Document zoYyX5qVp7rykmyezGvnngkMB

TECHNICAL SUPPORT DOCUMENT n PROPOSED IDENTIFICATION OF VINYL CHLORIDE AS A TOXIC AIR CONTAMINANT Part B Report State of California Air Resources Board Stationary Source Division July 1989 CMA 010336 D P. -r T Prepared by: California Department of Health Services Principal Editor: Norman Gravitz, Ph.D., MPH, Staff Toxicologist Reviewed by: George V. Alexeeff, Ph.D. Michael J. Lipsett, M.D. Douglas N. Cox, Ph.D. (California Public Health Foundation) Based in part on work submittted by: Carla C. Christensen and C. Tucker Helmes, Biological and Environmental Chemistry Department, SRI International, 333 Ravenswood Avenue, Menlo Park, California 94025, Under Contract 85-86676 (045A) and by: Deborah Grady, M.D., M.P.H. School of Medicine, University of California, San Francisco, and Allan Smith, M.D., PH.D. University of California, Berkeley CMA 0*0337 TABLE OF CONTENTS 1.0 EXECUTIVE SUMMARY 1.1 Vinyl Chloride Highlights 2.0 METABOLISM AND PHARMACOKINETICS 2.1 Summary 2.2 Absorption. Distribution, and Excretion 2.2.1 Inhalation Administration 2.2.2 Intragastric, Intraperitoneal, Intravenous, Dermal and Oral Administration 2.3 Metabolism 3.0 ACUTE TOXICITY 3.1 Summary 3-2 Acute Zaxic,,Lty 4.0 SUBCHRONIC AND CHRONIC TOXICITY 4.1 Human 4.2 Animals 5.0 DEVELOPMENTAL AND REPRODUCTIVE EFFECTS 5.1 Summary 5.2 Teratogenic Effects in Animals 5.2.1 Inhalation Studies 5.3 Reproductive Effects in Humans Pa&e 1-1 1-5 2-1 2-1 2-2 2-2 2-9 2-12 3-1 3-1 3-1 4-1 4-1 4-3 5-1 5-1 5-2 5-2 5-5 CMA 010338 i 5.0 CENOTOXICITY 6.1 Summary 6.2 Mutagenicity 6.2.1 Bacterial Assays 6.2.2 Eukaryotic Systems 6.2.3 Cultured Mammalian Cell Assays 6.2.4 la Vivo Mutagenicity Assays 6.3 Chromosomal Damage 6.3.1 Dominant Lethal Tests 6.3.2 Chromosome Aberration/Sister Chromatid Exchange Studies 6.3.2.1 Experimental Studies 6.3.2.2 Human Observations 6.3.3 Micronucleus Tests 6.3.4 DNA Damage/Unscheduled DNA Synthesis (UDS) Tests 6.4 Mammalian CellTransformation 7.0 CARCINOGENICITY 7.1 Animal Studies 7.1.1 Summary 7.1.2 Intraperitoneal, Subcutaneous, and Transplacental Administration 1 i Page 6-1 6.1 6-1 6-2 6-5 6-6 6-7 6-7 6-7 6-8 6-8 6-9 6-12 6-12 6-13 7 -1 7-1 7 -1 7-2 CMA 010339 ii I 7.1.3 Oral Administration 7-2 7.1.3.1 Studies by Maltoni and Associates 7-2 7.1.3.2 Studies by Feron and Associates 7-3 7.1.3.3 Studies by Til and Associates 7-5 7.1.4 Inhalation Administration 7-9 7.1.4.1 Studies in Rats 7-9 7.1.4.2 Studies in Mice 7-11 7.1.4.3 Studies on the Potential Effects of Age at Time of Exposure 7-14 7.1.4.4 Studies by Maltoni and Associates 7-20 Human Studies on the Carcinoeenic Effects of Vinvl Chloride 7-31 7.2.1 Introduction 7-31 7.2.2 General Design of Epidemiologic Studies 7-31 7.2.3 Difficulties in Interpreting the Epidemiologic Evidence 7.2.4 Mortality Studies 7-32 7-35 7.2.5 Cancer Risks Associated with Exposure to Vinyl Chloride 7-47 7.2.5.1 Liver Cancer 7-47 7.2.5.2 Other Cancers 7-50 7.2.5.2.1 Brain Cancer 7-50 7.2.5.2.2 Lung Cancer 7-51 7.2.5.2.3 Lymphoma 7.2.6 Exposure Information 7-52 7-52 7.2.7 Conclusions 7-55 iii CHA 010340 3.0 QUANTITATIVE CARCINOGENIC RISK ASSESSMENT 3.1 Introduction 3.2 Analysis of Data from Maltoni et al. 8.3 Analysis of Data from Bi et al. S.U Analysis of Data fromDrew et al. 8.5 Human Studies 8.6 Choice of Appropriate Risk Estimates 9.0 CONCLUSIONS 9.1 Acute Toxicity 9.2 Subchronic and Chronic Toxic1tv 9.3 Pharmacokinetics 9.4 Reproductive Toxicity 9.5 Mutagenicity 9.6 Carcinogenicity REFERENCES APPENDIX A: Abstracts of Maltoni et al. (1984) Bioassays APPENDIX B: Cancer Risk Assessment for Vinyl Chloride Based on Human Data APPENDIX C: Using the Average Exposure Rate of a Cohort in Risk Assessment Analysis 8-1 g.]_ 8-3 8-5 8-6 8-7 8-8 9-1 9-1 9-1 9-2 9-3 9-3 9-4 iv CMA 010341 LIST OF TABLES 4. i 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 Subchronic and Chronic Toxicity of Vinyl Chloride Administered by Inhalation to Animals 4-4 Incidence of Liver Tumors and Neoplastic Nodules in Wistar Rats Exposed Orally to Vinyl Chloride (Feron et al., 1981) 7-6 Incidence of Lung Angiosarcomas, Abdominal Mesotheliomas, and Mammary Tumors in Wistar Rats Exposed Orally to Vinyl Chloride (Feron et al., 1981) 7-7 Liver Tumor Incidence in Male and Female Wistar Rats Exposed to Vinyl Chloride by Oral Administration for 149 Weeks (Til et al., 1983) 7-10 Tumor Incidence Following Vinyl Chloride Exposure in Female Rats, Hamsters and Mice From the Study of Drew et al. (1983) 7-18 Experimental Protocol for Inhalation Studies (Maltoni et al., 1984) 7-23 Tumors Correlated to Inhalation Exposure to Vinyl Chloride in Rats, Mice, and Hamsters in the BT Experiments 7-25 Lowest Concentration at Which a Significant (p < 0.05) Excess of Tumors Was Reported by Maltoni and Associates in Inhalation Studies at Specific Sites in Sprague-Dawley Rats (Maltoni et al., 1984) 7-26 Incidence of Liver Angiosarcomas (LAS) in Male and Female Sprague-Dawley Rats Exposed for 52 Weeks to Vinyl Chloride (Maltoni et al., 1984) 7-27 Incidence of Mammary Gland Carcinomas in Female SpragueDawley Rats and Swiss Mice Exposed by Inhalation to Vinyl Chloride (Maltoni et al., 1984) 7-29 Incidence of Pulmonary Adenomas, Mammary Carcinomas, and Liver Angiosarcomas in Male and Female Swiss Mice Exposed to Vinyl Chloride by Inhalation (Experiment BT4) (Maltoni et al., 1984) 7-30 v CHA 010342 '.11 A Summary of Epidemiologic Data for Occupationally Exposed Vinyl Chloride Workers 7-12 A Summary of Tumor Incidences and Standardized Mortality Ratios (SMR) for Occupationally Exposed Vinyl Chloride Workers 7-13 A Summary of Epidemiologic Studies Which Examined Possible Correlations Between Occupational Vinyl Chloride Exposure and Primary Cancers of the Liver 7-14 A Summary of Epidemiologic Studies Which Examined Possible Correlations Between Occupational Vinyl Chloride Exposure and Brain Cancer 7-15 A Summary of Epidemiologic Studies Which Examined Possible Correlations Between Occupational Vinyl Chloride Exposure and Lung Cancer 7- 16 A Summary of Epidemiologic Studies Which Examined Possible Correlations Between Occupational Vinyl Chloride Exposure and Lymphoma 8- 1 Range of Cancer Potency Values for Vinyl Chloride Calculated from AnimalCarcinogenicity Studies 8-2 Range of Human Risks for Vinyl Chloride Exposure Estimated from Animal Carcinogenicity Studies 8-3 Selected Human Risk Estimates for Lifetime Exposure to 1 ppb Vinyl Chloride Page, 7-36 7-38 7-48 7-49 7-53 7 - 54 8-12 8-13 8-14 LIST OF FIGURES 2.1 Metabolism of Vinyl Chloride 2-14 vi CMA 010343 1u J U \i Executive .Summary Vinyl chloride is a short-chain halogenated hydrocarbon used predominantly in the manufacture of polyvinyl chloride and various packaging and construction products. Vinyl chloride has a very low degree of acute toxicity, with two-hour inhalation LD^ values ranging from 27,419 ppm in mice to 236,215 ppm in rabbits and guinea pigs. Exposure to high concentrations can lead to narcosis, cardiovascular and respiratory irregularity, convulsions, cyanosis and death. Several human deaths have been attributed to occupational exposure to very high levels of vinyl chloride, Autopsies of these patients revealed congestion of the liver, spleen and kidneys. Acute toxicity symptoms are thought to occur above 100 ppm. Chronic exposure of workers to vinyl chloride has been shown to lead to "vinyl chloride disease", characterized by occupational aero-osteolysis, vasospasm of the hands similar to Raynaud's syndrome, dermatitis, circulatory and central nervous system alterations, thrombocytopenia, splenomegaly and changes in liver function. Eight symptoms commonly reported by workers exposed to vinyl chloride (including dizziness, headaches and nausea) were observed even at dose levels below 50 ppm. Vinyl chloride has been shown to induce cancer in animals in utero. but has not been shown to cause any other reproductive or 1-1 CMA 010344 44 developmental effects in rats, mice and rabbits. Epidemiologic studies of families of vinyl chloride workers or communities having vinyl chloride processing facilities suggested the possibility of an increased incidence of birth defects and spontaneous abortions among people at risk; however, subsequent reviews of these studies have concluded that there is inadequate evidence to link ehYirbPBgntal__ or paternal .exposure__ to. vinvl chloride with birth defects or spontaneous abortions in humans. The noncarcinogenic effects occur at concentrations near or above 10 ppm, which is greater than four orders of magnitude above possible general ambient levels in California (0.5 ppb). The noncarcinogenic effects also occur at concentrations greater than 3 orders of magnitude above the highest concentrations measured near landfills (10 ppb). Consequently, DHS staff do not expect noncarcinogenic adverse health effects to occur from acute or chronic exposures to vinvl chloride in ambient air. Xh__ International Agency for Research on Cancer (IARC). the United State?__ Environmental Protection Agency__ LEPA1 SSdL thg, California Department of Health Services (CDHS) have identified vinvl chloride as a chemical for which there is sufficient evidence carcinogenicity In both humans and experimental animals. Chronic inhalation and oral exposures of rats, mice and hamsters to vinyl chloride have been associated with an increased incidence of malignant and benign tumors at several sites including the liver, lung, mammary gland and the nervous system. In humans, 1-2 CMA 010345 epidemiological studies of occupationally exposed workers have linked vinyl chloride exposure to development of a rare cancer, liver- angiosarcoma, and have suggested a relationship between exposure and lung and brain cancers. Although pharmacokinetic studies in humans exposed to vinyl chloride are rare, limited evidence indicates that, following inhalation of low levels of vinyl chloride (3 to 24 ppm), up to 71% (with a mean value of 42%) of the given dose may be absorbed. Vinyl chloride absorption appears to depend on its metabolism, which is a dose-dependent, saturable process. Due to saturation of the enzyme systems responsible for the metabolism of vinyl chloride (cytochrome P-450 and alcohol dehydrogenase), exposure to concentrations above approximately 250 ppm will not necessarily lead to an increasing incidence of tumor development. Metabolism of vinyl chloride leads to formation of chloroethylene oxide and chloroacetaldehyde, two reactive intermediates which undergo covalent binding to cellular macromolecules and are thought to be responsible for the toxic effects of vinyl chloride. These and other metabolites may be further metabolized and excreted in the urine. Unmetabolized vinyl chloride is eliminated primarily in exhaled air. Vinyl chloride is mutagenic in both prokaryotic and eukaryotic test systems, with significantly greater genotoxicity seen after metabolic activation. DHS staff have found no evidence of a carcinogenic threshold level and because__ vjltfl--ffhlQI.jde--is. 1-3 CMA 010346 mutagenic. the staff recommends that vinvl chloride be considered as ..nor having a threshold for carcinogenicity. Several animal carcinogenicity and human epidemiological studies of occupationally exposed workers have been analyzed for risk assessment purposes. Although actual exposure levels are not known, exposure estimates have been used to evaluate the Waxweiler et al. (1976) study of vinyl chloride workers. Based on these estimates, DHS staff has calculated that a lifetime exposure to 0,485 ppb might result in an incremental individual cancer risk of 1 x 10 ^(assuming liver, brain and lung cancer are all related to vinyl chloride exposure). This yields a risk estimate of 2.1 x 10'6/ppb. In the case that only liver cancer is assumed to be linked to exposure, a lifetime exposure to 1.0 ppb may be expected to result in a risk of 1.0 x 10 Due to inadequate exposure data, follow-up time and other methodological problems, DHS staff suggest that the human risk estimates be used only for comparative purposes. Evaluation of animal experiments by the linearized multistage model yields a range of human risks spanning from 1.8 x 10* /ppb to 3.9 x 10* /ppb, with most estimating a risk of between 10 -4 and 10 -5 /ppb. Evaluation of animal tumorigenicity data indicates that vinyl chloride's carcinogenic potency is dependent on sex, tumor site and age of exposure. Taking these factors into account, DHS staff believe that the human risk estimates are consistent with those obtained for laboratory animals. The staff of DHS recommends that the animal data be used to evaluate the risks resulting from vinyl chloride exposure. Consequently, the 1-4 CMA 010347 ni I rante of risks estimated from analyses of animal studies and recommended bv DHS for regulatory purposes lie between 3.9 x 10 -fi /pob and 1.8 x 10 - 3 /nob. Vinyl chloride has not been detected in the ambient air of California (limit of detection - 0.5 ppb) except at certain "hot spots". Air Resources Board (ARB) staff has monitored vinyl chloride emissions from the BKK hazardous waste site in West Covina and the Oil landfill in Monterey Park. Estimates of peak exposure concentrations for maximally exposed receptors range from 2 to 10 ppb at the BKK landfill and from 0.6 to 9 ppb at the Oil site. Air Resources Board staff has estimated that between 17,000 and 131,000 im individuals may be exposed to 1 ppb at the BKK site. A lifetime exposure of 131,000 residents to 1 ppb would be associated with an upper bound estimate of 0.5 to 236 excess cancer cases. The calculations represent the upper range of plausible excess cancer risk: the actual risk, which cannot be calculated, may be insignificant. Based on the finding of vinyl chloride-induced carcinogenicity and the results of the risk assessment, DHS staff lnd.S__ that' vinvl__ chloride is an air pollutant which mav cause or fegntribw__ to an increase in mortality or an increase in serious lUnggg.--or which mav nose a present or potential hazard to human CMA 010348 1 CMA 010349 DRAFT Vinyl Chloride Highlights uRkr1 I. National and International Evaluation (Other Agencies' Evaluation) A. International Agency for Research on Cancer (IARC) 1. Short-Term Tests: There exists sufficient evidence of mutagenic activity__ both__ wi__ without an exogenous abolic activation system. 2. Animal carcinogenicity bioassays: There exists sufficient evidence of animal carcinogenicity by oral administration or inhalation. 3. Human evidence: There__ exists__ sufficient evidence of carcinogenicity to humans. Occupational exposure to vinyl chloride has been linked with development of angiosarcoma of the liver, and has been associated with tumors of the brain and lung and of the hematopoietic and lymphatic systems. Vlnvl chloride is grouped under IARC category 1. meaning that it is__ causally associated with cancer in humans. B. U.S. Environmental Protection Agency (EPA) 1. Short-Term Tests: Sufficient___ evidence of mutagenic activity exists both with__ and without an__ SKgggpous 1-6 cMA 010350 metabolic activation system. for both DMA damage ar.d mutation. 2, Animal carcinogenicity bioassays: There exists sufficient evidence. of animal carcinogenicity bv administration orally pf bY. i-nfraletjon, 3. Human data: A number of epidemiological studies have linked vinvl chloride with angiosarcoma and .other forms of neoplasms.____ Sufficient evidence exists to indicate that vinvl chloride is a human carcinogen bv inhalation. C. Conclusions: Both EPA and 1ARC have concluded there is ample evidence that vinyl chloride is genotoxic and is carcinogenic in both animals and humans. II. Exposure Sources A. Air Levels 1. Ambient levels were monitored in the Los Angeles Basin area in 1983 and 1984 by the South Coast Air Quality Management District. All measurements were below their limit of detection of 0.5 ppb. 2. Ambient levels measured in "hot spots" by ARB staff: Estimates for the maximally exposed receptors downwind from and adjacent to hazardous waste sites ranged from 0* 10 ppb. 1-7 0103^1 II. Quantitative Risk Assessment A. Range of Extrapolation: Animal to human exposures in air for calculated lifetime daily exposure. 1. Experimental to ambient: Vinyl chloride has not been detected in ambient air, except at "hot spots". 2. Experimental to "hot spots": The lowest doses in the animal studies are approximately 10- to 20-fold higher than the highest residential exposures. B. Ranee of Risks: The human risks associated with a continuous, lifetime exposure to vinyl chloride have been estimated using the linearized multistage model from both animal carcinogenicity bioassays and epidemiological studies of exposed workers. Unit risks for humans estimated from animal data range from 1.8 x 10* /ppb to 3.9 x 10'/ppb, depending on experimental exposure levels, tumor type observed, and sex, species, and age of animal evaluated. A unit risk of 2.1 x 10*Vppb ft liver, lung, and brain cancer was derived from epidemiological studies for comparative purposes. 1-8 010352 010353 METABOLISM AMD PHARMACOKINETICS draft Summary Experimental evidence has suggested that vinyl chloride must undergo transformation to a reactive metabolite(s) by the liver to be toxic. Based on this information, the best dose-response data would consider the amount of vinyl chloride actually absorbed and metabolized rather than the reported exposure or administered dose concentrations. Reports of the vinyl chloride metabolism in humans are sparse, but limited evidence indicates that, after inhalation exposure to low concentrations, up to 71% (average - 42%) of a given dose was absorbed (Krajewski et al., 1980). Based on this study it is assumed that 71% of an inhaled vinyl chloride exposure may be absorbed by humans at ambient concentrations. Unmetabolized vinyl chloride is eliminated primarily via the lungs. Unlike in other species, absorption of vinyl chloride at the doses tested was not concentration-dependent in humans. Data from rodent studies suggest that the absorption of vinyl chloride depends on its rate of metabolism and the extent of metabolic saturation. Studies in rats and a single experiment in monkeys indicate that the metabolic pathways of vinyl chloride become saturated at exposure concentrations between 100 and 300 ppm. Metabolism of vinyl chloride involves the cytochrome P-450 mixedfunction oxidase system. The first step is thought to be epoxidation of the double bond to form the reactive epoxide 2-1 CMA 01035-4 DRAF chloroethylene oxide, which may undergo a number of further reactions, including binding to cellular macromolecules. Intramolecular rearrangement of the chlorine atom may also occur, resulting in the formation of chloroacetaldehyde, another reactive intermediate. In addition, alcohol dehydrogenase has a role in vinyl chloride biotransformation, since inhibitors of this enzyme can significantly reduce the amount of vinyl chloride metabolized. Section 2.3 of this report provides a detailed discussion of vinyl chloride metabolism. 2.2 Absorption. Distribution and Excretion 2.2.1 Inhalation The pharmacokinetics of vinyl chloride following inhalation has been studied in multiple species of experimental animals. The uptake of vinyl chloride at higher doses appears to depend on its metabolism. The metabolic breakdown of vinyl chloride in rats and monkeys (and perhaps in other species) is a dose-dependent, saturable process (Buchter et al., 1980, Filser and Bolt, 1979). Substantial species differences have been observed in the races of vinyl chloride clearance, with first-order metabolic clearance rates (in liters/hour/kg body weight) for the elimination of vinyl chloride decreasing in the order of mouse (25.6) > gerbil (12.5) > Wistar rat (11.0) > Rhesus monkey (3.55) > rabbit (2.74) > human (2.02) (Buchter et al., 1980). 2-2 CMA 010355 Results from inhalation exposure studies in humans, monkeys, and rats using direct and indirect test methods indicate that vinyl chloride is rapidly absorbed and metabolized, quickly distributed throughout the body, and excreted by the kidneys. Unmetabolized vinyl chloride is expired by the lungs and, to a limited extent, expelled in the feces. Several limited studies have been conducted in humans measuring vinyl chloride absorption following inhalation exposure. Krajewski et al. (1980) observed that five male volunteers exposed to 3, 6, 12, or 24 ppm vinyl chloride for six hours by a "face only" chamber absorbed an average of 42% of the dose regardless of concentration. Large interindividual variation in the degree of vinyl chloride retention was observed, with one individual retaining 71% of the dose at the time exposure was terminated; no other individual retained greater than 45%. This indicates a possible large range of interindividual variability. Concentration of vinyl chloride in expired air, measured for 90 minutes after cessation of exposure, decreased to negligible amounts after only 30 minutes post exposure. The quantity of unmetabolized vinyl chloride exhaled was considered negligible and constituted roughly 4% of the inhalation concentration of vinyl chloride to which subjects were exposed (Krajewski et al., 1980). Thus, humans metabolized up to 96% of the absorbed vinyl chloride dose. Buchter et al. (1978) reported that humans exposed to 2.5 ppm vinyl chloride retained 26-28% of the administered dose (Krajewski et 2-3 CMA 010356 draf al., 1980). Substantial interindividual differences were reported - in this study, these differences appear due to differences in the adipose tissue mass among individuals, although this hypothesis has not been confirmed in follow-up studies (Buchter, 1979; Buchter et al., 1978; Bolt et al., 1981). Pulmonary absorption of vinyl chloride by rats occurs rapidly. Blood levels of vinyl chloride increase with the dose. Blood concentrations quickly decline after cessation of exposure; unmetabolized vinyl chloride is exhaled (Withey, 1976; Hefner et al., 1975a; 1975b; 1975c). Evidence from both whole animal and "nose-only" inhalation studies in rats indicates that the rate of pulmonary uptake of vinyl chloride in a closed system is partially dependent on the extent of metabolism (Bolt et al., 1977; Hefner et al., 1975a; 1975b; Withey, 1976). In the "nose-only" exposure system used by Hefner et al. (1975a), pretreatment of rats with either pyrazole (a non-specific inhibitor of alcohol dehydrogenase) or 95% ethanol significantly reduced both the uptake (as calculated from the disappearance of vinyl chloride from the exposure chamber) and metabolism of vinyl chloride. This held true for both exposure levels. Pyrazolepretreated rats were exposed to either 65 or 1234 ppm, while ethanol-pretreated rats were exposed to 56 or 1034 ppm. Several groups of investigators have presented additional data concerning the uptake, metabolism and disposition of vinyl chloride 2-4 CMA 010357 DRArT following inhalation exposure (Bolt et al., 1976; 1977; Hefner et al., 1975a; 1975b; Buchter et al., 1977). In an investigation into 14 the disposition of C-vinyl chloride, Bolt and co-workers (1976) exposed male Wistar rats to initial concentrations of "less than 100 ppm" vinyl chloride (apparent range 1-50 ppm) in a closed system for six hours. The half-life for vinyl chloride disappearance from the chamber was about 68 minutes. From this study, the authors estimated that approximately 40% of the inspired vinyl chloride was absorbed by the lungs (Bolt et al., 1976). Pulmonary uptake of vinyl chloride by rats was completely blocked following pretreatment with the cytochrome P-450 inhibitors 6- nitro-1,2,3-benzothiadiazole or 3-bromophenyl-4(5)-imidazole (Bolt _ et al., 1976). Uptake of vinyl chloride appeared to be linked to its metabolism, since 24 hours after pretreatment with the relatively short-lived P-450 inhibitor 3-bromophenyl-4(5)-imidazole the uptake of vinyl chloride had returned to control levels. Following exposure, the liver and kidney contained the highest levels of vinyl chloride metabolites (Bolt et al., 1976). In an attempt to determine the exact minimal concentration of vinyl chloride in air necessary to achieve metabolic saturation. Bolt et al. (1977) exposed groups of rats to a wide range of vinyl chloride concentrations and showed that saturation occurred at 250 ppm. First-order kinetics occurred at exposures less than 250 ppm, while zero-order kinetics predominated at higher exposures. Hefner and colleagues (1975a; 1975b) exposed male Sprague-Dawley rats to initial vinyl chloride concentrations ranging from 50 to 2-5 CMA 010358 DRAF 1,167 ppm in a closed nose-only inhalation system. The rate of uptake of vinyl chloride by the animals (as calculated from the rate of disappearance of vinyl chloride from the chamber atmosphere) was approximately three times greater for doses less than 105 ppm (range 50 to 105 ppm) than for doses greater than 220 ppm (range 220 to 1,167 ppm). After an initial equilibration period and regardless of the administered concentration, vinyl chloride disappearance from the chamber apparently followed firstorder kinetics. The half-life for atmospheric vinyl chloride at concentrations below 100 ppm was 86 minutes compared with 261 minutes for concentrations greater than 220 ppm. Hefner et al. (1975b) concluded that the predominant pathway for metabolism of vinyl chloride by rats exposed to 100 ppm or less is saturable, and that this metabolism was due primarily to alcohol dehydrogenase (based on the inhibitor studies with pyrazole and ethanol). Studies in rats and monkeys suggest that, after absorption, vinyl chloride is rapidly distributed to all tissues reached by the bloodstream (Duprat et al., 1977; Buchter et al., 1980). Lipids or lipoproteins, rather than proteins, transport vinyl chloride in the blood (Bolt et al., 1977). Studies of the distribution of re labeled vinyl chloride in rats indicated that, immediately after inhalation administration, the liver (predominant site of metabolism) and the kidneys (site of excretion of polar metabolites) contained the highest concentrations of 14C activity, followed by lungs, spleen, and small intestine (Watanabe et al., 14 1976a; Bolt et al., 1976). However, C counts quickly decreased 2-6 CKA 010359 DRAFT after cessation of exposure. In one study, vinyl chloride metabolite concentrations decreased significantly in these tissues 48 hours