or both. In spite of these uncertainties, some of the published findings concerning
23      TCVG production would not predict any less susceptibility for humans than for rodents with
24      regard to renal toxicity.  The higher percentage of mercapturate found in rat versus human urine
25      does not indicate a higher level of production of toxic products in the  rat, because excreted
26      mercapturate allows no estimate of the amount of TCVC or N-acetyl TCVC being processed
27      through alternate routes.  Furthermore, it is not known whether sex-dependent variation of beta
28      lyase activity exists in humans as it does in rats.  Human variation might also explain study
29      differences in reported excretion rates (see Section 3.3.3.2).
30            Several enzymes of the oxidative and GSH metabolism, notably CYP2E1, CYP3A4,
31      GSTZ, GSTA, GSTM, and GSTT, show genetic polymorphisms with the potential for variation
32      in production of specific metabolites. The effect of metabolic variation, including potential
33      implications for differential toxicity, has not been well studied. The limited data available on
34      tetrachloroethylene metabolites show DCA to be a potent, irreversible inhibitor of GSTZ
35      activity, with greater inhibition of this enzyme in mice than in humans.  Studies show that

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-3       DRAFT - DO NOT CITE OR QUOTE

-------
 1      inducers of CYP enzymes, such as toluene, phenobarbital, and pregnenolone-16 alpha-
 2      carbonitrile, increase tetrachloroethylene metabolism, whereas CYP enzyme inhibitors such as
 3      SKF 525A, metyrapone, and carbon monoxide decrease tetrachloroethylene metabolism.
 4      Additionally,  chronic exposure to tetrachloroethylene has been shown to cause self-induction of
 5      metabolism (see Section 3.3.4).
 6
 7      6.1.3. Noncancer Toxicity in Humans and Laboratory Animals
 8             Targets of toxicity observed in human and animal studies include the liver, kidney, CNS,
 9      reproductive system, and developing fetus. Both occupational and residential epidemiologic
10      studies have examined the effects of tetrachloroethylene exposure via inhalation. Humans were
11      found to be particularly sensitive for neurological effects, including decrements in vision or
12      visuo-spatial function, and other neurobehavioral (cognitive) effects following inhalation
13      exposure.  These findings are supported by the consistency of the observations in a number of
14      epidemiologic studies of different designs, populations, and statistical analyses, despite study
15      flaws.  Altmann et al. (1995) identified a pattern of neurobehavioral  deficits in a study of
16      residents living in buildings co-located with a dry cleaning establishment that is similar to the
17      pattern observed in occupational populations with tetrachloroethylene exposures, thus providing
18      evidence of an association with nonoccupational exposure.  A second residential study
19      (Schreiber et al., 2002) also suggests that children may be uniquely susceptible to visuo-spatial
20      effects, but larger studies in humans and studies using animal models are needed to confirm this
21      observation as well as reports of color vision discrimination and contrast sensitivity (black-white
22      discrimination) changes. The large body of evidence assessing neurobehavioral effects and
23      tetrachloroethylene does not permit a  distinction between acute effects and effects of repeated
24      exposure.  Furthermore,  no studies are available to evaluate chronic disabling neurological
25      disease. Occupational studies have examined the effects of tetrachloroethylene on other
26      endpoints as well, with the strongest evidence being for markers of liver and kidney damage and
27      for reproductive/developmental  effects such as spontaneous abortion. The few studies
28      examining inhalation exposure to tetrachloroethylene and immune or endocrine system effects
29      are inadequate for fully evaluating potential associations.
30             The measure of the extent of exposure in many of the epidemiologic studies is imprecise,
31      and, in occupational situations, there are potential exposures to other solvents, although to a
32      lesser extent than with tetrachloroethylene. Relationships between exposure to
33      tetrachloroethylene and responses are not generally observed. Possible explanations for this are
34      exposure misclassification due to use  of current exposure measurements, an exposure or
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-4        DRAFT - DO NOT CITE OR QUOTE

-------
 1      response function that is above the increasing portion of the exposure-response curve, or, more
 2      unlikely, a response that does not increase with increasing tetrachloroethylene exposure.
 3             Epidemiologic studies of oral exposures to tetrachloroethylene have only examined
 4      adverse pregnancy outcomes or postnatal effects (see Section 4.7.1).  There is some evidence for
 5      growth retardation in infants born to mothers residing in housing with drinking water
 6      contaminated with tetrachloroethylene.
 7             Tetrachloroethylene exposure to animals by the inhalation or oral route results in toxicity
 8      to the liver, kidney, and nervous system; by inhalation, it also causes developmental and
 9      reproductive effects.  Specifically, several measures of toxicity have been observed in the liver,
10      such as, increased liver weight, infiltration of fat, necrosis, peroxisome proliferation, polyploidy
11      of hepatocytes, and increased triglycerides. In the kidney, increased weight, hyperplasia, hyaline
12      droplets, and protein cast formation in tubules have been observed.  In the CNS, alteration of
13      brain neurotransmitter levels, increased motor activity, and delayed reaction times to visual
14      stimuli have been observed.  Animals exposed in utero to tetrachloroethylene by inhalation
15      showed signs of fetal growth retardation, increased fetal mortality, and behavioral changes
16      occurring after birth.  There is little information on developmental or reproductive effects in
17      animals by the oral route of exposure.  There are very few studies of immune system toxicity,
18      and none of these studies are in intact animals. No information is available on the effects of
19      tetrachloroethylene on the endocrine system in animals.
20             Targets of toxicity are the same in animal and human studies (i.e., liver, kidney, CNS,
21      reproductive system, and developing fetus).  The effect domain in animals and humans indicates
22      that both cognition and visual function are affected by tetrachloroethylene. Affected organs are
23      all sites of high metabolic activity, and the CNS is also a lipid accumulation site, consistent with
24      the absorption, distribution, metabolism, and elimination profile of tetrachloroethylene.
25
26      6.1.4. Carcinogenicity in Humans and Laboratory Animals
27             Overall, the epidemiologic evidence has associated tetrachloroethylene exposure with
28      excess risks for a number of cancers, although a causal association has yet to be definitively
29      established.  Studies of tetrachloroethylene and cancer showed positive associations between
30      exposure and cancer of the lymphoid system, esophagus, and cervix, with more limited evidence
31      for cancer of the bladder, kidney, and lung. For both lymphoid and esophageal cancer, excess
32      risk was observed in studies of human populations exposed to tetrachloroethylene and other
33      solvents, including studies of exposures to dry cleaners or workers involved with  degreasing
34      metal parts. In these cases, average risks were doubled as compared with those of referents.
35      Furthermore,  studies of drinking water exposure also support an association between lymphoid

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-5        DRAFT - DO NOT CITE OR QUOTE

-------
 1      cancer and tetrachloroethylene and other solvents, as do case-control studies that assessed
 2      employment as dry cleaners or laundry workers.  Chance and confounding by smoking are
 3      unlikely the sole explanations for the observed excesses in risks. Information is lacking on life
 4      style and socioeconomic factors, which are indirect surrogates for human papilloma virus
 5      infection, a known risk factor for cervical cancer (see Section 4.8.1.2 for more details).
 6            The laboratory animal database includes 10 lifetime rodent bioassay data sets that
 7      demonstrate increased cancer incidence. (Two additional study data sets, in male and female rats
 8      exposed orally, were inconclusive due to excessive mortality caused by pneumonia or
 9      tetrachloroethylene-related toxic nephropathy). Hepatocellular adenomas and carcinomas in
10      mice and MCL in  rats occurred in multiple lifetime rodent bioassays, and
11      hemangioendotheliomas in male mice (JISA, 1993) and cancers of the kidney and brain (glioma)
12      in male rats (NTP, 1986) occurred in single lifetime bioassays. Also known as
13      hemangiosarcomas, hemangioendotheliomas are rare tumors of the epithelial lining of blood
14      vessels.  These tumors have been observed in a limited number of bioassays, including vinyl
15      chloride and 1,3-butadiene.  Although the dose-response relationships for kidney and brain
16      tumors observed in male rats were not as strong as for the preceding cancers, and the increasing
17      dose-response trend for kidney tumors was not statistically significantly, both tumor types were
18      considered tetrachloroethylene-related and biologically relevant (see Section 5.4.2.3).
19            The statistically significantly elevated incidences of hepatocellular  carcinomas and
20      adenomas in male and female mice and MCL in male and female rats are considered to be
21      indicators of potential human health hazard, despite questions regarding high background
22      incidences of these tumors in controls and MOA hypotheses (see Section 6.1.5.1).  The finding
23      of an increased incidence of hepatocellular carcinomas and adenomas in female mice with a low
24      background incidence of these tumors suggests tetrachloroethylene is the etiological agent and
25      supports an inference of tetrachloroethylene as a risk factor for liver tumors in male mice that
26      have a higher background incidence. Moreover, kidney cancer and MCL in rats as indicators of
27      a potential human cancer hazard appear reasonable, given the observations in the epidemiologic
28      studies.
29            Although there are segments of the population who may be especially susceptible to the
30      toxic effects of tetrachloroethylene, there are too few studies specifically on tetrachloroethylene
31      to examine this hypothesis directly.  A potential exists for infant exposures from several
32      pathways, including breast or other milk containing tetrachloroethylene. Infants younger than 6
33      months of age have slower renal clearance and less active liver metabolizing enzymes. The
34      nervous  system in the developing fetus and in infants matures later than other systems. Elderly
35      persons and those  with liver and kidney diseases also have slower clearance of toxic

                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  6-6        DRAFT - DO NOT CITE OR QUOTE

-------
 1      substances—especially lipophilic chemicals. Existing PBPK studies are not yet reliable for
 2      quantitative use for estimating pharmacokinetic susceptibility in infants or the elderly (see
 3      Section 4.9.1 for more details).
 4
 5      6.1.5. Mode-of-Action Information
 6            Although a wealth of new data related to understanding the toxic effects caused by
 7      tetrachloroethylene exposure has been published over the past decade, the MOA is not yet
 8      sufficiently characterized, tested, or understood for any one of these adverse effects. A number
 9      of alternative hypotheses are identified and examined as possible MO As for liver and kidney
10      toxicity. Hypothesized MO As for mononuclear cell leukemia, neurotoxicity, and
11      developmental/reproductive effects are indirect and are based on experimental observations of
12      exposures to agents other than tetrachloroethylene. The available evidence points to multiple
13      hypothesized MO As as being involved, and, in each case, no one MOA can be uniquely
14      identified (see Section 4.10.3 for more  details). The sections following immediately below
15      summarize the MOA information available for liver, kidney,  and other targets of
16      tetrachl oroethy 1 ene toxi city.
17
18      6.1.5.1. Liver Mode-of-A ction Information
19            The MOA for tetrachloroethylene-induced liver toxicity, including tumor induction, is
20      not known.  Tetrachloroethylene-induced liver tumors in mice are believed to result from
21      chloroacetic acid metabolites and other intermediate products of the oxidative pathway, with
22      MOA hypotheses  focused on the role of the major urinary metabolite TCA.  Because both
23      tetrachloroethylene and TCA have been shown to activate the PPAR-a, as evidenced by
24      peroxisome proliferation, the ability of PPAR receptors to trigger a number of cellular events
25      suggests a possible relationship with tumor induction. However, metabolism to TCA does not
26      obviously explain  tetrachloroethylene-induced liver tumors, suggesting that other metabolites or
27      intermediates contribute to tetrachloroethylene liver toxicity. Key  steps in one MOA hypothesis,
28      namely that TCA alters cell signaling processes through activation of PPAR-a, have yet to be
29      fully identified both in mice and in humans.
30            Experimental evidence does not support peroxisome proliferation, per se, as a proposed
31      MOA. Specifically, peroxisome proliferation does not correlate well with tumor incidence.
32      Peroxisomes are seen at exposure concentrations higher than those that induce liver tumors, and
33      peroxisome proliferation is also  seen in rat liver and mouse kidney, sites that do not show
34      carcinogenicity (see Section 4.10.3). The  ability of PPAR receptors to trigger nonperoxisomal
35      events suggests that toxicity and tumor induction may not be causally related to peroxisome
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                    6-7        DRAFT - DO NOT CITE OR QUOTE

-------
 1      proliferation, but that tumorigenesis may be only a concurrent happening with many other
 2      events. The relationship between these events and tetrachloroethylene tumor induction is not
 3      understood. At the current time, sufficient evidence does not exist to suggest that
 4      tetrachloroethylene or its oxidative metabolites could initiate hepatocarcinogenesis via a
 5      mutagenic MOA.
 6
 7      6.1.5.2. Kidney Mode-of-Action Information
 8             Several MO As for kidney toxicity are possible, although the supporting evidence is
 9      limited.  Induction of alpha-2|i-globulin occurs only at doses higher than the doses that induce
10      kidney cancer in male rat bioassays, and it is not likely to have an important role in toxicity or
11      tumor induction.  Peroxisome proliferation has been weakly detected in rat kidneys—which do
12      show carcinogenicity—but peroxisome proliferation is more extensive in mouse kidneys, which,
13      too, has not demonstrated cancer. Scientific evidence is more supportive of the  possibility that
14      reactive metabolites from the GSH conjugation pathway  are in some way responsible for kidney
15      toxicity. These metabolites are associated with cytotoxicity and are mutagenic in Salmonella.
16
17      6.1.5.3. Mode-of-Action Information for Other Targets of Toxicity
18            The MOA of tetrachloroethylene-induced leukemogenesis in rats is not well understood;
19      specifically whether the parent compound, a metabolite, or several metabolites are involved.
20      Metabolites from GSH metabolism may contribute to toxicity, as supported by the finding of
21      aplastic anemia and DNA changes in lymphatic  tissues in calves exposed to
22      S-(l,2-dichlorovinyl)-L-cysteine DCVC, which  is structurally similar to the TCVC that is
23      produced through tetrachloroethylene GSH metabolism, although the study  of TCVC in calves
24      was negative.
25            For neurotoxicity, the parent compound, rather than the metabolites, might be exerting an
26      anesthetic-like effect on the lipid membranes in  the nervous system  or interacting with several
27      neurotransmitter receptors. However, this hypothesis is not supported by specific studies on
28      tetrachloroethylene.
29            The MO As hypothesized for developmental toxicity differ according to effect.  The
30      neurobehavioral effects during development may be mediated by the same MOA as the
31      neurotoxic effects discussed above.  For fetal toxicity, TCA, an organic acid, lowers the pH of
32      the fetal compartment (see Section 4.7.4); this may be a contributing factor, given the finding of
33      developmental toxicity with TCA exposure.  These proposed hypotheses, however reasonable,
34      lack experimental support.
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  6-8        DRAFT - DO NOT CITE OR QUOTE

-------
 1            The binding of reactive metabolites of tetrachloroethylene to proteins in liver, kidney and
 2      serum, has the potential to contribute to the pathogenesis of several diseases, including cancer
 3      and autoimmune disease.
 4
 5      6.1.5.4. Mode-of-Action Conclusions and Implications for Dose-Response Analyses
 6            In summary, there is no obvious common MOA for the different toxicological effects of
 7      tetrachloroethylene, nor has a sequence of key events been identified for any of the individual
 8      adverse effects.  MOA information does not indicate in any instance that toxicity endpoints in
 9      animals are not relevant to humans, nor does it provide a basis for non-default procedures for
10      estimating risk or establishing reference values.  Specifically, hypothesized rodent-only MO As
11      are not sufficiently established, and it is reasonable to use animal tumors as an indicator of a
12      potential human cancer hazard.  Rodent tumors, leukemia, and cancer of the liver and kidney
13      have human analogues. For example, mononuclear cell leukemia in rats is also known as large
14      granular lymphocytic leukemia; large granular lymphocyte leukemia represents a well-
15      recognized group of lymphoid neoplasms in humans (Stromberg, 1985).
16            In the absence of a well characterized MOA that could explain dose-response
17      relationships at doses lower than those leading to observed effects, the cancer dose-response
18      modeling is carried out using  a linear extrapolation performed in accordance with default
19      recommendations in the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA,  2005a).
20      The available data on noncancer toxicity of tetrachloroethylene support using EPA's RfC/RfD
21      methodologies to derive noncancer toxicity values.  These approaches are detailed in Section 6.2.
22
23      6.1.6.  Weight-of-Evidence Descriptor for Cancer Hazard
24            Tetrachloroethylene is "Likely to be carcinogenic to humans" by all routes of exposure,
25      within the framework of the 2005  Guidelines for Carcinogen Risk Assessment (U.S. EPA,
26      2005a). As specified in the guidelines, the descriptor "Likely to be carcinogenic to humans"
27      expresses the conclusion regarding the weight of evidence for carcinogenic hazard potential and
28      is presented only in the context of a weight-of-evidence narrative. Although the term "likely"
29      can have a probabilistic connotation in other contexts, its use as a weight-of-evidence descriptor
30      does not correspond to a quantifiable probability of whether the chemical is  carcinogenic.  The
31      five recommended standard hazard descriptors are as follows:
32                •  "Carcinogenic to humans"
33                •  "Likely to be carcinogenic to humans"
34                •  "Suggestive evidence of carcinogenic potential"
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-9        DRAFT - DO NOT CITE OR QUOTE

-------
 1                •  "Inadequate information to assess carcinogenic potential"
 2                •  "Not likely to be carcinogenic to humans"
 3
 4             These descriptors are not unlike those used by the IARC, NTP, and other health agencies
 5      that weigh carcinogenicity evidence. If there are no or insufficient pertinent data, then the
 6      descriptors "Inadequate information to assess carcinogenic potential" or "Suggestive evidence of
 7      carcinogenic potential" are used.  If the evidence is stronger, as is the case with
 8      tetrachloroethylene, the descriptor "Likely to be carcinogenic to humans" is used; convincing
 9      evidence, usually conclusive demonstration of causality in epidemiologic studies, would support
10      "Carcinogenic to humans."  On the other hand, if the conclusion is negative (i.e., strong,
11      consistent and compelling information indicating the absence of human health hazard), the agent
12      would be described as "Not likely to be carcinogenic to humans." Thus, going down the list of
13      descriptors from "Carcinogenic to humans" to "Inadequate information to assess carcinogenic
14      potential" indicates a decrease in the level of evidence of a human health hazard. In summary,
15      use of the weight-of-evidence descriptor "Likely to be carcinogenic to humans" for
16      tetrachloroethylene is intended to communicate that the available information indicates the
17      presence of a human health hazard.
18             The weight-of-evidence conclusion represented by the top three levels of evidence is
19      related to, but distinct from, the quantitative dose-response assessment/conclusions in that the
20      judgment that an agent is a human carcinogen does not guarantee adequate data to quantitatively
21      estimate human risk.  Notably, evaluation of an agent that is judged a likely human carcinogen
22      may offer data conducive to estimating human risk. Indeed, dose-response assessments are
23      generally completed for agents considered "Carcinogenic to humans" and "Likely to be
24      carcinogenic to humans."  Section 6.2 provides the dose-response analyses for
25      tetrachloroethylene.
26             Three lines of evidence in the hazard database support the weight-of-evidence descriptor
27      for the cancer hazard for tetrachloroethylene: (1) tetrachloroethylene exposure is associated with
28      excess risks for a number of cancers in human epidemiologic studies, although a causal
29      association has yet to be sufficiently established; (2) tetrachloroethylene is a rodent carcinogen in
30      10 of 10 lifetime bioassay data sets, including by oral and inhalation routes; and (3) the available
31      information indicates that the cancer bioassay data are relevant to use as indicators of potential
32      human cancer hazard.  Briefly, the epidemiologic evidence has associated tetrachloroethylene
33      exposure with excess risks for a number of cancers including cancer of the lymphoid system,
34      esophagus, and cervix, with more limited evidence for cancer of the bladder, kidney, and lung.
35      For both lymphoid and esophageal cancer, excess risk was observed in studies of people who

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  6-10      DRAFT - DO NOT CITE OR QUOTE

-------
 1      work as dry cleaners or degreasers, populations who experience inhalation exposure to
 2      tetrachloroethylene and other solvents. In both cases, average risks were doubled as compared
 3      with those of referents. Furthermore, studies of drinking water exposure also support an
 4      association between lymphoid cancer and tetrachloroethylene and other solvents, as do case-
 5      control studies that assessed employment as a dry cleaner or laundry worker. Chance and
 6      confounding by smoking are unlikely explanations for the observed excesses in risks.
 7             As summarized in Section 6.1.4, the laboratory animal database includes 10 lifetime
 8      rodent bioassay data sets demonstrating increased cancer incidence. The findings include liver
 9      cancer in both sexes of mice and mononuclear cell leukemia in both sexes of rats following
10      either oral or inhalation exposures and, in single bioassays, male rat kidney and brain tumors
11      (gliomas) and mouse hemangiosarcomas of the liver or spleen. In addition, although not all
12      tetrachloroethylene metabolites have been tested  for carcinogen!city in rodents, the oxidative
13      metabolites TCA and DC A are hepatocarcinogens in one or more species.  Taken together, these
14      data support a weight-of-evidence descriptor of "Likely to be carcinogenic to humans" by all
15      routes of exposure.
16
17      6.2. DOSE-RESPONSE CHARACTERIZATION
18             Quantitative estimates of risk to humans are derived separately for noncancer and cancer
19      effects. RfD and RfC values are derived from  epidemiologic studies of residential populations
20      exposed to tetrachloroethylene from nearby dry cleaning facilities.  Residents in these studies
21      have shown an impaired ability to detect and respond to visual stimuli compared to responses of
22      controls (see Section 5.1.1). Inhalation cancer risk has been estimated from animal data on
23      malignant tumors induced in tests involving lifetime exposure to tetrachloroethylene at known
24      concentrations.
25
26      6.2.1. Noncancer Toxicity (Reference Concentration [RfC]/Reference Dose [RfD])
27             A broad range of animal toxicology and human epidemiologic data are  available for the
28      hazard  assessment of tetrachloroethylene.  The nervous system appears to be a  sensitive organ
29      system, particularly in human studies (see Section 4.6.1). Nevertheless, critical data gaps have
30      been identified and uncertainties associated with  data deficiencies are more fully discussed in
31      Chapter 5  and in the remainder of this section.  Even with these uncertainties, the database of
32      human and animal studies on inhalation and oral  toxicity of tetrachloroethylene can support
33      derivation of inhalation and oral reference values. A number of epidemiologic studies of
34      neurological effects in either occupational workers or residential subjects with
35      tetrachloroethylene exposure or toxicological studies in rodents are considered  for developing an
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                    6-11      DRAFT - DO NOT CITE OR QUOTE

-------
 1      RfC and RfD.  No single epidemiologic study is considered to be without flaws and
 2      uncertainties, although these are different among studies and studied populations. Among the
 3      epidemiologic studies, seven studies were considered for supporting an inhalation reference
 4      value, and a study of neurobehavioral deficits in people residing near dry cleaning facilities
 5      (Altmann et al., 1995) was identified as the principal study using a weight-of-evidence approach
 6      (see Sections 5.1. and 5.2. for more details).  The small number of subjects (14 exposed of 37
 7      subjects studied) can introduce uncertainties particularly regarding stability of statistical
 8      inferences. However,  statistically significant group differences between the adjusted mean
 9      scores of exposed and control subjects on three neurobehavioral tests (simple reaction time,
10      p< 0.05 for the first test and/? < 0.01 for the  second test; continuous performance,/' < 0.05; and
11      visual memory, p < 0.05) were observed after adjusting for covariates and possible confounders
12      of age, gender, and education.  In all cases, the exposed subjects had slower response times or
13      more errors than did the unexposed controls.  Other factors were also considered in the overall
14      weight-of-evidence analysis. Table 6-1 summarizes the rationale for selection of the principal
15      study (the  rationale is also addressed in Sections 5.1 and 5.2).
16
17      6.2.1.1. Assessment Approach Employed
18             Noncancer toxicity RfC and RfD are developed using EPA's RfC and RfD
19      methodologies (U.S. EPA, 1993,  1994). The RfC for tetrachloroethylene is derived through a
20      process of (1) considering all studies and selecting the critical effects that occur at the lowest
21      exposure concentration, (2) selecting the point of departure (POD) at which critical effects either
22      are not observed or would occur at a relatively low prevalence (e.g., 10%), (3) deriving the POD
23      in terms of the HEC, and (4) reducing this exposure concentration by UFs to account for
24      uncertainties in the extrapolation from the study conditions to an estimate of human
25      environmental exposure.
26                    The RfC is developed from the point of departure (POD) of 4.8 mg/m3 (0.7 ppm),
27      which was associated with impaired cognitive function and visual information processing in a
28      study of people residing near dry cleaning facilities (Altmann et al., 1995). The assumption that
29      the residents were continuously exposed to tetrachloroethylene eliminated the need for a duration
30      adjustment to the POD. There  is  sufficient evidence from occupational studies of higher
31      tetrachloroethylene concentrations to confirm that the nervous system is the primary target for
32      the effects of tetrachloroethylene, with several studies showing a similar pattern of effects in the
33      residential study (Seeber, 1989; Ferroni et al., 1992; Cavalleri et al., 1994, Echeverria et al.,
34      1994,  1995).  The median concentration in Altmann et al. (1995) is similar to the concentration
35      in a pilot residential study reporting deficits in visual contrast sensitivity (Schreiber et al., 2002),

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-12      DRAFT - DO NOT CITE OR QUOTE

-------
1
2
Table 6-1.  Summary of rationale for principal study selection
        Consideration/
           approach
               Type of data
                        Decision
       Quality of study
           Animal neurotoxicity
           studies.
Animal neurotoxicity studies are considered as supporting
studies. An RfC/RfD from human data, if available and of
adequate quality, is from the species of interest to EPA,
reduces interspecies extrapolation uncertainties and is
preferred.
       Quality of study
           Human neurotoxicity
           studies.
Both occupational and residential studies on
tetrachloroethylene exposure contain uncertainties regarding
their use for quantitative analysis. Some of these
epidemiology studies carry greater weight for quantitative
analysis than the less informative studies of Ferroni et al.
(1992), Echeverria et al. (1994, 1995), and Spinatonda et al.
(1997) with reporting or exposure assessment deficiencies.
       Measurement
       tool
           Standardized
           neurobehavioral
           battery.
Both occupational and residential epidemiology studies
assessed neurobehavioral function using a standardized
neurobehavioral battery.  The battery has been widely
administered to occupational populations in different
settings with a reasonably high degree of validity.  WHO
and ATSDR recommend these test methods to evaluate
nervous system deficits in adults and children.
       Endpoint
           Deficits in
           neurological domains
           such as attention,
           motor function,
           vigilance, or visuo-
           spatial function.
There is congruence of neurological effects observed in
studies of both residential and occupational populations.
These domains are also sensitive to acute tetrachloroethylene
exposure in controlled human studies.  The consistency of
observed effects between occupational and residential
populations and their persistence with lower
tetrachloroethylene concentration, as experienced by
residential populations, provide a strong rationale for a study
of lower-level residential exposures as the basis for the RfC.
       Relevance of
       Exposure
       Scenario
           Epidemiology studies
           of residential
           populations.
Tetrachloroethylene exposure to residential populations is of
lower concentration and of chronic duration compared to
acute duration and higher concentration exposure to
occupational populations.  Additionally, potential to
tetrachloroethylene peak or intensity concentrations is more
common with occupational exposures. A study of
residential exposure, if adequate and of similar quality as an
occupational epidemiology study, is preferred for supporting
the RfC because it represents exposure scenarios of interest
to EPA.
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                      6-13       DRAFT - DO NOT CITE OR QUOTE

-------
 1      and it is lower than the concentration associated with markers of kidney and liver damage and
 2      reproductive/developmental effects (spontaneous abortion and changes in sperm quality) in
 3      several epidemiologic studies of occupational exposure (Franchini et al., 1983; Gennari et al.,
 4      1984; Olsen et al., 1990; Eskenazi et al., 1991a, b; Mutti et al., 1992; Brodkin et al., 1995; Doyle
 5      et al., 1997; Verplanke et al., 1999; Trevisan et al, 2000). Effect levels in animal studies are,
 6      generally, similar to those in the occupational epidemiologic studies, with BMC modeling of
 7      liver toxicity showing effect concentrations only slightly higher  than the POD from residents
 8      (Kjellstrand et al., 1984; USA, 1993).
 9             A composite UF of 300 is adopted to account for human  variation (factor of 10),
10      extrapolation from an observed-effect level to a no-effect level (factor of 10), and uncertainties
11      in the database (factor of 3).  Limited data on tetrachloroethylene blood concentration among
12      human subjects indicated that a choice of 3 for the pharmacokinetic portion of the 10-fold human
13      variation UF is reasonable.  The rationale for a 3-fold database UF is based on critical data gaps
14      and takes into account a lack of animal studies designed to clearly investigate the human findings
15      in cognition and visual system dysfunction and a lack of cognitive testing in both
16      developmentally exposed animals and adult animals exposed to tetrachloroethylene for longer
17      than acute durations (see Sections 5.1, 5.2, and 5.3 for further discussion of these issues).  These
18      data are needed to allow for a fuller characterization of the exposure-response relationship.  The
19      RfC was calculated by dividing the POD by the composite UF = 4.8 mg/m3/300 = 1.6 x 10'2
20      mg/m3.
21             The database for oral exposure to tetrachloroethylene is limited to four subchronic gavage
22      studies, one subchronic drinking water study, and no human studies. In addition to using the
23      animal data on oral exposure, the assessment attempted to expand the database for derivation of
24      an RfD using relevant inhalation data and route-to-route extrapolation with the aid of a
25      pharmacokinetic model.  Route extrapolation of human inhalation data is considered  a
26      reasonable alternative to using the limited oral data in animals because tetrachloroethylene has
27      been shown to be rapidly and well absorbed by the oral and inhalation routes of exposure, and
28      the metabolic pathways and kinetics of excretion with oral exposure are similar to those of
29      inhalation exposure.  Furthermore, human data, when adequate,  are preferred for supporting the
30      RfD, and human data of inhalation exposure are available.
31             The residential inhalation study of Altmann et al. (1995)  of neurobehavioral deficits and
32      three acute and subchronic toxicological studies were examined  for supporting an RfD. The RfD
33      was derived from Altmann et al. (1995) with the aid of an extrapolation from the inhalation to
34      the oral route using pharmacokinetic modeling. The daily oral ingestion dose that results in the
35      same tetrachloroethylene blood concentration associated with the POD for inhalation, 4.8 mg/m3,

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-14       DRAFT - DO NOT CITE  OR QUOTE

-------
 1      is 1.1 mg/kg-day.  This value is equivalent to an oral LOAEL.  Blood tetrachloroethylene
 2      concentration is a well-validated dose metric, and the estimate varies little between models. The
 3      human LOAEL falls within the PODs from oral studies in animals. The UFs, as used for the
 4      inhalation RfC, are adopted for the RfD for oral exposure, namely, a composite factor of 300 (10
 5      for human variation, 10 for extrapolation from a LOAEL to a NOAEL, and 3 for database
 6      uncertainties; see also Sections 5.1, 5.2, and 5.3). The  oral RfD is, therefore, 1.1 mg/kg-day/300
 7      =4 xlQ'3 mg/kg-day.
 8            To show the range of tetrachloroethylene concentrations at which different neurotoxic
 9      effects and toxic effects in other organ systems have been observed, the points of departure and
10      reference values that could have been derived from these other studies were compared with that
11      of the principal study. These graphs allow a direct visualization of how the values compare to
12      the data from which the principal conclusions have been derived.  This has been done for both
13      inhalation reference concentrations and oral reference doses in Figures 6-1, 6-2, and 6-3.
14
15      6.2.1.2. Impact of Assumptions, Uncertainties and Alternatives on Reference Concentration
16              and Reference Dose
17            A number of uncertainties underlie the RfC  and RfD for tetrachloroethylene and are
18      discussed below in this section. A  quantitative characterization of the uncertainty in the RfC and
19      RfD for tetrachloroethylene is not feasible because  of the varied nature of the available database
20      and the limited data available for many of the studies.  Most significantly, the available chronic
21      toxicity studies of tetrachloroethylene exposure demonstrated varying degrees of support for a
22      POD for the RfC and RfD. A weight-of-evidence approach was adopted to identify principal or
23      critical studies, with the additional  studies supporting the principal studies  (see also Sections 5.1
24      and 5.2).
25
26      6.2.1.2.1. Point of departure. Most of the available studies did not provide enough data to
27      support benchmark dose modeling; they only supported PODs based on LOAELs,  especially for
28      the human studies, or LOAELs and NOAELs.  Such a POD has a number of shortcomings
29      relative to a POD obtained from benchmark dose-response modeling (i.e., a benchmark
30      concentration).  First, LOAELs and NOAELs are a reflection of the particular exposure  levels at
31      which a study was conducted, contributing some inaccuracy to the POD determination.  Second,
32      LOAELs and NOAELs  reflect the number of study subjects or test animals and typically are
33      dissimilar in detection ability and statistical power,  with smaller studies tending to identify
34      higher exposure levels as PODs relative to larger but otherwise similarly designed  studies. This
35      is an important consideration for studies with multiple  exposure groups and studies that  did not
36      identify LOAELs but has much less impact for the single-group studies that identified a LOAEL.
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-15       DRAFT - DO NOT CITE OR QUOTE

-------
o
Oi
               Inhalation  Neurotoxicity RfVs
O
oo
   I
   Co

H £
                                    Human      Human      Human      Human
                                                                                   Human
                                                                                                Rat
                                                                                                           Gerbil
                   Q.
                   Q.
                   o
                   5
                   •s
                   o
                   o
                      0.001
                     0.0001
                    0.00001
                       0.01
                                                                                                                        --100
                                                                                                                        -- 10
                                                                                                                        --1
                                                                                                                        - 0.1
                                                                                                                        -- 0.01
                                                                                             o
                                                                                             c
                                                                                             o
                                                                                             o
                                                                                                                        - 0.001
  Visual
 contrast
 (day care
employees)
 (Schreiber
etal.,2002)
                                                         Vision, cognitive  Vision, 5%  Visual function,   Visual
                                                            function,     change in  ~10% change   function
  Visual
 contrast
(apartment  vigilance,3 ~15%
 residents)    decrement
 (Schreiber     (Altmann
etal.,2002)    etal.,1995)
                                                                         color
                                                                        confusion
                                                                         index
                                                                        (Cavalleri
                                                                       etal., 1994)
  in neuro-
  behavioral
   tests
(Seeber, 1989)
 (Mattsson
etal., 1998)
   Brain
 chemistry
(Rosengren
etal., 1986)
                        Principal study.
KEY
  4  Point of departure
     (POD)
I    I UF, extrapolation from a
     lowest-observed-
     adverse-effect level
     (LOAEL) to a no-
     observed-adverse-effect
     level (NOAEL)
I    I UF, animal-to-human
     extrapolation

     UF, human variation
     UF, extrapolation from
     subchronic exposure
     duration to chronic
     exposure duration
 • UF, database
     deficiencies

  O  Reference value (RfV)
                  Figure 6-1.  Array of PODs and reference values for a subset of neurotoxic effects in inhalation studies.

-------
H  £
                             Inhalation Organ-Specific RfVs
              100
               10
                             Human
                                 Human
Rat
Mouse
o
•a
ra
is
c

8

O
o
              0.1
             0.01
             0.001
            0.0001

n i
II 1
^ 1 UUUUUU1
I I 1
• 111.
1

CMS, ~'\5% Kidney, increase in Developmental/ Liver, 10%
decrement in verbal mean urinary Reproductive, 1% increase in liver
function, cognitive concentration of increase in fetal angiectasis,
function, vigilance, several renal death, BMCLoi BMCUo/p as POD
and vision3 enzymes as POD (J ISA, 1993)
(Altmann et al., (Multti (Tinston
1995) etal.,1992) et al., 1994)
- 100

- 10 j~*
E
-1 I
p
:oncentra1
-0.01






KEY
A Point of departure
(POD)
I I UF, extrapolation
from a lowest-
observed-adverse-
effect level (LOAEL)
to a no-observed-
adverse-effect level
| 	 1 (NOAEL)
UF, animal-to-human
cmi extrapolation
^_ UF, human variation
^^ UF, database
deficiencies
Reference value
(RfV)

           Principal study.
   TO
              Figure 6-2.  Organ-specific RfVs for inhalation exposure to tetrachloroethylene.