after a single inhalation exposure (50 ppm for five hours) compared to measurements made immediately after exposure ended (Bolt et al., 1976). Watanabe and co-workers (1976a) also examined the fate of 14C-vinyl chloride following inhalation exposure in rats, Male SpragueDawley rats were exposed to 10 or 1,000 ppm vinyl chloride in whole-body metabolism cages for six hours and were observed for an additional 72 hours. After exposure to 10 ppm vinyl chloride, urinary radioactivity accounted for 68%, expired vinyl chloride for 2%, expired C02 for 12%, feces for 4%, and carcass and tissues for 14%, respectively, of the recovered radioactivity. After exposure to 1,000 ppm, urinary radioactivity accounted for 56%, expired vinyl chloride for 12%, expired C02 for 12%, feces for 4%, and carcass and tissues for 15% of the recovered radioactivity. The patterns of pulmonary elimination of unmetabolized vinyl chloride following exposure to 10 or 1,000 ppm were similar and could be described by first-order kinetics, with half-lives of 20.4 and 22.4 minutes, respectively. A corresponding biphasic elimination of urinary radioactivity following inhalation exposure to 10 or 1,000 ppm vinyl chloride was observed; the half-lives for the initial phase were 276 and 246 minutes, respectively. The liver and skin contained the highest concentrations of radioactivity 72 hours after exposure to either dose. The authors concluded that since "the rate of elimination of vinyl chloride per se from the lungs or 2-7 CMA 010340 DRAFT 14 C activity in the urine was not different in rats exposed to 10 or 1000 ppm," the dose-dependent fate (the relative amount of vinyl chloride excreted by the two different routes) was not attributable to saturation of the excretion pathways. The results are in agreement with the hypothesis that the metabolism of vinyl chloride becomes saturated at high exposure levels (Watanabe et al., 1976a). The pharmacokinetics of inhaled vinyl chloride in a closed system has also been examined in Rhesus monkeys (Buchter et al., 1980). Uptake of vinyl chloride appeared to depend on its metabolism and to be a dose-dependent, saturable process. When monkeys w re exposed to concentrations up to 200-300 ppm in a closed system, vinyl chloride disappearance from the chamber followed apparent first-order kinetics. At higher exposure levels (up to 800 ppm), zero-order kinetics were observed, implying metabolic saturation. The first-order clearance rate was 3.55 liters/hour/kg. The clearance rate fell by 90% after pretreatment with the alcohol dehydrogenase inhibitor disulfiram (Buchter et al., 1980). Thus, alcohol dehydrogenase appears to have a role in vinyl chloride metabolism along with cytochrome P-450. Liver microsomal enzyme activities and macromolecular covalent binding in rats following either single or repeated exposures to vinyl chloride were compared by Vatanabe et al. (1978a). One group of rats was exposed by inhalation to 5,000 ppm nonlabeled vinyl chloride 6 hours/day, 5 days/week for 7 weeks, and then exposed to 14 14 C-vinyl chloride on the last day. The fate of the C-vinyl 2-8 CMA 010341 chloride from these rats was compared with a separate group exposed 14 for a single 6 hour period to 5,000 ppm of C-vinyl chloride. The activities of aniline hydroxylase and p-nitroanisole O-demethylase were the same in rats exposed once or repeatedly or in unexposed control rats. Covalent binding to hepatic macromolecules was greater in rats repeatedly exposed as compared to those given a single exposure. Watanabe et al. (1978a) concluded that this "increase in hepatic macromolecular binding indicates that repeated exposure augments the reaction of electrophilic metabolites with macromolecules, and this may be expected to enhance potential toxicity, including carcinogenicity". Chronic exposure (28,000 ppm, seven hours/day, five days/week for 2, 4 or 6 weeks) was found to increase glutathione reductase activity, glutathione-S-epoxide transferase activity, glutathione* S-aralkyl transferase activities, and glutathione levels in rat liver and to depress cytochrome P-450 levels (Du et al., 1982). This suggests that a reactive metabolite of vinyl chloride can destroy cytochrome P-450 and disrupt several enzymes that may effect its chronic toxicity. 2.2.2 Intraeastric. Intraoeritoneal. Intravenous. Dermal, and Oral Administration Uptake and absorption of vinyl chloride administered by intragastric (IG), intraperitoneal (IP) and intravenous (IV) administration follows the patterns observed in Inhalation studies. 2-9 CMA 010362 DRAF appears from these studies that the quantity of vinvl chloride metabolized by these routes is dependent on the quantitv administered. Green and Hathway (1975) examined the excretion pattern of single doses of 0.25 and 450 mg/kg of radiolabeled l4C-vinyl chloride administered to rats by the IG, IP, and IV routes. More than 90% of the administered dose was excreted within the first 24 hours. Exhalation of unmetabolized vinyl chloride is the predominant route of excretion for each route of exposure at the high dose and for the low-dose intravenous exposure. After IG administration of the high dose, more than 90% of the dose was exhaled as unmetabolized vinyl chloride and leas than 1% as CO^, while 5% of the administered radioactivity was found in the urine. At the low dose, urinary excretion accounted for 72% of the dose, unchanged exhaled vinyl chloride for 4% of the dose, and COj for 13% of the dose. About 100 times more vinyl chloride was metabolized at the higher dose level than at the lower dose (an 1,800-fold difference in dose). These observations suggest chat the metabolism of vinyl chloride is saturable by administration of a single dose. In another experiment, chronic IG dosing with unlabeled vinyl chloride at 3, 30, or 300 mg/kg daily for 60 days did not affect the rate or route of elimination of a single dose of radiolabeled vinyl chloride from the body. Based on these results, the authors suggested chat vinyl chloride excretion data for a single dose may also apply for chronic exposure to vinyl chloride. DRAFT Watanabe and associates (1976b) examined the excretion of MC- labeled vinyl chloride following single oral doses of vinyl chloride in rats. Their results were similar to those of Green and Hathway (1975). After administration of a single oral dose of 0.05, 1, or 100 rag/kg of 14C-vinyl chloride to male rats, urinary metabolites accounted for 68, 59, and 11%, respectively, of the administered dose while the in expired air accounted for 9, 13, and 3%, respectively. Pulmonary elimination of unmetabolized vinyl chloride represented only 1 to 3% at the lower dose levels, but 67% at the higher dose level. Pulmonary clearance of the 0.05 and 1 mg/kg doses was monophasic, with half-lives of 53.3 and 57.8 minutes, respectively. Clearance of the 100 mg/kg dose was biphasic, with half-lives of 14.4 and 40.8 minutes for the fast and slow phases, respectively. Absorption of vinyl chloride after oral administration has been measured in rats, both in diet studies (Feron et al., 1981) and gavage studies (Withey, 1976; Watanabe, 1976b). In these reports, almost 100% of the administered dose was absorbed, suggesting extensive gastrointestinal uptake of vinyl chloride. Maximum blood concentrations of vinyl chloride were observed within 10-20 minutes following dosing with aqueous or vegetable oil solutions (dose range 12.5-28.2 mg per rat (Withey, 1976). Green and Hathway (1975) observed absorption of 98.7% from the gastrointestinal tract following an oral dose of 450 mg/kg. 2-11 CMA 01036* Limited percutaneous absorption (0.03% of dose) following whole body exposure (excluding the head) to either 800 or 7000 ppm of vinyl chloride has been demonstrated in monkeys (Hefner et al., 1975c). The usefulness of this study is limited, however, since only one monkey was exposed at each dose level. Exposure times were limited to 2.5 hours for the 800 ppm group and 2 hours for the 7000 ppm group. The majority of the absorbed vinyl chloride was eliminated in the expired air (Hefner et al., 1975c). 2.3 Metabolism Two main routes have been proposed for the metabolism of vinyl chloride. The first involves both microsomal and nonmicrosomal enzymes and results in the conversion of vinyl chloride to 2chloroethanol and subsequent oxidation to 2-chloroacetaldehyde and monochloroacetic acid. This pathway is believed to operate at low doses ( 100 ppm) and is saturable. It leads to the production of polar metabolites, which are predominantly excreted in the urine. The initial.studies of Hefner and colleagues (Hefner et al., 1975a; 1975b), discussed in Section 2.2, provide evidence for the role of alcohol dehydrogenase in the metabolism of vinyl chlorid . Following exposure of Sprague-Dawley rats to low concentrations (< 200 ppm), vinyl chloride was metabolized to 2-chloroethanol, chloroacetaldehyde, and monochloroacetic acid by an alcohol dehydrogenase (ADH)mediated pathway. Pretreatment of rats with pyrazole (an ADH inhibitor) or 95% ethanol significantly reduced 2-12 CMA 010365 both the uptake and metabolism of inhaled vinyl chloride (Hefner et al., 1975a), The second proposed pathway involves only microsomal enzymes and is believed to result in the formation of chloroethylene oxide, which may then spontaneously rearrange to form 2-chloroacetaldehyde and, subsequently, monochloroacetic acid (Kilbey, 1981). The epoxide, chloroacetaldehyde, and monochloroacetic acid can then undergo conjugation with glutathione. Further metabolism of these glutathione conjugates can produce a number of compounds, some of which have been identified in the urine of animals treated with vinyl chloride (Figure 2.1). Specifically, monochloroacetic acid, S-(carboxymethyl)cysteine, N-acetyl-S-(2-hydroxyethyl) cysteine, N- acetyl-vinylcysteine, and thiodiglycolic acid have been found in the urine of rats exposed to vinyl chloride by the inhalation and oral routes (Green and Hathway, 1975; 1977; Watanabe et al., 1976a; 1976b). Thiodiglycolic acid and chloroacetic acid have been detected in the urine of workers exposed to atmospheric vinyl chloride (Muller et al., 1978; Heger et al., 1982). The generation of CO^ from vinyl chloride has been postulated to occur through the tricarboxylic acid cycle or the one- or two-carbon pools, with chloroacetic acid or chloroethylene glycol as the starting intermediate (Woo et al., 1985). Studies by Bolt and co-workers (1976) indicate that the cytochrome P-450 system is involved in vinyl chloride metabolism. Their results demonstrated that the uptake of 50 ppm vinyl chloride in a 2-13 CMA 010366 figure 2.1 fletabolism of Vinyl Chloride CI!2-CHC1 ilxcd-functloii% TM r_ ^ oxldase9 ll2C CIIC1 \/ ->cicn2-aio aldehyde dehydrogennnc vinyl chloride chioroethylcneoxide* chloroac etuldehydo^ CH'H -COOJl mnunchlnrai el lr ,n 1<1 i t CSli ' CMA 0 1 0 3 6 7 Mrio cys-S-CH2-CH201l <---------2 S-(2-hydroxyethyl)cysteine t N-Ac-cy*-S-CH2-CH2OH N-acetyl-S-(2-hydroxyethyl)cystelne 1 Source: I ARC (1979) 1 Detected In vivo 2 Detected _ln vitro 1 esil * glutathione S-fornylnethyl- 1glutathione cys-S-CI^-CIIO S-formylmethyl- cysteine i aldehyde oxlda9e xanthine oxidase > c-s-rti -rum M-catLuxymethyl *glut all) lone --cy;.-s-t;n2-f LS-cnrboxymethylcysi efiu ` Nil. (traosamlnat Jo:i) Lit (oxidative decarboxy I at Inn) H00C-Cil?-.S-CJI -(.non thlodlpl >'i n II i n I it * (thlodiacet 1 ; nrl i'i --n closed system was completely blocked by inhibitors of cytochrome P- 450, such as 3-bromophenyl-4(5)-imidazole or 6-nitro-1,2.3- benzothiodiazole. Pretreatment with the insecticide dichlorodiphenyl trichloroethane (DDT), an inducer of cytochrome P- 450, was effective in enhancing uptake and absorption. However, phenobarbital, another P-450 inducer, has shown no effect on vinyl chloride metabolism (Guengerich and Watanabe, 1979), possibly due to selective induction of different cytochrome P-450 isozymes by the two compounds. Chronic ethanol treatment has been shown to potentiate the carcinogenic effect of vinyl chloride in male Sprague-Dawley rats (Radike et al., 1981). Animals were exposed by inhalation to 600 ppm vinyl chloride four hours/day, five days/week, for one year. Ingestion of 5% ethanol in water (volume/volume, v/v) a 1ihimm was begun four weeks prior to vinyl chloride exposure and continued for life or until the termination of the experiment, 2.5 years after the first vinyl chloride exposure and 1.5 years after vinyl chloride exposure was terminated. The incidence of liver angiosarcoma in rats exposed to vinyl chloride and ethanol was 50% (40/80) versus 23% (18/80) in rats exposed to vinyl chloride alone and 0% (0/80) in animals treated only with ethanol. Radike and associates have suggested that this potentiation of tumor formation may be due to the effect of alcohol on vinyl chloride metabolism and a shared step in the oxidation of ethanol and vinyl chloride. The acetaldehyde product in ethanol metabolism may compete with chloroacetaldehyde for ADH. This would result in 2-15 CMA 010368 DRAFT higher levels of chloroacetaldehyde. However, this metabolite may not be the ultimate carcinogen. Chloroacetaldehyde buildup may result in a decrease in epoxide-to-aldehyde conversion, leading to epoxide buildup and increased interaction with cellular macromolecules. Radiolabeled vinyl chloride has been shown to bind covalently to cellular macromolecules ia vivo and in vitro (Watanabe et al., 1978b; Woo et al., 1985; International Agency for Research on 14 Cancer [IARC] 1979). Watanabe et al. (1978b) exposed rats to Cvinyl chloride (range 1-5000 ppm) for six hours, and measured covalent binding of radioactivity to hepatic macromolecules, RNA and DNA, along with levels of hepatic glutathione. Binding of vinyl chloride metabolites to liver macromolecules did not increase proportionately with dose, but was instead related to the total amount of vinyl chloride metabolized. Binding appeared to plateau above 500 ppm, while below 100 ppm binding was approximately proportional to the increase in exposure. Depression of hepatic glutathione occurred only at exposure levels of 100 ppm or higher. Covalent binding to RNA or DNA was not detected for any exposure group (Watanabe et al., 1978b). However, a subsequent study found covalently bound vinyl chloride metabolites attached to proteins and nucleic acids isolated from the livers of rats exposed to either 10 or 250 ppm vinyl chloride for two hours. (Guengerich and Watanabe, 1979). Rat liver DNA isolated from the two groups of exposed animals contained 0.04 and 0.9 pg of total bound metabolites per gram of wet liver, respectively. Pretreatment with 2-16 CHA 010369 ( phenobarbital had no apparent effect on metabolism or DMA-binding of metabolites, but did increase binding to protein and RNA at the 10-ppm dose level. In vitro binding of 14C-vinyl chloride to proteins and nucleic acids appeared to be dependent on the thiol content of the proteins and the presence of reduced nicotinamide adenine dinucleotide phosphate (NADPH), oxygen, and microsomal enzymes (Guengerich and Watanabe, 1979). Both chloroethylene oxide and chloroacetaldehyde have been studied as possible reactive intermediates that could act as the "ultimate" mutagen or carcinogen formed from vinyl chloride. The epoxide is considered to be the most biologically active metabolite (Bartsch et al., 1975; Laib and Bolt, 1977). Other researchers hav proposed that chloroacetaldehyde may be a more effective alkylating agent (Woo et al., 1985). in vivo and in vitro studies by Guengerich and Watanabe (1979) suggest that the mechanism for activation and binding of vinyl chloride involves the release of the chloride atoms as chloride ions, either in the actual activation mechanism or in rearrangment of the metabolite or adduct. However, Guengerich and Strickland (1977) have demonstrated that neither chloroethylene oxide or 2-chloroacetaldehyde appear to be the vinyl chloride metabolite responsible for destruction the heme group of cytochrome P-450, and that other mechanisms (or reactive metabolites) may account for this. 2-17 CMA 010370 DRAFT CMA 010371 3.0 ACUTE TOXICITY DRAFT The toxic effects of acute exposure to vinyl chloride have been reported by several investigators (Selikoff and Hammond, 1975; Torkelson and Rowe, 1981; EPA, 1984). Relevant findings are summarized below. 3.1 Summary The acute effects of vinyl chloride are similar for humans and animals: central nervous system depression (anesthesia) and cardiac, circulatory, and respiratory irregularities. Frostbite from contact of skin with liquid vinyl chloride has been reported. Repeated inhalational exposure of humans to high concentrations of vinyl chloride has been associated with narcosis, damage to the liver, spleen, and circulatory system, and a complex of symptoms identified as occupational acro-osteolysis. With the exception of acro-osteolysis, the occurrence of these toxic symptoms has also been confirmed in experimental animals. The exact occupational exposure levels associated with these symptoms are not known, but are thought to be above 100 ppm. 3.2 Acute Toxicltv The doses causing 50% lethality (LDjq) in groups of animals exposed to vinyl chloride by inhalation for two hours has been reported to 3-1 CMA 010372 DRAFT be 27,419 ppm in mice, 47,640 ppm in rats, 236,215 ppm in guinea A pigs, and 263,215 ppm in rabbits, indicating a very low order of acute toxicity. Toxic symptoms following exposure included narcosis accompanied by respiratory and circulatory disturbances. Death was caused by respiratory failure. Microscopic examination of all animals indicated damage to the lungs, liver, and kidneys (Prodan et al., 1975a). Several human deaths following very high exposure (concentrations unreported) to vinyl chloride have been reported. Autopsies revealed congestion of the liver, spleen, and kidneys (Danziger, 1960, cited in Maltoni et al., 1984), Lester and co-workers (1963) estimated that the short-term (five minutes) exposure limit (STEL) of vinyl chloride to which a human could be exposed without symptoms of acute toxicity was between 8,000 and 13,000 ppm. Suciu et al. (1975) reported that workers exposed to vinyl chloride (levels not given) experienced euphoria, intoxication, and narcosis. They also reported generalized transient contact dermatitis after dermal exposure. 3-2 CMA 010373 J.O SUBCHRO-'IG AND CHRONIC TOXICITY draft Several studies on the toxic effects resulting from subchronic and chronic exposure to vinyl chloride have been published (Selikoff and Hammond, 1975; Torkelson and Rowe, 1981; EPA, 1984), The relevant findings are summarized below. 4 Hyman Reports on the adverse effects of repeated occupational exposure to vinyl chloride are based mainly on the observations of workers who have been the most heavily exposed. These individuals were involved in occupations such as cleaning autoclaves and centrifuges, or engaged in drying and shifting processes. They experienced a wide range of symptoms: a vasospastic disorder in the hands similar to Raynaud's syndrome; occupational acro- osteolysis, which included clubbing-like swellings and loss of bone from the terminal phalanges, scleroderma-like skin changes, and dermatitis; acrocyanosis, consisting of vascular changes and impaired thermoregulation; positive cold test reactions; capillaroscopic alterations; paresthesias; and central nervous system symptoms. These clinical symptoms (classified as "vinyl chloride disease") were accompanied by circulatory disturbances, thrombocytopenia, splenomegaly, and changes in the liver. The period of exposure before the first sign of symptoms was as short as one month to as long as three years. A year or two after removal from exposure, most of the abnormalities disappeared 4-1 CMA 010374 DRAF (Veltman ec al., 1975; Wilson et al,, 1967; Harris and Adams, 1967; Lilis et al., 1975). Several studies have reported hepatotoxicity and impaired liver function in humans resulting from exposure to vinyl chloride at concentrations ranging from 1 to 470 ppm (Marstellar and Lelbach, 1975; Lilis et al., 1975; Thomas and Popper, 1975; Suciu et al., 1975). Repeated occupational exposure to vinyl chloride has also been noted to result in impaired pulmonary function (Miller et al., 1975; Gamble et al., 1976). Interstitial pulmonary fibrosis has been reported, but these particular workers were also exposed to polyvinyl chloride dust. It has been proposed, but not satisfactorily demonstrated, that interstitial pulmonary fibrosis may be caused by vinyl chloride-altered immune status (Lilis et al., 1975; Ward et al., 1976). In a study of present and past workers affected with vinyl chloride disease, Ward et al. (1976) observed a range of symptoms associated with immune system dysfunction in 19 of the 28 affected workers. From their study of occupationally exposed workers, Spirtas et al. * (1975) concluded chat a dose-response relationship existed between exposure to vinyl chloride and certain acute (primarily neurological) symptoms. The investigators examined the frequency of eight symptoms indicative of central nervous system disturbance, peripheral neuromuscular and neurovascular disturbance, and local CMA 010375 4-2 irritation. Vinyl chloride doses were estimated from company data describing probable exposure scenarios for different job descriptions. Exposure concentrations appeared to range from 0 to 200 ppm. They observed a statistically significant dose relationship in the occurrence of five of the eight symptoms (dizziness, nausea, headache, tingling sensation in arms and legs, and fatigue). These symptoms occurred after exposures to less than 50 ppm. These data support other observations in humans that indicate vinyl chloride can produce adverse health effects even at levels below 50 ppm (Spirtas et al., 1975). 4.2 Animals Repeated inhalation exposure to vinyl chloride has been reported to result in osteoporosis and toxicity to the liver, kidney, spleen, lung, and testes in certain animals. The results of some of these studies are reported in Table 4.1. 4-3 CMA 010376 Species Guinea Pig Rat Rat Rat Rat Rat, rabbit TABLE 4.1 SUBCHRONIC AND CHRONIC TOXICITY OF VINYL CHLORIDE ADMINISTERED TO ANIMALS BY INHALATION -Asms_____ 100,000 ppm 2 hr/day 200, 100, 50 ppm 7 hr/day TOO ppm 2 hr/day 20,000 ppm 0 hr/day 5,000 ppm 7 hr/day Duration (months) 3 6 6 3 12 0.03-0.04 g/L 4 hr/day 6 Observations Liver, kidney, spleen toxicity. Increase in liver weights; 100 ppm MOAEl. No effects observed. Increased mean liver and spleen weight. Growth retardation; shortened blood clotting time; increased kidney, heart, spleen weight; increased mortality; degenerative and hyper plastic changes in the liver. Cardiovascular disorders, changes in the bioelectric activity of the hypothalamus, hyperadrenalemia, osteoporosis. Reference Prodan et al., 1975b Torkelson et al., 1961 Torkelson et al., 1961 Lester et al., 1963 Feron et al., 1979a,b Basalaev et al., 1972 CMA 0 1 0 3 7 7 l table 4.1 continuted Species Rat Rat Nice Mice Dose 20.W0, 500, 50 ppm 5 hr/day 10, 100, 3000 ppm 6 hr/day 60/wk 1,000, 250, 50 ppm 6 hr/day 6,000, 2,500 ppm 5 hr/day Duration {months} 10 up to 12 up to 12 5 to 6 Observations Liver and testes lesions at exposures of 50 and 500 ppm, respectively; depression of body weight gain at dose levels. Increased kidney, liver, spleen, and heart weight; decreased testis weight in all within 6 months,testis daatage. Deaths at high dose caused by hepatitis; at 50 ppm, lethargy, weight loss, rough coat, hepatitis. Proliferation and hypertrophy of terminal bronchiolar cells at both dose levels. Reference Sokal et al., 1900 Bi et at., 1985 Lee et at., 1977 Suzuki 1980, 1981 01037Q O a n DRAFT* CMA 010379 DRA 5.0 DEVELOPMENTAL AND REPRODUCTIVE EFFECTS 5.1 Summary The developmental and reproductive toxicity of vinyl chloride has been investigated in several experimental and epidemiologic studies (Barlow and Sullivan, 1982; Bardin et al., 1982; Hemminki and Vineis, 1985). Vinyl chloride crosses the placenta of experimental animals. Some data indicates it may act as a transplacental carcinogen. No teratogenic effects were observed when vinyl chloride was administered by inhalation at maternally toxic doses. A single unconfirmed report disclosed a teratogenic effect in rats after vinyl chloride exposure as low as 2.5 ppm. Evidence that vinyl chloride causes male reproductive damage has been presented in one experimental study and in a few human case studies. Epidemiologic analysis of communities located close to polyvinyl chloride plants have suggested an association between those locations and an increased risk of birth defects, but none of the studies have adequately controlled for all confounding variables, and no positive correlation has been made conclusively linking vinyl chloride exposure with harmful reproductive effects. 