-------
o
Oi
O
oo
   I
   Co

00
   I
   Co
   §
H I
O >
HH Oq
H TO
si
                 100
                                Oral  Organ-Specific RfVs
                        Human            Mouse           Mouse             Rat
           01
           1/1
                0.01
               0.001
              0.0001
                  CMS, -15% decrement
                     in verbal function,
                     cognitive function,
                   vigilance, and vision8
                   (Altmann et al., 1995)
   Neuro-
developm ental
 (Fredriksson
 et al., 1993)
  Liver, one S.D.
change in the m ean
liver weight over the
 control response
   (Buben and
 O'Flaherty, 1985)
  System ic
  toxicity
(Hayes et al.,
   1986)
aPrincipal study.
KEY
 4   Point of departure
     (POD)
\    \ UF, extrapolation from a
     lowest-observed-
     adverse-effect level
     (LOAEL) to a no-
     observed-adverse-effect
     level (NOAEL)
^^ UF, animal-to-human
     extrapolation

""" UF, human variation
I    I UF, extrapolation from
     subchronic exposure
     duration to chronic
     exposure duration
^H UF, database
     deficiencies

  O  Reference value (RfV)
    Figure 6-3  Oral organ-specific reference values for exposure to tetrachloroethylene.
H
W

-------
 1      Third, LOAELs and NOAELs represent different response rates, as noted on Figures 6-1, 6-2,
 2      and 6-3, and qualitative and quantitative comparisons are not possible lacking characterization of
 3      the underlying dose-response curve.
 4             PODs identified from fitting benchmark dose models overcome some of the deficiencies
 5      associated with LOAELs and NOAELs. Benchmark dose models were fit to five data sets
 6      (Buben and O'Flaherty, 1985; JISA,  1993; NTP, 1986; Brodkin et al., 1995; Tinston, 1994) with
 7      sufficient information. The choice of benchmark dose model did not generally lead to significant
 8      uncertainty in estimating the POD since benchmark effect levels were within the range of
 9      experimental data.  While this examination of a subset of chronic toxicity studies on
10      tetrachloroethylene exposure provides some insight on study and endpoint differences in PODs,
11      lacking characterization of dose-response curves for all studies, especially the more critical
12      studies, uncertainty associated with the PODs cannot be adequately quantified in this database.
13             Effects in the CNS and in other organ systems in  occupational populations and in animals
14      are observed at higher average tetrachloroethylene concentrations than the Altmann et al. (1995)
15      residential study. Uncertainties in other studies of neurotoxicity and of other organ systems
16      differ from those of Altmann et al. (1995). For both occupational and residential populations,
17      studies do not describe a NOAEL and human variation is not well characterized in study
18      subjects.  Uncertainties associated with the occupational  studies include (1) potential for
19      neurobehavioral effects at lower exposures and (2)  exposure pattern differences between
20      occupational and residential studies with peaks characterizing occupational exposures. Using an
21      occupational study to support the RfC may not be fully protective of neurological effects as has
22      been observed in populations co-located near dry cleaners (Altmann et al., 1995; Schreiber et al.,
23      2002; and NYSDOH, 2005a, c). For animal  studies, uncertainties are associated with
24      extrapolating high concentration exposure, typically of subchronic duration to genetically inbred
25      rodents, to infer a concentration of tetrachloroethylene that is likely to be without an appreciable
26      risk of adverse health effects over a lifetime to a diverse human population.
27
28      6.2.1.2.2. Extrapolation from laboratory animal studies to humans. Extrapolating from
29      animals to humans embodies further issues and uncertainties. First, the effect and its magnitude
30      associated with the concentration at the point of departure in rodents is extrapolated to human
31      response. Pharmacokinetic models are useful to examine species differences in pharmacokinetic
32      processing. This was possible for liver toxicity where limited MOA information suggests
33      metabolism as important to toxicity.  The ranges of BMCLs presented for liver effects (a 10-fold
34      range of estimates of tetrachloroethylene metabolism) demonstrate the uncertainty in
35      tetrachloroethylene pharmacokinetic models.  The discrepancies among the models and

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                    6-19      DRAFT - DO NOT CITE OR QUOTE

-------
 1      experimental data may point to large uncertainties in the parameters used in these models.
 2      Because the accuracy of the models has been evaluated only against blood and breath
 3      concentrations of the parent compound, their reliability for predicting total metabolites is
 4      unknown.
 5
 6      6.2.1.2.3. Human variation. Heterogeneity among humans is another uncertainty associated
 7      with extrapolating doses from animals to humans. Uncertainty related to human variation needs
 8      consideration, also, in extrapolating dose from a subset or smaller sized population, say of one
 9      sex or a narrow range of life stages typical of occupational epidemiologic studies, to a larger,
10      more diverse population.
11             In the absence of tetrachloroethylene-specific data on human variation, a factor of 10 was
12      used to account for uncertainty associated with human variation.  Human variation may be larger
13      or smaller; however, tetrachloroethylene-specific data for examining the potential magnitude of
14      over- or under-estimation are few. The pharmacokinetic model of Clewell et al. (2004) of mean
15      physiological parameters used to explore age-dependent pharmacokinetic differences suggests a
16      2-fold variation in blood tetrachloroethylene levels (Chapters 3 and 5).  Bois et al. (1996) and
17      Chiu and Bois (2006) have examined uncertainty and variation in a tetrachloroethylene
18      pharmacokinetic model describing the amount of tetrachloroethylene metabolism.  This analysis
19      suggests large uncertainty is associated with estimating the quantity of tetrachloroethylene
20      metabolism in humans.
21
22      6.2.1.2.4. Database uncertainties. Critical data gaps have been identified with uncertainties
23      associated with database deficiencies on developmental, immunologic, and neurotoxic  effects,
24      particularly data to characterize dose-response relationships and chronic visuo-spatial functional
25      deficits and cognitive effects of tetrachloroethylene exposure under controlled laboratory
26      conditions.  Several halogenated organic solvents  have been linked with altered immune system
27      function in both animals and humans (e.g., toluene, trichloroethylene).  Additional data from
28      inhalation, oral, and dermal exposures at different durations are needed  to assess the potential
29      immunotoxicity of tetrachloroethylene.  This lack of data, combined with the concern that other
30      structurally related solvents, has been associated with immunotoxicity and contributes to
31      uncertainty in the database for tetrachloroethylene.
32             Data from acute studies in animals (Warren et al., 1996; Umezu et al.,  1997) suggest that
33      cognitive function is affected by exposure to tetrachloroethylene.  These studies do not address
34      the exposure-response relationship for subchronic and chronic tetrachloroethylene exposures on
35      cognitive functional deficits observed in humans (e.g., Seeber,  1989; Echeverria et al.,  1994; and

                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                    6-20       DRAFT - DO NOT CITE OR QUOTE

-------
 1      Altmann et al., 1995).  Even more importantly, cognitive testing is lacking in both
 2      developmentally exposed animals and adult animals following exposures to tetrachloroethylene
 3      that are longer than acute durations of exposure.  Visual system dysfunction and processing of
 4      visuo-spatial information are sensitive endpoints in human studies. The exposure-response
 5      relationship of these functional deficits could be evaluated more definitively with studies using
 6      homologous methods that examine retinal and visual function in experimental animals.
 7      However, there has been a limited evaluation of visual function in rodents, with the exception of
 8      the evoked potential studies by Mattsson et al. (1998). These types of studies could help
 9      determine whether there are both peripheral and central nervous system effects of
10      tetrachloroethylene exposure on visual perception, and they could be used as an animal model to
11      better define the exposure-response relationships.
12            Subjects in the  epidemiologic studies comprise adults, and some characterization of the
13      response of children to tetrachloroethylene exposure was found in limited data for a similar
14      neurological (visual system) parameter (Schreiber et al., 2002) and in a larger number of subjects
15      (NYS DOH, 2005 a,c) using other visually based testing paradigms.  Additionally, in a postnatal
16      neurotoxicity  study in  mice (Fredriksson et al., 1993), persistent neurological effects (i.e.,
17      increased locomotion and decreased rearing behavior at 60 days of age, measured 43 days after
18      exposure ceased) were observed at an oral dose of 5 ppm, with no NOAEL, although this study
19      did not conform to traditional toxicity testing guidelines (see  Section 4.6.2.2). These results
20      suggest that if adequate, robust, dose-response data using the most appropriate
21      neurophysiological and cognitive tests were available, the exposure eliciting an adverse response
22      (and hence the POD for the reference value)  could be lower than that established based on
23      deficits in visuo-spatial and cognitive function following tetrachloroethylene exposure in healthy
24      adults (Altmann et al.,  1995).
25
26      6.2.2. Cancer Risk Estimates
27            Following the scientific principles and procedures outlined in EPA's  Guidelines for
28      Carcinogen Risk Assessment (U.S. EPA, 2005a), the cancer risk values are based on the 95%
29      lower confidence limits on the air concentrations associated with a 10% extra risk of cancer
30      incidence (LECios). The LECio values were calculated from  data on MCL in male rats, the most
31      sensitive species/gender in the rodent cancer bioassay conducted at the lowest concentration
32      range, using the multistage dose-response model. A linear low-dose extrapolation was then used,
33      in accordance with default recommendations in EPA's Guidelines for Carcinogen Risk
34      Assessment (U.S. EPA, 2005a).  The approach and associated choices and assumptions are
35      described in Sections 6.2.2.1.
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  6-21       DRAFT - DO NOT CITE OR QUOTE

-------
 1             A broad range of animal toxicology and human epidemiologic data are available for the
 2      identification of a carcinogenic hazard from exposure to tetrachloroethylene. Nevertheless,
 3      critical data gaps have been identified, and uncertainties associated with data deficiencies are
 4      more fully discussed in Chapter 5 and Section 6.2.2.2.  Given the choices of tumor type, point of
 5      departure, and low-dose extrapolation approach necessary to provide a plausible upper bound
 6      risk estimate, there are additional considerations that contribute to uncertainty in the cancer risk
 7      values. These uncertainties have a varied impact on risk estimates.  Some (i.e., the bioassay or
 8      cross-species scaling approach)  suggest risks could be higher than estimated while others would
 9      decrease estimates or have an impact of uncertain direction (i.e., the human population
10      variability, dose metric, and model-based uncertainty at the POD).  While some uncertainties
11      could be quantitatively characterized, it is likely that the residual uncertainties remain the largest,
12      yet can only be  qualitatively expressed: i.e., low dose extrapolation, MOA, and human
13      sensitivity and variability.  Even if experimental data could further elucidate these uncertainties,
14      extrapolation of animal bioassay data to human (done here using allometric scaling) will  remain
15      a substantial and unknown uncertainty.  The tetrachloroethylene unit risk estimate, calculated
16      using three PBPK models, ranges from 2 x 10"6 to 2 x 10"5 per |ig/m3. From this range, the upper
17      end unit risk of 2 x 10"5 per |ig/m3 is the most public health protective value for the upper bound
18      risk estimate.
19             The tetrachloroethylene oral slope factor, using the three PBPK  models for route-to-route
20      extrapolation from the experimental data to humans,  ranges from 1 x 10"2 to 1 x 10"1 per mg/kg-
21      day.  From this  range, the upper end  slope factor of 1 x 10"1 per mg/kg-day is the most public
22      health protective value for the upper bound risk estimate. With the exception of the route-to-
23      route extrapolation step, the uncertainties associated with the slope factor estimation are the same
24      as for the unit risk estimation.
25             Section  6.2.2.2 describes the uncertainties outlined above, their impact on cancer  risk
26      estimation, the choices made and justification for each, and the associated data gaps. Section
27      6.2.2.3 provides a quantitative analysis of the potential numeric impact  of three of these sources
28      of uncertainty on the unit risk estimate (the statistical uncertainty, PBPK model, and tumor site)
29      using the multistage model in the observed range and linear low-dose extrapolation. Section
30      6.2.2.4 and the table therein provide a summary of the cancer risk estimate.
31
32      6.2.2.1. Assessment Approach Employed
33             Animal bioassay data are used to derive quantitative cancer risk estimates for humans due
34      to the lack of quantitative exposure information in the occupational epidemiology studies.  The
35      cancer dose-response analysis considers three bioassays but relies on the JISA (1993) study
                   This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-22      DRAFT - DO NOT CITE OR  QUOTE

-------
 1      results. This is primarily because the JISA (1993) study included lower exposures than did the
 2      two earlier bioassays for both species tested, and, therefore, its use reduces extrapolation
 3      uncertainty slightly. For mice, the lowest exposure concentration of 10 ppm was 10-fold lower
 4      than the lowest exposure concentration in the NTP inhalation study (NTP, 1986). For rats, the
 5      low-exposure concentration of 50 ppm was fourfold lower than in the NTP (1986) study.
 6            Dose-response analyses for hepatocellular tumors in male and female mice,
 7      hemangiosarcomas in male mice, and mononuclear cell leukemia and kidney tumors in male and
 8      female rats are carried out using the rate of total metabolite production as estimated from the
 9      three recently developed toxicokinetic models as the dose metric. Dose-response analyses are
10      also carried out using administered dose as the dose metric to allow comparison to the
11      pharmacokinetic model-based risk estimates. EPA's methodology for cross-species scaling was
12      applied for relevant tumor sites to address toxicological equivalence across species (U.S. EPA,
13      1992).  This methodology  is based on the observation that equal average lifetime concentrations
14      or AUC of the toxic moiety has been associated with toxicological equivalence across species.
15      This cross-species relationship has been shown to accommodate the general species variation in
16      pharmacokinetics and the carcinogenic response to internal doses. Although the available
17      pharmacokinetic data for tetrachloroethylene do not allow estimates of AUC, the use of
18      metabolized tetrachloroethylene scaled to mg/kg3/4-day in order to estimate the dose resulting in
19      the same lifetime risk in animals and humans is consistent with the EPA methodology and
20      further substantiated in the present document (see Section 5.4.4.2.1). This consideration of
21      cross-species scaling and toxicological equivalence is consistent with EPA's other carcinogen
22      assessments and its treatment of pharmacokinetic dose metrics.
23            The steps involved in generating the unit risk from the dose-response data are illustrated
24      using the male rat MCL data, as follows:
25         (1) A fit of the tumor incidence versus total metabolite curve using a multistage model
26            (BMDS, version 1.3.2) gave an LECio,  or 95% lower confidence bound on the exposure
27            associated with 10% extra risk,  of 0.81  per mg-eq/kg-day (Figure 5-8a);
28         (2) The point of departure (LECio)  was then transformed to a human equivalent value by
29            dividing the animal value by (human body weight /animal body weight)0'25 = (70/0.45)°'25
30            = 3.53 to give a human equivalent value of 0.23 mg-eq/kg-day of metabolite formation;
31         (3) Three different models (see Section 3.5) of total human metabolite formation from
32            tetrachloroethylene exposure were used to estimate the environmental exposure that
33            would correspond to the human equivalent LECio (Bois et al., 1996; Rao and Brown,
34             1993; Reitz et al., 1996). The lowest human equivalent LECio resulting from the three
35            models (Rao and Brown) is 47,000 |ig/m3. The highest human equivalent LECio
36            resulting from the models is 4,700 |ig/m3 (Bois et al.);
                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                  6-23       DRAFT - DO NOT CITE OR QUOTE

-------
 1         (4) The unit risk calculated using three PBPK models ranges from 2 x 10"6 to 2 x 10"5 per
 2            Hg/m3.  From this range, the upper-end unit risk of 2 x 10"5 per |ig/m3 is the most public
 3            health protective value for the upper bound risk estimate.
 4
 5            Age adjustment factors for early life exposures as discussed in the Supplemental
 6      Guidance for Assessing Susceptibility for Early-Life Exposure to Carcinogens (U. S. EPA,
 7      2005b) are not  recommended because little evidence exists to indicate that tetrachloroethylene or
 8      its oxidative metabolites directly damage DNA, information about genotoxicity of GSH
 9      metabolites in cell assays other than Salmonella or in in vitro experiments are lacking, and the
10      MO A for tetrachloroethylene has not been established.
11
12      6.2.2.2. Impact of Assumptions, Uncertainties and Alternatives on Unit Risk Estimates
13            A number of uncertainties underlie the cancer unit risk for tetrachloroethylene.  These are
14      discussed in the following paragraphs. Specifically addressed is the impact on the assessment of
15      issues such as the use of models and extrapolation approaches, the reasonable alternatives, the
16      choices made, and the data gaps identified. In addition, the use of assumptions, particularly
17      those underlying the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), is explained
18      and the decision concerning the preferred approach is given and justified.  Several of the
19      uncertainties with the largest impact cannot be considered quantitatively, such as human
20      population variability and the most relevant dose metric.  Thus, an overall integrated quantitative
21      uncertainty analysis is not presented.
22
23      6.2.2.2.1. Human population variability. The extent of inter-individual variability in
24      tetrachloroethylene metabolism has not been characterized.  As noted in Section 6.1.2, several
25      enzymes of the oxidative and GSH metabolism, notably CYP2E1,  CYP3A4, GSTZ, GSTA,
26      GSTM, and GSTT show genetic polymorphisms with the potential for variation in metabolite
27      production. The limited data available on tetrachloroethylene metabolites show DCA to be a
28      potent, irreversible inhibitor of GSTZ activity, with greater inhibition of this enzyme in mice
29      than in humans. Tetrachloroethylene metabolism has been shown  to increase by inducers of
30      CYP enzymes such as toluene, phenobarbital, and pregnenolone-16 alpha-carbonitrile, whereas
31      CYP enzyme inhibitors such as  SKF 525A, metyrapone, and carbon monoxide have been shown
32      to decrease tetrachloroethylene metabolism. Additionally, chronic exposure to
33      tetrachloroethylene has been shown to cause self-induction of metabolism.  Human population
34      variability  is summarized above (see Section 6.2.1.2.3) and in covered in more detail in
35      Chapters.