01039 5-1 'sratozenjc Effects in Animals DRAFJ Inhalation Studies Rats: John ec al. (1977) reported that no developmental toxicity or defects occurred when pregnant Sprague-Dawley rats were exposed to either 500 or 2500 ppm vinyl chloride for seven hours daily on days 6 through 15 of gestation. These concentrations proved toxic to the mothers, however. In a separate experiment, pregnant rats exposed to 2500 ppm vinyl chloride by inhalation and 15% ethanol in drinking water experienced greater maternal and fetal toxicity than animals exposed only to vinyl chloride, but no teratogenic response, was observed. However, fetal body measurements were lower among those rats that received ethanol and vinyl chloride. These effects on fetuses were similar to those reported following administration of ethanol only (John et al., 1981). Ungvary et al. (1978) exposed groups of three pregnant CFY rats to 1500 ppm vinyl chloride continuously on days 1 through 9, 8 through 14, or 14 through 21 of gestation. An increased number of resorbed fetuses was found in the group exposed to vinyl chloride during the first 9 days (p < 0.05), but no significant effects were observed in rats exposed at other stages of gestation. In a recent study reported in abstract form, Radike et al. (1988) reported that vinyl chloride was a transplacental carcinogen capable of causing perinatal oncogenesis. An increase in the 5-2 CHA 010381 DRAFT numbers of liver carcinomas and angiosarcomas in the offspring of pregnant rats exposed to 600 ppm for four hours/day from day 9 to day 21 of gestation was observed. Post-natal exposure of the pups to 600 ppm increased the incidence of liver tumors. Coadministration of 5% ethanol with vinyl chloride did not increase the incidence of treatment-related malignancies. A single Russian study has reported an association between vinyl chloride exposures of as low as 2.5 ppm during pregnancy and embryo lethality, teratogenicity, and fetotoxicity in rats (ftirkova et al., 1978, cited in Barlow and Sullivan, 1982). The study and its results were reported only qualitatively and no statistical data were published. Adverse effects reported included doubling of embryo mortality, a high incidence of cerebral malformations, and fetotoxicity. Bi et al. (1985) examined the effects of vinyl chloride on testicular seminiferous tubules in rats. Groups of 75 rats were exposed by inhalation to either 0, 10, 100 or 3000 ppm vinyl chloride for six hours/day, six days/week for three, six, nine or twelve months. Eight to thirty rats were sacrificed after each exposure period, with remaining animals killed 18 months after the initial exposure (i.e., six months after terminating exposure). Incidence of seminiferous tubule damage for the control, 10, 100 and 3000 ppm group were 19, 30, 37 and 56%, respectively. Changes included cytoplasmic vacuolation, nuclear condensation, fusion of spermatids and spermatocytes, and epithelial necrosis and 5-3 CMA 010382 DRAFT degeneration. Seminiferous tubule damage in the two higher dose groups was significantly greater than for the control group (Bi et al., 1985). Mice: Groups of 30 to 40 pregnant CF-1 mice were exposed by inhalation to either 50 or 500 ppm vinyl chloride for seven hours/day on days 6*15 of gestation. Exposure to 500 ppm caused maternal toxicity while no maternally toxic effects were observed at 50 ppm. No developmental defects were reported in fetuses exposed to either concentration. An increased number of resorptions and decreases in litter size and fecal body weight were seen in mice exposed to 500 ppm, but these effects were considered secondary to the toxic effects of vinyl chloride in the mother (John et al., 1977; 1981). Rabbits: No teratogenic or embryotoxic effects were observed in the offspring of pregnant rabbits (15 to 20 per group) exposed by inhalation to either 500 or 2500 ppm vinyl chloride for seven hours per day on days 6-18 of gestation. The Incidence of resorptions was significantly increased in rabbits exposed to 2500 ppm vinyl chloride, a dose that produced other adverse effects in the dam (John et al., 1977; 1981). Simultaneous administration of 15% ethanol in the drinking water and 500 ppm vinyl chloride in air resulted in increased toxicity to the mother and produced defects in the developing embryo not observed in animals exposed to vinyl chloride alone. 5-4 CMA 010363 Reproductive Effects in Humans DRAFT Several epidemiologic studies have been conducted to assess potential reproductive and developmental effects in the families of vinyl chloride workers (reviewed in Wagoner and Infante, 1980; Clemmesen, 1982). Infante (1976) analyzed birth certificate data obtained from a group of Ohio communities, three of which contained vinyl chloride polymerization plants. Although a statistically significant increase (p < 0.01) in birth defects was observed in the towns with vinyl chloride facilities (compared with the birth defect rate for the entire State of Ohio), several other cities without vinyl chloride factories exhibited rates equally high and higher. Spontaneous abortion rates were also elevated in wives of vinyl chloride workers (Infante, 1976). Edmonds et al. (1975; 1978) conducted two case-controlled studies evaluating CNS malformations among offspring of vinyl chloride workers and families living near polyvinyl chloride facilities in Painesville, IN and Kanawha County, WV. More cases than controls lived within three miles of the polyvinyl chloride plants (p < 0.02). In reviewing these three studies, Hemminki and Vineis (1985) concluded that there was inadequate evidence linking environmental or paternal exposure to vinyl chloride with birth defects in humans. Theriault et al. (1983) measured the incidence of birth defects in infants bom to residents of Shawinigan, Canada between 1966 and 1979. A vinyl chloride polymerization plant had been operating in the town since 1943. Although the authors stated that some 5-5 01038^ CHA DRAFT descriptive data suggested an association between ambient exposure to vinyl chloride and birth defects in the exposed community, no significant increases in either still births or birth defects were observed (Theriault et al., 1983). 5-6 CMA 01038s 6.0 GENOTOXICITY DRAFT 5.1 Summary The genotoxicity of vinyl chloride has been reviewed by several authors (IARC, 1979; Duverger et al., 1981; Bartsch et al., 1975; SRI International, 1983; Fabricant and Legator, 1981). Vinyl chloride causes genetic damage in many test systems, including bacteria, fungi, higher plants, and in vitro mammalian systems, as well as in vivo in Drosophila (fruit fly), rodents, and humans. Previous reviews have suggested that a metabolite of vinyl chloride is the major cause of the observed genotoxicity. However, vinyl chloride has been observed to be mutagenic in some in vitro test systems without an exogenous activation system. This particular effect may be the result of endogenous cellular metabolizing enzymes, or the molecule itself may be genotoxic. From experiments in laboratory animals vinyl chloride does not appear to cause genetic damage to germ cells, but does transform mammalian cells and enhances virally-induced mammalian cell transformation In vitro. This strong evidence of the genotoxicity of vinyl chloride suggests that its reported carcinogenicity proceeds by genotoxic mechanisms. Data that support this suggestion are summarized below. 6.2 Mutagenicity Vinyl chloride is mutagenic in most major short-term tests. Its activity is enhanced in the presence of exogenous or endogenous 6-1 CMA 010386 DRAFT metabolic activation, suggesting that a metabolite may be more mutagenic than the vinyl chloride molecule itself. 6.2.1 Bacterial Assays Several studies of vinyl chloride have been conducted using the Ames' SalfflppeUfr tvphimurium (. fvphimurium) assay (McCann et al., 1975; Bartsch and Montesano, 1975; Bartsch et al., 1975; Garro et al., 1976). These studies indicate that vinyl chloride apparently acts as a mutagen whose effect is significantly enhanced in the presence of liver microsomal enzyme preparations from mice, rats, or humans, and NADPH. For example, Bartsch and Montesano (1975) investigated the mutagenicity of vinyl chloride in air at concentrations of 0, 0.2, 2, or 20% in both the absence and the presence of S-9 fraction obtained from livers of uninduced or phenobarbitone-induced rats. In the absence of metabolic activation, a dose-related Increase of up to 15 times background was observed in . tvphimurium strains TA1535 and G46. In the presence of S-9 from uninduced rats, the frequency of revertants was increased up to 23 times above background in strain TA1530, and up to 16 and five times above background in strains TA1535 and G46, respectively. The frequency of revertants Increased to approximately 28 times above background in strain TA1530, and to 18 and six times above background in strains TA1535 and G46, respectively, when S-9 from phenobarbitone-induced rats was used. In the same study, chloroacetaldehyde, a metabolite of vinyl chloride, proved mutagenic (15 times above background) in strain 6-2 CMA 010387 DRAFT TA1530 in the absence of exogenous metabolic activation. Chloroethylene oxide was less toxic than chloroacetaldehyde, but was also mutagenic (nine times above background) when tested without exogenous metabolic activation. The authors proposed that the increase in revertants in the absence of an exogenous metabolic activation system was either the result of nonenzymatic breakdown products of vinyl chloride or a result of compounds formed by bacterial enzymes. However, the answer to this question was not effectively resolved by this study (Bartsch and Montesano, 1975). Salmonella tvphimurium strains TA1536, TA1537, and TA1538, which are specifically reverted by frameshift mutagens, were unaffected by concentrations of up to 20% vinyl chloride in air, even in the presence of liver fractions from rats or mice (Bartsch et al., 1975). Vinyl chloride in water or methanol was not mutagenic when tested in . tvphimurium strains TA100, TA1530, TA1535 or G46, even with S-9 liver fractions from phenobarbital-induced mice. The authors hypothesized that the inactivity of vinyl chloride might have been caused by the rapid diffusion of vinyl chloride from the solution into the atmosphere (Bartsch et al., 1975). Other experiments have confirmed the mutagenic activity of vinyl chloride in Salmonella. Vinyl chloride was mutagenic in S. tvphimurium strain TA1530, both with and without activation, after incubation in a vinyl chloride/ethanol medium. The mutation rate increased when cells were incubated in the presence of ultraviolet light and decreased when hydroquinone, a radical-trapping agent, 6-3 010388 DRAFT was added to the incubation medium. These results suggest that radical metabolites may be important determinants of mutagenic activity (Duverger-Van Bogaert et al., 1982). The results of other studies support this hypothesis: for example, a free-radical generating system, riboflavin irradiated with ultraviolet light, doubled the number of vinyl chloride-induced revertants in . strain TA1530 (Garro et al., 1976). In at least one study, the increases in vinyl chloride-induced mutagenicity in . tVDhimurium strain TA1530 observed with the addition of liver fractions obtained from untreated or PCB-induced animals were similar, although PCB would be expected to increase, the mixed-function oxidase content of the liver and, consequently, the potential number of mutants (Garro et al., 1976). Vinyl chloride was mutagenic in strain TA1530 in the presence of rat or mouse liver S-9 fraction from Aroclor-induced animals. Mutagenicity was observed even in the absence of an NADPH- generating system. Heat-inactivation of the mixed-function oxidase system did not result in decreased mutagenicity of vinyl chloride. These results suggest that the mutagenic activity observed with vinyl chloride in the Ames' test is not necessarily due to enzymatic activation by a mixed-function oxidase system. Vinyl chloride induced forward and reverse mutations in Escherichia coll (. coli) strain 343/113 (Mohn, 1981) and forward mutations in . coli strain K12 with, but required metabolic activation with mouse liver microsomes (Greim et al., 1975, cited in 1AS.C, 1979). 6-4 CMA 010389 DRAFT Chloroethylene oxide at concentrations of 2.5 mmol was more cytotoxic and mutagenic than chloroacetaldehyde at concentrations of 100 mmol when tested in . coli strain K12A (Perrard, 1985). These results are consistent with those obtained in the Salmonella tvphimurium assay (Bartsch et al., 1975). 6.2.2 Eukaryotic Systems Vinyl chloride induced forward mutations in the yeast Schizosaccharomyces pombe following either a host-mediated assay in mice or in vitro after metabolic activation with mouse liver microsomes (Loprieno et al., 1976; Bartsch and Montesano, 1975). Chloroethylene oxide was mutagenic without activation in the same system (Loprieno et al., 1976). In Saccharomvces cerevisiae strain D4> vinyl chloride (in concentrations of either 16 or 48 mM) induced gene conversion at the adenine-2 and tryptophan-5 loci only in the presence of mouse liver microsomes (Loprieno et al., 1976). Vinyl chloride, both as a gas and as an ethanol solution, was tested for potential mutagenicity in two strains of the fungus Haurosoora crassa. There was no detectable mutagenic effect, either with or without metabolic activation. The authors suggested this was because vinyl chloride could not penetrate the conidia (spore) (Drozdowicz and Huang, 1977). 6-5 CMA 010390 6.2.3 ultured Mammalian Cell Assays DRAFT Vinyl chloride was tested in the Chinese hamster ovarv/hypoxannhine guanine phosphoribosyl transferase (CHO/HGPRT) system, an assay designed to detect mutations in the gene coding for the HGPRT locus. Vinyl chloride (at concentrations of 10% in air) was mutagenic only in the presence of complete S-9 mixtures from Aroclor-induced rat livers. When various cofactors used to activate the liver enzymes (for example, NADPH) were not included in this test system, vinyl chloride was inactive even at higher concentrations (Krahn, 1979). Forward mutations were induced in V79 Chinese hamster lung cells in the presence of phenobarbital-pretreated rat liver supernatant (15,000 x g) (Drevon et al., 1977, cited in IARC, 1979), Huberman et al. (1975) reported that at concentrations of 6-13 nmo\ the vinyl chloride metabolites chloroethylene oxide and 2- chloroacetaldehyde caused a dose-dependent induction of 8- azaguanine (four to eight times above background) and ouabain- resistant (up to 23 times above background) mutants in Chin se hamster V79 cells in vitro. Both 2-chloroethanol and aonochloroacetic acid (at concentrations of up to 2500 ^mol) w re found to be inactive (Huberman et al., 1975). 6-6 01039l In Vivo Mutagenicity Assavs drafi A significant increase in recessive lethal mutations in Drosophila melanogaster was observed after exposure to 850 ppm vinyl chloride for two days. Exposure to 30 ppm for 17 days also caused an increase in recessive lethal mutations. Although vinyl chloride was tested at concentrations ranging from 30 to 50,000 ppm, the mutation frequency rate reached a plateau at 10,000 ppm, a finding the authors attributed to saturation of metabolizing enzymes (Verburgt and Vogel, 1977). However, vinyl chloride did not cause any significant increase in dominant lethal mutations, translocations, or entire or partial sex*chromosome loss following, exposure to 30,000 ppm for 2 days (Verburgt and Vogel, 1977). Chromosomal Damage Dominant Lethal Tests Vinyl chloride failed to produce dominant lethal mutations in offspring of male CD-I mice exposed by inhalation to concentrations of 3,000, 10,000, or 30,000 ppm, six hours/day for five days, and Chen mated with successive pairs of untreated females over an eight-week period (Anderson et al., 1977). There was no evidence that vinyl chloride had any mutagenic effect on any maturation stage of spermatogenesis. In addition, no significant increase in the number of post-implantation early fecal deaths, no evidence of 010392 6-7 DRAFT preimpiantation egg loss, and no reduction in fertility were observed in this study (Anderson et al., 1977). Male rats were exposed to 0, 50, 250, or 1000 ppm vinyl chloride by inhalation for six hours/day, five days/week for 11 weeks (Short et al., 1977). During the eleventh week of exposure, the rats were housed with two untreated females for seven evenings or until matings occurred in both females. Although there was a significant reduction in the number of females who became pregnant when housed with males exposed to 1000 ppm vinyl chloride, there was no significant effect on total implants/female or dead implants/female in those females that became pregnant (Short et al., 1977). No dominant lethal mutations were produced in Drosophila melanoeaster following exposures of up to 30,000 ppm for two days (Verburgt and Vogel, 1977). 6.3.2 Chromosome. Aberration/Sister Chromatid Exchange Studies 6.3.2.1 Experimental Studies Sister chromatid exchanges (SCE) and aberrant metaphases were increased in chromosomes of bone marrow cells of Chinese hamsters exposed to either 1.25, 2.5 or 5% (v/v) vinyl chloride in air for 6, 12, or 24 hours. The greatest number of SCEs were seen after exposure to 2.5% vinyl chloride for 24 hours. The greatest number 6-8 CMrt 010393 draft of aberrant metaphases was observed after exposure to 5% vinyl chloride for 24 hours (Basler and Rohrborn, 1980). The mutagenic potential of vinyl chloride was evaluated in the mammalian spot test. Female C57B1/6J Han mice were mated to male Han/T mice, then exposed to 4600 ppm vinyl chloride in air for five hours on day 10 of gestation. No effect on litter size or coat color was seen in offspring (Peter and Ungvary, 1980). No significant increase in any specific genetic effect was observed in bone marrow cells obtained from male tfistar rats exposed to vinyl chloride at 1500 ppm, six hours per day for five days. There was a significant increase in total "abnormalities" (including chromatid gaps, breaks, and fragments) following this exposure scenario; no increases in any of these parameters (including total "abnormalities") were observed when vinyl chloride exposure was extended to three months (Anderson and Richardson, 1981). 6.3.2.2 Human Observations Several studies of chromosomal abnormalities in the peripheral lymphocytes of workers exposed to vinyl chloride were reported in the IARC monograph (1979). Aberrations most frequently reported were fragments, dicentrics and rings, and breaks and gaps. These earlier studies were of limited value, involving small groups of workers with inadequate controls. For example, Leonard and 6-9 CMA 010394 DRAFT associates (1977) examined lymphocytes from seven men working in a vinyl chloride plant and 11 workers in a vinyl chloride polymerization plant. The incidence of such chromosome aberrations as chromatid breaks and gaps were comparable in all groups, but the degree of severity of the abnormalities observed was more severe in ten of the 11 polymerization plant workers than in the seven workers from the other vinyl chloride factory. The lack of controlled conditions greatly reduces the usefulness of this study. Vinyl chloride levels were less than 10 ppm at the time of the study, but were estimated to have been as high as 500 ppm in earlier years. Also, several of the polymerization plant workers had been given X-ray treatment on the hands, but no controls had been exposed to similar X-rays (Leonard et al., 1977). Another study of 56 workers in the polyvinyl chloride industry suggested that occupational exposure to vinyl chloride could have a measurable effect on the induction of chromosomal aberrations in cultured lymphocytes obtained from these workers (Purchase et al., 1975). Exposure levels were not measured. Workers from both the test and .control groups who had been exposed to X-rays or had had prolonged drug treatment or recent viral infections were excluded from the study. However, the results from this study and their significance were not discussed (Purchase et al., 1975). Kucerova and colleagues (1979) found that the frequency of SCE and other chromosomal aberrations was significantly higher in workers exposed to 20-150 ppm vinyl chloride in air than in unexposed controls matched for sex and age. Chromatid and chromosome breaks were 6-10 CMA 010395 draft detected in the greatest frequency; chromatid and chromosome exchanges occurred only sporadically. Some subsequent studies have verified these findings. The majority suggest that the frequency of occurrence of aberrations decreased with decreasing occupational exposure levels. For example, polyvinyl chloride workers (N - 52) exposed to mean concentrations of 2.34 ppm vinyl chloride had significantly greater numbers of chromosome breaks and chromosomal aberrations than did unexposed controls (N - 74) (Suskov and Sazonova, 1982). However, in another study, workers exposed to low levels of vinyl chloride showed no differences from controls in the number of SCE or chromosome breaks. Significant differences had been seen in the same population previously when occupational vinyl chloride exposures had been higher (Hansteen et al., 1978). A study of a large number of polyvinyl chloride workers suggested that vinyl chloride exposures below 15 ppm did not induce chromosomal aberrations (Ficciano et al., 1977). When lymphocyte cultures from a group of 109 workers who had worked in the plant (exposure periods ranged from one to 332 months) were compared with cultures from a control group of 295 pre-employment examinees, no significant chromosomal differences were observed. The workers had been exposed to levels of 15.2 ppm vinyl chloride before 1960, 11.4 ppm from 1960 to 1972, and 8.7 ppm between 1973 and 1974. The subjects and controls were not matched for age or for exposure to X-rays, however. 6-11 0*'6 cA DRAFT Cytogenetic studies performed on lymphocytes isolated from 39 workers from a polyvinyl chloride plant and 16 control males demonstrated a significant increase in chromosome-breakage frequency for the exposed workers (3.41% versus 1.79%, respectively). This study was repeated for 37 of the 39 workers 2-2.5 years later, during which time the workers had only a minimal exposure to vinyl chloride. More appropriate in-plant matched controls were selected for the follow-up study. In the repeat study no difference was found in mean chromosome-breakage frequency between the workers and their controls (Hansteen et al., 1978). 6.3.3 Micronucleus Tests .. In CBA male mice exposed to 5% vinyl chloride In air, nearly a four-fold increase in micronucleated cells was observed (Jenssen and Ramel, 1980). 6-3.4 DNA Damage/Unscheduled DNA Synthesis (UPS) Tests Vinyl chloride has been reported to induce unscheduled DNA synthesis in adult rat hepatocytes, but no experimental details were provided in the publication (Probst et al., 1981). Differential killing was induced in the repair-deficient . coll strain polA in assays using the standard disc and liquid suspension methods (Rosenkranz, 1981). 6-12 CMA 010397 6.4 Mammalian Cell Transformation DRAFT Vinyl' chloride, 20 to 50% in air, has been reported to transform BHK cells exposed (Styles, 1980). A clear positive transformation response was obtained in BALB/C-3T3 mouse cells exposed to vinyl chloride; in addition, vinyl chloride (chamber concentrations 01024 ppm) caused a dose-dependent cytotoxicity (Tu et al., 1985). An increased sensitivity to transformation by SA-7 virus was observed in primary Syrian hamster embryo (SHE) fibroblasts exposed 3 to vinyl chloride concentrations up to 194 /jg/cm (75,781 ppm) (Hatch et al., 1981). 6-13 CMA 010398 DRAFT* CMA 010399 CARCINOGENICITY Animal Studies DRAFT Summary The evidence for the carcinogenicity of vinyl chloride in laboratory animals has been reviewed by Kalmaz and Kalmaz, 1984, IARC, 1979, SRI, 1983, Kuzmack and McGaughy, 1975, and Purchase et al., 1987. Adequate experimental evidence exists to indicate that vinyl chloride is carcinogenic in mice, rats, and hamsters when given orally and by inhalation. Vinyl chloride has been found to cause tumors in a dose-related manner at several sites, including liver, lung and mammary gland. The oncogenic response appears to be a function of the site, vinyl chloride concentration, tumor type, species of animal, and route of administration. Although some evidence of vinyl chloride-induced carcinogenesis has been observed by all routes of administration and in all species tested, important discrepancies in the protocols of many studies has limited their usefulness in quantitative risk assessment. These discrepancies include the lack of appropriate control groups, insufficient exposure time, or incomplete histopathology of the animals. Studies that have been used previously in risk assessment include feeding studies (Feron et al., 1981; Til et al., 1983) and a series of inhalation studies (Maltoni et al., 1984). In the Feron studies, liver angiosarcomas and hepatocellular tumors (the 7-1 CMA 010400 DRAFT primary sice) were produced after chronic oral administration of vinyl chloride. In the studies by Maltoni et al. (1984) a wider variety of tumor types was observed. These studies and others are reviewed below. 7.1.2 Intraperitoneal. Subcutaneous, and Transplacental Vinyl chloride has been tested in experimental animals by intraperitoneal, subcutaneous, and transplacental administration, but for various reasons all of these studies were deemed inadequate for the evaluation of the carcinogenic risk of vinyl chloride. These reports and the reasons for their inadequacy are described in Appendix A. 7.1.3 Oral Administration 7.1.3.1 Studies Jbv Maltoni and Associates Rats: Liver angiosarcomas and morphologic alterations of the liver were induced in groups of 40 male and 40 female Sprague-Dawley rats after gastric intubation of 0, 3.33, 16.65, or 50 mg/kg vinyl chloride in olive oil five days/week for 52 weeks. These animals were then observed for the remainder of their lives (Experiment BT11, Maltoni et al., 1984, IARC, 1979). Dose-related increases in the incidence of several types of tumors were observed, including liver angiomas and angiosarcomas, nephroblastomas, and mammary 7-2 CMA 010401 tumors. DRAFT In a subsequent experiment, 0, 0.03, 0.3, or 1.0- mg/kg was administered by the same protocol, except that the dose groups contained 75 animals of each sex. Liver angiosarcomas were found in one female in 0.3 mg/kg group and two females and one male in the 1.0 mg/kg group. No such tumors were observed in controls (Experiment BX27, Maltoni et al., 1984). No statistical analyses were reported for any of these experiments. .1.3.2 Studies by Feron and Associates Vinyl chloride in soybean oil was administered by gastric intubation at a dose of 300 mg/kg once daily, five days/week for 83 ~ weeks, to 60 male and 60 female Wistar rats; no vehicle controls were used. Of the 109 animals examined, 56 had angiosarcomas of the liver and 52 had angiosarcomas of the lung (Feron et al., 1981). Although vinyl chloride was clearly demonstrated to be carcinogenic in this study, the data are not suitable for use in quantitative risk assessment because of the lack of vehicle-treated controls. In conjunction with the above experiment, groups of 60-80 male and 60-80 female five-week old Wistar rats were fed polyvinyl chloride powder (10% of diet) with or without a high vinyl chloride monomer content (0 to 4000 ppm) in the diet for their lifetimes (Feron et al., 1981). The actual doses of vinyl chloride given to rats in the feed were 0, 1.7, 5.0, and 14.1 mg/kg/day. Access to food for controls and treated animals was limited to four hours per day; an 7-3 CMA 010402 DRAFT additional control group was fed ad libitum. Gross pathology was ^ performed on all animals that died or were killed; complete histopathology of all organs was performed on only 20 males and 20 females from the controls and 20 males and 20 females from each of the two highest dosage groups. The animals chosen for complete histopathology were those that lived the longest before being killed. Histopathology of all other rats was restricted to the liver, zymbal glands, lungs, kidneys, spleen, pituitary, thyroid, adrenals, grossly visible tumors, and organs containing lesions suspected of bearing tumors. Statistical significance of tumor incidence was determined by the Chi-square test. Vinyl chloride caused a dose-related increase in the death rate in the 5.0- and 14,1-mg/kg groups; all animals receiving the highest dose were dead by week 134, with females dying earlier than males (Feron et al., 1981). In the low-dose group the mortality of male rats was comparable with that of controls; the death rate in female rats was slightly higher than chat in controls. Death of created animals was attributed to pulmonary or hepatic insufficiency due to neoplastic or nonneoplastic lesions in these organs. Liver angiosarcomas were reported In 27/59 (p < 0.001) and hepatocellular carcinomas in 8/59 (p < 0.01) male rats receiving 14.1 mg/kg/day. Incidences of angiosarcomas and hepatocellular carcinomas were 9/59 (p < 0.01) and 29/59 (p < 0.001), respectively, in females receiving the highest dose (Table 7.1) (Feron et al. 1981). Necrosis, centrllobular degeneration and 7-4 CMA 010403 DRAFT mitochondrial damage were also seen in the hepatic parenchyma of rats administered vinyl chloride. The incidence of angiosarcoma of the lung was also significantly increased in high-dose males (19/59, p < 0.001) and females (5/57, p < 0.05) (Table 7.2). Lowdose males and females showed necrotic damage of the liver and 26/58 low-dose females (p < 0.01) had neoplastic nodules of the liver (Table 7.1) (Feron et al., 1981). It is possible that underreporting of tumors at all sites occurred because of the incomplete histopathology performed and the fact that only the longest-surviving high-dose animals were chosen for complete histopathology. 7.1.3.3 Studies by Til and Associates As a follow-up to the study of Feron and co-workers (1981), groups of 100 male and 100 female Vistar rats (except for the top-dose group, which was composed of 50 animals of each sex) were fed polyvinyl chloride (up to 1% of diet) with a high content of vinyl chloride monomer for up to 149 weeks (Til et al., 1983). Levels of vinyl chloride administered in the powder were 0, 0.017, 0.17, and 1.7 mg/kg/day for 149 weeks. Actual oral exposure to vinyl chloride monomer (calculated by measuring the evaporative loss of vinyl chloride during the four-hour feeding periods, the rate of food intake, and the level of vinyl chloride in the feces) was estimated to be 0.014, 0.13, or 1.3 mg vinyl chloride/kg/day for the low, middle, and high dose groups, respectively. Access to food was limited to four hours per day. An additional control 7-5 CMA 010404 TABLE 7.1 DRAFT INCIDENCE OF LIVER TUMORS AND NEOPLASTIC NODULES IN VISTAR RATS EXPOSED ORALLY TO VINYL CHLORIDE (Feron et al., 1981) ,lumoy Typs/Ssy,_______ Liver Angiosarcoma Male Female Hepatocellular Carcinoma Male Female Neoplastic Nodules Male Female --------------------------------- Incidence_________________ ___________ Vlnvl Chloride (mg/kg/dav) _Q 1.7 5.0 14.1 0/55 0/57 0/58 0/58 6/56** 2/59 27/59***' 9/57** 0/55 0/57 1/58 4/58 2/56 19/59*** 8/59** 29/57*** 0/55 2/57 1/58 26/58** 7/56** 39/59*** 23/59*** 44/57*** Number in denominator - number of animals necropsied. Values marked with asterisks differ significantly from controls according to the Chi-square test: * p < 0.05 ** p < 0.01 *** p < 0.001 7-6 CMA 010405 TABLE 7.2 DRAFT INCIDENCE OF LUNG ANGIOSARCOMAS. ABDOMINAL MESOTHELIOMAS AND MAMMARY TUMORS IN VISTAR RATS EXPOSED ORALLY TO VINYL CHLORIDE (Feron et al.. 1981) Tumor Tvoe/Sex Lung Angiosarcoma Male Female Abdominal Mesotheliomas Male Female Mammary Adenoma or Adenocarcinoma or Anaplastic carcinoma Female Incidence1 Vinvl Chloride (mz/kz/dav) 0. 1.7 14.1 0/55 0/57 0/58 0/58 A/56*2 1/59 19/59*** 5/57* 3/55 1/57 1/58 6/58* 7/56 3/59 8/59 3/57 3/57 2/58 5/59 9/57 Number in denominator - number of animals necropsied. Values marked with asterisks differ significantly from controls according to the Chi-square test: * p < 0.05 ** p < 0.01 *** p < 0.001 OlOAO* 7-7 group, comprised of 100 rats of each sex, received food ad 1id 1 turn and were housed in a separate room. Gross pathology was performed on all animals and was restricted to the liver, all grossly visible tumors or presumable tumors in the abdominal cavity, zvmbal gland, and mammary glands. No clinical signs of toxicity attributable to vinyl chloride were observed. In the lowest- and mid-dose group, body weight and survival of treated rats were not significantly different from those of controls. In the high-dose group, mortality was slightly increased. The results of this study demonstrated significant increases in the incidences of hepatic foci of cellular alteration, neoplastic nodules, hepatocellular carcinomas, liver-cell polymorphism, and cysts in the highest dose group. Two females and one male in this group developed liver angiosarcomas. Females, but not males, of the low- and mid-dose groups developed a higher incidence of hepatic basophilic foci of cellular alteration. No pathologic effects in other organ systems were attributed to vinyl chloride exposure (Table 7.3) (Til et al., 1983). Til and co-workers reported that a threshold of 0.17 mg vinyl chloride/kg/day for the induction of tumors in rats was observed. In fact, a threshold cannot be demonstrated. Vinyl chloride induced hepatocellular alterations at all concentrations tested. Histopathology of all organs was not performed on all animals; therefore, tumors not grossly observable or palpable could have CMA 010407 7-8 been hissed. Because of the shortcomings of the study, its utilitv for the evaluation of carcinogenic risk is limited. . 1. -+ Inhalation Exposure Several researchers have investigated the potential carcinogenicity of vinyl chloride administered by inhalation (Viola, 1977; Caputo et al., 1974; Keplinger et al., 1975; Lee et al., 1977; Hong et al., 1981; Suzuki, 1981; Groth et al., 1981; Drew et al., 1983; Maltoni et al., 1984). All experiments confirm the carcinogenicity of vinyl chloride, although only a few of the studies are adequate for a quantitative evaluation of carcinogenic risk. 7.1.4.1 Studies in Rats The earliest information on the experimental carcinogenicity of vinyl chloride administered by inhalation was reported by Viola (1971). Wistar rats were exposed to 30,000 ppm by inhalation (four hours/day, five days/week) for twelve months. At the end of the treatment period, the surviving animals were killed at 20-day intervals and "the most important tissues and organs examined histologically by standard methods". The primary tumors observed were located in the zymbal gland (found only in rodents), with metastases to the skin, bone, and lung. Caputo and associates (1974) exposed Wistar rats to 50-20,000 ppm vinyl chloride four hours/day, five days/week for 12 months. Liver 7-9 CMA 010408 TABLE 7.3 DRAFT LIVER TUMOR INCIDENCE IN MALE AND FEMALE VISTAR RATS EXPOSED TO VINYL CHLORIDE BY ORAL ADMINISTRATION FOR 149 WEEKS (Til et al.t 1983) Tumor Tvoe/Sex Liver Angiosarcoma Male Female Hepatocellular Carcinoma Male Female Neoplastic Nodules Male Female J2__ MrUVg,_________ Vinvl Chloride (me/ke/dav) 0.014 0J.3, 13 0/99 0/98 0/99 0/99 0/99 0/96 1/49 2/49 0/99 1/98 0/99 0/99 0/99 1/96 3/49 3/49 0/99 0/99 0/99 1/99 0/99 0/99 1/49 9/49 ^Number in denominator - number of animals necropsied. Vinyl chloride intake data was adjusted to compensate for loss of vinyl chloride during the four-hour feeding periods. The initial levels of vinyl chloride administered in the diet were 0, 0.017, 0.17, and 1.7 mg/kg/day. 7-10 CMA 0*0409 DRAFT angiosarcomas and skin carcinomas were observed in animals exposed . zo 500 ppm or greater and lung adenomas in those exposed to 2.000 ppm or more. Bi et al. (1985) evaluated the tumorigenic potential of vinyl chloride in rats following inhalation exposure to 0, 10, 100 or 3000 ppm (six hours/dav, six days/week). The incidence of liver angiosarcomas was 0/19, 0/20, 7/19 and 17/19 for the four exposure groups, and 0/19, 0/20, 2/19 and 9/20 for lung angiosarcomas, respectively. The authors failed to discuss the specific types of tumors or their significance, focusing instead on the testicular effects of vinyl chloride (discussed in Section 5 of this document) _ (Bi et al., 1985). .1.4.2 Studies in Mice In a preliminary paper reviewed by IARC (1979), Keplinger and co workers (1975) reported results from ongoing tests on mice, rats, and hamsters. Vinyl chloride was carcinogenic in all three species; the female mouse was the most sensitive of the animals tested-. CD1 Swiss mice were exposed to 0, 50, 200, or 2,500 ppm vinyl chloride seven hours/day, five days/week for nine months, then observed for another nine months. Primary tumors found in animals that died included liver angiosarcomas, lung adenomas, and mammary adenocarcinomas. At the time of the IARC report, histological evaluation had been carried out only on grossly visible tumors, but no final report has been published. 7-11 CMA 010410 DRAFT Consequently, we cannot accurately quantify tumor incidence in the s tudy, Lee and co-workers (Lee et al. , 1977; IAB.C, 1979) reported that female mice were more responsive to vinyl chloride exposure chan rats. Two month-old male and female CD-I mice were exposed by inhalation to 0, 50, 250, or 1,000 ppm vinyl chloride for six hours/day, five days/week for 52 weeks (end of experiment). Vinyl chloride induced primary tumors in mice at multiple sites after exposure to 50 ppm or more. Liver cell angiosarcomas, bronchiolo alveolar adenomas, mammary ductular adenocarcinomas, and squamous and anaplastic cell carcinomas (with meeastases to the lung) were observed in treated animals. Vinyl chloride induced tumors at all dose levels, with the incidence and severity of the tumors increasing with dose. The total tumor incidence may have been underestimated because of the short duration of the study. Hong and colleagues (Hong et al., 1981), as a follow-up of the studies of Lee and associates (Lee et al., 1977), examined the development and incidence of vinyl chloride-related carcinogenic effects during a post-exposure follow-up period. Groups of eight to 28 two month-old male and female CD-I mice were exposed to 0, 50, 250, or 1,000 ppm for one, three or six months and subsequently observed for 12 months before being sacrificed. Although the number of animals used in the experiment was inadequate for risk assessment purposes, four of sixteen female mice exposed to 50 ppm vinyl chloride for one month (and autopsied 7-12 CMA 010411 DRAFT one vear later) exhibited mammary gland adenocarcinomas or carcinomas. In mice, the combined (male and female) incidences of hemangiosarcomas for the 250 and 1,000 ppm groups were significantly higher than in controls (p - 0.05). Tumor incidence was related to dose and duration of exposure. Bronchiolo-alveolar tumors were also significantly increased in the high-dose group (p - 0.05), but no clear trend for the other dose levels was observed (Hong et al., 1981). In rats, tumor incidence rates following exposure for one or three months did not differ significantly from control values. After a six or ten month exposure, the combined (male and female) - cumulative incidences of hemangiosarcomas, hepatocellular carcinomas, and neoplastic liver nodules in rats exposed to 250 or 1,000 ppm differed significantly from those in combined male and female control animals (statistics not reported) (Hong et al., 1981). Suzuki (1981a) exposed male CD-I mice (between 30 and 40 per group) to 1, 10, 100, 300, or 600 ppm vinyl chloride six hours/day, five days/week for four weeks. The animals were then observed for up to 41 weeks after cessation of exposures. One mouse in the 10 ppm group had a subcutaneous hemangiosarcoma in the left ear 29 weeks after exposure; one mouse in the 600 ppm group developed a hepatic hemangiosarcoma 65 weeks after exposure. In a separate study, Suzuki (1981b) exposed 27 mice to either 2500 or 6000 ppm vinyl chloride for five or six months. Additional mice were exposed to 7-13 CMA 010412 DRAFT 1. 10 or 100 ppm for four weeks, and sacrificed forty weeks after exposure. All animals were evaluated for pulmonary tumors. Twenty-six of the 27 high dose animals possessed "alveologenic" tumors. Animals in the lower dose groups exhibited a dose-related trend for pulmonary tumor formation (Suzuki, 1981b). Although this study cannot be used to quantify risk due to study design (for example, inadequate number of test animals), it did demonstrate a carcinogenic response to vinyl chloride after exposure to relatively low concentrations for short durations. 7.1.4.3 Studies on the Potential Effects of Age at Time of SxBpsure ., Groch et al, (1981) exposed groups of 110-128 male and female Sprague-Dawley rats to 948 ppm vinyl chloride in air seven hours/day, five days/week for 29 weeks, beginning at ages varying from six weeks to 52 weeks. Animals were sacrificed after termination of exposure. On the basis of this testing regime, those researchers concluded that vinyl chloride-induced liver angiosarcomas occurred with the greatest frequency in rats whose exposure period began at 52 weeks of age, with females more susceptible than males. The data and study methodology are inadequate for making this conclusion, however. If liver angiosarcomas are expressed at a later age in the rat's life cycle, animals exposed at an early age and sacrificed early in their life cycles would not have had time to express the same tumor incidence as they would if they had lived their full lifetimes. The animals 7-14 CHA 010-413 DRAFT exposed later in their life cycles would then seem to have the highest tumor incidence. Drew et al., (1983) looked at the effect of age and exposure duration on vinyl chloride oncogenicity in females of several different species of rodents. Groups of female CD-I Swiss mice, B6C3F1 mice, Fischer 344 rats, and Golden Syrian hamsters (N - 34 for mice, N - 56 for rats and hamsters) were exposed to vinyl chloride for six hours/day, five days/week for six, 12, 18, or 24 months, beginning at eight weeks of age, and observed for their lifespans. Other groups were held until six or 12 months of age, exposed for six or 12 months, and then observed for the remainder of their lifespans. The exposures were conducted at a single dose level for each species; mice, rats and hamsters were administered 50, 100, and 200 ppm* respectively. All animals exposed to vinyl chloride at age eight weeks (the start of the experiment) exhibited decreased survival relative to controls (Drew et al., 1983). B6C3F1 mice experienced the most significant life-shortening regardless of the age at which exposure was begun. No significant decrease in survival was observed in rats, hamsters, or Swiss mice initially exposed after six months of age. Other clinical signs of vinyl chloride toxicity were not evident and liver necrosis was not observed. In rats, exposure to vinyl chloride was associated with hemangiosarcomas, mammary gland adenocarcinomas and adenomas, and hepatocellular carcinomas (Table 7.4) (Drew et al., 1983). The 7-15 CMA 010414 draft. incidence of hemangiosarcomas was a function of the duration of * exposure: the longer the exposure period the greater the incidence of hemangiosarcomas. A six-month exposure produced a low incidence of hemangiosarcomas and hepatocellular carcinomas only if begun early in life. One-year exposures produced a significant incidence of tumors, especially if begun early in life. The incidence of mammary gland adenocarcinomas and fibroadenomas was not always related to exposure duration, but the incidence was higher in rats whose exposure began at eight weeks of age. Hepatocellular carcinomas were induced in a dose-related manner in rats when exposures began at eight weeks. In hamsters, hemangiosarcomas, mammary gland carcinomas, stomach adenomas, and skin carcinomas were associated with vinyl chloride exposure (Table 7.4) (Drew at al. 1983). The highest incidence of hemangiosarcomas and stomach adenomas occurred In animals exposed early in life for only six months. The highest incidence of mammary gland carcinomas was seen in animals exposed at an early age for up to twelve months. Exposure beginning at or after eight months of ' age resulted In a markedly lower tumor incidence, possibly because the lifespans of chronically exposed hamsters were significantly reduced to the point that late-appearing tumors would not be expressed. Mice, especially the B6C3F1 strain, appeared to be the species most sensitive to the carcinogenic effects of vinyl chloride (Table 7.4) (Drew et al., 1983). Hemangiosarcomas and mammary gland carcinomas 7-16 CMA 010415 DRAiFT in boch strains and lung carcinomas in Swiss mice were associated with vinyl chloride exposure. In B6C3F1 mice, exposure to vinvl chloride for six months resulted in 60-70% incidence of hemangiosareomas, regardless of the age at exposure initiation. The incidence of mammary gland carcinomas in B6C3F1 mice was greatest when the animals were exposed early in life. Lower incidences of this tumor were seen when initial exposure occurred at a later age. In Swiss mice, exposure to vinyl chloride at an early age resulted in the highest incidence of hemangiosareomas, mammary gland carcinomas, and lung carcinomas, regardless of duration of exposure. Lower incidences of all tumors were observed in animals exposed later in life. The patterns of tumorigenicity produced by vinyl chloride in the study by Drew et al. (1983) are consistent with patterns reported in other inhalation studies. However, the results reported by these investigators apparently contradict those of Groth et al. (1981). This apparent contradiction can be explained by the fact that Groth et al. reported only the incidence of hemangiosareomas, a tumor shown in the Drew study to be a relatively late'appearing tumor that developed regardless of either the age at initial exposure or the duration of exposure. In the Groth et al. study, animals exposed at a young age were also sacrificed at a young age, thereby decreasing the probability of hemangiosarcoma development relative to the older exposed animals who were allowed to live. 7-17 CM6 010416 TABLE 7.4 TUMOR INCIDENCE FOLLOWING VINYL CHLORIDE EXPOSURE IN FEMALE RATS, HAMSTERS AND MICE FROM THE STUDY OF DREW ET AL. (1983) Tumor Type Length of Exposure (Months) LDE (ppm)^- Female Fisher 344 Rat: Experimental Exposure 100 ppm Liver Hemangiosarcomas control 6 12 18 24 0 4.46 8.93 13.40 17.86 Mammary Gland Adenocarcinoma control 6 12 18 24 0 4.46 8.93 13.40 17.86 Hepatocellular Carcinoma control 6 12 18 29 0 4.46 8.93 13.40 17.86 Female B6C3F1 Mice: Experimental Exposure 50 ppm Hemangiosarcoma (all sites) control 6 12 18 0 2.23 4.46 '- Mammary Gland Carcinoma control 6 12 18 0 2.23 4.46 Tumor Frequency (%) 0.9 (1/112) 5.3 (4/76) 20.0 (11/55) 23.6 (13/55) 34.7 (19/55) 4.5 (5/112) 7.9 (6/76) 19.6 (11/56) 16.4 (9/55) 9.1 (5/55) 0.9 (1/112) 4.0 (3/75) 7.1 (4/56) 14.8 (8/54) 16.4 (9/55) '- 5.8 (4/69) 68,7 (46/67) 76.7 (69/90) *- 4.3 (3/69) 43.2 (29/67) 41.1 (37/90) ** 7-18 01041? .A3LE " . - continued DRAFT Tu.nor Type Length of Exposure (Months) LDE (ppm)^ Tumor1 Frequency (%) Female CD-I Swiss Mice: Exoerimental Exoosure 50 Dom Hemangiosarcoma (all sites) control 6 12 18 0 2.23 4.46 6.69 1.4 (1/71) 43.3 (29/67) 63.8 (30/47) 44.4 (20/45) Mammary Gland Carcinoma control 6 12 18 0 2.23 4,46 6.69 2.8 49.3 46.8 48.9 (2/71) (33/67) (22/47) (22/45) Lung Carcinoma control 6 12 18 0 2.23 4.46 6.69 12.7 27.7 31.9 24.4 (9/71) (18/65) (15/47) (11/45) Female Golden Svrian Hamster: Exoerimental Exposure 200 ppm Hemangiosarcoma (all sites) Mammary Gland Carcinoma control 6 12 18 0 6 12 18 0 8.93 17.86 26.79 0 8.93 17.86 26.79 0.0 14.8 7.7 1.9 (0/143) (13/88) (V52) (2/103) 0.0 (0/143) 32.2 (28/87) 59.6 (31/52) 46.1 (47/102) Skin Carcinoma 0 6 12 18 0 8.93 17.86 26.79 0 2.5 18.8 3.3 (0/133) (2/80) (9/47) (3/90) Hj)E - Lifetime Daily Exposure (in ppm) 7-19 CMA 010418 Studies bv Maltoni and Associates draft Maltoni and co-workers performed a series of chronic inhalation studies on rats, mice, and hamsters in the Bentivoglio Laboratories (BT) or the Bologna Institute of Oncology (Maltoni et al.t 1984). The investigators studied the effects of exposure to 14 concentrations of vinyl chloride (1-30,000 ppm) in male and female rats and six concentrations of vinyl chloride in male and female mice and male hamsters. A summary of some of these experiments are included both in this section and in Appendix A. In each experiment, animals were exposed to vinyl chloride for four hours daily, five days per week for various durations, and observed for the rest of their lives. A number of the experimental procedures were not described or were inadequately described in the report by Maltoni et al. (1984). A full necropsy was performed on each animal and the following tissues reportedly were routinely excised for histopathology: brain, zymbal glands, interscapular brown fat, salivary glands, tongue, thymus, lungs, liver, kidneys, adrenal glands, spleen, pancreas, esophagus, stomach, intestine, bladder, uterus, gonads, and any organ in which pathologic lesions were observed. Details of the experimental protocol for the BT experiments are provided in Table 7.5 (Maltoni et al., 1984). Data on noncarcinogenic toxic effects of vinyl chloride were sparsely reported in the Maltoni BT experiments. Vinyl chloride appeared to be toxic at the higher concentrations, but reportedly the high mortality at these dose levels was due to a high incidence 7-20 CHA 10419 DRAFT of vinyl chloride - induced tumors. The available information or. survival, including Kaplan-Meier survival curves, indicates that vinyl chloride decreased survival in a dose-dependent manner. In the Maltoni experiments, exposure to vinyl chloride was associated with an increased incidence of malignant tumors at a variety of tissue sites in all of the species tested. A summary of these tumor sites is provided in Table 7.6 (Maltoni et al., 1984). A direct relationship between exposure levels and tumor incidence was apparently demonstrated, although no statistical tests for trends were performed. Results of experiments on Sprague-Dawley rats exposed to vinyl chloride for 52 weeks were statistically analyzed using the Fischer exact probability test. Correspondence analysis was also performed on the relationship of the incidence of liver angiosarcomas, zymbal gland carcinomas, nephroblastomas, and forestomach papillomas and acanthomas to vinyl chloride exposure (Tassignon, 1980, cited in Maltoni et al., 1984), The results of this analysis were not discussed by Maltoni et al. (1984). A summary of the lowest concentrations at which a statistically significant excess of tumors was observed is given in Table 7.7. Experiment BT1. Most previous risk assessments have been based on the data from experiment BTl (Maltoni et al., 1984). In this study, 30 Sprague-Dawley rats of each sex were exposed to concentrations of vinyl chloride ranging from 50 to 10,000 ppm for four hours daily, five days per week for 52 weeks, beginning at 13 weeks of age. A positive control group received 2,500 ppm of vinyl 7-21 CMA 010420 acetate. DRAFT After treatment the animals were observed for their lifespans up to 135 weeks. Survival of both males and females decreased in a dose-related manner, especially at concentrations above 500 ppm. Vinyl chloride appeared more toxic to females than to males in this experiment. Vinyl chloride was associated with an increased incidence of liver angiosarcomas in a dose-related fashion. These results are presented in Table 7.8 (Maltoni et al., 1984). In addition to liver angiosarcomas, vinyl chloride (at concentrations above 2500 ppm) caused an increased incidence of zymbal gland carcinomas, nephroblastomas, hepatomas, and neuroblastomas. The incidence of liver angiosarcomas was probably underestimated at the higher exposure levels due to mortality resulting from tumors at other sites. Experiment BT15. Groups of 60 male and 60 female Sprague-Dawley rats were exposed to 0, 1, 5, 10, or 25 ppm of vinyl chloride for four hours daily, five days per week for 52 weeks, beginning at 13 weeks of age (Maltoni et al., 1984). Following exposure the animals were observed for the remainder of their lives (up to 147 weeks). Available data, including Kaplan-Meier survival curves, indicated that vinyl chloride did not affect survival at the concentrations tested. No statistical analyses of mortality and body weight data were reported. Mortality was greater in the male concrol group than in the treated groups: the time at which 50% of the male concrol group had died was week 72, compared with week 100 in the 25-ppm vinyl chloride group. No explanation was given for this decreased survival. 7-22 CHft 010421 Ixper iment: Number TABLE 7.5 DRAFT EXPERIMENTAL PROTOCOL FOR INHALATION STUDIES MALTONI AND CO-WORKERS (1984) Dose (DDm) Exposure Duration (weeks) Species/ Strain Age at Start of Exposure (weeks) Number of Animals pe^ Dose Level, BTl 3T2 BT6 BT9 BT15 BT3 BT14 BT4001 BT4006 BT5 BT7 BT17 0, 50, 250, 500, 2,500, 6,000, 10,000 1, 100, 150, 200 30,000 0, 50 o, 1, 5, 10, 25 0, 50, 250, 500, 2,500, 6,000, 10,000 6,000, 10,000 0, 2,500 0, 2,500 6,000, 10,000 52 52 52 52 52 17 5 5 76 69 15 1 0, 50, 250, 500, 2,500, 6,000, 10,000 0, 1 52 52 Rat/SD 13 30 M, 30 F (30 M, 30 F) Rat/SD Rat/SD Rat/SD Rat/SD Rat/SD 13 60 M, 60 F (85 M, 100 F) 17 30 M, 30 F (no controls) 13 150 M, 150 F (50 M, 50 F) 13 60 M, 60 F (60 M, 60 F) 12 30 M, 30 F (30 M, 30 F) Rat/SD 21 (parents) 1 day (offspring) Rat/SD 13 1 day Rat/SD 1 day Rat/SD 19 (fetus) Rat/Wistar 11 6F (no controls) 21-22 M, F (no controls) 54 F (60 F) 68 M, 64 F (158 M, 149 F) 60 M, 60 F (60 M, 60 F) 30 F 13-29 M, F (no controls) 30 M (40 M) Rat/Wistar 13 120 M (130 M) 7- CMA 010422 Table ".5 continued DRAFT Experiment Number Dose 'tom) BT4 0, 50, 250, 500, 2,500, 6,000, 10,000 BT8 0, 50, 250 500, 2,000, 6,000, 10,000 Exposure Duration ("weeks')1 30 Species/ Strain Age at Start of Exposure <weeks 1 Number of Animals per Mouse/Swis s 11 30 M, 30 F (80 M, 70 F) 30 Hamster/ 11 30 M Syrian golden (62 M) ^"Exposures were for four-hours daily, fivei days per week. 2 "Number in parentheses - number of control animals for experiment. 