                  This document is a draft for review purposes only and does not constitute Agency policy
        06/06/08                                   6-24      DRAFT - DO NOT CITE OR QUOTE

-------
 1           A separate issue is that the human variability in response to tetrachloroethylene is also
 2    poorly understood. The effect of metabolic variation, including potential implications for
 3    differential toxicity, has not been well studied.  Although a mutagenic MOA would indicate
 4    increased early-life susceptibility, there are no data exploring whether there is differential
 5    sensitivity to tetrachloroethylene carcinogenicity across life stages. Thus, this lack of
 6    understanding about potential differences in metabolism and susceptibility across exposed
 7    human populations represents a source of uncertainty. Nevertheless, the existing data support
 8    the possibility of a heterogeneous response that may function additively to ongoing or
 9    background exposures, diseases, and biological processes. As noted in Section 4.9.5., some
10    evidence shows certain subpopulations may be more susceptible to tetrachloroethylene exposure.
11    As discussed under (2) below, these considerations strengthen the scientific support for the
12    choice of a linear non-threshold extrapolation approach. In summary, the human equivalent risk
13    estimates for tetrachloroethylene, therefore, do not reflect this source of uncertainty.
14
15    6.2.2.2.2.  Choice of low-dose extrapolation approach. A key consideration in clarifying how
16    risks should be estimated for low-dose exposure is the MOA. As noted above in Section 6.1.5,
17    MOA data are lacking or limited for all of the candidate cancer endpoints for tetrachloroethylene
18    (i.e., rat MCL and kidney tumors, mouse hepatocellular tumors and hemangiosarcomas). When
19    the MOA cannot be clearly defined, EPA uses a linear approach to estimate low-dose exposure
20    risk, based on the following broad and long-held scientific assumptions, which supported
21    development of EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a):
22       •   A chemical's carcinogenic effects may act additively to ongoing biological processes,
23           given that diverse human populations are already exposed to other agents and have
24           substantial background incidence of various  tumors.  Under these conditions, a nonzero
25           slope of the response as a function of chemical exposure is expected.
26       •   A broadening of the dose-response curve in the human population (less rapid fall-off with
27           dose) and, accordingly, a greater potential for risks from low-dose exposures (see Zeise et
28           al., 1987; Lutz et al., 2005) would result for two reasons.  First, even if there is a
29           threshold concentration at the cellular level, that threshold is likely to be different among
30           different individuals. Second, greater variability is anticipated in response to exposures
31           in the heterogeneous human population than in controlled laboratory species and
32           conditions (due to, e.g., genetic variability, disease states, age).
33       •   The general use of linear extrapolation provides plausible upper-bound risk estimates that
34           are believed to be health-protective (U.S. EPA, 2005a) and  also provides consistency
35           across assessments.
36
37           The extent to which the overall uncertainty in low-dose risk estimation could be reduced
38    if the MOA for tetrachloroethylene were known with a high degree of confidence is of interest,
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   6-25       DRAFT - DO NOT CITE OR QUOTE

-------
 1    but clear data on the MOA of tetrachloroethylene is not available, and even if it were,
 2    incorporation of MOA into dose-response modeling might not be straightforward and might not
 3    significantly reduce the uncertainty about low-dose extrapolation. This is because the MOA, as
 4    well as other factors, especially human response variability, are determinants of the dose-
 5    response function in humans.
 6           This chemical assessment also evaluates the extent to which a collection of mathematical
 7    functions fit to one of the tetrachloroethylene bioassay data sets and extrapolated down to low
 8    doses, could inform uncertainty. There is not sufficient information regarding the MOA to
 9    support a chemical-specific inference about dose-response behavior at low dose for
10    tetrachloroethylene. Thus, it is of interest to observe how different functions fit to the tumor data
11    may diverge when extrapolated downward.  Much previous experience has supported a general
12    mathematical property that different curves, though fitting observed experimental data well,
13    often diverge widely when extrapolated to doses well outside the observed range.  Indeed, the
14    inability of curve-fitting procedures to provide useful compound-specific information about low-
15    dose risks has been a principal motivation for the "model free" approach of straight line
16    extrapolation from a point of departure within the observed range of the data (Krewski and van
17    Ryzin,  1981; NRC, 1983).
18           Calculations here encompassed four alternative functional forms frequently used for
19    noncancer dose-response assessment in the observable range of experimental data (multistage,
20    Weibull, log-logistic, and log-probit) that can accommodate a wide variety of dose-response
21    shapes, including threshold-like behavior. These models were fit to the mononuclear cell
22    leukemia data in male rats using the EPA BMDS program, and distributions of model results
23    were evaluated (see Appendix 5B for more details).  These calculations confirm the expected
24    finding that alternative functional forms fit to this tetrachloroethylene tumor data set are
25    consistent with a wide range of numerical values for probability of response when extrapolated
26    down to low dose, as illustrated in Table 6-2.
27           With such large spreads in confidence intervals, the extrapolated models, in effect,
28    provide little information about actual low-dose risks.  These results are not presented as the
29    basis for alternative estimates of human risk, because they do not provide sound or useable
30    scientific estimates for the compound-specific risks from tetrachloroethylene. As noted
31    previously, such results serve to underscore the EPA Cancer Guidelines' rationale for the use of
32    a consistent model-independent approach.
33           A number of different biological motivations have been put forward to support functional
34    forms that might be used to estimate risks from low-dose exposure to carcinogens or other toxic
35    substances. For cancer, the most prominent class of models treats tumorigenesis as a multi-event

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   6-26       DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
            Table 6-2. Summary of dose-specific extra risks (means and 95% confidence
            limits) for four dose-response models fit to incidence of leukemias in male
            rats exposed to tetrachloroethylene via inhalation (JISA, 1993)
Model
Log-probit
Multistage
Log-logistic
Weibull
Estimated extra risk, (mg-eq/kg-day)"1,
corresponding to 1.5 x 10"5 mg-eq/kg-day internal dose in ratsa
5th percentile
2 x 10-170
4.546 x 10'13
2 x 10'21
0
Mean
8.172xlO'4
9.172 x 10'7
1.078 x 10'3
1.339 x 10'3
95th percentile
1.008 x 10'5
1.819 x 10'6
7.776 x 10'4
1.893 x 10'3
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
      From Appendix 5B.

     process and characterizes the probability of accumulation of a series of changes (conceptualized
     as mutations or other events) that, together, will result in formation of a malignant tumor. In
     particular, EPA utilized the multistage model for low dose extrapolation of cancer risks in many
     assessments. Risk estimates utilizing EPA's application of the multistage model have been
     shown to be similar to the linear (straight line) risk estimation procedure now used by EPA
     (Subramaniam et al., 2006). More complex multi-event models allow for the modeling of
     formation and growth of populations of initiated and transformed cells and are still well
     recognized tools for investigating biologically based dose response modeling for carcinogenesis.
           The concept of a distribution of individual thresholds is a second approach used to
     motivate functional forms for dose-response modeling. Such models assume that there is an
     "individual  threshold" for each member of the human population, and interindividual variation in
     these thresholds determines the  dose-response curve for a population. A recent National
     Research Council report on risk assessment issues for TCE (NRC, 2006) included a discussion of
     models based on distributions of thresholds. That report noted that if one  assumes a normal or
     logit distribution for individual thresholds this leads to a probit or logistic dose-response function
     for the population and suggests that a variety of other distributions for thresholds would also lead
     to sigmoidal shaped dose-response functions.  The NRC report expressed the view that,
     "Although linear extrapolation has been advocated as an intentionally conservative approach to
     protect public health, there  are some theoretical reasons to think that sublinear nonthreshold
     dose-response models may  be more relevant for human exposure to toxicants, regardless of the
     mode of action" (p. 319). On the other hand, the same report also noted that a very broad class
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   6-27      DRAFT - DO NOT CITE OR QUOTE

-------
 1    of dose-response functions can be obtained using distributions of thresholds models: "In fact any
 2    monotonic dose-response model, including the linearized multistage model, can be defined
 3    solely in terms of a tolerance distribution without resorting to mechanistic arguments.  These
 4    considerations suggest that one must consider both the role of mode of action and the role of
 5    response variability among humans in determining the likely shape of the dose-response
 6    function" (p. 323).
 7           The discussion from the NRC TCE document emphasizes some key points in risk
 8    assessment.  Variability in the human population will have an important influence on the shapes
 9    of the dose-response relationships for that population. This is distinct from the amount of
10    variability that may be observed in inbred animal strains.  As noted in the NRC report, "One
11    might expect these  individual tolerances to vary extensively in humans depending on genetics,
12    coincident exposures, nutritional status, and various other susceptibility factors..." (p. 320).
13    Thus, if a distribution-of-thresholds approach is considered for a carcinogen risk assessment,
14    application would depend on the ability of modeling to reflect the degree of variability in
15    response in human  populations. By design, most cancer bioassays are conducted in inbred
16    rodent strains; accordingly, the parameters provided by curve fits of distribution-of-thresholds
17    models to bioassay  data would not be predicted to reflect the dose-response patterns in diverse
18    human populations. It is important to note that the NRC text has no recommendation for an
19    approach where a tolerance distribution model for humans is estimated by a statistical fit to
20    rodent bioassay data.
21           The question of whether a tolerance distribution model is indeed an appropriate basis for
22    a risk assessment also warrants consideration. Low-dose linearity can arise in other contexts
23    distinct from effects of population variability and may be directly appropriate to a MOA. Low-
24    dose linearity can also arise due to additivity of a chemical's effect on top of background
25    chemical exposures and biological processes.  In the case of chemicals such as
26    tetrachloroethylene, basic biological data do not exist to support the appropriateness of an
27    individual threshold model above models having inherent low-dose  linearity.  However, if
28    distribution of thresholds modeling were supported, it would need to be developed based on an
29    examination of predicted variability within in human  population.
30           Given the current state of scientific knowledge about tetrachloroethylene carcinogen!city,
31    the straight-line-based risk estimates presented above form the preferred recommendation for
32    estimating a plausible upper-bound estimate of potential human risks from tetrachloroethylene.
33    This approach is supported by both general scientific  considerations, including those supporting
34    the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),  as well as chemical-specific
35    findings.  The former include the scientific principles articulated previously (ie,  the expectation

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   6-28       DRAFT - DO NOT CITE OR QUOTE

-------
 1    that a chemical functions additively to background exposures, diseases, and processes; that
 2    variability within the human population would broaden the dose-response curve and eliminate
 3    individual thresholds if present; and that the approach provides consistency across assessments
 4    facilitating direct comparison of the derived risk values). The latter include evidence that, within
 5    the dose range of the cancer bioassays, the observable tumor response data are consistent with a
 6    linear model and do not suggest occurrence of a threshold, and that variability in the human
 7    response across the population is expected (eg, Bois et al., 1996; Clewell et al., 2004).
 8
 9    6.2.2.2.3.  Dose metric. Tetrachloroethylene is metabolized to several intermediates with
10    carcinogenic potential. Although much data exist for the metabolite TCA, several analyses
11    indicate that TCA alone is not able to explain  the toxicity associated with tetrachloroethylene
12    exposure; therefore, at least one other toxic agent appears to be involved. It is unclear whether
13    total metabolism—either as a measure of a precursor or intermediate, or as a surrogate directly
14    proportional to the toxic agent(s)—is an adequate indicator of potential risk. Since the
15    experimental evidence supports a role for metabolism in tetrachloroethylene's toxicity, use of
16    total metabolism (the only measure of metabolism available) to estimate cancer risk is germane
17    to this assessment. Use of administered dose  (without use of a PBPK model) yields risk
18    estimates intermediate between those based on the higher and lower PBPK models.
19
20    6.2.2.2.4.  Choice of species/gender. Table 6-3 summarizes the factors influencing the choice of
21    rodent tumor data set for human risk characterization. It is assumed that the observed rodent
22    tumors are relevant to humans, an assumption supported by a number of factors. Primary among
23    these factors is that a carcinogenic response is also observed in humans. Human-rodent site
24    concordance is not generally assumed (e.g., due to potential differences in pharmacokinetics,
25    DNA repair, other protective systems across species and tissues [U.S. EPA, 2005a]).  In keeping
26    with this view, certain  tumors associated with tetrachloroethylene exposure in human mortality
27    studies (e.g., cervix and esophagus) were not observed in rodents; cancer of the lymphoid system
28    was associated with tetrachloroethylene exposure in humans, with  some evidence for an
29    association with bladder, kidney, and lung cancer. In addition, rat  and mouse tumor types also
30    differ from each other. Finally, conclusive MOA data are lacking for the observed rodent and
31    human tumors.
32           MCL is the cancer response of highest magnitude and is reproducible in two bioassays
33    and in both genders. Although MCL has a high and variable incidence in unexposed F344 rats, a
34    biologically and statistically significant increase over background was observed (see
35    Section 5.4.1). Section 4.8.2.2.1.4 addresses the qualitative similarities among MCL to certain

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  6-29       DRAFT - DO NOT CITE OR QUOTE

-------
 1
 2
       Table 6-3. Summary of considerations for each rodent tumor type

Magnitude of
response
Specificity of
response
Replication of
findings in multiple
bioassays
Other considerations
contributing to
uncertainty in rodent
data
Relevance to humans:
a) qualitative
(biologic) site
concordance
b) occurrence in
human studies
c) confidence in
MOA
Overall
considerations for
choice of tumor type
Mononuclear cell
leukemia
+++
Rats, both genders
4/4 study datasets
High background
rate
Yes
Yes, but exact
match of tumor
classification is not
found/may not be
possible
No data
Rodent response of
highest magnitude,
reproducible; no
MOA data
Liver
++
Mice, both
genders
6/6 study datasets
High background
rate in male mice
Yes
Yes, but
association is
weak
PPAR-a
activation may
contribute, but is
not sole MOA
Rodent response
of considerable
magnitude,
reproducible,
some MOA data
Kidney
++
Male rats only
1/2 study datasets
Rare tumor
(unlikely to be due
to chance, but low
incidence)
Yes
Yes, but
association is weak
Multiple MO As
may play a role
Rare tumor in
rodents and
humans; MOA
data are strongest
Hemangiosarcoma
+
Male mice only
1/3 study datasets
Rare tumor (unlikely
to be due to chance,
but low incidence)
Yes
No, but tumor type is
rare
No data
Rare tumor in rodents
and humans; no
MOA data
 4
 5
 6
 7
 8
 9
10
11
12
13
lymphoid cancers, and the implications regarding human relevance. This section also addresses
the association of elevated lymphoma mortality with tetrachloroethylene exposure in humans.
The MOA for MCL remains unexplored.
       Male rats had the higher response level of MCL as estimated using the multistage model.
Occasionally, if the multistage model does not adequately fit a data set, an alternate model can be
used to determine the POD.  In the case of female rat MCL data, the best-fitting model (Weibull)
yielded central tendency risk estimates 10-fold higher than those from the multistage model fit of
the male rat MCL data. Consequently, there is some uncertainty in characterizing the magnitude
of MCL response, with the use of the male rat MCL data possibly underestimating risk. While
           This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                  6-30      DRAFT - DO NOT CITE OR QUOTE

-------
 1    the female rat MCL data suggest a supralinear fit extending into lower exposure levels (also
 2    apparent for female rat MCL data from the NTP bioassay), the multistage model fit was
 3    technically adequate (p = 0.48). In keeping with EPA's past practice of preferring the multistage
 4    model in order to provide some measure of consistency across different carcinogen assessments,
 5    the more linear multistage fit to the female rat MCL data supports the risk estimate derived from
 6    the male rat MCL data.
 7          The mouse liver tumor is a robust (i.e., of significant magnitude) finding in several
 8    studies, including in both sexes.  As is the case with MCL, the background for this tumor type is
 9    high—especially in males.  A biologically and statistically significant increase over background
10    was observed in males and females. There is evidence that activation of the PPAR-a receptor by
11    the tetrachloroethylene metabolite TCA contributes to the induction of mouse liver tumors.
12    However, it is not the only operative MOA involved in hepatocellular tumorigenesis. Thus, the
13    MOA remains unresolved.
14          Two tumor types were observed in only one bioassay. Kidney  tumors rarely occur in
15    unexposed rodents but were significantly elevated with tetrachloroethylene exposure in the male
16    rat NTP bioassay. The MOA is better understood for kidney tumors than for the other sites.
17    Hemangiosarcoma is another rare tumor associated with tetrachloroethylene exposure in the
18    male mouse JISA (1993) study. There are no MOA data for hemangiosarcomas.
19
20    6.2.2.2.5. PBPK model. Toxicokinetic models are used in this assessment for deriving dose
21    metrics to support dose-response analyses.  The evidence suggests that the by-products of
22    tetrachloroethylene metabolism are responsible for liver and kidney toxicity and for
23    carcinogenicity. Inhaled concentration of the parent compound is, therefore, not an appropriate
24    dosimeter for these effects, and pharmacokinetic modeling of daily overall metabolized dose is
25    expected to be an improvement in spite of the many attendant uncertainties in the modeling.  Of
26    the available toxicokinetic models on tetrachloroethylene, the assessment considers three
27    recently developed models that describe  parent tetrachloroethylene and overall metabolism of the
28    parent compound in humans. These models do not describe the kinetics and transformation of
29    total metabolic products or any individual metabolite.  All three models provide reasonably good
30    predictions of exhaled breath and blood tetrachloroethylene concentrations, so there is no
31    particular basis for preferring one model over another.  A 10-fold difference is shown in model
32    predictions of the rate of metabolism in humans, a reflection of model  differences in the values
33    for the metabolic parameters. Because the accuracy of the models has been evaluated only
34    against blood and breath concentrations of the parent compound—quantities that are insensitive
35    to these parameters—the reliability of these models for predicting the rate of total metabolism in

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   6-31      DRAFT - DO NOT CITE OR QUOTE

-------
 1    humans is unknown. Data on total metabolite levels are not available in humans, and the use of
 2    available urinary and blood TCA data is problematic.  The overall difference in risk estimates
 3    using these three models is approximately 10-fold.
 4
 5    6.2.2.2.6. Cross-species scaling.  An adjustment for cross-species scaling (BW3/4) was applied
 6    to address toxicological equivalence of internal doses between each rodent species and humans,
 7    consistent with the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a); the
 8    approach is detailed in Section 5.4.4.2.1. It is assumed that, without data to the contrary, equal
 9    risks result from equivalent constant exposures.  While the true correspondence of equipotent
10    tetrachloroethylene exposures across species is unknown, the use of BW3/4 scaling is expected
11    neither to over- or underestimate human risk (U.S. EPA, 1992).
12
13    6.2.2.2.7. Choice ofbioassay. The JISA (1993) inhalation bioassay provides data on the lowest
14    experimental exposures, and its use, therefore, reduces extrapolation uncertainty slightly.  For
15    mice, the lowest exposure concentration of 10 ppm was 10-fold lower than the lowest exposure
16    concentration in the NTP inhalation study (NTP, 1986). For rats, the low-exposure concentration
17    of 50 ppm was fourfold lower than in the NTP study.  Although the JIS A and NTP inhalation
18    bioassays used similar rodent strains, differences in the animals used (in addition to  other
19    unidentified factors) may have contributed to the twofold higher incidence of hepatocellular
20    tumors and MCL in the NTP study.  Consequently, the estimated risks are twofold lower than
21    previous EPA assessments which relied on the NTP bioassay (U.S. EPA, 1991).
22
23    6.2.2.2.8. Statistical uncertainty at the point of departure. Parameter uncertainty within the
24    chosen model reflects the limited sample size of the cancer bioassay.  For the multistage model
25    applied to this data  set, there is a relatively small degree of uncertainty at the 10% extra risk level
26    (the point of departure for linear low-dose extrapolation).
27
28    6.2.2.3.  Quantitative Analysis of Multiple Uncertainties on Cancer Unit Risk
29          Figure 6-4 and Table 6-4  show the central estimates and upper and lower confidence
30    limits of the inhalation risk per unit concentration for the rodent data sets under consideration, as
31    determined using BMDS (version 1.3.2). The upper bound inhalation risk per unit concentration
32    has been calculated as the ratio of the benchmark response (10% extra risk for all data sets
33    except the kidney tumors, which was 5%) to the 95% lower confidence limit of the benchmark
34    dose (the LECio). These results show that the lower bound of risks ranges from 20% to 40% of
35    the upper bound of risks. The values at the right end of each bar represent the unit risk estimates

                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                   6-32      DRAFT - DO NOT CITE OR QUOTE

-------
                    * ta—i
                                    * o
Male  rat, MCL

Male mouse, hepatocellular tumors
 and hemangiosarcomas

Female rat, MCL

Female mouse, hepatocellular tumors
                                            Kidney tumors, male rat (NTP)
                                            Hemangiosarcomas, male mouse
  IE-OS      1E-07   0.000001  0.00001   0.0001     0.001      0.01       0.1

       Human equivalent tetrachloroethylene risk estimates (per ug/cu.m)
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
      Figure 6-4. Cancer risk estimates for tumor sites associated with
      tetrachloroethylene exposure in rodent bioassays, using the multistage model.
      The four gender/species data sets are provided in the upper section of the graph
      while two tumor types observed in single bioassays are provided in the lower
      section. The symbols denote the slopes (to background risk) from the mean
      estimate of exposure corresponding to 10% extra risk, using the Rao and Brown
      (1993) PBPK model (n) and the Bois et al. (1996) model (o) to extrapolate to
      human equivalent exposures.  The bars indicate the slopes from the lower and
      upper bounds on the mean estimates. * indicates lower bounds that could not be
      estimated.
          This document is a draft for review purposes only and does not constitute Agency policy
06/06/08                                 6-33      DRAFT - DO NOT CITE OR QUOTE

-------
1
2
3
4
             Table 6-4.  Combined impact on tetrachloroethylene cancer risk estimates
             (per ug/m3) of statistical uncertainty,3 PBPK model and tumor site(s), using
             multistage model in observed range and linear low-dose extrapolation


Gender/species (tumor type)
Male rat (MCL)


Male mouse (hepatocellular tumor and
hemangiosarcoma)

Female rat (MCL)


Female mouse (hepatocellular tumor)


LB
C
UB
LB
C
UB
LB
C
UB
LB
C
UB
PBPK model
Rao and Brown
(1993)
5 xlO'7
1 x 10'6
2 x 10'6
*
3 x 10'7
1 x 10'6
*
5 x 10'7
9 x 10'7
1 x 10'7
2 x 10'7
3 x 10'7
Bois et al.
(1996)
5 xlO'6
1 x 10'5
2 x 10 5
*
3 x ID'6
1 x 10'5
*
5 x ID'6
9 x 10'6
1 x 10'6
2 x 10'6
3 x 10'6
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
     a In some cases, the lower bounds on risk could not be estimated.
     LB = Lower risk estimate derived from upper statistical confidence limit on the POD concentration (UECio).
     C  = Central risk estimate derived from the MLE estimate of EQ 0, and from the mean of the bootstrap distribution
          of BMR/ECio values (equal to each other in this case, see Appendix 5B).
     UB = Upper bound risk estimate derived from the lower bound statistical confidence limit POD concentration
          (LECio).
     Bolded value is used to derive assessment's unit risk estimate.
     supported by each data set. Figure 6-4 and Table 6-4 also show the range of upper bound
     inhalation risks due to the highest and lowest metabolic rate pharmacokinetic models used to
     describe the rate of metabolism of tetrachloroethylene, as described in Sections 6.2.2.2 and 3.5
     and Tables 5-8 and 5-9. A third model (Reitz et al., 1996), not shown in Figure 6-4, yields
     results between the other two. For each PBPK model, unit risk estimates based on the male
     mouse and female rat are similar, each about twofold lower than the male rat MCL unit risk
     estimate. The unit risk estimate based on the least sensitive species/gender (female mouse,
     hepatocellular tumor) is about eightfold less than that given by the male rat MCL estimate. Two
     tumor types, each seen in only one bioassay, would respectively give  unit risk estimates eightfold
     lower (male mouse hemangiosarcoma in the JISA [1993] bioassay) and fivefold lower (male rat
     kidney tumor in the NTP bioassay) than the JISA male rat MCL unit risk estimate.  Unit risk
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                                  6-34       DRAFT - DO NOT CITE OR QUOTE

-------
 1    estimates based on administered dose (without use of a PBPK model) do not correlate with those
 2    from any particular PBPK model, but they generally fall between the higher and lower unit risk
 3    estimates derived from the PBPK models.
 4          Because rodent carcinogenicity is consistently evident in all data sets, with multiple types
 5    of tumors occurring, the concern for human carcinogenic risks is increased. This supports
 6    selection of the most sensitive observation as a basis for risk estimation. For tetrachloroethylene,
 7    MCL in male rats is  the basis for risk estimation. While this tumor type tends to have a high
 8    background response in rats, it was not as high in the JISA (1993) rats, at about 20%, compared
 9    with the male NTP rats at about 56%.  Consistent with EPA's 2005 Guidelines for Carcinogen
10    Risk Assessment, it is still important to communicate the potential for increased cancer incidence
11    over background. In addition, while there is no exact analogue of MCL in humans, as noted
12    earlier, the EPA 2005 Guidelines for Carcinogen Risk Assessment notes that site concordance is
13    not necessary for assessing potential carcinogenic risk to humans.
14
15    6.2.2A.  Conclusions
16          Tetrachloroethylene is "Likely to be carcinogenic to humans" by all routes of exposure,
17    using the framework specified in the Guidelines for Carcinogen Risk Assessment (U.S. EPA,
18    2005a).  Three lines  of evidence in the hazard database support this weight-of-evidence
19    descriptor for the cancer hazard for tetrachloroethylene:
20       (1) Tetrachloroethylene is a carcinogen in rodents in  10 of 10 lifetime bioassay data sets—
21          including by  oral  and inhalation routes
22       (2) It is reasonable to use these animal tumors as indicators of potential human cancer hazard
23       (3) Tetrachloroethylene exposure is consistently associated with excess risks for a number of
24          cancers in human epidemiologic  studies, although a causal association has yet to be
25          definitively established.
26
27          The laboratory animal database includes 10 lifetime rodent bioassay data sets
28    demonstrating increased cancer incidence (two more study data sets were inconclusive due to
29    excessive mortality from pneumonia or tetrachloroethylene-related toxic nephropathy). The
30    findings include liver cancers in both sexes of mice following either oral or inhalation exposures,
31    and following inhalation  exposures, mononuclear cell leukemias in both sexes of rats (multiple
32    bioassays), as well as male rat kidney and brain tumors (gliomas) and male mouse
33    hemangioendotheliomas of the liver or spleen (single bioassays). In addition,  although not all
34    tetrachloroethylene metabolites have been tested for carcinogenicity in rodents, the oxidative
35    metabolites TCA and DC A are hepatocarcinogens in one or more species.  Although insufficient
36    to establish causality, the epidemiologic  evidence has consistently shown a positive association
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  6-3 5       DRAFT - DO NOT CITE OR QUOTE

-------
 1    of inhalation and oral tetrachloroethylene exposure with excess risks for a number of neoplasms.
 2    These include cancer of the lymphoid system, esophagus, and cervix, with more limited evidence
 3    for cancer of the bladder, kidney, and lung. Taken together, these data support a weight-of-
 4    evidence descriptor of "Likely to be carcinogenic to humans" by all routes of exposure for
 5    tetrachloroethylene.  Use of the weight-of-evidence descriptor "Likely to be carcinogenic to
 6    humans" for tetrachloroethylene is intended to communicate that the available information
 7    indicates the presence of a human health hazard.
 8           Consistent with this view, dose-response assessments are generally completed for agents
 9    considered "Likely to be carcinogenic to humans." The unit risk is intended to be a plausible
10    upper bound estimate of risk and, accordingly, all  such estimates described below are based on
11    the following:  (1) the most sensitive tumor type in rodents, with regard to species, gender, and
12    type of malignancy; (2) the POD based on the upper confidence bound on risk derived from
13    statistical modeling of the observed dose-response data; and (3) a linear low-dose extrapolation
14    approach. A linear extrapolation was performed in accordance  with default recommendations in
15    the Guidelines of Carcinogen Risk Assessment (U.S. EPA, 2005a) because of the lack of
16    substantial biological basis for doing otherwise (particularly, the lack of knowledge about the
17    MO A for any of the observed tumors), and other approaches to  estimate upper bounds on risk
18    were not considered informative for risk estimation.  Table 6-5  gives a summary of the impact
19    and justification of these choices. On the other hand, alternative choices for these approaches,
20    while providing a perspective as to the overall uncertainty in human cancer risk, would not
21    provide upper bounds on risk.
22           Given the choices of tumor type, point of departure, and low-dose extrapolation approach
23    described in Table 6-5, there are additional considerations that contribute to uncertainty in the
24    plausible upper bound unit risk, which are summarized in Table 6-6. These uncertainties have a
25    varied impact on  risk estimates.  Some (i.e., the bioassay or cross-species scaling approach)
26    suggest risks could be higher than estimated, while others would decrease estimates or have an
27    impact of uncertain direction (i.e., the human population variability, dose metric, and model-
28    based uncertainty at the POD).  While some uncertainties could be quantitatively characterized,
29    it is likely that the residual uncertainties represent the largest and can only be qualitatively
30    expressed.  Such uncertainties pertain to MOA and human sensitivity and variability. Even if
31    these could be further elucidated by additional data, extrapolation of animal bioassay data to
32    humans (done here using allometric scaling) will remain a substantial and unknown uncertainty.
33    The PBPK model uncertainty is the only one for which there is  no basis for preferring one
34    alternative to another, so the tetrachloroethylene unit risk estimate, calculated using three PBPK
35    models, ranges from 2 x 10"6 to 2 x 10"5 per |ig/m3.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  6-36       DRAFT - DO NOT CITE OR QUOTE

-------
o
Oi
Table 6-5. Considerations leading to the determination of a reasonable upper bound on risk
This document is a draft for review purposes only and does not constitute Agency policy
06/08 6-37 DRAFT: DO NOT CITE OR QUOTE

Consideration/
approach
Most sensitive
tumor response
Statistical sampling
uncertainty in
observable range
Dose-response
relationship at low-
dose

Impact on estimated risk
8-fold range from most to least sensitive rodent
tumor type. For example, Bois PBPK model
gives the following risk estimates using the
LECio and linear low-dose extrapolation in
(ug/m3)-1:
Rats, MCL
Males
Female
Mice, liver tumors
Males
Female
2 x 10'5
9 x 10'6
1 x 1Q-5
3 x IQ-6
J, 1.6-fold if ECio used rather than LECio, e.g.
Bois PBPK model gives the following risk
estimates in (ug/m3)-1 using linear low-dose
extrapolation with male rat MCL data:
Using UECi0: 5x 10'6
Using ECi0: 1 x 10'5
Using LECi0: 2x 10'5
Could J, to a negligible
unknown extent, and is
uncertainties.
value or t to an
among the largest
Decision
Male rats, MCL.
Lower bound on 10% risk
concentration (LECio)-the
EPA's default approach for
calculating a plausible upper
bound.
Linear low-dose extrapolation
from POD (default approach).
Justification
MCL had the greatest response and is reproducible
across studies.
Limited size of bioassay results in sampling
variability; lower bound (LECio) is the lower 95%
confidence interval on the concentration yielding a
10% risk and thus makes it less likely the estimate
underestimates risk due to small sample size.
Low-dose linear approach is supported by general
considerations (additivity to background,
population heterogeneity) and tetrachloroethylene-
specific data (male rat MCL data are linear in
observed range).


-------
Table 6-6. Considerations that impact uncertainty in reasonable upper bound risk estimates
Consideration/
approach
Human population
variability in
metabolism and
response/ sensitive
subpopulations
Dose metric
(uncertainty about
active moiety)
PBPK model
Cross-species
scaling
Bioassay
Model-based
uncertainty in POD
Impact on estimated risk
There may be subpopulations at higher risk and
there may be individuals for whom risk is
negligible.
Alternatives could t or J, risk estimate by an
unknown extent.
10-fold range among the three available models,
e.g. see below estimates of risk per unit
concentration (ug/m3)"1 using linear
extrapolation from LECio
Male Rats, MCL Bois: 2 x 10"5
Reitz: 9 x 10"6
Rao: 2 x 10'6
Alternative generic scaling approaches could J,
or t risk estimate (e.g., 3.5-fold J, [scaling by
BW] or t 2-fold [scaling by BW2/3 ]). Residual
uncertainty in scaling may t or J, risk estimate
by an unknown extent.
t 2-fold if NTP study used.
1.4-fold range in models that were explored.
Decision
Considered qualitatively.
Considered total metabolism
and administered
concentration.
The highest value provides a
reasonable upper bound
estimate of potential human
cancer risk; a range of
estimates is provided.
BW3/4 (default approach).
JISA study.
Multistage model used.
Justification
No data to support alternative estimates.
Experimental evidence supports a role for
metabolism in toxicity, but actual responsible
metabolites are not clearly identified.
There is no scientific basis for choosing among
pharmacokinetic results for estimating total
metabolism of tetrachloroethylene given
limitations in available data.
BW3/4 limits bias in estimate.
JISA study used the lowest experimental
exposures (reduces extrapolation uncertainty).
Flexible model; limited additional uncertainty
based on comparison with three other models.

-------
 1          As addressed in the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
 2    derivation of the central and lower bound risk estimate can be of value in some settings, such as
 3    screening analyses and for ranking agents as to their carcinogenic hazard. For such purposes, the
 4    cancer risk values based on the ECio represent the central and lower bound estimates of risk,
 5    respectively, for a particular data set. For tetrachloroethylene, a range of central estimates (based
 6    on the ECio using male rat MCL data from the JISA study and linear low-dose extrapolation and
 7    based on three PBPK model choices, as addressed in Table 6-4) is from 1 x 10"6 to 1 * 10"5 per
 8    ng/m3. The corresponding range of lower bound estimates (derived from the UECio and based
 9    on the same choices of tumor type, low-dose extrapolation approach and using the three
10    available PBPK model choices, as addressed above) is from 5 x 10"7 to 5 x 10"6 per |ig/m3.
11          To summarize, tetrachloroethylene is "Likely to be carcinogenic to humans" by all routes
12    of exposure. A  lack of human carcinogenicity, while not ruled  out, is considered unlikely.
13    Existing data indicate that (1) tetrachloroethylene is a rodent carcinogen in 10 of 10 lifetime
14    bioassay datasets, including by oral and inhalation routes (2) the observed animal effects are
15    relevant to use as indicators of human carcinogenic risk; and (3) tetrachloroethylene exposure is
16    associated with  excess risks for several cancers in human epidemiological studies, although a
17    causal relationship has yet to be established. In addition, the carcinogenicity of
18    tetrachloroethylene is also supported by other lines of evidence, including data on its metabolism
19    and pharmacokinetics and the demonstrated hepatocarcinogenicity of the oxidative metabolites
20    TCA and DC A in one or more species.  In view of the likely carcinogenicity, a dose-response
21    assessment was  undertaken with the purpose of identifying a plausible upper bound estimate of
22    risk. A range of unit risk estimates for tetrachloroethylene is from 2 x 10"6 to 2 x 10"5 per |ig/m3,
23    with the upper-end unit risk of 2 x 10"5  per |ig/m3 being the most public health protective value
24    for the upper bound risk estimate.
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                  6-39       DRAFT - DO NOT CITE OR QUOTE