7*24 CMA 010423 DRAFT TABLE 7.6 TUMORS CORRELATED TO INHALATION EXPOSURE TO VINYL CHLORIDE IN RATS, MICE, AND HAMSTERS IN THE BT EXPERIMENTS1 Tumors Liver angiosarcomas Hepatomas Encephalic neuroblastomas Lung adenomas Lymphomas/leukemias Angiosarcomas at other sites Zymbal gland epithelial tumors Nephroblastomas Cutaneous epithelial tumors Mammary adenocarcinomas Forestomach papillomas, acanthomas Rat Mouse Hamster ++ + + (+) + + ++ + + (+) (+) (+) (+) ++ + (+) (+) + ^Data from Maltoni et al., 1984 + - Tumor incidence was statistically significant (P < 0.05) by the Fisher exact test. (+) - Association was not statistically significant, but was considered biologically significant. 7-25 CM* 010^24 TABLE 7 . 7 DRAFT# LOWEST CONCENTRATION AT WHICH A SIGNIFICANT (p < 0.05) EXCESS OF TUMORS WAS REPORTED BY MALTONI AND ASSOCIATES1 IN INHALATION STUDIES AT SPECIFIC SITES IN SPRAGUE-DAWLEY RATS2 T-Umor Forestomach papilloma Zymbal gland carcinoma Neuroblastoma Nephroblastoma Liver angiosarcoma Mammary adenocarcinoma Vinyl Chloride Concentration (coni') 30.000 (male, female) 10.000 (male, female) 10,000 (female) 250 (female) 100 (male) 200 (male) 25 (female)2 5 (female) 1Data are from Maltoni et al., 1984. 2 Significant at this dose level when specific corrected tumor incidence is used, p - 0,047. Analysis by Fisher exact probability test. 7-26 CMA 010425 TABLE 7.8 DRAFT ' INCIDENCE OF LIVER ANGIOSARCOMAS (LAS) IN MALE AND FEMALE SPRAGUE-DAWLEY RATS EXPOSED FOR 52 WEEKS TO VINYL CHLORIDE (Maltoni et al., 1984) S tudv BT1 BT2 Experimental Dose Level (com) 0 50 250 500 2,500 6,000 10,000 0 100 150 200 LAS Incidence Mai? Female 0/30 0/30 1/30 0/30 6/30 3/30 3/30 0/30 1/30 2/30 6/30 7/30 10/30 4/30 0/85 0/60 1/60 7/60 0/100 1/60 5/60 5/60 BT6 30,000 5/30 13/30 BT9 BT15 0 0/50 0/50 50 1/150 12/150 0 0/60 0/60 1 0/60 0/60 5 0/60 0/60 10 0/60 1/60 25 1/60 4/60 LAS Incidence in Historical Controls: 1/1179 2/1202 Corrected ,, LAS Incidence" Male Female 0/22 0/26 1/28 0/22 6/26 3/17 3/21 0/61 0/37 1/36 7/42 0/29 1/29 2/26 6/28 7/24 10/25 4/25 0/68 1/43 . 5/46 5/44 5/22 13/24 0/29 2/70 0/25 0/48 0/43 0/42 1/41 0/38 12/110 0/44 0/55 0/47 1/46 4/40 1/364 2/541 Number in denominator * number of animals necropsied. Number in denominator * number of animals alive when first liver angiosarcoma was observed. 7-27 CMA 010426 DRAF The incidence of mammary gland carcinomas in created females was higher chan in controls at all concentrations of vinyl chloride exposure. However, the differences from control values were statistically significant only at concentrations above 1 ppm. The mammary gland adenocarcinoma incidence for this and the other relevant BT experiments are presented in Table 7.9 (Maltoni et al., 1984). Experiment BT4. Thirty male and 30 female Swiss mice were exposed to 0, 50, 250, 500, 2,500, 6,000, or 10,000 ppm of vinyl chloride four hours daily, five days weekly for 30 weeks, beginning at 11 weeks of age (Maltoni et al., 1984). The study was terminated 81 weeks after the exposure period began. Vinyl chloride was highly toxic to both males and females, but males appeared more sensitive than females to the toxic effects of vinyl chloride. Survival decreased in a dose-related manner, although statistical analysis apparently was not performed on the data presented. A very high incidence of lung adenomas was observed in vinyl chloride-treated male and female mice. A statistically significant increase in the incidence of liver angiosarcomas was seen in male and female mice exposed to vinyl chloride, but a dose response was not seen in the male animals. In addition, a high incidence of mammary gland adenocarcinomas occurred in treated female mice. These results are presented in Table 7.10 (data from Maltoni et al., 1984). 7-28 CMA 010*27 TABLE 7.9 INCIDENCE OF MAMMARY GLAND CARCINOMAS IN FEMALE SPRAGUE-DAWLEY RATS AND SWISS MICE EXPOSED BY INHALATION TO VINYL CHLORIDE (Maltoni et al., 1984) S cudv No. Experimental Dose Level (onm) ^ Tumor Incidence4" 3T1 (Rat) 0 50 250 500 2,500 6,000 10,000 0/30 2/30 2/30 1/30 2/30 0/30 3/30 BT2 (Rat) 0 100 150 200 2/60 4/60 6/60 5/60 BT6 (Rat) 30,000 2/30 BT9 (Rat) 0 9/50 50 59/150 BT15 (Rat) 0 6/60 1 14/60 5 22/60 10 21/60 25 16/60 Tumor Incidence in Historical Controls 100/1202 BT4 (Mice) 0 50 250 500 2,500 6,000 10,000 1/80 12/30 13/30 10/30 9/30 9/30 14/30 Tumor Incidence in Historical Controls 21/554 Corrected Tumor Incidence 0/29 2/30 2/27 1/28 2/25 0/28 3/29 2/100 4/60 6/60 5/60 2/30 9/43 59/142 6/60 14/60 22/60 21/60 16/60 100/1202 1/673 12/30:: 13/29^ 10/28^ 9/30^ 9/28^ 14/28"5 21/5543 Number in denominator - number of animals examined. Number in denominator - number of animals alive when first malignant mammary tumor was observed (type unspecified). Number in denominator - number of animals alive when first mammary tumor was observed (type unspecified). 7-29 CMA 010428 TABLE 7.10 INCIDENCE OF PULMONARY ADENOMAS, MAMMARY CARCINOMAS, AND LIVER ANGIOSARCOMAS IN MALE AND FEMALE SWISS NICE EXPOSED TO VINYL CHLORIDE BY INHALATION (EXPERIMENT BI4)1 ----------- Organ/Sex fylrnonary (lung) adenomas* Hates Females Liver anaiosarccams'* o 8/75 7/67 Males Fetaales Mammary adenocarcinomas* 0/62 0/62 Hates Feamles 0/74 1/67 50 3/27 3/30 1/18 0/26 0/27 12/30 Cancer Incidence Vinyl Chloride CocaI* 4 250 500 2.500 6.000 10.000 Historical Control 24/29 17/29 24/29 26/29 18/25 22/30 23/27 24/29 20/24 26/28 34/491 27/533 9/23 9/21 6/17 8/26 6/13 10/24 2/12 11/21 1/9 9/20 0/545 0/554 0/29 13/29 1/28 10/28 0/23 9/30 0/24 9/28 0/22 14/28 1/521 22/545 Vata fromHaltoni et al., 1984. Ikaber in denominator * number of animals alive when first pulmonary (lung) adenoma was observed (11 weeks). ^Number in denominator * nuaber of animals reportedly alive when first liver angiosarcoma was observed (32 weeks). 4 Nimber in denominator * nuaber of animals reportedly alive when first lung mammary tumor (type unspecified) was observed (16 weeks). DRAFT CMA 0 1 0 4 2 9 1 Human Studies or. the Carcinogenic Effects of Vir.vi Chloride .Z.i Introduction In 1974, Creech and Johnson described three cases of angiosarcoma of the liver (LAS) among workers at the B.F. Goodrich Tire and Rubber Co. in Louisville, Kentucky. Because LAS is a very rare cancer (20-25 cases per year in the United States), the clustering of three cases in one vinyl chloride polymerization facility indicated an abnormally high incidence of this cancer. Based on this report, as well as data indicating that vinyl chloride is carcinogenic in laboratory animals, multiple studies of workers exposed to this agent were conducted. By 1985, at least 17 epidemiologic studies relating vinyl chloride exposure to the incidence of various cancers had been completed. 7.2.2 General Design of Epidemiologic Studies Most of the epidemiologic studies have been retrospective cohort designs. Groups of workers in the vinyl chloride industry were selected by reviewing employment records. Few baseline data other than age, job classification, and length of employment were obtained. The concentrations of vinyl chloride to which these workers were exposed were generally not available, since ambient levels of vinyl chloride were not routinely measured before 1975. Almost all of the investigators estimated vinyl chloride exposure 7-31 CMA 010430 based on some combination of job classification and length of exposure. Exposure information included measured levels of vinyl chloride in only two studies (Ott et al., 1975; Buffler et al., 1979). In all reports, workers were traced to determine the number of deaths that had occurred in the defined cohort. The cause of death was based on information available from death certificates. In the studies from Sweden and Norway, national cancer registries also provided data to assess incidence of cancer (Byren et al., 1976; Heldaas et al., 1984). The expected numbers of deaths were estimated using population-based mortality statistics. Finally, a standardized, mortality ratio (SMR) was calculated from the proportion of observed to expected deaths from each cause and the statistical significance of these ratios tested. Difficulties In Interpreting the Epidemiologic Evidence There are several problems involved in the interpretation of these studies: 1. Inadequate information on worker outcome. In several of the studies reviewed, outcome data on approximately 10% of the original workers were not obtained (Duck et al., 1979). Since the tumor incidence in humans exposed to vinyl chloride is relatively low, the loss of 10% of the data base could have a 7-32 CMA 010431 DRAFT significant effect on the observed tumor rate, and possiblv allow for an underestimation of risk. 2. Inadequate exposure data. Specific exposure data did not exist in any of the studies reviewed with the exception of Ott et al. (1975) and Buffler et al. (1979). In some cases, no attempt was made to evaluate exposure. In most studies, exposure was estimated from odor levels, acute toxicity levels, job classification, or length of exposure - methods all considered unreliable for accurate exposure estimation. However, gross differences in exposure levels based on the type of job and length of exposure may have occurred, particularly before 1975, when very high levels of vinyl chloride were common in the industry (up to 500 ppm with rare excursions up to 4,000 ppm) (Ott et al. 1975). After 1975, ambient workplace levels were drastically reduced to an average of about 1 ppm, so that differences in dose estimated by job classification became small. 3. Inadequate follow-up time. In this review, "follow-up time" is the time period between the onset of exposure and the point at which evaluation is completed. The average latency period for LAS among vinyl chloride workers worldwide was determined to be 22.1 years (Stafford, 1983). Several of the epidemiologic studies either do not report length of follow-up or include a large number of subjects who had recently become employed in the vinyl chloride industry and who, therefore, had very short 7-33 OiOA^ draft follow-up time periods. Inclusion of subjects with shorter average follow-up time tends to obscure an exposure-response relationship. Use of death certificates to determine outcome. Cancer is often not listed as a cause of death on death certificates. Even when cancer is reported, the type of cancer may not be given and may not be verified. Healthy worker effect. In retrospective cohort studies of mortality among occupational groups, the number of deaths are compared to general population rates. A disproportionate number of newly hired employees demonstrating self-selection for good health, or selected for better general health through pre-employment examinations, skews the mortality experience of the worker population. Ott and co-workers (1975) have estimated that the average mortality rate for workers in the plastics industry is about 80% of population-based rates. Since observed deaths in the study cohorts are compared to estimates based on the incidence of death in the general population (not in healthy workers), the SMR becomes an underestimate of risk. Unfortunately, only one of the epidemiologic studies reviewed used a control group of workers not exposed to vinyl chloride (Theriault and Allard, 1981). The failure of more epidemiologic studies to use a proper control cohort increases the difficulty in interpreting the effects of vinyl chloride on the exposed worker. 7-34 CMA 010433 Inadequate statistical power. The term ""sstt.atistical power" represents the probability of detecting an association or an excess risk if that association really exists. For example, if the power of a study to detect a specified change in risk is 80%, there is an 80% chance that the study will show an increased risk if that risk really exists. Negative findings may have two explanations: there may actually be no increased risk, or the study may have had inadequate power to show a risk that really exists. Power depends on the level of statistical significance used (usually p < 0.05), the number of outcomes expected (in these studies cause-specific deaths), and the strength of the association being investigated (effect size). Mortality Studies A summary of the important characteristics of individual epidemiologic studies is given in Tables 7-11 and 7-12. Each study should be evaluated keeping in mind the difficulties noted above. Soon after the initial case reports by Creech and Johnson (1974), describing the identification of liver angiosarcomas in vinyl chloride workers, Monson et al. (1974) published a proportionate mortality analysis of the deaths of 161 vinyl chloride workers at two plants in the United States. A statistically significant 50% excess mortality for all cancers and an 11-fold increase in mortality from cancer of the digestive system, including five angiosarcomas of the liver (LAS), were observed. In addition, 7-35 CMA 010434 STUDY COHORT F/IKX) TABLE 7-11 A SUMMARY OF EPIDEMIOLOGIC DATA FOR OCCUPATIONALLY EXPOSED VINYL CHLORIDE WORKERS BEATHS(X) EXPOSURE(YRS) F/U TIHES(YRS) DOSE SNR DEATH ALL SITES LIVER(LAS) BRAIN LUNG LYMPHOMA 1. Tabershau t Gaffey1 8,384 USA (1974) 2. Duck et al. 2,122 U.K. (1975) 1258(15X) 7(0.3X) 352(4.7X) 152(7.2X) >1 10.2X>20yrs >0 >1 27X>19yrs EST 75 EST 96 110 943<6) 96 993(0> 155* 112 103 106 3. Nicholson et al. 257 2(0.8X) 24(9.3X) >5 >10 EST 126 231 - (3) USA (197|) 4. Ott et al. USA(1975) 594 0(0X) 79(13.3X) >0 > 0 Measured 89 81 - (0) 77 5. Byren et al. 771 21(2.A) 58(75X) >0 55X>10yrs EST 413a(2) 6123 168 Sweden (1976) -0 6. Uaxweiler et al. 1,294 13(1X) 136(10.5X) >5 t jJ USA (1976) 7' >10 EST 108 1155^(11) a5 329 156 159 >15 189b 1606b 4989 194 "* 176 7. Fox and Collier U K (1977) 7,717 393(5.IX) 409(5.3X) >0 8X20yrs >0 EST 75.4 90.7 1408a(2) 54.6 89.8 90.9 8. EEH USA (1975) 10,173 496(4.8X) 707(6.9X) >1 19.3X20yrs 32X>20yrs EST 89 104 753(5> a 203 107 112 9. Buffler et al. Texas (1979) 464 0(0X) 28(0X) >0 > 0 Measured 87 138 - (0) a 208 10. Bertazii et at. Italy (1979) 4,777 659(13.SX) 62(1.3X) >0.5 >0.5 EST 44 97 8008(3> 125 81 133 11. Masuda et al. Japan (1979) 304 1(0.3X) 26(8.5X) >1 >1 EST 138 500S(0) 125 CHA 0 1 0 4 3 5 TABLE 7-11 (con't) stuot COHORT F/UIX) DEATHSU) EXPQSURE(YRS) F/U TIHES(TRS) DOSE SUB DEATH ALL SITES LI VER(LAS) BRAIN LONG LYHPHOMA 12. Weber, Reinl, Greiser Germany (1981) production 7,021 processing 4,007 unexposed 4,910 700(4.4X) 414(5.9%) 360(9%) 417(8.5%) 1 13. Cooper USA (1981) 14. Hakassira Japan (1983) 10,173 496(4.8%) 707(6.9X) 4,524 29(0.6%) 209(4.6%) >0 >0 >0 >1 >1 >0 EST 95 >0 EST 95 >0 EST 78 33.4X EST 89 Mean EST 87 16.3yrs 112 85 83 104 a 138 b 1525 434 40l" 753(8) 236a(3) 162 a 535 184 a 203 107 86 214* 34 77 112 15. Heldass et al. Norway (1984) 454 0(0%) 50(11%) >1 >1 EST 84 114 (1) 180 .......................................... Relative Risk ............... J J 16. Theriault t Allard Canada (1981) exposed 451 0(0X) 59(2.6X) >5 81X115yrs EST 1.07 1.48 6.25S(10) unexposed 871 233(26.8X) .36 1. The studies of Cooper and EEH are reanalyses of the Tabershau and Gaffey Cohort 2. SHR subjects also in the Tabershau and Gaffey Cohort 3. SHR is for "digestive system cancer", not liver cancer 4. SHR is for "other and unspecified cancer*, 40X of uhich were brain cancer 5. SHR is for cancer of CHS, not Brain F/U = Follow up tine (years) EST = Esti stated dose *p < 0.05 p < 0.01 CMA 010-436 f ORAFI CMA 0 1 0 4 3 7 STUDY TABLE 7-12 A SUMMARY OF TUMOR INCIDENCES AND STANDARDIZED MORTALITY RATIOS (SMR) FOR OCCUPATIONALLY EXPOSED VINYL CHLORIDE WORKERS DEATH ALL CANCER LIVER CANCER BRAIN CANCER LUNG CANCER 0 E SHR 0 E SMR 0 E SMR 0 E SMR 0 E SHR 1. Hanson el (. 2. Tabershaw t 161-161-100 Gaffey 352-667-75* 3. Duck et *1. 136 142.2 96 A. Nicholson et si. 24--19--126 5. Ott et el. 79-69.1-89 6. Byren et at. 7. Uaxweiler et al. 15 year B. Fox l Collier 9. EEH 136-126.3-108 393 521.2-75.6 707-795--89b 10. Buftler et al. It. Bertazzi et al. 12. Hasuda et al. 13. Weber et al. Production Processing Control 14. Cooper 15. HeIdaas 16. Theriault and Allard 17. Nakamura 128 167.6-87 41-27.9-150 8 0.7 1100 79-77- 110 35-36.4-96 9--3 9......... 13--16--81 35 23.5-149* 31 16.9-184b 115-126.8-90.7 139-141.4-104 8-5.2-154 30-30.9-97 8 5.8-138 19-21.7-94 11-11.1-99 31............ 4--.97--413 7-0.6-1155b 7 0.4-1606** 1-.71-140.8 29-40.8-75 0--.2......... a--i--aoob 1-.6--167 -1523 -636 16-5.4-259.3 6-2.54-236 % < 0.05 5-1.2-420 17-11.78 155 13-7.9 160 25-23.9-112 16-15.5 103 2 .38 -612* 3-0.9-329* 3-0.6-498* 2-3.66-54.6 12 5.9 203* 1-0.1 1-0.8-125 0--.15----- 3--1.8 108 12-7.7-156 11-5.7-194* 46-51.2-90 45-44.3-107 5 -1.7--289* 7 -7.7 91 1 .8-125 45-44.3-107 5-2.8- 180 2- 5.8 34.6 2-2.3-0.86 p < 0.01 o 3 observed e 3 expected SMR - standardized mortality ratio LYMPHOMA 0t SHH 5 3.4 -150 6--6.1-106 21................. 4-2.5-159 3-1.7-176 9 9.0 99.9 11-10.4 112 0 0.5......... 4 -3 133 0--.5........... 11-10.4 112 DRAFT increases in the proportionate mortality ratios (FMR) for brain cancer, lung cancer, and lymphoma were noted. Proportionate mortality ratios do not represent a specific measure of risk, but the consistent PMR excesses for neoplasms found in this study suggests that vinyl chloride may operate as a multisystem carcinogen. Tabershaw and Gaffey (1974) published a large cohort study of 8,384 vinyl chloride workers at 33 plants in the United States, which demonstrated a statistically significant increase in angiosarcoma of the liver and a nonsignificant positive trend correlating vinyl chloride exposure with lymphoma and cancers of the buccal cavity and pharynx, CNS (primarily brain), and lung. The SMRs for all these tumor types were greater in the high exposure groups after the cohort was stratified by high and low exposure indices (estimates based on job classification and length of exposure), but the differences in SMRs for the high- and low-exposure groups were not statistically significant. Follow-up in this study was only 85% complete. The workers for whom follow-up was incomplete were mostly older workers, and Tabershaw and Gaffey (1974) suggested that these workers, who experienced a long latent period after exposure, might show a somewhat different mortality pattern from workers who were followed up. Another significant problem is that the authors reported only digestive system cancer and did not distinguish cancer of the liver from other cancers in this classification. Information on this cohort has been updated and 7-39 CMA 010438 reanalyzed by Cooper (1931) . The final report included 10.173 vinyl chloride workers from 37 plants in the United States (Cooper, L9'81) . Follow-up had increased to 95.1% of the cohort and extended more than 20 years for 33.4% of the cohort. Statistically significant excess mortality was shown for LAS and for CNS cancers (primarily brain). Again, SMRs for lung cancer and lymphoma were elevated but not statistically significant. Duck and co-workers published an analysis of 2,122 vinyl chloride workers in Great Britain (1975). In this study, no excess of total or cause-specific mortality occurred. There were no cases of LAS, although one was recorded in the cohort after the study period ended. Only 16% of the cohort in this study was followed more than 15 years from the time of initial exposure, which undermines the reliability of the negative results of this study. Nicholson and colleagues reported on 257 workers in the United States who were exposed to vinyl chloride for at least five years and whose initial exposure occurred more than ten years before the end of the study (1975). These inclusion criteria are important because this is the first study that attempted to limit the cohort to workers who had significant vinyl chloride exposure and follow up time. Three cases of LAS were observed and the SMRs for all deaths and deaths due to cancer were elevated. Because LAS is otherwise exceedingly rare, the increased incidence of this tumor was statistically significant, but the study lacked power to detect significant increases in other classifications of malignancy. 7-40 CMA 010439 A larger cohort based on similar criteria was assembled bv Vaxweiier et ai. (1976) for the National Institute for Occupational Safety and Health (NIOSH). This study followed an adequate number of workers (1,294) for more than 10 years, with all having had more than five years of exposure. Separate analyses were also performed for those workers with more than 15 years of follow-up time. Significant excesses in the SMB. of exposed workers were found for all deaths due to cancer, liver cancer (11 cases of LAS), and CNS cancers. Standard mortality ratios for lung cancer and lymphoma were elevated, but were not significant at the p < 0.05 level. Workers with more than 15 years of follow-up time showed higher mortality rates compared to those with ten years of follow-up time, _ The SMR for lung cancer reached statistical significance in the group with a 15-year follow-up. This cohort provides the strongest evidence for the association between length of time since exposure to vinyl chloride and the subsequent development of cancers of the liver, CNS, and lung (Waxweiler et al., 1976). Ott and associates completed a study of 594 Dow Chemical workers in Michigan (1975). Many of these workers were also included in the study by Tabershaw and Gaffey (1974). The best available vinyl chloride exposure data are included in this study. Automated sampling of air levels began in one plant as early as 1959. Unfortunately, a large number of workers had less than one year of vinyl chloride exposure at the time of this report. Stratifying the cohort into low, medium, and high exposure groups resulted in less than 200 subjects per group, with only 20, 18, and 22 deaths per 7-41 010440 CM* group. respectively, .\'o cases of LAS and no significant increase in mortality from any cause for the entire cohort were noted. However, total deaths and deaths due to cancer were significantly higher in the high vinyl chloride exposure group compared to all other dose groups. These data are insufficient to develop any human dose-response relationship. The only other study providing quantified human exposure data is by Buffler and co-workers (1979). Area sampling began after 1971 for 464 Dow Chemical vinyl chloride workers in Texas, but data on exposure levels were not available for those workers (the majority) exposed before monitoring began. No cases of LAS were observed among these subjects. There was a statistically significant excess only for lung cancer in exposed workers. The number of deaths (N - 28) in this cohort was very small, making it impossible to perform statistical assessment of many of the causes of death. Buffler and associates have published the only information on the smoking habits of vinyl chloride workers. Even after adjustment for smoking habits, the excess of lung cancers in this group remained significant. Byren et al. (1976) and Heldaas et al. (1984) reported on Swedish and Norwegian vinyl chloride workers, respectively. Although these studies included all vinyl chloride workers in the respective countries and had the advantage of access to excellent cancer registries, the total number of workers in both studies was still quite small. Nevertheless, both studies show a significant excess 7-42 010441 CH* DKAFT of LAS in exposed workkeerrss.. The Swedish study indicates a significantly increased mortality for CNS cancers. Both reports also demonstrate a trend toward increased mortality due to lung cancer. Fox and Collier (1977) studied all 7,717 workers in Britain who may have been occupationally exposed to vinyl chloride between 1940 and 1974, Four cases of liver cancer were found; two of these were angiosarcomas. No other tumor type showed a significant increase (statistical methods not reported). Because workers were added to this cohort as they entered the industry, the study included a large proportion of workers with brief exposure and short follow-up time. Approximately 75% of the subjects had been employed in the vinyl chloride industry for less than ten years and only 8% of the workers had been employed for more than 20 years. Inadequate length of exposure and follow-up make this study's negative results of questionable validity. Bertazzi et al. (1979) examined the mortality rates among 5,441 Italian vinyl chloride workers. This study showed a significant increase in mortality among exposed workers only for liver cancer (three cases of LAS). Follow-up was less than optimal (14% of the total remained untraced), and person-years at risk were calculated as if the workers unavailable to follow-up were all alive and well, which contributed to the very low SMR for all causes of death. 7-43 CMA 010442 Masuda. and co-workers studied 304 Japanese vinyl chloride workers This cohort was too small to determine statistical significance for any cause of death. Weber, Reinl, and Greiser (1981) reported on mortality information from three cohorts of German chemical industry workers: 7,021 vinyl chloride and polyvinyl chloride production workers (usually considered a high exposure area), 4,007 polyvinyl chloride processing workers (a lower exposure area), and 4,910 chemical workers not exposed to vinyl chloride (1981). The SMRs were determined for causes of death in each of the three groups but no statistical comparisons were made. A significant increase in mortality from liver cancer was observed in all three of the groups evaluated, most notably for the vinyl chloride processing workers (SMR - 1523). A significant increase in malignancies of the lymphatic and hematopoietic tissues was noted among the production workers, while a significant increase in brain tumors was observed among the processing personnel. Analysis of the mortality experience of 4,524 Japanese vinyl chlorida workers by Nakamura (1983) revealed a significant increase in the mortality ratio for death from all cancers and from liver cancer alone (three cases of LAS). Cancer of the lung was not elevated; cancers of the CNS and lymphoma were not reported in this study. 7-44 CMA 010443 Theriault and Allard (1981) studied Canadian vinyl chloride workers in the only cohort to employ an occupational control group for evaluation of relative risk in workers exposed to vinyl chloride. The control cohort consisted of 870 chemical workers not exposed to vinyl chloride, while the study group comprised 585 vinyl chlorideexposed workers, with 454 of these workers exposed for more than five years. Exposure levels were not quantified. Very few deaths (59 cases) occurred in the exposed group, compared with 233 in the control group. The only significantly increased relative risk was for liver cancer (eight cases of LAS). The SMR for digestive cancer (which includes liver cancer) among workers exposed for greater than five years was 259, significantly greater (p < 0.01) than for the general population. The authors suggested that the small size of the study reduced the power of the study with respect to finding an excess of CNS cancer or lymphoma that may have been present. Theriault (1983) published an extended follow-up on this same cohort in 1983 with no significant changes in the initial findings. Heldaas and co-workers reported a study of cancer incidence and mortality in a cohort of 454 male workers exposed to vinyl chloride and polyvinyl chloride between 1950 and 1969 (Heldaas et al. 1984). This cohort was divided into three exposure groups, as estimated from job classification, and the study population followed for 27 years. The investigation demonstrated an increased incidence of malignant melanoma, and cancer of the lung, colon, and thyroid 7-45 CMA 010444 DRAF in -he exposed cohort. This study reported cancer incidence, as well as mortality, unlike most ocher studies. The observation of an increased incidence of malignant melanoma is the first reported in humans (Heldaas et al., 1984), Four malignant melanomas of the skin were identified in the study population where only 0.8 were expected. Three of four cases of malignant melanomas occurred in the high exposure group, where 0.5 cases were expected. The fourth case was in the medium exposure group with 0.18 cases expected. After the observation period, one more case was diagnosed in the medium exposure group. The authors noted one additional case of incipient malignant melanoma in the medium exposure level group that was diagnosed in 1977 but not included in the study (Heldaas et al. 1984). Dahar et al. (1988) recently published an update to the vinyl chloride mortality study of Ott et al. (1975). There was no statistically significant excess for any neoplasm or disease of interest among the exposed cohort. Rinsky et al. (1988) evaluated the mortality rate and cause of death for a cohort of chemical workers in West Virginia. A statistically significant increase in liver cancer and lympho- and reticulo-sarcoma was seen among the workers. The mortality rate of the 29,319 male workers studied was similar to that of the U.S. white male population (Rinsky et al,, 1988). Smulevich et al. (1988) reported that mortality resulting from tumors of the digestive organs, respiratory system, bone and connective tissues, brain and skin was higher in workers exposed to 7-46 CMA 010445 vinyl chloride chan for che general population (USSR). ho cases of liver angiosarcoma were reported in the study cohort during the follow-up period (Smulevich et al,, 1988). ".2.5 Cancer Risks Associated with Exposure..to. Vinyl Chloride "2.5.1 Liver Cancer Between 1961 and 1977, 23 cases of LAS were reported among approximately 20,000 vinyl chloride workers in the United States (Lelbach and Marsteller, 1981; Spirtas and Kaminski, 1978). The expected incidence of LAS is 0.014 cases per 100,000 per year in the general population in the United States (Heath et al., 1975). Based on analysis of these data, the relative risk for developing LAS following vinyl chloride exposure among this country's vinyl chloride workers is 483. The epidemiologic studies also demonstrate a strong and consistent association between vinyl chloride exposure and primary cancer of the liver. All eight of the studies that assessed risk for primary liver cancer note a statistically significant increase in standardized mortality ratios (SMR). The average relative risk for liver cancer among vinyl chloride workers is five to six times greater than the incidence of that seen in the general population. Strong evidence suggests that exposure to vinyl chloride can cause liver cancer. All reports published to date indicate that the standardized mortality ratios of exposed workers are elevated, and 7-47 CMA 010446 TABLE 7-13 A SUMMARY OF EPIDEMIOLOGIC STUDIES WHICH EXAMINED POSSIBLE CORRELATIONS BETWEEN OCCUPATIONAL VINYL CHLORIDE EXPOSURE AND PRIMARY CANCERS OF THE LIVER STUDY INCREASING INCREASING SMR RESULT DOSE1 F/U TIME2 3 Byren et al. Waxweiler et al. Fox & Collier Bertazzi et al. Masuda Weber et al. 3 Theriault & Allard Nakamura 413a 1155b 141a 800a 500a 1523b 6.25a 236a Significant Significant Significant Significant Significant Significant Significant Significant -- -- Yes -- -- Yes -- Yes Yes Yes ... ... Yes No Yes 1 Does risk increase with higher estimated dose? 2 - F/U time - Follow-up time (years) Does risk increase with longer latency? 3 - Relative risk, not SMR ap < 0.05 bp < 0.01 7-48 CMA 010447 TABLE 7-14 draft A SUMMARY OF EPIDEMIOLOGIC STUDIES WHICH EXAMINED POSSIBLE CORRELATIONS BETWEEN OCCUPATIONAL VINYL CHLORIDE EXPOSURE AND BRAIN CANCER STUDY INCREASING INCREASING stm RESULT DOSE1 F/U TIME2 3 Byren et al. Waxweiler et al. Fox & Collier Bertazzi et al. Weber et al. 3 Cooper 612a 329a 55 125 535a 203a Significant Significant + Significant Significant -- - -Yes -- No -- --- Yes ... ... No --- 1 Does risk increase with higher estimated dose? 2 - F/U Time - Follow up time (years) Does risk increase with longer latency? 3 Cooper's data are used in the most recent reevaluation of the Tabershaw, Gaffey and EEH cohort. ap < 0.05 + * non-significant positive trend for increased risk (p < 0.05) 7-49 CMA 010448 DRAF v-"11 5pta 1r-* of liver cancer was seen to increase wirh both and a .longer follow-up time (Table 7-13) creased dose 2.5.2Other Cancers The association between vinyl chloride exposure and increased risk for other cancers is not as clear as that for liver cancer. Some evidence associates exposure to vinyl chloride with increased mortality ratios for brain cancer, lung cancer, and lymphoma. Since these cancers appear more commonly in the general population than LAS and primary liver cancer, it becomes more difficult to show increased risk. 7.2.5.2.1 Brain Cancer Workers exposed to vinyl chloride appear to be at greater risk for brain cancer than do non-exposed populations. Of the six studies that assessed the risk of brain cancer, five showed a positive trend for increased risk of this cancer type following exposure to vinyl chloride, with four demonstrating statistical significance (p < 0.05) (Table 7-14). Cancer risk increased an average of four times above that expected in the general population in those studies that exhibited a significantly increased risk. Of the two studies not showing a significant increase in risk for brain cancer, statistical power in the Bertazzi and associates study was only about 35% (Bertazzi et 7-50 CMA 010449 al. . 1979) , while that of Fox and Collier (1977) was approximately 80% (Beaumont and Breslow, 1981) . In the Fox and Collier study, the number of deaths overall was low and, most importantly, a large percentage of workers in the cohort was verv recently employed in the vinyl chloride industry and thus had a short follow-up time. These factors may partially explain why this study failed to detect an association between vinyl chloride exposure and brain cancer. 7.2.5.2.2 Lung Cancer The evidence linking vinyl chloride exposure with lung cancer - remains inconclusive. Analyses of SMRs for cancer of the lung were performed in 12 studies (Table 7-15). Of these, seven studies showed an increased risk for lung cancer, but only one was statistically significant at the 5% level (Buffler et al., 1979). This increased risk persisted after adjusting for personal smoking habits (for this particular cohort). However, this cohort was small and the study was unable to demonstrate an increased risk for any other cancer. The Waxweiler et al. cohort (which had a follow-up period greater than 15 years) also used a small group (1976). Of all the studies that examined the risk for lung cancer, only those of Fox and Collier (1977) and Cooper (1981) have greater than 80% power to detect increased relative risks (of 1.5 or 2) for lung cancer following exposure to vinyl chloride. Both of these studies found no statistically significant increased risk for lung cancer. 7-51 CMA 010450 -vraphoma An association between vinyl chloride exposure and lymphoma has not been established. Five studies evaluated the risk of lymphoma development among workers occupationally exposed to vinyl chloride (Table 7-16). Four of the studies showed a positive trend for lymphoma among vinyl chloride workers, but statistical significance was noted only by Weber et al. (1981). However, the statistical power in all of these studies was less than 80% to demonstrate a relative risk of two, and less than 40% to show a relative risk of 1.5. Exposure Information Most of the published epidemiologic studies did not present quantified exposure data. Levels of exposure were estimated by job classification and length of employment. Only the studies by Ott and co-workers (1975) and Buffler and associates (1979) contain measured industrial hygiene data. After the workers were classified according to exposure levels, the cohorts were too small to yield any statistically significant correlations. A doseresponse relationship cannot be constructed based on these kinds of human data. This conclusion was also reached by the United States Environmental Protection Agency (1986), CMA 010451 DRAFTcable A SUMMARY OF EPIDEMIOLOGIC STUDIES WHICH EXAMINED POSSIBLE CORRELATIONS BETWEEN OCCUPATIONAL VINYL CHLORIDE EXPOSURE AND LUNG CANCER sim INCREASING INCREASING SMR RESULT DOSE1 TIME2 3 4 Duck et al. Ott et ai. Byren et al. Waxweiler et al. Fox & Collier Buffler et al. Bertazzi et al. Masuda et al. ,,Cooper 3 Heldass et al. 4 Theriault & Allard Nakamura 103 77 168 156 90 268a 91 125 107 180 .36 86 + + + Significant + + + - m -- -- No -- -- -- Yes -- -- ... No ... ... Yes ... ... -- -- No ... 1 - Does risk increase wich higher estimaced dose? 2 - F/U time - Follow-up time (years) Does risk increase with longer latency? 3 - Cooper's data is used in the most recent revaluation of the Tabershaw, Gaffey and EEH cohorts. 4 - Relative risk, not SMR + - non-significant positive trend for increased risk (p < 0.05) ap < 0.05 7-53 CM* 010-452 TABLE 7-16 ORAFt A SUMMARY OF EPIDEMIOLOGIC STUDIES WHICH EXAMINED POSSIBLE CORRELATIONS BETWEEN OCCUPATIONAL VINYL CHLORIDE EXPOSURE AND LYMPHOMA STUDY Waxweiler et al. Fox & Collier Bertazzi et al. Weber et al. 3 Cooper SMR 159 100 133 214a 112 RESVLI INCREASING DOSE1 INCREASING F/U TIME2 3 -f+ Significant + ... -- -- + ... Yes ---- ""* 1 - Does risk increase with higher estimated dose? 2 - F/U time - Follow-up time (years) Does risk increase with longer latency? 3 - Cooper's data are used in the most recent revaluation of the Tabershaw, Gaffey and EEH cohorts. ap < 0.05 + - non-significant positive trend for increased risk (p < 0.05) 7-54 CMA 010453 Conclusions Epidemiologic studies of workers exposed to high levels of vinyl chloride indicate that this chemical is a human carcinogen. Strong evidence suggests that vinyl chloride causes an increased risk for angiosarcoma of the liver. Vinyl chloride appears to be associated with a moderately increased risk for brain cancer, but current epidemiologic evidence does not demonstrate a significant correlation between vinyl chloride exposure and subsequent development of lung cancer or lymphoma. Exposure data in humans are inadequate to construct a dose-response curve for any of the cancers studied. However, exposure levels can be estimated, and appendices B and C discuss the potential human health risk for vinyl chloride using exposure estimates for the Waxweiler et al. (1976) study of occupationally-exposed workers. 7-55 01045* DRAFT ChA 010455 QUANTITATIVE CARCINOGENIC RISK ASSESSMENT DRAFT Introduction Vinyl chloride has been demonstrated to be a carcinogen in multiple species of laboratory animals, including rats, mice and hamsters. Inhalation exposure of rats to low doses of vinyl chloride (<250 ppm) led predominantly to formation of hepatocellular carcinomas, neoplastic nodules of the liver, mammary tumors and nephroblastomas, while higher exposures were associated with liver and lung angiosarcomas, neuroblastomas, and zymbal gland carcinomas. Feeding studies have confirmed these results. Vinyl chloride has also been* identified as a human carcinogen by IARC (1979), the EPA (1984b), and the State of California (CDHS, 1985). Epidemiologic evidence has linked occupational exposure to vinyl chloride with the development of liver angiosarcomas in chronically exposed workers. Several aspects of vinyl chloride metabolism (reviewed in Section 2 of this document) are relevant to its carcinogenicity. First, the oncogenicity of vinyl chloride appears to be due to one or more reactive metabolites, rather than the parent molecule; and secondly, the metabolism of vinyl chloride is a saturable, dose-dependent process. The rate of formation of the carcinogenic metabolites is limited by the metabolism of the parent compound, and once the enzyme systems responsible (alcohol dehydrogenase and cytochrome P-450) become saturated, administering a larger dose of vinyl chloride may not necessarily result in a significantly greater tumor incidence. CMA 010456 8-1 Exposure concentrations above 250 ppm may exceed the metaoolic caoacicv of these enzymes, and thus exposures to higher concentrations have to be evaluated with this caveat in mind. There are many inherent difficulties and uncertainties involved in using animal data to determine human risk. Because of this, risk estimates have been calculated from different species, sexes, experiments and tumor types. Several adjustments need to be made to the experimental exposure data to calculate the lifetime daily exposure (LDE) levels. Thus, by: for inhalation exposures, the reported dose must be multipliedm H/24: where H is the hours of exposure per day. This converts the exposure period to a time weighted average for 24 hours daily continuous exposure. D/7: where D is the number of days exposed per week. This converts the dosing schedule to a time- weighted average for a seven day/week continuous exposure. Le/L: where Le is the length of the experiment and L is the lifespan of the animal (the longer of Le or 24 months). This converts the experimental protocol to a continuous lifetime exposure. The staff of DHS has used the linearized multistage computer program GLOBAL86 to calculate potential risks associated with vinyl chloride exposure. The multistage model may be expressed as: P(d) 1 - e + qjd + q2d 8-2 + q,ak> CHA 010437 vhere P(d) is the lifetime probability of cancer for a given dose d of carcinogen, is a constant that accounts for the background incidence of cancer occurring in the absence of carcinogen, and q q7, ... are coefficients that allow the data to be expressed to various powers of the dose of carcinogen to obtain the best fit of the model to the data. Analysis of Data from Maltoni et al. The most sensitive site, sex and species in which a significant increase in tumor formation was observed following exposure to vinyl chloride appear to be the mammary gland in female Sprague-Dawley rats, from the Maltoni et al. BT15 experiment (Table 7-8). A statistically significant increase in malignant mammary gland tumors was observed following a one year exposure to levels as low as 5 ppm. This experiment (BT15) has several potential problems associated with it: 1) the observed mammary tumor incidence in both treated and control animals was elevated compared to the results obtained from other Maltoni et al. vinyl chloride experiments (Table 7.8); and 2) there is no observable dose-response trend over the entire dose range of the experiment (0-25 ppm); tumor incidence at the 5 ppm dose level was greater than at the 25 ppm level. Risk estimates can be calculated by the linearized multistage model for both the entire dose range as well as for the first three doses (0, 1, and 5 ppm) where the tumor incidence increases with the dose. The 95% upper confidence interval of the cancer potency slope for mammary tumors for the entir~ xposure protocol in BT15 suggests a lifetim risk of 8-3 CMA O104S8 DRAFT 2.-+ x _0 ,/ppo; analyzing just the 0, 1, and 5 ppm dose levels yields .3 a risk estimate of 1.9 x 10 /ppb. However, there do not appear to be any scientifically justifiable reasons for dropping the two upper doses value from the analysis; DHS staff therefore recommends that the -4 of 2.4 x 10 /ppb derived from the incidence of mammary gland carcinomas in female rats for the entire exposure range from the BT15 experiment be used as an estimate of risk from vinyl chloride exposure. Risk estimates for mammary gland carcinomas for other Maltoni et al. experiments which demonstrated a significant increase in tumor formation are lower than that derived from the BT15 data (Table 8.1). Maltoni et al. also observed significant increases in the induction of liver angiosarcomas (LAS) following a one-year exposure of rats to vinyl chloride (Table 7.7). Although there is substantial variability in the different data sets, all experiments show a clear trend towards an increased incidence of LAS, a relatively rare tumor response in laboratory animals. The incidence of IAS in control groups of the experiments evaluated herein is consistent with the low historical control value of 2/541 claimed by Maltoni et al. (1984). These tumors were observed in female rats at experimental exposures as low as 10 ppm (1/46). Female rats were more sensitive than males to the oncogenic effects of vinyl chloride in all Maltoni et al. experiments (Table 7.7). This increased sensitivity is reflected in the risk estimates (Table 8.1). Risk values for IAS range from 9.0 x 10'Vppb for male rats in the BT1 experiment to 7.4 x 10 Vppb for female rats from the BT15 experiment. CMA 010459 8-4 draft for some of the inhalac ion experiments, exposures greatly exceeded the metabolic capability of the enzyme systems involved. Based on the studies of Bolt et al. (1977) and Hefner et al. (1975b), it appears that metabolic saturation is reached at vinyl chloride concentrations of approximately 250 ppm. Maltoni et al. exposed animals to doses as high as 30,000 ppm. Exposure levels in the Bi et al. study reached 3,000 ppm. Tumor incidence and risk estimates derived from the groups exposed to concentrations greater than 250 ppm may not accurately reflect the oncogenic potential of vinyl chloride. Thus, although risk estimates for the entire range of exposures from the Maltoni et al. BT1 experiment and the Bi et al. study have been calculated, DHS staff recommends that only the risk estimates based on exposures of less than 250 ppm be considered valid estimates of cancer risk. This choice preempts the need to adjust for saturation in the risk calculations since many experimental results using animals exposed to concentrations below 250 ppm are available. Table 8.1 gives risk estimates calculated by the linearized multistage model for LAS and other tumor types from both male and female rats for inhalation experiments done by Maltoni et al., Bi et al., and Drew et al. for the entire exposure range, as well as for exposures of less than 250 ppm. 8.3 Analysis of Data from BI et al. Bi et al. (1985) measured liver and lung angiosarcomas in rats following an 18-month exposure to either 0, 10, 100, or 3000 ppm vinyl chloride. A significant increase in tumor induction at both 8-5 CMA 010-4dO sices was observed, especially for the high-dose group. The vinvl chloride concentration in the high-dose group would almost certainly have saturated the metabolizing enzymes involved. Since lung cancers have been linked to vinyl chloride in occupationally exposed workers, risk estimates for lung angiosarcomas have been estimated using the data of Bi et al. (1985). Risk estimates for liver and lung angiosarcomas for exposure scenarios of 0 to 100 ppm and 0 to 3000 ppm have been calculated, and are included in Table 8.1. Risks based on the 0 to 100 ppm exposure range are substantially greater. The risk for liver angiosarcomas calculated from the 0 to 100 ppm data was 3.6 x 10 Vppb, while the risk for lung angiosarcoma for the same group was 1.6 x 10*^/ppb. Analysis of Data from Drew et al Drew et al. evaluated the effects of age and length of exposure on vinyl chloride-induced carcinogenicity in female rats, mice and hamsters (Drew et al., 1983). Female Golden Syrian hamsters, F-344 rats, CD-I Swiss mice, and B6C3F1 mice were exposed to 200, 100, 50 and 50 ppm vinyl chloride, respectively, six hours/day, five days/week for 6, 12 and 18 months. These doses are established carcinogenic levels of vinyl chloride for each species. The rats and hamsters also underwent exposure periods of up to 24 months. The risk assessments were based on those exposure durations demonstrating the greatest tumorigenic response in each species. Because exposure scenarios were limited to a single concentration per species, lifetime daily exposure equivalents were calculated using the 8-6 CMA 010461 different exposure lengths ORAr as the primary variable (instead of the dose, as is more customary). Hemangiosarcoma was the most sensitive tumor type, and the most sensitive sex and species were female B6C3F1 mice. Except for female F-344 rats, the authors made no distinction between liver hemangiosarcomas and hemangiosarcomas found at other sites. Hemangiosarcomas in B6C3F1 10 -4 /ppb. The 95% upper mice yielded a risk estimate of 4.2 x confidence interval for mammary gland carcinomas 10 -4/ppb. in female CD-I Swiss mice suggest a risk of 6.1 x The most sensitive tumor site in female F-344 rats was the liver hemangiosarcoma, with an estimated risk value of 3.7 x 10'Vppb; for female Syrian golden hamsters, mammary gland carcinoma was the -5 10 /ppb. most sensitivesite, with a risk value of 5.3 x The risk estimates for these and the remainder of the Drew et al. study are found in Table 8.1. 85 Human Studies The prior review of the epidemiological studies (Section 7.2 of this document) strongly suggests a causal association between vinyl chloride and several different types of cancer, including liver, lung, and brain. Although exposure data from the occupational cohort studies were inadequate to derive a dose-response curve, other historical industrial hygiene data can be used to reconstruct a range of likely exposures from which risk estimates can be derived (Appendix B). Using this approach the exposure level associated with CMA 0104A2 8-7 DRAf one million risk for these carcinogenic endpoints vas estimated as 0.4-85 ppb and the estimated risk per ppb as 2.1 x 10 /ppb (Appendix , Table D). If the assumption is made that just liver cancer (the tumor type for which the epidemiologic evidence is strongest) is associated with vinyl chloride exposure, the risk per ppb is estimated to be 1.0 X 10 Vppb, In a more recent study of vinyl chloride workers, Heldaas et al. (1984) linked vinyl chloride exposure with malignant melanomas of the skin, lung cancer, colon cancer, and thyroid cancer, strengthening the association between vinyl chloride exposure and the development of human cancers. 8.6 Choice of Appropriate Risk Estimates table 8.1 has identified a range of human and animal cancer potency values (q*) from vinyl chloride inhalation carcinogenicity studies in laboratory animals. These potency values can be converted to risk estimates by the equation: dose x potency - risk. Therefore, 1 ppb x fa.9 x 10 ^ (ppb) * a risk of 3.9 x 10 for a lifetime exposure to 1 ppb vinyl chloride. Human risk was estimated from rodent data as follows: Estimated Human Risk - Rodent CHA Oi0463 8-8 where ii and I., are rhe inhalation rates of rodents and humans, respectively, and and W are the body weights of rodents and humans. respectively. Humans are assumed to weigh 70 kg and inhale 20 m'Vdav. The inhalation rates (I) for mice and rats were estimated using the following formulas (EPA, 1985) : For mice: I - 0.0345 12/3 3 For rats: I - 0.0105 wt (kg)/0.113 (kg) m /day 3 The inhalation rate for hamsters was assumed to be 0.086 m /day (Biology Data Book, 1974). Rodent bodyweight values for the studies of Maltoni et al. (1984) and Bi et al. (1985) were derived from data provided in the respective publications. Rodent bodyweights were not given for the Drew et al. (1983) study, and were estimated to be 300 g for rats, 30 g for mice, and 92 g for hamsters. Human risks associated with vinyl chloride exposure as estimated from -6 -3 the animal data range from 2.4 x 10 /ppb to 5.0 x 10 /ppb, as shown in Table 8.2. Due to reasons previously discussed (which include metabolic saturation and experimental variability), the staff of DHS believes that a more appropriate range of risks lies between 3.9 x 10 /ppb and 1.8 x 10* /ppb (Table 8.3). Using data from Maltoni and Lefemine (1975), the EPA (1984) calculated a unit risk of 6.8 x 10 6 (ppb) -1 This was converted to a human q of 2.5 x 10 -2 (mg/kg/day) -1 , equivalent to a q| of 1.8 x 10 -5 (ppb) -1 The human q* 's from the analysis herein range from 3.9 x 10 (ppb) ^ to 1.8 x 8-9 CMA 010464 10 ^(ppb) a range which includes the EPA's potency value of 1.3 x 10 (ppb)"1. Appendices B and C detail the calculation of human risk from estimates of occupational vinyl chloride exposure, determined to be 2.1 x 10 -6 /ppb. Considering the potential differences in exposure duration, oncogenic sensitivity of different species, age of exposure, sex, and levels of exposure, the estimated unit risk values for the human epidemiologic data and those calculated from animal data are consistent with one another. Since many of the tumors associated with vinyl chloride exposure (particularly LAS) exhibit a long latency period, exposure at an early age would produce a greater risk. The average latency period for the development of LAS in occupationally exposed vinyl chloride workers was determined to be 22.1 years (Stafford, 1983). Drew et al. (1983) demonstrated that in rats, mice and hamsters, the highest incidence of neoplasms was observed when vinyl chloride exposure was started early in life. Exposures early in life may produce up to a 10-fold greater incidence in tumors compared to exposures late in life. The BT15 experiment by Maltoni et al. (1984) suggests that mammary gland carcinoma may be a more sensitive indicator of vinyl chloride exposure than LAS, with . . ^ risk estimates of 6.3 x 10 /ppb and 1.9 x 10 /ppb for the two tumor types, respectively. Also, examination of the Drew et al. data (Table 8.3) indicates that the mammary gland is a more sensitive endpoint than LAS. Thus, it is possible that an epidemiologic study of adult males measuring the incidence of LAS may underestimate the actual risk associated with vinyl chloride exposure. Although there 8-10 CHA 0104A5 are significant difficulties associated with extrapolating between animals and humans, the animal studies have the benefit of examining both sexes for a greater proportion of their lifetimes. The human epidemiology studies have been conducted primarily in adult male vinyl chloride workers, which may not adequately estimate risks for the general population. 