-------
  1                                    REFERENCES FOR CHAPTER 6
  2
  O

  4     Altmann, L; Neuhann, HF; Kramer, U; et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
  5     low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops. Environ Res 69:83-89.
  6
  7     Birner, G; Rutkowska, A; Dekant, W. (1996) N-acetyl-S-(l,2,2-trichlorovinyl)-L-cysteine and 2,2,2-
  8     trichloroethanol: two novel metabolites of tetrachloroethene in humans after occupational exposure. Drug Metab
  9     Disp 24:41-48.
10
11     Bogen, KT; McKone, TE. (1988) Linking indoor air and pharmacokinetic models to assess tetrachloroethylene risk.
12     Risl Anal 8:509-520.
13
14     Bois, FY; Gelman, A; Jiang, J; et al. (1996) Population toxicokinetics of tetrachloroethylene. Arch Toxicol
15     70(6):347-55.
16
17     Brodkin, CA; Daniell,  W; Checkoway, H; et al. (1995) Hepatic ultrasonic changes in workers exposed to
18     perchloroethylene. Occup Environ Med 52:6796685.
19
20     Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
21     and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.
22
23     Cavalleri, A; Gobba, F; Paltrinieri, M; et al. (1994) Perchloroethylene exposure can induce colour vision loss.
24     Neurosci Lett 179:162-166.
25
26     Chiu, WE; Bois, FY. (2006) Revisiting the population toxicokinetics  of tetrachlorotethylene. Arch Toxicol
27     80:382-385.
28
29     Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004)  Evaluation of the potential impact of age- and gender-specific
30     pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79:381-393.
31
32     Doyle, P; Roman, E; Beral, V; et al. (1997) Spontaneous abortion in dry cleaning workers potentially exposed to
33     perchloroethylene. Occup Environ Med 54:8486853.
34
35     Echeverria, D; Heyer, N; Checkoway, H; et al. (1994)  Behavioral investigation of occupational exposure to solvents:
36     perchloroethylene among dry cleaners, and styrene among reinforced fiberglass laminators.  Final Report. Report
37     prepared for the Centers for Disease Control and Prevention under Grant No. 5 R01 OHo2719-03. Battelle Centers
3 8     for Public Health Research and Evaluation.
39
40     Echeverria, D; White, RF;  Sampaio, C. (1995) A behavioral evaluation of PCE exposure in patients and dry
41     cleaners: a possible relationship between clinical and preclinical effects. J Occup Environ Med 37:667-680.
42
43     Eskenazi, B; Fenster, L; Hudes, M; et al. (1991a) A study of the effect of perchloroethylene exposure on the
44     reproductive outcomes of wives of dry-cleaning workers.  Am J Ind Med 20:5936600.
45
46     Eskenazi, B; Wyrobek, AJ; Fenster, L; et al. (1991b) A study of the effect of perchloroethylene exposure on semen
47     quality in dry cleaning workers.  Am J Ind Med 20:5756591.
48
49     Ferroni, C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
50     perchloroethylene. Neurotoxicology 13:243-247.
51
52     Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators  of renal damage in workers exposed to organic
53     solvents. Int Arch Occup Environ Health 52:1-9.
54

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                         6-40        DRAFT - DO NOT CITE OR QUOTE

-------
  1     Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
  2     tetrachloroethylene as neonates. Toxicol Lett 66:13Bl9.
  O

  4     Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
  5     tetrachloroethylene. Am J Ind Med 21:661-671.
  6
  7     Green, T; Odum, J. Nash, JA; et al. (1990) Perchloroethylene-induced rate kidney tumors: an investigation of the
  8     mechanisms involved and their relevance to human. Toxicol Appl Pharmacol 103:77B89.
  9
10     Hayes, JR; Condie, LW, Jr.; Borzelleca, JF.  (1986) The subchronic toxicity of tetrachloroethylene
11     (perchloroethylene) administered in the drinking water of rats. Fundam Appl Toxicol 7:119-125.
12
13     JISA (Japan Industrial Safety Association). (1993) Carcinogenicity study of tetrachloroethylene by inhalation in rats
14     and mice. Data No. 3-1. Available from: EPA-IRIS Information Desk.
15
16     Kjellstrand, P; Holmquist, B; Kanje, M; et al. (1984) Perchloroethylene: effects on body and organ weights and
17     plasma butyrylcholinesterase activity in mice. Acta Pharmacol Toxicol (Copenh) 54:414-424.
18
19     Krewski, D; vanRyzin, J. (1981) Dose response models for quantal response toxicity data, in statistics and related
20     topics. In: Csorgo, M; Dawson, DA; Rao, JNK; Saleh, ADMdE; eds.  North-Holland Publishing Company.
21
22     Lutz, WK; Gay lor, DW; Conolly, RB; et al.  (2005) Nonlinearity and thresholds in dose-response relationships for
23     carcinogenicity due to sampling variation, logarithmic dose scaling, or small differences in individual susceptibility.
24     Toxicol Appl Pharmacol 207:565-569.
25
26     Mattsson, JL; Albee,  RR; Yano, BL; et al. (1998) Neurotoxicologic examination of rats exposed to 1,1,2,2-
27     tetrachloroethylene (perchloroethylene) vapor for 13 weeks. Neurotoxicol Teratol 20:83B98.
28
29     Mutti, A; Alinovi, R; Bergamaschi, E; et al.  (1992) Nephropathies and exposure to perchloroethylene in dry-
30     cleaners.  Lancet 340:1896193.
31
32     Nagano, K; Nishizawa, T; Yamamoto, S; et  al. (1998) Inhalation carcinogenesis studies of six halogenated
33     hydrocarbons in rats and mice. In: Chiyotani, K; Hosoda, Y; Aizawa, Y, eds. Advances in the prevention of
34     occupational respiratory diseases: proceedings of the 9th international conference on occupational respiratory
35     diseases, Kyoto, Japan, October 13-16, 1997. Amsterdam: Elsevier; pp. 741-746.
36
37     National Toxicology  Program (NTP). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
38     (perchloroethylene) (CAS No. 127-18-4) in F344/N rats and B6C3Flmice. National Institutes of Health, Public
39     Health Service, U.S. Department of Health and Human Services.  Available online at http://ntp.niehs.nih.gov.
40
41     NRC (National Research Council). (1983) Risk assessment in the Federal government: managing the process.
42     Committee on the Institutional Means for Assessment of Risks to Public Health, Commission on Life Sciences,
43     NRC. Washington, DC; National Academy Press.
44
45     NYS DOH (New York State Department of Health). (1997) Tetrachloroethylene ambient air criteria document.
46     Final Report. Albany, NY.
47
48     NYS DOH (New York State Department of Health). (2005a) Improving human risk assessment for
49     tetrachloroethylene by using biomarkers and neurobehavioral testing. U.S. EPA Star Grant #R827445. Grant
50     #R827446. Available online at
51     http://cfpub.epa.gov/ncer  abstracts/index.cfm/fuseaction/displav.abstractDetail/abstract/977/reprort/O.
52
53     NYS DOH (New York State Department of Health). (2005c) Pumpkin patch day care center follow-up evaluation.
54     Final Report. New York State Department of Health, Bureau of Toxic Substance Assessment, Division of
5 5     Environmental Health Assessment, Center for Environmental Health.
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        6-41        DRAFT - DO NOT  CITE OR QUOTE

-------
 2     Olsen, J; Hemminki, K; Ahlborg, G; et al. (1990) Low birthweight, congenital malformations, and spontaneous
 3     abortions among dry-cleaning workers in Scandinavia.  Scand J Work Environ Health 16:163B168.
 4
 5     Rao, HV; Brown, DR. (1993) A physiologically based pharmacokinetic assessment of tetrachloroethylene in
 6     groundwater for a bathing and showering determination. Risk Anal 13:37^9.
 7
 8     Reitz, RH; Gargas, ML; Mendrala, AL; et al. (1996) In vivo and in vitro studies of perchloroethylene metabolism
 9     for physiologically based pharmacokinetic modeling in rats, mice, and humans. Toxicol Appl Pharmacol 136:289-
10     306.
11
12     Rosengren, LE; Kjellstrand, P; Haglid, KG. (1986) Tetrachloroethylene: levels of DNA and S-100 in the gerbil CNS
13     after chronic exposure. Neurobehav Toxicol Teratol 8:201-206.
14
15     Schreiber, JS; Hudnell, HK; Geller, AM; et al. (2002) Apartment residents' and day care workers' exposures to
16     tetrachloroethylene and deficits in visual contrast sensitivity. Environ Health Perspect 110:655-664.
17
18     Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene. Neurotoxicol Teratol
19     11:579-583.
20
21     Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
22     of subjects chronically exposed to organic solvents (II)]. G Ital Med Lav Ergon 19:85-88.
23
24     Stromberg, PC. (1985) Large granular lymphocyte leukemia in F344 rats.  Model for human T gamma lymphoma,
25     malignant histiocytosis, and T-cell chronic lymphocytic leukemia.  Am J. Pathol 119:517-519.
26
27     Tinston, DJ. (1994) Perchloroethylene: A multigeneration inhalation study in the rat.  CTL/P/4097. Available from:
28     EPA IRIS Information Desk.
29
3 0     Trevisan, A; Macca, I; Rui, F; et al. (2000) Kidney and liver biomarkers in female dry-cleaning workers exposed to
31     perchloroethylene.  Biomarkers 5:399B409.
32
33     Umezu, T; Yonemoto, J; Soma, Y; Miura, T. (1997) Behavioral effects of trichloroethylene and tetrachloroethylene
34     in mice. Pharmacol Biochem Behav 58:665-671.
35
36     U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
37     carcinogenicity of tetrachloroethylene (perchloroethylene). Office of Health and Environmental Assessment,
38     Washington, DC; EPA/600/6-91/002F.  Available from: National Technical Information Service, Springfield, VA.
39
40     U.S. EPA (Environmental Protection Agency). (1992) Draft report: a cross-species scaling factor for carcinogen risk
41     assessment based on equivalence of mg/kg3/4/day.  Federal Register 24152-24173.
42
43     U.S. EPA (Environmental Protection Agency). (1993) Reference Dose (RfD): Description and Use in Health Risk
44     Assessments Background Document 1A, March 15, 1993.
45
46     U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
47     and application of inhalation dosimetry.  Office of Health and Environmental Assessment, Environmental Criteria
48     and Assessment Office, Cincinnati, OH; EPA/600/8-90/066F. Available from: National Technical Information
49     Service, Springfield, VA; PB2000-500023, and online at http://www.epa.gov/ncea.
50
51     U.S. EPA (Environmental Protection Agency). (2005a). Guidelines for carcinogen risk assessment. Federal Register
52     70(66)17765-17817. Available online at http://www.epa.gov/cancerguidelines.
53
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                        6-42        DRAFT - DO NOT CITE OR QUOTE

-------
 1    U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
 2    early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F. Available
 3    online at http://www.epa.gov/cancerguidelines.
 4
 5    Verplanke, AJ; Leummens, MH; Herber, RF. (1999) Occupational exposure to tetrachloroethene and its effects on
 6    the kidneys.  J Occup Environ Med 41:11B16.
 7
 8    Warren, DA; Reigle, TG; Muralidhara, S; et al. (1996) Schedule-controlled operant behavior of rats following oral
 9    administration of perchloroethylene: time course and relationship to blood and brain solvent levels. J Toxicol
10    Environ Health 47:345-362.
11
12    Zeise, L; Wilson, R; Crouch, EA. (1987) Dose-response relationships for carcinogens: a review. Environ Health
13    Perspect 73:259-306.
                   This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                        6-43        DRAFT - DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14                                      APPENDIX A:
15
16                        SUMMARY OF EARLIER ASSESSMENTS
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                             A-l          DRAFT—DO NOT CITE OR QUOTE

-------
 1                 APPENDIX A: SUMMARY OF EARLIER ASSESSMENTS
 2
 3   A.I.         ORAL INGESTION ASSESSMENTS	A-4
 4          A.I.I. U.S. EPA Oral Ingestion Assessments	A-4
 5                A.I.1.1.  IRIS Database, U.S. EPA, 2005	A-4
 6                A. 1.1.2.  Health Assessment Document for Tetrachloroethylene,
 7                         U.S. EPA, 1985	A-4
 8          A.I.2. Oral Ingestion Assessments Conducted by Non-U.S. EPA Agencies	A-4
 9                A. 1.2.1.  California Environmental Protection Agency, 2000, Draft
10                         Public Health Goal for Tetrachloroethylene in Drinking Water	A-4
11                A. 1.2.2.  Agency for Toxic Substances and Disease Registry (ATSDR)
12                         Toxicological Profile for Tetrachloroethylene (Update),
13                         September 1997	A-5
14          A. 1.3. Summary of Ingestion Risk Estimates	A-6
15
16   A.2.   INHALATION ASSESSMENTS	A-7
17          A.2.1. U.S. EPA Inhalation Assessments	A-7
18                A.2.1.1.  Cleaner Technologies Substitutes Assessment: Professional
19                         Fabricare Processes, Office of Pollution Prevention and Toxics,
20                         U.S. EPA, 1998	A-7
21                A.2.1.2.  Addendum to the Health Assessment Document for
22                         Tetrachloroethylene, U.S. EPA, 1986	A-7
23                A.2.1.3.  Health Assessment Document for Tetrachloroethylene,
24                         U.S. EPA, 1985	A-8
25          A.2.2. Inhalation Assessments Conducted by Non-U.S. EPA Agencies	A-8
26                A.2.2.1.  California Environmental Protection Agency, 2002	A-8
27                A.2.2.2.  Massachusetts'Derivation of Inhalation Unit Risk for
28                         Tetrachloroethylene, Massachusetts Department of
29                         Environmental Protection, 1998	A-9
30                A.2.2.3.  ATSDR Toxicological Profile for Tetrachloroethylene (Update),
31                         September 1997	A-9
32                A.2.2.4.  Tetrachloroethane—Ambient Air Criteria Document, New
33                         York State Department of Health, 1997	A-10
34                A.2.2.5.  Priority Substances List Assessment Report:
3 5                         Tetrachloroethylene, Canada Health and Welfare, 1993	A-11
36          A.2.3. Summary of Inhalation Risk Estimates	A-ll
37   A.3.   QUALITATIVE RISK ASSESSMENTS	A-ll
38          A.3.1. U.S. EPA Qualitative Risk Assessments	A-13
39                A.3.1.1.  Response to Issues  and Data Submission on the Carcinogenicity
40                         of Tetrachloroethylene, U.S. EPA, 1991	A-13
41          A.3.2. Qualitative Risk Assessments Conducted by Non-U.S. EPA Agencies	A-13
42                A.3.2.1.  Organization for Economic Cooperation and Development
43                         (OECD), Screening Information Data Set (SIDS) Initial
44                         Assessment Report: Comprehensive Risk Assessment Report
45                         for Tetrachloroethylene, 1996	A-13

                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                             A-2          DRAFT—DO NOT CITE OR QUOTE

-------
 1                                   CONTENTS (continued)
 2
 O
 4                 A.3.2.2. IARC Monographs on the Evaluation of Carcinogenic Risks to
 5                         Humans, Volume 63, 1995	A-13
 6                 A.3.2.3. Report on Carcinogens, Eleventh Edition, NTP, 2005	A-14
 7
 8   References	A-15
 9
10
11
12
13
14
15
16                                      LIST OF TABLES
17
18   Table A-l.    Estimates of ingestion unit risk using different methods	A-7
19
20   Table A-2.    Ingestion, noncancer endpoints	A-7
21
22   Table A-3.    Estimates of cancer inhalation unit risk using different methods	A-14
23
24   Table A-4.    Inhalation, noncancer endpoints	A-14
25
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-3          DRAFT—DO NOT CITE OR QUOTE

-------
 1    A.l.  ORAL INGESTION ASSESSMENTS
 2    A.1.1. U.S. EPA Oral Ingestion Assessments
 3    A.l.1.1.  IRIS Database, U.S. EPA, 1988
 4          In 1988, the U.S. Environmental Protection Agency (EPA) established a reference dose
 5    (RfD) for the ingestion of tetrachloroethylene (U.S. EPA, 2005). An RfD is an estimate (with
 6    uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
 7    population (including sensitive subgroups) that is likely to be without an appreciable risk of
 8    deleterious noncarcinogenic effects during a lifetime. For the oral RfD, EPA used the Buben and
 9    O'Flaherty (1985) gavage study.  A no-observed-adverse-effect level (NOAEL) of 20mg/kg-day
10    was determined, based  on hepatotoxicity in mice.  This value was duration adjusted and an
11    uncertainty factor of 1,000 was applied (10 for intraspecies variability, 10 for interspecies
12    variability, 10 for extrapolation from a subchronic study). EPA places medium confidence in the
13    RfD derivation because the data set lacks information on reproductive and teratology endpoints.
14    An RfD of 0.01 mg/kg-day was derived.
15
16    A.I.1.2.  Health Assessment Document for Tetrachloroethylene, U.S. EPA, 1985
17          Dose response data for hepatocellular carcinomas observed in female mice in the
18    National Cancer Institute gavage study (NCI, 1977) were used to derive the unit risk. The
19    potency estimate for tetrachloroethylene was calculated using the linearized multistage model
20    and the dose metabolized and eliminated in urine.  The unit risk was derived by multiplying the
21    assumed daily intake of 2 L of water contaminated with 1 [ig/L tetrachloroethylene for a person
22    (2.9 x 10"5 mg/kg-day)  by the potency estimate for tetrachloroethylene (5.1 x 10"2) to derive the
23    unit risk. The upper-bound estimate of the incremental lifetime risk due to consuming water
24    contaminated with 1  |J,g/L of tetrachloroethylene was calculated to be 1.5  x 10"6.
25
26    A.1.2. Oral Ingestion  Assessments Conducted by Non-EPA Agencies
27    A.l.2.1.  World Health Organization, Concise International Chemical Assessment
28    Document 68, 2006
29          "...The available information on oral exposure was inadequate for derivation of a TDI by
30    the oral route. However, as tetrachloroethene is well absorbed after inhalation or ingestion and
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-4         DRAFT—DO NOT CITE OR QUOTE