8-11 CMA 010466 CMA 0 1 0 4 6 7 Experiment Maltoni et al. BTI BT2 BT9 BTI 5 TABLE 8.1. RANGE OF CANCER POTENCY VALUES EUR VINYL CHLORIDE CALCULATED FROM ANIMAL CARCINOGENICITY STUDIES Species and Sex Tumor Type Experiment al Exposures (ppm) IDE (ppm) rat, male rat, female rat, female rat, female rat, female rat, male rat, female rat, female rat, male rat, female rat, female rat, female LAS LAS LAS LAS mammary gland LAS LAS mammary gland LAS LAS mammary gland mammary gland 0-10,000 0-10,000 0-250 0-200 0-200 0-50 0-50 0-50 0-25 0-25 0-25 0-5 0-595 0-595 0-14.9 0-11.9 0-11.9 0-3 0-3 0-3 0-1.5 0-1.5 0-1.5 0-0.3 ~k `h ;uiilira \' (ppb> . 1019.0 2.7 X X 1i0 '!6 1.5 X 1011.5 X 10 5 1.3 X 2.6 X 10 5 6.0 X 10-? 1011.6 X 10's 2.9 X 7.4 X 10 4 2.4 X 10 3 1.9 X 10 3 Bi et al. rat, male rat, male rat, male rat, male us US lung angiosarcoma lung angiosarcoma 0-3000 0-100 0-3000 0-100 0-482.1 0-16.1 0-482.1 0-16.1 8.1 X ">1 1013.6 X 10 6 2.3 X 1.6 X io'5 Drew et al. rat, female rat, female rat, female B6C3F1 mouse, female B6C3F1 mouse, female CD-I Swiss mouse, female CD-I Swiss mouse, female CD-I Swiss mouse, female hamster, female hamster, female hamster, female us mammary gland hepatocellular carcinoma hemangiosarcoma mammary gland hemangiosarcoma mammary gland lung carcinoma hemangiosarcoma skin carcinoma mammary gland 0-100 0-100 0-100 0-50 0-50 0-50 0-50 0-50 0-200 0-200 0-200 Lifetime Daily Exposure (in ppm) -4- --------------------2 95% UpperConfidence Interval 0-17.9 0-8.9 0-17.9 0-4.5 0-2.2 0-4.5 0-2.2 0-4.5 0-2.2 0-17.9 0-17.9 3.7 X 105 3.1 X 10 5 1. / X 10 k b 4.2 X 10 5 1012.5 X 10 4 1.5 X 6.1 X 10, 1011.2 X 10 *1 2.5 X 2.0 X lo'l 5.3 X 10 5 3 LAS - Liver Angiosarcoma <37 --n TABLE 8.2. RANGE OF HUMAN RISKS FOR VINYL CHLORIDE EXPOSURE ESTIMATED FROM ANIMAL CARCINOGENICITY STUDIES Experiment Species and Sex Estimated Weight (kg) Estimated Inhalation Rate (m /day) Estimated Human Risk, Maltoni et al. BT1 rat, male rat, female rat, female BT2 rat, female rat, female BT9 rat, male rat, female rat, female BT15 rat, male rat, female rat, female rat, female 03 w Bi et al. rat, male rat, male rat, male rat, male .425 .275 .275 .275 .275 .600 .400 .400 .600 .400 .400 .400 .300 .300 .300 .300 .254 .190 .190 .190 .190 .320 .244 .244 .320 .244 .244 .244 .201 .201 .201 .201 - (\ 2.4 x 10" 71 * 10-6 3.9 x 10 " 3.9 x 10 ^ 3.4 x 10 ^ 6.8 x 10 1.6 x 10 h 4.2 x 10 7.6 x 10 1.9 x 10 ' 6.3 n 10 5.0 x 10 J 2.1 x 10 ^ 9.5 x 10 ' 6.0 x 10~ 4.2 x 106 CMrt 0104A 8 Drew et al. rat, female rat, female rat, female .300 .300 .300 .201 .201 .201 B6C3F1 mouse, female B6C3F1 mouse, female CD-I Swiss mouse, female CD-I Swiss mouse, female CD-I Swiss mouse, female hamster, female hamster, female hamster, female .030 .030 .030 .030 .030 .092 .092 .092 .039 .039 .039 .039 .039 .086 .086 .086 ---------------- 1 See text for discussion of estimating the inhalation rate and human risk. 9.7 X 10 5 8.1 X 10-5 4.5 X 10 b 1.2 X 10-5 7.3 X 10-4 4.4 X 10 3 1.8 X 10-4 3.5 X 10-5 7.0 X 10-5 5.6 X 10-4 1.5 X 10 50 I DRAF TABLE 8.3. SELECTED HUMAN RISK ESTIMATES FOR LIFETIME EXPOSURE TO 1 PP3 VINYL CHLORIDE1 experiment Species and Sex Maltoni et al. BT1 BT2 BT9 BT15 rat, female rat, female rat, female rat, male rat, female rat, female 3i ec al. rat, male rat, male Most Sensitive Tumor Site 2 LAS LAS LAS LAS mammary gland LAS LAS lung angiosarcoma 95% Upper Confidence Interval (risk/ppb) 3.9 x 10'* 3.9 x 10' 1.6 x 10 4 6.8 x 10' 6.3 x 10'* 5.0 x 10 J 9.5 x 10'^ 4.2 x 10'6 Drew et al. rat, female B6C3F1 mouse, female CD-I Swiss mouse, female hamster, female Appendices B and C human, male human, male LAS hemangiosarcoma mammary gland mammary gland liver liver, lung and brain 9.7 X 10 4 1.2 X 10-3 1.8 X 10-4 1.5 X 10 4 * -6 1.0 X 10 6 2.1 X 10 6 ''`Estimates were selected on the basis of the most sensitive species, sex, and tumor site for experimental exposure scenarios of less than 250 ppm 2 LAS - Liver Angiosarcoma 8-14 CMA 010469 0? . . CONCLUSIONS DRAr Acute Toxicitv Vinyl chloride has a relatively low degree of acute toxicity in experimental animals; two-hour inhalation LD^q values are greater than 200,000 ppm in several species. Exposure to high concentrations can lead to narcosis, cardiovascular and respiratory irregularity, convulsions, cyanosis and death. Several human deaths have been attributed to occupational exposure to very high levels of vinyl chloride. Autopsies of these patients revealed congestion of the liver, spleen and kidneys. Subchronic and Chronic Toxicitv Chronic exposure of workers to vinyl chloride has been shown to lead to "vinyl chloride disease", characterized by occupational acro-osteolysis, vasospasm of the hands similar to Raynaud's syndrome, dermatitis, circulatory and central nervous system alterations, thrombocytopenia, splenomegaly and changes in liver function (Veltman et al., 1975). Spirtas et al. (1975) measured the frequency of eight symptoms commonly reported by workers exposed to vinyl chloride (including dizziness, headaches and nausea) and observed a dose-response relationship using exposure levels estimated from job classifications. These symptoms were observed even at dose levels below 50 ppm. CMA 010470 9-1 Pharmacokinetics DRAFT Approximately 4-2% (but up to 71%) of an inhaled dose of vinyl chloride was absorbed by both man and rats. Oral exposure results in more complete absorption. Radiolabeled vinyl chloride metabolites have been detected in a range of tissues, suggesting thorough distribution. Most of the metabolized vinyl chloride is excreted by the kidney, often as glutathione conjugates. Unmetabolized vinyl chloride is eliminated primarily by pulmonary excretion. Both alcohol dehydrogenase and cytochrome P-450 are involved in the metabolism of vinyl chloride. The evidence suggests that reactive metabolites may be responsible for the toxic effects of vinyl chloride, with the most likely candidates thought to be chloroethylene oxide and chloroacetaldehyde. The metabolism of vinyl chloride appears to be dose-dependent and saturable, with higher doses incompletely metabolized. It has been suggested chat saturation of the metabolizing enzymes occurs at exposure scenarios between 100 and 300 ppm, with 250 ppm chosen herein as the concentration necessary to achieve metabolic saturation. CMA 10471 9-2 Reurocue rive Toxicity DRAFT N'o teratogenic or embryotoxic effects were observed in mice, rats or rabbits exposed to vinyl chloride at maternally toxic doses during gestation. A recent study has suggested that vinyl chloride can cross the placental barrier of exposed pregnant female rats and cause liver cancer and angiosarcoma in the offspring. Epidemiologic studies have suggested a possible increased rate of fetal deaths in women whose husbands were occupationally exposed to vinyl chloride. However, additional studies have concluded that there was no association between vinyl chloride exposure and fetal deaths or birth defects. 9.5 Mutagenicity Vinyl chloride has been identified as a mutagen in bacteria, yeast and animal systems, both with and without addition of an exogenous metabolic activation system. Chloroacetaldehyde and chloroethylene oxide, the putative toxic metabolites of vinyl chloride, were also mutagenic. Levels of chromosomal aberrations and sister chromatid exchanges were higher in workers exposed to vinyl chloride (20 to 150 ppm) than for unexposed control groups. Workers exposed to less than 15 ppm showed no differences in chromosome breaks or aberrations from controls. 9-3 CMA 010472 3 Carcinogenicity Both experimental animal studies and epidemiological studies of worker populations have demonstrated chat vinyl chloride is carcinogenic. The International Agency for Research on Cancer (IARC) reviewed the literature on vinyl chloride mutagenicity and carcinogenicity and concluded that vinyl chloride is a proven human carcinogen (IARC, 1979) and placed vinyl chloride in its carcinogenicity group 1. Substances assigned to this category have demonstrated sufficient evidence to support a causal association between exposure and cancer in humans. IARC also noted that, "...several independent but mutually confirmatory studies have shown that exposure to vinyl chloride results in an increased carcinogenic risk in humans, involving the liver, brain, lung and hemolymphopoietic systems in man." They also noted in "two proportionate mortality studies ... there appeared to be an increased proportion of cancer of the digestive system in both sexes and possibly of the urinary system and of the breast in woman," and "there is no evidence that there is an exposure level below which no increased risk of cancer would occur in humans" (IARC, 1979). The Environmental Protection Agency (EPA, 1984b) has likewise reviewed the data and also concluded that vinyl chloride is a proven human carcinogen. The EPA placed vinyl chloride in its group A as a proven human carcinogen. 9-4 CMA 010473 DRAFT Although both EPA and the National Academy of Science have concluded that there were insufficient data to base a quantitative carcinogenic risk assessment from epidemiological studies, DHS staff have included a human risk assessment using an estimation of exposure made by Barnes (1976) and Paddle (1986) (see Appendices B and C) . The estimated incremental lifetime risk of liver, lung, and brain tumors as a result of vinyl chloride exposure is 2.1 x 10'6/ppb. This risk is several orders of magnitude lower than that calculated from data for female mouse liver angiosarcomas or mammary gland carcinomas (Drew et al., 1983). The animal studies demonstrated a relationship between tumor formation and the sex and age of the animal at first exposure. Fetuses, newborns, younger animals, and females exhibited the highest carcinogenic sensitivity (Drew et al., 1983). In the epidemiological studies of vinyl chloride workers, who were predominantly male, the average age at first exposure was 29.7 years. Thus, to protect all members of the general population, it is more appropriate to base risk assessment calculations on the animal inhalation studies, which were based on a more representative portion of the population actually at risk. The staff of the Department of Health Services conclude that: 1- viPyl--chloride is__ mutagenic and a__ proven animal and human gflisinpggn- 9-5 CMA 010474 2. Since vinyl chloride is genotoxic and there is no experimental evidence that vinyl chloride has a carcinogenic threshold, it should not be considered to have one. Animal evidence has demonstrated that vinyl chloride is carcinogenic at a lifetime daily exposure of 0.06 ppm. Potential human residential exposures may be only from six to 60-fold lower than those in the animal studies. 3- Vinyl___ chloride__has__been demonstrated to cause a number of malignant tumor types in animals, including, angiosarcoma of both the liver and lung, hepatocellular carcinomas, several different lung tumors. hrain tumors, and other types of cancers. Vinyl chloride has been shown to cause liver angiosarcoma in humans and epidemiological evidence suggests that vinyl chloride may induce lung, breast, and brain tumors. Vinyl chloride has been demonstrated to be multisite carcinogen, and the risk analysis performed by the staff of DHS reflects this finding. 4- Quantitative__ risk assessments of the relevant animal inhalation aaifljftg--of vinvl chloride using the linearized multistage model hflXfi--a.uggcatsfl--a--lan&S--?f Potential__ human risks from 1.8 X --2--(Tafrlg 8,2)- The human risk from UgCUPfltlonal--vlnxl__ chloride exposure has been estimated herein -6 2__ k--2.1 x 10 /ppb (Appendices B and C). Thus, although the human risk estimate is based on a historical reconstruction of occupational exposures, it is close to the range estimated from animal studies. Vinyl chloride has not been detected in 9-6 CMA 010475 the ambient air of California (.limit of detection - 0.5 ppb). The California Air Resources Board has monitored vinyl chloride emissions from the BKK landfill in West Covina and the Oil landfill in Monterey Park. Estimates of peak exposure concentrations for maximally exposed receptors range from 2 to 10 ppb at the BKK landfill and 0.6 to 9 ppb at the Oil site. The Air Resources Board has estimated that between 17,000 and 131,000 individuals may be exposed to 1 ppb at the BKK site. A lifetime exposure of 131,000 residents to 1 ppb would be associated with an upper bound estimate of 0.5 to 236 excess cancer cases. These calculations represent an upper range of plausible excess cancer risk: the actual risk, which cannot be calculated, may be insignificant. CMA 010476 draf; CMA 010477 REFERENCES Anderson D, Hodge MCE and Purchase IFH (1977) Dominant lethal studies with the halogenated olefins vinyl chloride and vinylidene dichloride in male CD-I mice. Environ Health Perspect 21:77-78. Anderson D and Richardson CR (1981) Issues relevant to the assessment of chemically induced chromosome damage in vivo and their relationship to chemical mutagenesis, Mutat Res 90:261-272. Bardin CW, Taketo T, Gunsalus GL, Koide SS and Mather JP (1982) The detection of agents that have toxic effects on the testis and male reproductive tract. In: The Banburv Report. Vol. 11: Environmental Factors in Human Growth and Development, pp. 337-355. Barlow SM and Sullivan FM (eds.) (1982) Reproductive Hazards of Industrial Chemicals. Academic Press, London, pp. 566-582. Barnes AW (1976) Vinyl chloride and the production of PVC. Proc Roy Soc Med 69:278-281. Bartsch H, Malaveille C and Montesano R (1975) Human, rat and mouse livermediated mutagenicity of vinyl chloride in S. typhimurium straias. Int J Cancer 1: 429-437. Bartsch H and Montesano R (1975) Mutagenic and carcinogenic effects of vinyl chloride. Mutat Res 32:93-114. Basalaev AV, Vazine AN *and Kochetkov AG (1972) On the pathogenesis of changes developing due to a long-term exposure to the effect of vinyl chloride. Gig Tr Prof Zabol 1:24-27. Cited in EPA, 1984. Basler A and Rohrborn G (1980) Vinyl chloride: an example for evaluating mutagenic effects in mammals in vivo after exposure to inhalation. Arch Toxicol 45:1-7. Beaumont JJ and Breslow NE (1981) Power consideration in epidemiologic studies of vinyl chloride workers. Am J Epidemiol 114:723-734. Bertazzi PA, Villa A, Foa V, Saia B, Fabbri L, Mapp C, Mercer C, Manno M, Marchi M, Marlani F and Bottasso F (1979). An epidemiological study of vinyl chloride exposed workers in Italy. Arg Hig Rada Toksikol 30:403-409 (Suppl). Bi W, Wang Y, Huang M and Meng D (1985) Effect of vinyl chloride on testis in rats. Ecotoxicol Environ Safety 10:281-289. Biology Data Book (1974) P. L. Altman and D. S. Dittmer, Eds. Published by the Federation of American Societies for Experimental Biology, Bethesda, MD. Volume III, p. 1741. Bolt HM, Filser JG and Buchter A (1981) Inhalation pharmacokinetics based on gas uptake studies. III. A pharmacokinetic assessment in man of "peak concentrations" of vinyl chloride. Arch Toxicol 48:213-228. R-l CMA 010478 3olc HM. Kappus Disposition of 162. H, Buchter A and Bolt %' 14 (1,2- C) vinyl chloride (1976) Original in the rat. Arch Investigations. Toxicol 35:153- 3olt HM, Laib RJ, Kappus H and Buchter A (1977) Pharmacokinetics of vinyl chloride in the rat. Toxicol 2:1-79-188. Buchter A, Bolt HM, Kappus H and Bolt W (1977) Die gewebsverteilung von 1,2- 14 C-vinylchlorid bei der ratte, Int Arch Occup Environ Health 39:2732. Buchter A, Bolt HM, Filser JG, Goergens HW, Laib RJ and Bolt W (1978) Pharmakokinetik und karzinogenese von vinyl chloride. Arbeitsmedizinische Risikobeurteilung. Verh Dtsch Ges Arbeitsmedizin (Gentner, Stuttgart) 18: 111-124. (cited in Bolt et al, 1981). Buchter A (1979) Erarbeitung einer speziellen arbeitsmedizinischen Uberwachungsuntersuchung in korrelation zur individvellen VinylchloridExposition. Forschungsbericht des Landes Nordrhein-Westfalen. Fachgruppe Medizin, Nr. 2813, Westdeutscher Verlag, Opladen. (cited in Bolt et al,, 1981). Buchter A, Filser JG, Peter H and Bolt HM (1980) Pharmacokinetics of vinyl chloride in the Rhesus monkey. Toxicol Lett &:33-36. Buffler PA, Wood S, Eifler C, Suarez L and Kilian DJ (1979) Mortality experience of workers in vinyl chloride monomer production plant. J Occup Med 21:195-202. Byren D, Engholm G, Englund A and Westerholm P (1976) Mortality and cancer morbidity in a group of Swedish VCM and PCV workers. Environ Health Perspect 12:167*170. California Department of Health Services (CDHS, 1985) Guidelines for Chemibfrl__ Carcinogen__ Risk__ Assessments__ an__ Their__ Scientific Rationale. California Department of Health Services, Health and Welfare Agency, Sacramento, CA. California Health and Safety Code (1988) Chapter 3.5: Toxic air contaminants; section 39650. In: California Air Pollution Control Laws. Calfomia Air Resources Board, Sacramento, CA. Caputo A, Viola PL and Bigotti A (1974) Oncogenicity of vinyl chloride at low concentrations in rats and rabbits. Int Res Comnrun 2:1582. Clemmesen J (1982) Mutagenicity and teratogenicity ofvinyl chloride monomer (VCM). Epidemiologic evidence. Mutat Res 98:97-100. Cooper WC (1981) Epidemiologic study of vinyl chloride workers: mortality through December 31, 1972. Environ Health Perspect 41:101-106. Creech JL Jr., and Johnson MN (1974) Angiosarcoma of liver in manufacture of polyvinyl chloride. J Occup Med 14:150-151. the R-2 CHA 010479 n*fvt ."wi r Dahar, VS, 3ond, GG, McLaren, EA, Sabel, FL, Lipps, TE and Cook, RR (1988) Update to vinyl chloride mortality study. J Occup Med 30:648-649. Danziger H (1960) Accidental poisoning by vinyl chloride: Report of two cases. Can Med Assoc J 82 : 828-830, Drevon C, Kuroki T and Montesano R (1977) Microsome mediated mutagenesis of a Chinese hamster cell line by various chemicals (abstract). In: 2nd International Conference on Environmental Mutagens. Edinburgh, 1977, p. 150. Drew RT, Boorman GA, Haseman JK, McConnell EE, Busey WM and Moore JA (1983) The effect of age and exposure duration on cancer induction by a known carcinogen in rats, mice, and hamsters. Toxicol Appl Pharmacol 68:120130. Drodzdowicz BZ and Huang PC (1977) Lack of mutagenicity of vinyl chloride in two strains of Neurosoora crassa. Mutat Res 48:43-50. Du JT, Tseng MT and Tamburro CH (1982) The effect of repeated vinyl chloride exposure on rat hepatic metabolizing enzymes. Toxicol Appl Pharmacol 62:1-10. Duck BW, Carter JT and Coombes EJ (1975) Mortality study of workers in a polyvinyl-chloride production plant. Lancet 11:1197-1199. Duprat P, Fabry JP, Gradiski D and Magadur JL (1977) Metabolic approach to industrial poisoning: blood kinetics and distribution of 14C vinylchloride monomer (VCM). Acta Pharmacol Toxicol Suppl 41:142-143. Duverger M, Lambotte E, Malvoisin E, deMeester C, Pancelet F and Mercier M (1981) Metabolic activation and mutagenicity of vinylic monomers (vinyl chloride, styrene, acrylonitrile, butadiene). Toxicol Eur Res 2:131-140. Duverger-Van Bogaert M, Lambotte-Vandepaer M, deMeester C, Mercier M and Pancelet F (1982) Vinyl chloride and acrylonitrile: activation mechanism and mutagenicity. Toxicologie Experimentale 4:35-37, Edmonds, LD, Falk H and Nissim JE (1975) Congenital malformations and vinyl chloride. Lancet 11:1098. Edmonds LD, Anderson CE, Flynt JW Jr and James LM (1978) Congenital central nervous system malformations and vinyl chloride monomer exposure. A community study. Teratology 17:137-142. Environmental Protection Agency (EPA) (1984a) Ambient Water Quality Criteria for Vinyl Chloride. Draft. U.S. Environmental Protection Agency, Health Effects Branch, Criteria and Standards Division. Office of Drinking Water, Washington, D.C. Environmental Protection Agency (EPA) (1984b) Health Effects Assessment for Vinyl Chloride. Office of Health and Environmental Assessment, Cincinnati, Ohio. R-3 CJ1A 010480 Environmental Protection Agency (EPA) (1985) Addendum to the Health Assessment Document for Dichloromethane (Methylene Chloride). Updated Carcinogenicity Assessment of Dichloromethane (Methylene Chloride). EPA 600-8 -82/004F. Equitable Environmental Health, Inc. (EEH) (1978) Epidemiological study of vinyl chloride workers. Prepared for Manufacturing Chemists Assoc. Fabricant JD and Legator MS (1981) Mutagenicity studies of vinyl chloride. Environ. Health Perspect 41:189-193. Feron VJ, Kruysse A and Til HP (1979a) One year time-sequence inhalation toxicity study of vinyl chloride in rats. X. Growth, mortality, hematology, clinical chemistry and organ weights. Toxicology 13:25-28 Feron VJ, Spit BJ, Ammel HR and Kroes R. (1979b) One-year time-sequence inhalation toxicity study of vinyl chloride in rats. III. Morphological changes in the liver. Toxicology 1^:145-154. Feron VJ, Hendriksen CFM, Speek AJ, Til HP and Spit BJ (1981) Lifespan oral toxicity study of vinyl chloride in rats. Food Cosmet Toxicol 19:317-333. Filser JG and Bolt HM (1979) Pharmacokinetics of halogenated ethylenes in. rats. Arch Toxicol 42:123-136. Fox AJ and Collier PF (1977) Mortality experience of workers exposed to vinyl chloride monomer in the manufacture of polyvinyl chloride in Great Britain. Br J Ind Med 34:1-10. Gamble J, Liu S, McMichael AJ and Waxweiler RJ (1976) Effect of occupational and nonoccupational factors on the respiratory system of vinyl chloride workers. J Occup Med 18:659-670. Garro AJ, Guttenplan JB and Milvy P (1976) Vinyl chloride dependent mutagenesis: effects of liver extracts and free radicals. Mutat Res 38:81-88. Green T and Hathway DE (1975) The biological fate in rats of vinyl chloride In relation to its oncogencity. Chem-Biol Interact 11:545-562. Green T and Hathway DE (1977) The chemistry and biogenesis of the Scontaining metabolites of vinyl chloride in rats. Chem-Biol Interact 17:137-150. Greim H, Bonse G, Radwau Z, Reichert D and Heuschler D (1975) Mutagenicity in vitro and potential carcinogenicity of chlorinated ethylenes as a function of a metabolic oxirane formation. Biochem. Pharmacol 24:20132017. Groth DH, Coate WB, Ulland BM and Homung RW (1981) Effects of aging on the induction of angiosarcoma. Environ Health Perspect 41:53-57. R-4 CMA 010481 ,^ Ci !