-------
 1    there is little evidence of first-pass metabolism, a PBPK model was used to derive a TDI. The
 2    model predicted that tetrachloroethene consumed in drinking-water at a dose level of 0.047
 3    mg/kg body weight per day would yield an AUC in plasma similar to that from continuous
 4    exposure to tetrachloroethene at 0.2 mg/m3 in inhaled air. This oral  figure was rounded to give a
 5    TDI of 50 jig/kg body weight."
 6           An oral cancer risk value was not derived.
 7
 8    A. 1.2.2. California Environmental Protection Agency, 2001, Draft Public Health Goal for
 9    Tetrachloroethylene in Drinking Water
10           The California Environmental Protection Agency (Cal EPA) developed a Public Health
11    Goal (PHG) for tetrachloroethylene in drinking water on the basis of hepatocellular carcinomas
12    observed in male and female mice orally exposed to tetrachloroethylene (Cal EPA, 2001). PHGs
13    are based solely on health effects impacts and are set at levels that do not pose any significant
14    health risk, as determined by the California Office of Environmental Health Hazard Assessment.
15    For water-derived inhalation exposures, estimates were derived from studies showing
16    hepatocellular adenoma or carcinoma in male mice and mononuclear cell leukemia in both male
17    and female rats exposed by inhalation to tetrachloroethylene (NTP,  1986).  The pharmacokinetic
18    model described by Bogen et al. (1987) was used to estimate the "effective" dose for use in
19    quantitative calculations. The Bogen et al. study was chosen over other studies (Bois et al.,
20    1990; Chen and Blancato, 1987) because it provided dose estimates for mice and rats exposed
21    orally and by inhalation.
22           Tetrachloroethylene was treated as a directly acting genotoxic carcinogen, and a linear
23    low-dose extrapolation model was used.  The PHG established by Cal EPA is 0.056 ng/L. This
24    value corresponds to a unit risk estimate of 1.3 x 10"5 (fig/L)"1. This health-protective
25    concentration includes an estimate of inhalation exposure from showering in tetrachloroethylene-
26    contaminated water, flushing toilets, and other household activities  involving tap water.
27           Chronic toxicity, excluding cancer, was evaluated on the basis of neurobehavioral
28    endpoints (delayed  reaction time) observed in epidemiological studies of exposed humans.
29    These studies evaluated persons who were exposed to inhaled tetrachloroethylene.  Cal EPA
30    concluded that no single study was sufficiently reliable to be used as the primary basis for a
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                              A-5          DRAFT—DO NOT CITE OR QUOTE

-------
 1    health-protective standard; both the Altmann et al. (1995) and the Spinatonda et al. (1997)
 2    studies were quite small (14 and 35 subjects, respectively) and the Ferroni et al. (1992) study
 3    lacked details.  Therefore, the geometric mean from the three studies was used to derive an
 4    estimated health-protective concentration in drinking water.
 5          In calculating the mean, each study provided a lowest-observed-adverse-effect level
 6    (LOAEL) value, and a study-specific uncertainty factor was used (10 to account for the use of a
 7    LOAEL and 10 or 3 to account for potentially sensitive human subpopulations). A factor of 3%
 8    was applied for the relative source contribution because drinking water supplies only 3% of the
 9    total tetrachloroethylene exposure, and a water intake of 6.31 L/day was the calculated
10    equivalent drinking water ingestion rate that would supply the total tetrachloroethylene dose
11    from inhalation via showering and direct ingestion. The geometric mean  of these safe
12    concentrations calculated from the three studies is 1.1 * 10 "2 mg/L (11  ng/L). The investigators
13    concluded that this is the health-protective drinking water concentration for noncarcinogenic
14    effects.  With an assumption of 100% absorption from drinking water and an intake of 6.31
15    L/day, the equivalent dose corresponding to 11 |J,g/L is 1 |j,g/kg-day. This can be used to
16    compare the California safe limits to the RfD of other organizations.
17
18    A. 1.2.3. Agency for Toxic Substances and Disease Registry (ATSDR) Toxicological Profile
19   for Tetrachloroethylene (Update), September 1997
20          ATSDR has established a minimal risk level (MRL) for the acute ingestion of
21    tetrachloroethylene.  MRLs  are estimates of the daily human exposure to a hazardous substance
22    that is likely to be without appreciable risk of adverse noncancer health effects  over a specified
23    duration of exposure. These values are based only  on noncancer effects and are generally based
24    on the most sensitive endpoint considered to be of relevance to humans. The acute oral MRL
25    was derived from studies that showed hyperactivity in 60-day-old male mice that were treated
26    with tetrachloroethylene for 7 days beginning at 10 days of age (Fredriksson et al., 1993). The
27    MRL is based on a LOAEL (5 mg/kg-day) that was adjusted by an uncertainty factor of 100 to
28    account for the use of the LOAEL (10) and extrapolation from animals to humans (10).  For
29    tetrachloroethylene, the acute oral MRL is 0.05 mg/kg-day.
30
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-6          DRAFT—DO NOT CITE OR QUOTE

-------
 1
 2
 3
 4
 5
 6
 7
A.1.3.  Summary of Ingestion Risk Estimates

       The following tables summarize the quantitative risk estimates that have been developed

by EPA and other agencies.  Table A-l shows the cancer risk values. Table A-2 depicts the risk

estimates developed for noncarcinogenic endpoints.
Table A-l. Estimates of ingestion unit risk using different methods
Agency guideline
Public Health Goal (Cal EPA,
2000), based on metabolized
dose
Dose metabolized (U.S. EPA,
1985)
Unit risk value
(Hg/L)1
1.3 x 1Q-5
1.5 x 1Q-6
Studies used
NCI (1977),
NTP (1986)
NCI (1977)
Critical target effect(s)
Liver adenomas and carcinomas
in male and female mice,
mononuclear cell leukemia in
male and female rats
Liver carcinomas in female mice
 9
10
11
12
13
14
Table A-2. Ingestion, noncancer endpoints
Agency guideline
Cal EPA (1999)
Minimal risk level
(ATSDR, 1997)
Reference dose (U.S.
EPA, 1988)
Standard
1 1 |o,g/L (water
concentration)
0.001 mg/kg-day
equivalent dose
0.05 mg/kg-day
0.01 mg/kg-day
Studies used
Altmannetal. (1995),
Spinatondaetal. (1997),
and Ferroni et al. (1992)
Fredrikssonetal. (1993)
Buben and O' Flaherty
(1985)
Critical target effect(s)
Delayed reaction times in
humans.
Hyperactivity in male
mice
Liver toxicity in mice
15
16
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                               A-7          DRAFT—DO NOT CITE OR QUOTE

-------
 1   A.2. INHALATION ASSESSMENTS
 2   A.2.1. U.S. EPA Inhalation Assessments
 3   A.2.1.1.  Cleaner Technologies Substitutes Assessment: Professional Fabricare Processes,
 4   Office of Pollution Prevention and Toxics, U.S. EPA, 1998
 5          The Office of Pollution Prevention and Toxics (OPPT) developed the Cleaner
 6   Technologies Substitutes Assessment (CTSA) to provide comparative cost, risk, and
 7   performance information on professional fabricare processes. As part of this assessment, the
 8   health risks associated with the use of tetrachloroethylene in dry cleaning establishments was
 9   evaluated.  For carcinogenic effects, the CTSA used human-equivalent metabolized doses using
10   mouse and rat tumor data from the National Toxicology Program (NTP) (1986) study. The
11   approach used was similar to that used by EPA (U.S. EPA,  1986), but the mouse carcinoma-only
12   data set was omitted from the assessment to avoid double-counting of animals with adenomas
13   and carcinomas.  The analyses are based on taking the geometric mean of the unit risk from four
14   data sets that evaluated the incidence of male and female mouse liver adenomas and/or
15   carcinomas and male and female rat mononuclear cell leukemia. Using a linear-at-low-doses
16   approach, the unit risk was estimated to be 7.1 x 10"7 per |J,g/m3 of tetrachloroethylene in air.
17   The CTSA report states that the unit risk should not be used for lifetime average daily exposures
18   greater than 1.4 x 104 |j,g/m3.
19          Noncarcinogenic effects were also evaluated in the CTSA report.  A provisional RfC was
20   derived on the basis of mild renal tubule damage seen in a cross-sectional occupational study
21   (Franchini et al., 1983).  The average level of tetrachloroethylene exposure was equivalent to 10
22   mg/m3, and this value was used as the LOAEL. The LOAEL was adjusted to account for
23   duration  of exposure, and an uncertainty factor of 10 was applied to account for the use of a
24   LOAEL. An uncertainty factor to account for sensitive individuals was not applied, because the
25   derived RfC was to be used in the CTSA screening to evaluate occupational populations.  The
26   provisional RfC was established at 0.17 mg/m3.
27
28   A.2.1.2.  Addendum to the Health Assessment Document for Tetrachloroethylene, U.S. EPA,
29   1986
30          This assessment was conducted to reevaluate tetrachloroethylene carcinogenicity on the
31   basis of the released NTP (1986) inhalation animal bioassay.  On the basis of the evidence of
32   carcinogenicity in rats and mice, together with the inconclusive epidemiologic evidence,
33   tetrachloroethylene was recategorized as a Group B2 probable human carcinogen.
34          A new inhalation unit risk value was derived using the NTP (1986) inhalation study.  The
35   NTP bioassay  doses for rats and mice were converted to metabolized doses using the previously
                This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                               A-8          DRAFT—DO NOT CITE OR QUOTE

-------
 1    established dose-metabolism relationship (U.S. EPA, 1985). A linearized multistage model was
 2    used for the low-dose extrapolation.  Six different data sets from the NTP study were used to
 3    derive unit risk estimates. These data included endpoints on leukemia in male and female rats,
 4    liver carcinoma in male and female mice, and liver carcinomas and adenomas in male and female
 5    mice.  The revised upper-bound estimate of the incremental cancer risk due to lifetime exposure
 6    to 1 |J,g/m3 of tetrachloroethylene in air was determined to range from 2.9 x 10"7 to 9.5 x 10"7.
 7    This range includes the value determined using the NCI gavage study (NCI, 1977).  The unit risk
 8    was stated to be applicable only for low-level exposures where the relationship between ambient
 9    air concentrations and metabolized dose is linear.
10
11    A.2.1.3. Health Assessment Document for Tetrachloroethylene, U.S. EPA, 1985
12          Tetrachloroethylene was categorized as a Group C possible human carcinogen on the
13    basis of limited evidence of carcinogenicity in animals and inconclusive epidemiologic data.
14    Dose response data for hepatocellular carcinomas observed in female mice in the NCI gavage
15    study (NCI, 1977) were used to derive the unit risk. The potency estimate for
16    tetrachloroethylene was calculated using the linearized multistage model and the dose
17    metabolized and eliminated in the urine. Urinary metabolites were considered to account for
18    80% of total metabolites, as in Buben and O'Flaherty (1985). Unit risk was then calculated
19    using human body burden data from Bolanowska and Golacka (1972). This study provided
20    information on the relationship between the air concentration and the amount metabolized in
21    urinary excretion in human subjects.  The amount metabolized was assumed to be proportional to
22    the air concentration  and the duration of exposure. The upper-bound estimate of the incremental
23    cancer risk due to 1 ng/m3 of tetrachloroethylene in air was determined to be 4.8 x 10"7.
24
25    A.2.2. Inhalation Assessments Conducted by Non-U.S. EPA Agencies
26    A.2.2.1. World Health Organization, Concise International Chemical Assessment
27    Document 68, 2006
28          "In occupationally exposed cohorts, the most consistent adverse finding was
29    neurotoxicity; therefore, the most informative study on neurotoxic effects in exposed workers
30    was used to derive a TC. The mean exposure level (83 mg/m3) was taken as a LOAEC. This
31    was converted to an equivalent concentration for continuous exposure (20 mg/m3), and two
32    uncertainty factors of 10 were applied (one to account for interindividual differences, the other
33    because the selected concentration was a LOAEC rather than a NOAEC), to derive a TC of 0.2
34    mg/m3. For comparative purposes, a similar approach was used for studies reporting
35    nephrotoxicity. The most informative study yielded a mean occupational exposure of 100
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-9          DRAFT—DO NOT CITE OR QUOTE

-------
 1    mg/m3, which generated a TC of 0.24 mg/m3, a value in good agreement with the TC protective
 2    against neurotoxic effects. Available data indicate that liver toxicity would occur only at
 3    exposures higher than those that affect the CNS and kidney. A TC for spontaneous abortions
 4    was not derived. However, the TC of 0.2 mg/m3 is more than 3 orders of magnitude lower than
 5    the exposure concentration that induced mild adverse effects in laboratory animals, and so it was
 6    considered to be protective against reproductive toxicity in humans."
 7          "Tetrachloroethene has induced several types of tumour in rats and mice. Currently,
 8    there is no convincing evidence that these tumours arise via modes of action that operate only in
 9    rodents, and hence their relevance to humans cannot be  dismissed. Therefore, a BMC approach
10    was used, and a BMC and its lower confidence limit (BMCL) were calculated for each animal
11    tumour.  Of the tumours observed in experimental animals, hepatocellular adenomas and
12    carcinomas in male mice yield highest predicted risks.  The TC derived above, 0.2 mg/m3,
13    corresponds to a cumulative lifetime risk of 0.4  x 10~3 when a linear extrapolation is applied to
14    the BMCio as the point of departure."
15
16    A.2.2.2.  California Environmental Protection Agency, 2002
17          The California EPA Air Toxics Hot Spots program has derived an inhalation unit risk
18    value for tetrachloroethylene (Cal EPA, 2002).  The value was determined from data on
19    hepatocellular adenomas and carcinomas in male mice reported in the NTP (1986) bioassay
20    study. Two pharmacokinetic models were used to estimate the human inhaled concentrations
21    equivalent to the bioassay concentrations. These two models were described only as (1) a
22    steady-state model and (2) a physiologically based pharmacokinetic (PBPK) model. An
23    assumption that 18.5% of the applied dose is metabolized in humans was incorporated.  The
24    cancer potency values expressed in terms of human dose rates and derived using the two
25    different models and the rat and mouse studies ranged from 0.0025 to 0.093 per mg/kg-day.
26    Considering the quality of the cancer bioassay s and the uncertainty in human metabolism, Cal
27    EPA decided that the best value for the inhalation unit risk was 5.9 x 10"6 per |ig/m3.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-10         DRAFT—DO NOT CITE OR QUOTE

-------
 1    A.2.2.2. Massachusetts' Derivation of Inhalation Unit Risk for Tetrachloroethylene,
 2    Massachusetts Department of Environmental Protection, 1998
 3          The Massachusetts Department of Environmental Protection (MA DEP) classifies
 4    tetrachloroethylene as a Group B2 carcinogen with suggestive evidence for mutagenicity.  The
 5    unit risk was calculated on the basis of male and female liver tumors found in mice in the NCI
 6    gavage study (NCI, 1977). The dose calculations used were similar to those used by EPA (U.S.
 7    EPA, 1985), with two differences. MA DEP adjusted the lifetime average dose, which is based
 8    on urinary metabolites, to a dose of total metabolites. This adjustment was made  on the
 9    assumption that the urinary metabolites are 80% of the total metabolism.  To convert the
10    carcinogenic potency value to an inhalation exposure, MA DEP assumed that the  metabolized
11    dose is equal to 70% of the inhaled dose. This differs from the 1985 EPA assessment, where
12    0.66% of the total inhaled dose in humans is assumed to be metabolized to urinary metabolites.
13    MA DEP also calculated a unit risk using the NTP inhalation study, but did not consider this to
14    provide a reasonable quantitative estimate due to uncertainty  in the calculations of the
15    metabolized dose.  Using the NCI study, 5.5 x 10"5 per ng/m3 is recommended as  the unit risk.
16
17    A.2.2.3.  ATSDR Toxicological Profile for Tetrachloroethylene (Update), September 1997
18          ATSDR has promulgated both acute and chronic MRLs for the inhalation  of
19    tetrachloroethylene. The acute inhalation MRL was derived from studies where male volunteers
20    were exposed to 50 ppm tetrachloroethylene for 4 hrs/day for 4 days. The volunteers showed
21    increased pattern reversal visually evoked potential (VEP) latencies and deficits for vigilance and
22    eye-hand coordination (Altmann et al., 1992). Deficits were not seen at 10 ppm, and this value
23    was used as the NOAEL. This value was duration adjusted to extrapolate from intermittent
24    exposure, and an uncertainty factor of 10 was used to account for human variability.  The acute
25    inhalation MRL was established at 0.2 ppm (1.36 mg/m3).
26          The chronic duration MRL for the inhalation of tetrachloroethylene was based on a study
27    that showed increased reaction times in neurobehavioral tests given to female workers exposed
28    to tetrachloroethylene in dry cleaning shops (Ferroni et al., 1992). Air exposures  averaged 15
29    ppm tetrachloroethylene for an average of 10.1 years. The LOAEL in this study was 15 ppm.
30    This value was adjusted from an occupational exposure to a continuous exposure; an uncertainty
31    factor of 10 was used to account for the use of a LOAEL, and an additional factor of 10 was used
32    to account for human variability.  The chronic inhalation MRL was established at 0.04 ppm (0.27
33    mg/m3).
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-l 1         DRAFT—DO NOT CITE OR QUOTE

-------
 1    A.2.2.4. Tetrachloroethane-Ambient Air Criteria Document, New York State Department of
 2    Health, 1997
 3          The New York State Department of Health agrees with the International Agency for
 4    Research on Cancer (IARC) classification of tetrachloroethylene as a Group 2A (known animal)
 5    carcinogen. A linearized multistage model was applied to metabolized dose data for liver tumors
 6    in mice and mononuclear cell leukemia in rats (NTP, 1986). Estimates of the metabolized dose
 7    were based on predictions of a physiologically based pharmacokinetic (PBPK) model (mice) and
 8    on experimental data on the production of urinary metabolites (mice and rats). For mononuclear
 9    cell leukemia in rats, estimates based on air concentrations were also derived. ED 10 procedures
10    were also used to calculate unit risk values.  Due to uncertainty regarding which method provides
11    a better estimate, the central tendency of all these estimates (linearized multistage model, ED 10,
12    metabolized dose, and air concentration) was used to derive an upper-bound unit risk value. The
13    central tendency estimate for liver tumors in mice established an upper bound unit risk value of
14    0.88  x 10"6 (i^g/m3)"1; for mononuclear cell leukemia in rats, the central tendency estimate was
15    1.3 x 10"6 (i^g/m3)"1. An upper-bound risk estimate of  1 x 10"6 (jig/m3)"1  is the central tendency
16    of the mouse- and rat-based unit risk estimates and is the recommended criteria for evaluating
17    the excess  human carcinogenic risk associated with chronic exposure to l|j,g/m3
18    tetrachloroethylene in ambient air.
19          New York State has determined that the strength of human evidence on the
20    noncarcinogenic effects of tetrachloroethylene exposure support the use of human data for
21    determining an ambient air criteria for noncarcinogenic effects. A weight-of-evidence approach
22    was used, and multiple endpoints and epidemiologic studies were evaluated. Endpoints used in
23    the derivation of the ambient air criterion included evidence of central nervous system (motor
24    and cognitive effects) (Seeber, 1989), kidney (Mutti et  al., 1992), and liver dysfunction (Gennari
25    et al., 1992).  The lack of reproductive and developmental studies was identified as a significant
26    data gap because epidemiologic studies did not provide sufficient exposure data for criteria
27    evaluation.
28          Lowest-observed-effect level (LOEL) data were provided in the epidemiologic studies
29    listed above.  LOELs were duration adjusted to account for continuous exposure using EPA
30    inhalation guidelines (U.S. EPA, 1994).  For adult criteria, an uncertainty factor of 100 was
31    applied to each duration-adjusted LOEL (10 for variation in sensitivity among humans and 10 for
32    the use of a LOEL from a  subchronic study). For criteria protective of children, the appropriate
33    scaling and uncertainty factors were used. Child-adjusted LOELs were derived using physical
34    and physiological data for children. An uncertainty factor of 100 was applied to the child-
                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                                A-12         DRAFT—DO NOT CITE OR QUOTE

-------
 1    adjusted LOELs (10 for variation in sensitivity among humans, 3 for the use of a LOEL, and 3
 2    for concerns about the increased sensitivity of children to tetrachloroethylene toxicity).
 3          Listed below are the results of the safe ambient air level derivations:
 4
 5                    Effects                             Adults            Children
 6          Central nervous system (Seeber, 1989)        0.30 mg/m3        0.12mg/m3
 7          Kidney (Mutti et al., 1992)                   0.36 mg/m3        0.14 mg/m3
 8          Liver (Gennari et al., 1992)                  0.28 mg/m3        0.10 mg/m3
 9
10          New York State estimated that an ambient air criterion of 0.1 mg/m3 would provide the
11    general population, including sensitive subpopulations of infants, children, the infirm and
12    elderly, a sufficient margin of exposure over the air levels of tetrachloroethylene associated with
13    noncarcinogenic effects in humans and animals. The ambient air criterion was established at 0.1
14    mg/m3.
15
16    A.2.2.5. Priority Substances List Assessment Report:  Tetrachloroethylene, Canada Health
17    and Welfare Agency, 1993
18          Tetrachloroethylene has been classified in  Group 3 (possibly carcinogenic to humans) of
19    the classification scheme developed for use in the  derivation for the guidelines for Canadian
20    drinking water quality.  A tolerable daily intake (TDI) was derived using data from the NTP
21    (1986) study. It was assumed that 100% of the inhaled tetrachloroethylene was retained in the
22    mice. A LOAEL of 100 ppm for reduced survival and hepatotoxic effects in male mice and lung
23    congestion and nephrotoxic effects in male and female mice was used.  An uncertainty factor of
24    5,000 was applied to account for intraspecies variation (10), use of a LOAEL (10), interspecies
25    variation (10), and limited evidence of carcinogenicity (5). A TDI of 34 |J,g/kg bw/day was
26    derived. Using standardized conversion assumptions (EPA), this value is equivalent to 0.018
27    ppm (0.12 mg/m3).
28
29    A.2.3. Summary of Inhalation Risk Estimates
30          The following tables summarize the quantitative risk estimates that have been developed
31    by EPA and other agencies. Table A-3 shows the  cancer risk values. Table A-4 depicts the risk
32    estimates developed for noncarcinogenic endpoints.
33
34    A.3. QUALITATIVE RISK ASSESSMENTS
3 5          This section contains a brief review of documents that included only a qualitative
36    assessment of tetrachloroethylene toxicity and risk.

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-13         DRAFT—DO NOT CITE OR QUOTE

-------
1
2
Table A-3. Estimates of cancer inhalation unit risk using different methods
Dose metric surrogate
(agency)
Dose metabolized (Cal
EPA, 2002)
Dose metabolized (U.S.
EPA, 1998)
Total metabolized dose
(MA DEP, 1998)
Dose metabolized and
administered (NYS DOH,
1997)
Dose metabolized (U.S.
EPA, 1986)
Dose metabolized (U.S.
EPA, 1985
Unit risk value
(Hg/m3)-1
5.9 x 1Q-6
7.1 x ID'7
5.5 x 1Q-5
1 x 1Q-6
2.9 x 10'7to9.5
x 10'7
4.8 x 1Q-7
Studies used
NTP (1986)
NTP (1986)
NCI (1977)
NTP (1986)
NTP (1986)
NCI (1977)
Critical target effect(s)
Liver adenomas and carcinomas in male
mice
Liver adenomas and carcinomas in male
and female mice, mononuclear cell
leukemia in male and female rats
Liver tumors in male and female mice
Liver tumors in male and female mice,
mononuclear cell leukemia in male and
female rats
Liver adenomas in male and female
mice, combined liver adenomas and
carcinomas in male and female mice,
mononuclear cell leukemia in male and
female rats
Liver carcinomas in female mice
4
5
6
7
Table A-4. Inhalation, noncancer endpoints
Agency guideline
Provisional RfC (U.S.
EPA, 1998)
Ambient air criterion
NYS DOH, 1997)
Acute inhalation
minimal risk level
(ATSDR, 1997)
Chronic inhalation
minimal risk level
(ATSDR, 1997)
Tolerable daily intake
(CHWA, 1993)
Standard
0.025 ppm(0. 17 mg/m3)
0.015 ppm(0.1mg/m3)
0.2ppm(1.36mg/m3)
0.04 ppm (0.27 mg/m3)
0.0 1 8 ppm (0.12 mg/m3)
Studies used
Franchini et al.
(1983)
Seeber(1989),
Muttietal. (1992),
Gennari et al.
(1992)
Altmann et al.
(1992)
Ferroni et al.
(1992)
NTP (1986)
Critical target effect(s)
Renal tubule damage in
workers
Motor and cognitive effects,
kidney dysfunction and liver
dysfunction in workers
Neurological effects (VEP,
vigilance, and eye-hand
coordination) in males
Increased reaction times in
female workers
Reduced survival and
hepatotoxic effects in male
mice, lung congestion and
nephrotoxic effects in male
and female mice
               This document is a draft for review purposes only and does not constitute Agency policy
     06/06/08                               A-14        DRAFT—DO NOT CITE OR QUOTE

-------
 1    A.3.1. U.S. EPA Qualitative Risk Assessments
 2    A.3.1.1.  Response to Issues and Data Submission on the Carcinogenicity of
 3    Tetrachloroethylene, U.S. EPA, 1991
 4          This document discusses issues relating to the classification of tetrachloroethylene as a
 5    B2 carcinogen. Lengthy deliberation is given to the specific mechanisms of action that may
 6    explain all the tumor endpoints observed after exposure to tetrachloroethylene.  In conclusion,
 7    EPA stands behind the B2 classification and concludes that sufficient evidence of cancer in
 8    animals does exist.
 9
10    A.3.2. Qualitative Risk Assessments Conducted by Non-U.S. EPA Agencies
11    A.3.2.1.  Organization for Economic Cooperation and Development (OECD), Screening
12    Information Data Set (SIDS) Initial Assessment Report: Comprehensive Risk Assessment
13    Report for Tetrachloroethylene, 1996
14          The current European Union classification for tetrachloroethylene is Carcinogen
15    Category 3. Category 3 indicates "a substance which causes concern for man owing to possible
16    carcinogenic effect but in respect of which the available information is not adequate for making
17    a satisfactory assessment.  There is some evidence from appropriate animal studies, but this is
18    insufficient to place the substance in category 2. "
19          In the summary of carcinogenicity, the report (OECD, 1996) concludes that the liver
20    tumors found in mice and the kidney tumors found  in rats following repeated inhalation exposure
21    are almost undoubtedly not of significance in relation to human health.  This is based on believed
22    differences in metabolic pathways  and mechanisms of action. OECD does not believe that
23    peroxisome proliferation in mice is relevant to human cancer. Similarly, it believes that the
24    human renal beta lyase activity in humans is negligible compared to that in rats.  In evaluating
25    human carcinogenicity, OECD determined that the  epidemiological studies do not show evidence
26    supporting an increased risk of carcinogenicity in humans.
27
28    A.3.2.2.  IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 63,
29    1995
30          IARC determined that tetrachloroethylene is probably carcinogenic to humans and has
31    classified tetrachloroethylene as a Group 2A carcinogen. This judgment is based on limited
32    evidence in humans and sufficient  evidence of carcinogenicity in experimental animals. In
33    evaluating tetrachloroethylene, the following evidence was considered: although
34    tetrachloroethylene is known to induce peroxisome proliferation in mouse liver, poor quantitative
35    correlation was seen between peroxisome proliferation and tumor formation in the  liver after

                 This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                              A-l 5         DRAFT—DO NOT CITE OR QUOTE

-------
 1    administration of tetrachloroethylene by inhalation; the spectrum of mutations in proto-
 2    oncogenes in liver tumors from mice treated with tetrachloroethylene is different from that in
 3    liver tumors from mice treated with trichloroethylene; tetrachloroethylene induced leukemia in
 4    rats; and several epidemiological studies showed elevated risks for esophageal cancer, non-
 5    Hodgkin's lymphoma and cervical cancer.
 6          Evidence of cancer in animal studies is supported by studies that included both oral and
 7    inhalation exposures.  Cancer endpoints included increases in hepatocellular carcinoma in male
 8    and female rats after oral administration of tetrachloroethylene (NCI, 1977), increases in
 9    hepatocellular adenoma and carcinoma in male and female mice and increases in mononuclear-
10    cell leukemia  in male and female rats after inhalation exposure (NTP, 1986).  IARC does not
11    point to a single epidemiological study as being critical, but rather summarizes many studies that
12    support a relationship between cancer and tetrachloroethylene exposures.  IARC relies on the
13    consistent positive associations between human exposures to tetrachloroethylene and the risks
14    for esophageal cancer, non-Hodgkin's lymphoma, and cervical cancer.
15
16    A.3.2.3.  Report on Carcinogens, Eleventh Edition, NTP, 2005
17          The NTP lists carcinogenic substances in one of two categories: (1) known to be a human
18    carcinogen and (2) reasonably anticipated to be a human carcinogen.  They present a brief two-
19    page summary of the evidence for their classification. They classified tetrachloroethylene in
20    Category 2, "reasonably anticipated to be a human carcinogen."  It was first listed in the 5th
21    Annual Report on Carcinogens (1989). They based their classification on sufficient evidence of
22    carcinogenicity in experimental animals and limited evidence in humans.
                This document is a draft for review purposes only and does not constitute Agency policy
      06/06/08                               A-16         DRAFT—DO NOT CITE OR QUOTE

-------
  1                                   REFERENCES FOR APPENDIX A

  2
  3     Altmann, L; Weigand, H; Bottger, A; et al. (1992) Neurobehavioral and neurophysiological outcomes of acute
  4     repeated perchloroethylene exposure. Applied Psychology: An International Review 41:269-279.

  5     Altmann, L; Neuhann, HF; Kramer, U;  et al. (1995) Neurobehavioral and neurophysiological outcome of chronic
  6     low-level tetrachloroethene exposure measured in neighborhoods of dry cleaning shops.  Environ Res 69:83-89.

  7     ATSDR (Agency for Toxic Substance and Disease Registry). (1997) Toxicological profile for tetrachloroethylene
  8     (update). Prepared for Sciences, International under subcontract to Research Triangle Institute, ATSDR, Atlanta,
  9     GA.

10     Bogen, KT; Hall, LC; McKone, TE; et al. (1987) Health risk assessment of tetrachloroethylene (PCE) in California
11     drinking water. DE87-013493. Available from National Technical Information Services, Springfield, VA.

12     Bois, FY; Zeise, L; Tozer, TN. (1990) Precision and sensitivity of pharmacokinetic models for cancer risk
13     assessment: tetrachloroethylene in mice, rats, and humans. Toxicol Appl Pharmacol 102:300-315.

14     Bolanowska, W; Golacka, J. (1972) Absorption and elimination of tetrachloroethylene in humans under
15     experimental conditions (English translation).  Medycyna Pracy 23:109-119.

16     Buben, JA; O'Flaherty, EJ. (1985) Delineation of the role of metabolism in the hepatotoxicity of trichloroethylene
17     and perchloroethylene: a dose-effect study. Toxicol Appl Pharmacol 78:105-122.

18     Cal EPA (California Environmental Protection Agency). (2001) Public health goal for tetrachloroethylene in
19     drinking water. Office of Environmental Health Hazard Assessment.  Available from online at
20     http://oehha.ca.gov/water/shg/8310 IPHG.htm.

21     Cal EPA (California Environmental Protection Agency). (2002) Technical support document for describing
22     available cancer potency factors. Air Toxics Hot Spots Program Risk Assessment Guidelines Part II.  Available
23     online at http://www.oehha.ca.gov/air/hot_spots/pdf/TSDNov2002.pdf.

24     Chen, C; Blancato, J. (1987) Role of pharmacokinetic modeling in risk assessment: perchloroethylene as an
25     example. In: Safe Drinking Water Committee SoPNRC (ed) Pharmacokinetics in risk assessment. Washington, DC:
26     National Academy Press; pp. 367-388.

27     CHWA (Canada Health and Welfare Agency). (1993) Tetrachloroethylene: priority substances list report. Cat. No.
28     En 40-215/28E, Health-Related Sections: Canada Communication Group: Ottawa.

29     Ferroni,  C; Selis, L; Mutti, A; et al. (1992) Neurobehavioral and neuroendocrine effects of occupational exposure to
30     perchloroethylene. Neurotoxicology 13:243-247.

31     Franchini, I; Cavatorta, A; Falzoi, M; et al. (1983) Early indicators of renal damage in workers exposed to organic
32     solvents. Int Arch Occup Environ Health 52:1-9.

33     Fredriksson, A; Danielsson, BR; Eriksson, P. (1993) Altered behaviour in adult mice orally exposed to tri- and
34     tetrachloroethylene as neonates. Toxicol Lett 66:13-19.

3 5     Gennari, P; Naldi, M; Motta, R; et al. (1992) gamma-Glutamyltransferase isoenzyme pattern in workers exposed to
36     tetrachloroethylene. Am JIndMed 21:661-671.
                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                    A-17          DRAFT—DO NOT CITE OR QUOTE

-------
  1     IARC (International Agency for Research on Cancer). (1985) Tetrachloroethylene. In: IARC monographs on the
  2     evaluation of carcinogenic risks to humans; vol. 63: dry cleaning, some chlorinated solvents and others.  Lyon,
  3     France.

  4     MA DEP (Massachusetts Department of Environmental Protection). (1998) Inhalation unit risk for
  5     perchloroethylene. Memorandum from Marion Harnios, ORS, to Carol West, December 1988.

  6     Mutti, A; Alinovi, R; Bergamaschi, E; et al. (1992) Nephropathies and exposure to perchloroethylene in dry-
  7     cleaners. Lancet 340:189-193.

  8     NCI (National Cancer Institute). (1977) Bioassay of tetrachloroethylene for possible carcinogenicity. DHEW Pub.
  9     (NIH) 77-813. U.S. Department of Health, Education and Welfare, Public Health Service.

10     NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of tetrachloroethylene
11     (perchloroethylene) (CAS No. 127-18-4) inF344/N rats andB6C3Fl mice. 311,1-190. U.S. Department of Health
12     and Human Services. Technical Report Series.

13     NTP (National Toxicology Program). (2005) Report on carcinogens, eleventh edition. U.S. Department of Health
14     and Human Services, Public Health Service, Research Triangle Park, NC. Available online at http://ntp-
15     server.niehs.nih.gov.

16     NYS DOH (New York State Department of Health). (1997) Tetrachloroethene ambient air criteria document.  Final
17     Report. Albany, NY.

18     OECD (Organization for Economic Co-operation and Development). (1996) Comprehensive risk assessment report
19     on tetrachloroethylene, SIDS initial assessment report. OECD Secreteriat. Paris.

20     Seeber, A. (1989) Neurobehavioral toxicity of long-term exposure to tetrachloroethylene.  Neurotoxicol Teratol
21     11:579-583.

22     Spinatonda, G; Colombo, R; Capodaglio, EM; et al. (1997) [Processes of speech production: Application in a group
23     of subjects chronically exposed to organic solvents (II)].  G Ital Med Lav Ergon 19:85-88.

24     U.S. EPA (Environmental Protection Agency). (1985) Health assessment document for tetrachloroethylene
25     (perchloroethylene). National Center for Environmental Assessment, Washington, DC; EPA/600/8-82/005F.
26     Available from: National Technical Information Service, Springfield, VA; PB-85-249696/AS.

27     U.S. EPA (Environmental Protection Agency). (1986) Addendum to the health assessment document for
28     tetrachloroethylene (perchloroethylene) [review draft]. National Center for Environmental Assessment, Washington,
29     DC; EPA/600/8-82/05FA.

30     U.S. EPA (Environmental Protection Agency). (1991) Response to issues and the data submissions on the
31     carcinogenicity of tetrachloroethylene (perchloroethylene). Office of health and Environmental Assessment,
32     Washington, DC; EPA/600/6-91/002F.  Pages 1-73.

33     U.S. EPA (Environmental Protection Agency). (1994) Methods for derivation of inhalation reference concentrations
34     and application of inhalation dosimetry. Environmental Criteria and Assessment Office, Office of Health and
3 5     Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F.

36     U.S. EPA (Environmental Protection Agency). (1998) Cleaner technologies substitutes assessment: professional
3 7     fabricare processes. Office of Pollution Prevention and Toxics, Washington, DC; EPA 744-B-OO1.

38     U.S. EPA (Environmental Protection Agency). (2005) Integrated Risk Information System (tetrachloroethylene file).
39     National Center for Environmental Assessment, Washington, DC. Available online at http://www.epa.gov/iris.

                   This document is a draft for review purposes only and does not constitute Agency policy
       06/06/08                                   A-l 8          DRAFT—DO NOT CITE OR QUOTE

-------