\ n Guengerich FP and SCrickland TW (1977) Metabolism of vinyl chloride; destruction of the heme of highly purified liver microsomal cytochrome P-450-by a metabolite. Mol Pharmacol 13:993-1004. Guengerich FP and Watanabe PG (1979) Metabolism of [^C]- and [ ^Cl] - labeled vinyl chloride in vivo and in vitro. Biochem Pharmacol 28 :589-596. Hansteen IL, Hillestad L, Thiis-Evensen E and Heldaas SS (1978) Effects of vinyl chloride in man. A cytogenetic follow-up study. Mutat Res 51:271278. Harris EA and Adams WGF (1967) Acro-osteolysis occurring in men engaged in the polymerization of vinyl chloride. Brit Med J 2:712-714. Hatch GG, Mamay PD, Christensen CC, Casto BC, Langenbach R and Nesnow S. (1981) in vitro transformation of hamster embryo cells exposed to gaseous or volatile chlorinated hydrocarbons. Proc Am Assoc Cancer Res 22:119 (abstract). Heath CW Jr, Falk H and Creech JL Jr (1975) Characteristics of cases of angiosarcoma of the liver among vinyl chloride workers. Ann NY Acad Sci 246:231-236. Hefner RE Jr, Watanabe PG and Gehring PJ (1975a) Preliminary studies on the fate of inhaled vinyl chloride monomer (VCM) in rats. Environ Health Perspect 11:85-95. Hefner RE Jr, Watanabe PG and Gehring PJ (1975b) Experimental Studies. Preliminary studies of the fate of inhaled vinyl chloride monomer in rats, Ann NY Acad Sci 246:135-148. Hefner RE Jr, Watanabe PG and Gehring PJ (1975c) Percutaneous absorption of vinyl chloride. Toxicol Appl PharmacoL 34:529-532. Heger M, Muller G and Norpoth K (1982) Untersuchunger zur beziehung zwischen vinylchlorid (VCM)-aufnahme und metabolitenausscheidung bei 15 VCMexponierten. II. Messunger zur ausscheidung des VCM-metaboliten thiodiglykosaure im harn. Int Arch Occup Environ Health 50:187-196. Hehir RM, McNamara BP, McLaughlin J, Willigan DA, Bierbower CT and Hardisty JF (1981) Cancer induction following single and multiple exposures to a constant amount of vinyl chloride monomer. Environ Health Perspect 41:6372. Heidelberger C, Freeman AE, Pierta RJ, Sivak A, Bertram JS, Casto BC, Dunkel VC, Francis MW, Kakunaga T, Little JB and Schectman LM (1983) Cell transformation by chemical agents--a review and analysis of the literature. Mutat Res 114:283-285. Heldaas SS, Langard SL and Anderson A (1984) Incidence of cancer among vinyl chloride and polyvinyl chloride workers. Br J Ind Med 41:25-30. R-5 CMA 010482 Hemminki K and Vine is ? (1985) Extrapolation of the evidence on the teratogenicity of chemicals between humans and experimental animals; chemicals other than drugs. Teratogen Carcinogen Mutagen 1:251-318. Hong C3, Winston JM, Thornburg LP, Lee CC and Woods JS (1981) Follow-up study on the carcinogenicity of vinyl chloride and vinylidene chloride in rats and mice; tumor incidence and mortality subsequent to exposure. J Toxicol Environ Health 7:909-924. Hopkins J (1979) Vinyl Chloride. Part 1; Metabolism. Food Cosmet Toxicol 12:403-412. Huberman E, Bartsch H and Sachs L (1975) Mutation induction in Chinese hamster V-79 cells by two vinyl chloride metabolites, chloroethylene oxide and 2-chloroacetaldehyde. Int J Cancer 16:639-644. Infante PF (1976) Oncogenic and mutagenic risks in communities with polyvinyl chloride production facilities, Ann Ny Acad Sci 271:49-57. International Agency for Research on Cancer (IARC) (1979) IARC Monographs on h__ Evaluation of Carcinogenic__ Risk of Chemicals to Human?: Volume 19, Some Monomers, Plastics and Synthetic Elastomers, and Acrolein. International Agency for Research on Cancer, Lyon, France, pp. 377-438. Jenssen D and Ramel C (1980) The micronucleus test as part of a short-term mutagenicity test program for the prediction of carcinogenicity evaluated by 143 agents tested. Mutat Res 75:191-202. John JA, Smith FA, Leong BKJ and Schwetz BA (1977) The effects of maternally inhaled vinyl chloride on embryonal and fetal development in mice, rats, and rabbits. Toxicol Appl Pharmacol 39:497-513. John JA, Smith FA and Schwetz BA (1981) Vinyl chloride: inhalation teratology study in mice, rats and rabbits. Environ Health Perspect 41:171-177. Kalmaz EE and Kalmaz GO (1984) Carcinogenicity and epidemiological profile analysis of vinyl chloride and polyvinyl chloride. Regul Toxicol Pharmacol 4:13-27. Keplinger ML, Goode JW, Gordon DE and Calandra JC (1975) Interim results of exposure of rats, hamsters, and mice to vinyl chloride. Ann NY Acad Sci 246:219-224. Krahn DF (1979) Utilization of the CHO/HGPRT system: metabolic activation and method for testing gases. In: Hsie, A. W., O'Neill, J. P., and McElkeny, I. K. (eds.). Mammalian Cell Mutagenesis: The Maturation of Test Systems. Banbury Report No. 2, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 251-261. Krajewski J, Dobecki M and Gromiec J (1980) Retention of vinyl chloride in the human lung. Br J Ind Med 37:373-374. R-6 CMA 010483 s r' T 1 .I Rucerova M. Polivkova Z and Bacora J (1979) Comparative evaluation of the frequency of chromosomal aberrations and the SCE numbers in peripheral lymphocvtes of workers occupationally exposed to vinyl chloride monomer. Mutar'Res 62:97-100, Kuzmack AM and McGaughy RE (1975) Quantitative Risk Assessment for Community Exposure to Vinvl Chloride. U.S. Environmental Protection Agency, Washington, D.C. Laib RJ and Bolt HM (1977) Formation of imidazole derivatives of nucleic acid bases (DNA and RNA) by metabolites of vinyl chloride in vivo and in vitro. In: Proceedings of Jthe 4th Meeting of the European Association for Cancer Research. Lyon, France, p. 84. Lee CC, Bhandari JC, Winston JM, House WB, Peters PJ, Dixon RL and Woods JS (1977) Inhalation toxicity of vinyl chloride and vinylidene chloride. Environ Health Perspect 21:25-32. Lee CC, Bhandari JC, Winston JM and House WB (1978) Carcinogenicity of vinyl chloride and vinylidene chloride. J Toxicol Environ Health 4:15-30. Lelbach WK and Marsteller HJ (1981) Vinyl-chloride associated disease. In: Advances in Internal Medicine and Pediatrics. Springer-Verlag, Berlin. Leonard A, Decat G, Leonard ED, Leferver MJ, Decuyper LJ and Nicaise C (1977) Cytogenetic investigations on lymphocytes from workers exposed to vinyl chloride. J Toxicol Environ Health 2:1135-1141. Lester D, Greenberg CA and Adams WR (1963) Effects of single and repeated exposure of humans and rats to vinyl chloride. J Am Ind Hyg Assoc 24:265275. Lilis R (1981) Review of the pulmonary effects of poly(vinyl chloride) exposure. Environ Health Perspect 41:167-169. Lilis R, Anderson H, Nicholson WJ, Daum S, Fischbein AS and Selikoff IJ (1975) Prevalence of disease among vinyl chloride and polyvinyl chloride workers. Ann NY Acad Sci 246:23-41. Loprieno N, Barale R, Baroncelli S, Bauer C, Bronzetti G, Camelii A, Cerciangani G, Corsi C and Gervasi-Turcki G (1976) Evaluation of the genetic effects induced by vinyl chloride monomer under mammalian metabolic activation studies is vitro and in vivo. Mutat Res 40:85-96. Maltoni C and Lefemine G (1975) Carcinogenicity assays of vinyl chloride: current results. Ann NY Acad Sci 246:195-224. Maltoni C, Lefemine G, Ciliberti A, Cotti G and Caretti D (1984) Experimental Research__ on Vinvl Chloride Carcinogenesis. In: Maltoni C, Mehlman MA (eds.) Archives of Research on Industrial Carcinogenesis. Vol. II. Princeton Scientific Publishers, Inc., Princeton, New Jersey. CMA 010484 R-7 Marsteller HJ and Lelbac'n VK (1375 ) Unusual splenomegalic liver disease as evidenced by peritoneoscopy and guided liver biopsy among polyvinyl chloride production workers. Ann NY Acad Sci 246:95-134. Masuda Y (1979) Long term mortality study of vinyl chloride and polyvinyl chloride workers in a Japanese plant. Arh Hig Rada Toksikol 30:403-409 (Suppl). McCann J, Simmon V, Streitwieser D and Ames BN (1975) Mutagenicity of chloroacetaldehyde, a possible metabolic product of 1,2-dichloroethane (ethylene chloride), chloroethanol (ethylene chlorohydrin), vinyl chloride, and cyclophosphamide. Proc Natl Acad Sci USA 72:3190-3193. Miller A, Teirstein AS, Chuang M, Selikoff IJ and tfarshaw R (1975) Changes in pulmonary function in workers exposed to vinyl chloride and polyvinyl chloride. Ann NY Acad Sci 246:42-52. Mirkova E, Minailova A and Noska M. (1978) Embryotic and teratogenic action of vinyl chloride, Khigiena i Zdraveopazvane 23:440-443. Mohn GR (1981) Mutagenicity of selected chemicals in Escherichia coli test systems. In: de Serres, F. J., and Shelby, M. D. (eds.) Comparative Chemical__ Mutagenesis.___Environmental__ Science Research. Volume 24, Plenum Press, New York, pp. 69-108. Monson RR, Peters JM and Johnson MN (1974) Proportional mortality among vinyl-chloride workers. Lancetii pp. 397-398. Muller G, Norpoth K, Kusters E, Herweg K and Versin E (1978) Determination of thiodiglycolic acid in urine specimens of vinyl chloride exposed workers, Int Arch Occup Environ Health 41:199-205. Nakamura J (1983) A mortality study of vinyl chloride workers in Japan. J UOEH 1:49-57 (Suppl). 4 Nicholson WJ, Hammond EC, Seidman H and Selikoff IJ (1975) Mortality experience of a cohort of vinyl chloride-polyvinyl chloride workers. Ann NY Acad Sci 246:225-230. Ott MG, Langner RR and Holder BB (1975) Vinyl chloride exposure in a controlled industrial environment. Arch Environ Health 30:333-339. Paddle G (1986) Personal correspondence to A. H. Smith. Perrard MH (1985) Mutagenicity and toxicity of chloroethylene oxide and chloroacetaldehyde. Experientia 41:676-677. Peter S and Ungvary G (1980) Lack of mutagenic effect of vinyl chloride monomer in the mammalian spot test. Mutat Res 77:193-196. Picciano DJ, Flake RE, Gay PC and Kilian DJ (1977) Vinyl chloride cytogenetics. J Occup Med 19:527-530. R-8 CMA 010485 r Probst C-S, McMahon RE, Hill CE, Thompson CZ, Epp JK and Meal SB (1981) Chemically-induced unscheduled DNA synthesis in primary rat hepatocvte cultures: a comparison with bacterial mutagenicity using 218 compounds. Environ Mutagen 3.11-32. Prodan L, Suciu I, Pislaru V, Ilea E and Pascu L (1975a) Experimental acute toxicitv of vinyl chloride (monochloro-ethene). Ann NY Acad Sci 246:154158. Prodan L, Suciu I, Pislaru V, Ilea E and Pascu L (1975b) Experimental chronic poisoning with vinyl chloride. Ann. NY Acad. Sci 246:159-163. Purchase IFH, Richardson CR and Anderson D (1975) Chromosomal and dominant lethal effects of vinyl chloride. Lancet, August 30, 1975. Purchase IF, Stafford J and Paddle GM (1987) Vinyl chloride: an assessment of the risk of occupational exposure. Food Chem Toxicol 25:187-202. Radike MJ, Stemmer KL and Bingham E (1981) Effect of ethanol on vinyl chloride carcinogenesis. Environ Health Perspect 41:59-62. Radike MJ, Warkany J, Stemmer K, and Bingham E (1988) Perinatal carcinogenesis induced by inhaled vinyl chloride. Toxicologist 8:166 (abstract 660), Rinsky RA, Ott G, Ward E, Greenberg H, Halperin W and Leet T (1988) Study of mortality among chemical workers in the Kanawha Valley of West Virginia. Am J. Ind Hyg 13:429-438. Rosenkranz HS (1981) Mutagenicity of selected chemicals in Escherichia coli DNA repair deficient assays. Id: deSerres, F. J., and Shelby, M, D. (eds.), Comparative Chemical Mutagenesis. Environmental Science Research. Volume 24, Plenum Press, New York, pp. 5-18, Rossner P, Sram RJ, Novakova J and Lambl V (1980) Cytogenetic analysis in workers occupationally exposed to vinyl chloride. Mutat Res 73,:425-427. Selikoff IJ and Hammond EC (eds.) (1975) Toxicity of vinyl chloridepolyvinyl chloride, Ann NY Acad Sci 246. Short RD, Minor JC, Winston JM and Lee CC (1977) A dominant lethal study in male rats after repeated exposures to vinyl chloride or vinylidene chloride. J Toxicol 965-968. Smulevich VB, Fedotova IV and Filatova VS (1988) Increasing evidence of the rise of cancer in workers exposed to vinyl choride. Br J Ind Med 45:9397. Sokal JA, Baranski B and Majka J (1980) Experimental studies on the chornic effects of vinyl chloride in rats. J Hyg Epidemiol Microbiol Immunol 24:285-294. R-9 CMA 010486 A;\v l\ Spengier S and Singer B (1981) Transcriptional errors and ambiguity 64 resulting from the presence of 1,N -ethenoadenosine or 3,N -ethenocytidine in polyribonucleotides. Nucleic Acids Res 9:365-373. Spirtas R and Kaminski R (1978) Angiosarcoma of the liver in vinyl chloride/polyvinyl chloride workers. Update of the NIOSH Register. J Occup Med 20:427-429. Spirtas R, McMichael AJ, Garuble J and VanErt M (1975) The association of vinyl chloride with morbidity symptoms. Am Ind Hyg Assoc J 16:779-789. Stafford J (1983) Liver angiosarcoma case registry, ICI, (cited in Maltoni et al., 1984). Stanford Research Institute (SRI) (1983) Vinyl chloride. Xq: Monographs on Organic Air Pollutants. Final Report, prepared by SRI International for the National Cancer Institute, contract No. N01-CP-26004-Q2, pp. 375-395. Styles JA (1980) Studies on the detection of carcinogens using a mammalian cell transformation assay with liver homogenate activation. Xq: Norpoth KH and Garner RC (eds.), Shall__ Term__ Test Systems for Detecting Carcinogens. Springer-Verlag, New York, pp. 226-238. Suciu I, Prodan L, Ilea E, Paduraru A and Pascu L (1975) Clinical" manifestations in vinyl chloride poisoning. Ann NY Acad Sci 246:53-69 Suskov II and Sazonova LA (1982) Cytogenetic effects phenolformaldehyde and polyvinylchloride resins in man, 104:137-140. of epoxy, Mutat Res Suzuki Y (1980) Non-neoplastic effects of vinyl chloride in mouse lung. Environ Res 21:235-253. Su2uki Y (1981a) Electron microscopic observations of hepatic and subcutaneous hemangiosarcomas induced in mice exposed to vinyl chloride monomer. Am J Ind Med 2:103-117, Suzuki Y (1981b) Neoplastic and nonneoplastic effects of vinyl chloride in mouse lung. Environ Health Perspect 41:31-52. Tabershaw IR and Gaffey WR (1974) Mortality study of workers in the manufacture of vinyl chloride and its polymers. J Occup Med 16:509-518. Tassignon JP (1980) Methodologie du traitement des donnAes de 1'etude du Profeseur Maltoni. Report to the XX Meeting of "Le Club de Cancferogfenfese chimiaue" (Institut der Radium-Fondation Curie, Paris, 1979). Publications Essestielles, pp. 113-125. Testa B and Jenner P (1981) Inhibitors of cytochrome P-450s and their mechanism of action. Drug Metab Rev 12:1-117. Theriault G and Allard P (1981) Cancer mortality of a group of Canadian workers exposed to vinyl chloride monomer. J Occup Med 23:671-676. R-10 CMA 010487 Theriault G, Iturra H and Gingras S (1983) Evaluation of the association between birth defects and exposure to ambient vinyl chloride. Teratology 22:359-370, Thomas 13 and Popper H (1975) Pathology of angiosarcoma of the liver among vinyl chloride-polyvinyl chloride workers. Ann NY Acad Sci 246:268-277. Til HP, Inane1 HR and Feron VJ (1983) Lifespan Oral Carcinogenicity Study of 7invl Chloride in Rats. Prepared by the Netherlands Organization for Applied Scientific Research, Division for Nutrition and Food Research (CIVO Institute TNO) Report No. V 83.285/291099. Torkelson MS, Oyen F and Rowe VK (1961) The toxicity of vinyl chloride as determined by repeated exposure of laboratory animals. J Am Ind Hyg Assoc 22:354-361. Torkelson TR and Rowe VK (1981) Halogenated aliphatic hydrocarbons containing chlorine bromine and iodine. In: Patty's Industrial Hygiene and Toxicology. 3rd Revised Edition, vol. 2B, Chapter 48, pp. 3433-3601. Tu AS, Murray TA, Hatch KM, Sivak A and Milraan HA (1985) In vitro transformation of BALB/C-3T3 cells by chlorinated ethanes and ethylenes. Cancer Lett 28:85-92. Ungvary G, Hudik A, Tratui E, Lorincz M and Folly G (1978) Effects of vinyl chloride exposure alone and in combination with Trypan Blue--applied systematically during all thirds of pregnancy on the fetuses of CFy rats. Toxicology 11:45-54. Veltman G, Lange CE, Juhe S, Stein G and Bachner U (1975) Clinical manifestations and course of vinyl chloride disease. Part I. Toxicological investigation of vinyl chloride-polyvinyl chloride. Ann NY Acad Sci 246:7-17. Verburgt FG and Vogel E (1977) Vinyl chloride mutagenesis in Drosophila melanoeaster. Mutat Res 48:327-336. Viola PL, Bigotti A and Caputo A (1971) Oncogenic response of rat skin, lungs, and bones to vinyl chloride. Cancer Res 31:516-522. Wagoner JK, Infante PF (1980) A review of the methodological approaches in the assessment of an association between vinyl chloride exposure and reproductive hazards. In Infante PF, Legator MS (1980) "Proceedings of a Workshop on Methodology for Assessing Reproductive Hazards in the Work place "NIOSH-DHHS Publication No. 81-100, Washington, DC, pp. 43-49. Ward AM, Udnoon S, Watkins J, Walker AE and Darke CS (1976) Immunological mechanisms in the pathogenesis of vinyl chloride disease. Br Med J 1:936938. Watanabe PG, McGowan GR, Madrid EO and Gehring PJ (1976a) Fate of (^C)vinyl chloride following inhalation exposure in rats. Toxicol Appl Pharmacol 37:49-59. R-ll CMA 010488 DRAFT Watanabe PG, McGowan GR and Gehring PJ (1976b) Fate of (^C) vinyl chloride after single oral administration in rats. Toxicol Appl Pharmacol 36:339- Watanabe PG, Zempel JA and Gehring PJ (1978a) Comparison of the fate of vinyl chloride following single and repeated exposure in rats. Toxicol Appl Pharmacol 44:391-399. Watanabe PG, Zempel JA, Pegg DG and Gehring PJ (1978b) Hepatic macromolecular binding following exposure to vinyl chloride, Toxicol Appl Pharmacol 44:571-579. Waxweiler RJ, Stringer W, Wagoner JKC, Falk H and Jones J (1976) Neoplastic risk among workers exposed to vinyl chloride. Ann NY Acad Sci 271:40- 48. Weber H, Reinl W and Greiser E (1981) German investigation on morbidity and mortaility of workers exposed to vinyl chloride. Environ Health Perspect 41:95-99. Wilson WH, McCormick WE and Tatum CF (1967) Occupational acroosteolysis: report of 31 cases. J Am Med Assoc 201:83-87. Withey JR (1976) Pharmacodynamics and uptake of vinyl chloride monomer administered by various routes to rats. J Toxicol Environ Health 1:381394. Woo YT, Lai DY, Arcos JC and Argus MF (1985) Chemical Induction of Cancer. Structural Bases and Biological Mechanisms, volume IIIB, Aliphatic and Polyhalogenated Carcinogens. Academic Press, Inc., pp. 3-129. Zuccato E, Marcuci F and Mussiui E (1980) The role of respiration in vinyl chloride excretion in rats. Toxicol Lett 1:213-217. R-12 CMA 010489 DRAFT Appendix A Abstracts of Maltoni et al. (1984) Bioassays CMA 010490 A-1 Intranetitoneal Administration Rats: Groups of 30 male and 30 female 13-week-old Sprague-Dawley rats received an intraperitoneal injection of 4.25 mg vinyl chloride in olive oil on 1, 2, 3, or 4 occasions over a two-month period and observed for the duration of their lives (145 weeks). One nephroblastoma and one subcutaneous angiosarcoma were found. No difference in survival or body weight was observed between test animals and controls. This experiment was considered inadequate for the determination of the carcinogenic potential of vinyl chloride because of the unconventional dosing protocol used (Experiment BT12, Maltoni et al., 1984). Subcutaneous Administration Rats: In a separate study, a group of 75 male and female Sprague-Dawley rats was administered a single subcutaneous injection of 4.5 mg vinyl chloride in 1 ml olive oil at 21 weeks of age and observed for the remainder of their lifetime (145 weeks after injection). Body weight and survival were not significantly different between controls and treated animals. One nephroblastoma in a treated male was observed (Experiment BT13, Maltoni et al., 1984). The insufficient protocol prevents any assessment of the carcinogenicity of vinyl chloride from this experiment. Transplacental Exposure Rats: Groups of pregnant female Sprague-Dawley rats were exposed from day 12 to day 18 of gestation to 6,000 or 10,000 ppm vinyl chloride. The A-2 CMa OlO-49i females and offspring were observed for "heir lifetimes (143 weeks after start pf experiment). Survival of the offspring was poor after week 95 of the experiment. Several animals from both groups exposed in utero had mammary tumors, zymbal gland carcinoma, leukemias and nephroblastomas; no hepatic angiosarcomas or hepatomas were reported. No results from control animals were reported, thus statistical evaluation of these results is not possible. Only a few tumors were found in the female breeders (Experiment BT5, Maltoni et al., 1984; IARC, 1979), Transplacental-Inhalation Exposure Rats: Groups of 12-week-old pregnant Sprague-Dawley rats were exposed to either 0 or 2,500 ppm vinyl chloride four hours/day, five days/week for seven weeks, then seven hours/day for 69 weeks, after which time all animals died. One group of offspring was first exposed transplacentally from day 12 of gestation, then exposed by inhalation after birth using the same protocol. A second group of offspring was also exposed transplacentally from day 12 of gestation but was exposed by inhalation four hours/day, five days/week for seven weeks, then seven hours/day, five days/week for eight weeks. Vinyl chloride was toxic at all concentrations tested: all animals exposed to vinyl chloride for 76 weeks died by that time, whereas the control animals survived for up to 150 weeks. The poor survival of treated animals almost certainly diminished the number of observed tumors, especially tumors with long latency periods, such as liver angiosarcomas. An increased incidence of zymbal gland tumors (8/54), liver angiosarcomas (27/54), hepatomas (5/54), and neuroblastomas (32/54) were reported for the CA 010492 A- 3 breeding females exposed to vinyl chloride, compared to 1/60, 0,60, 0/60, 1/60. respectively, in Che controls. In the male offspring exposed to vinyl chloride for 76 weeks, 9/63 had zymbal gland carcinomas, 36/63 had liver angiosarcomas, 27/63 had hepatomas, and 31/63 had neuroblastomas, compared to 2/158, 0/158, 1/158, and 0/158, respectively, in the controls. In the female offspring exposed to vinyl chloride for 76 weeks, 6/64, 28/64, 38/63, and 28/64 were reported for these above tumors respectively compared to zero tumor incidence in the controls. The incidence of these same tumors in the male offspring exposed to vinyl chloride for only 15 weeks was 7/59, 24/59, 42/59, and 7/59 for the same tumors respectively, compared to 2/158, 0/158, 1/158, and 0/158, respectively, in the controls. In female offspring exposed for only 15 weeks, the incidence was 2/60, 28/60, 43/60, and 11/60 for the same tumors, respectively, compared to a zero incidence of these tumors in controls. These studies (BT4001, BT4006) were cited by Maltoni and colleagues (1984) as an example of transplacentally*induced-tumorigenesis, but was, in effect, an investigation of the increased sensitivity of young experimental animals to the toxic effects of vinyl chloride. The tumor incidence in breeders and offspring exposed to vinyl chloride for 76 weeks did not appear to differ significantly, nor did the increased tumor incidence in offspring exposed to vinyl chloride for 15 weeks appear to differ substantially from the tumor incidence in exposed breeders. However, no explicit statistical comparison of these parameters was made in the report (Maltoni et al., 1984; Experiments BT4001, BT4006). A-4 `Ol93 Inhalation Exposure Hamsters: Groups of 30 male Syrian golden hamsters were exposed to 0, 50, 250, 500, 2,500, 6,000, or 10,000 ppm vinyl chloride, four hours daily, five days weekly for 30 weeks, beginning at 11 weeks of age. The hamsters were then observed for their lifespan (109 weeks). Two liver angiosarcomas were observed in the group exposed to 500 ppm vinyl chloride and one liver angiosarcoma was observed in the group exposed to 6,000 ppm. The increased incidence of forestomach epithelial tumors in hamsters exposed to 500 ppm or more of vinyl chloride appeared to be biologically significant but no statistics were reported (Experiment BT8, Maltoni et al., 1984). A-5 CMA 010494 CMA 010495 DRAFT Appendix B Cancer Risk Assessment for Vinyl Chloride Based On Human Data B-l CMA 010496 introduction The main problem with using epidemiological data for health risk assessment is the frequent lack of suitable quantitative exposure data. Exposure data from animal bioassays are, by comparison, markedly more accurate. However, a major source of potential error in using animal data for human health risk assessment is the animal-to-human extrapolation of risk. It follows that the decision to use epidemiological data in health-based risk assessments should not be based solely on contrasting the quality of exposure data available for humans with that for animal studies, but should also take into account other important variables such as the observed tumor type and the latency period of the tumors under study. In fact, the decision concerning use of human versus animal data in a health-based risk assessment should take into account the main source of error from each approach: the exposure data for human studies and the animal-to-human extrapolation for animal studies. In the latter case, one would hope that the error in extrapolation to humans was within one order of magnitude. However, the adjustments made to correct for body surface area differences (when extrapolating from animal bioassays to humans) can in itself cause an order of magnitude difference in estimates when compared to estimates which use a simple dose per body weight extrapolation. Since other sources of error exist, the uncertainty of animal-to-human extrapolation probably lies within one or two orders of magnitude, except in the case where there may be no effect in an animal model but an effect in humans (or vice versa). In this case, the error would be infinite. B-2 CMA 010497 'Then human studies show a clear relationship between an outcome and an exposure. the question concerning their use in risk assessment should therefore be whether or not one can estimate exposure within one to two orders of magnitude. If sound human studies that enable exposure estimation within one to two orders of magnitude are available, then these studies should be utilized in addition to animal studies in evaluating the human health risk. In the case of vinyl chloride, there are unequivocal epidemiological data that identify vinyl chloride as a human carcinogen. Although detailed exposure data for the relevant time periods are lacking, estimates can be made that should be well within one order of magnitude of error. While the methods for estimating past exposure are crude, the true exposure rates are considered unlikely to be more, than five times higher or lower than estimated. It follows that the human data may be appropriate for vinyl chloride cancer risk assessment. Identifying the Most Appropriate Study for Use in Vinvl Chloride Risk Assessment A literature review was undertaken to assess available information regarding the risk to humans posed by exposure to vinyl chloride. The results of this review, summarized in Tables B-l and B-2, suggest that the cohort study reported by Waxweiler and co-workers contains the most thoroughly documented information for risk assessment purposes (Waxweiler et al., 1976). Eleven cases of angiosarcoma of the liver were identified among the 1,294 exposed workers. Thre~ additional cass of biliary cancer were seen. Significant B-3 CrtA 010498 excesses or both brain cancer and lung cancer were also observed. The Waxweiler report gives SMRs for liver cancer, brain cancer, and lung cancer. It is apparent from Table B-2 that the SMRs from Waxweiler et al. are consistent with some of the other studies. The cumulative risk of liver, lung and brain cancer following vinyl chloride exposure is greatest in the Waxweiler et al. (1976) report. Thus cancer risks to vinyl chloride workers are unlikely to be substantially underestimated by a risk assessment based on this study. The results reported for the Waxweiler cohort of polyvinyl chloride (PVC) workers involved a subset who had worked for at least five years between 1942 and 1973 and who had commenced work at least ten years before follow-up was completed. Follow-up for mortality was to the end of 1973. This subset of the cohort was comprised of 1294 workers. There were 136 deaths during the follow-up period of which 35 were due to cancer. The SMR for biliary and liver cancer was 1155, for brain cancer, 329, and for lung cancer, 156. The remainder of this appendix discusses the risk of liver, lung, and brain cancer for the workers in the Waxweiler et al. (1976) cohort. In addition, assessments evaluating the risks for each cancer have been recalculated, using the lifetime exposure estimations and assumptions described herein. Exposure Data As in many retrospective cohorts, individual exposure data were not available (Waxweiler et al., 1976). However, several reports have attempted to reconstruct the magnitude of exposure among vinyl chloride workers since B-4 CMft 010499 tine 1940's (Oct et al, , 1976; Jones, 1974; Paddle, 1986). Table 3-3 summarizes proposed estimates of exposure for several countries. Most of the data available for specific United States exposure levels utilize measurements determined at a single plant operated by Dow Chemical Company (Jones, 1974), Although some job class exposures were quite high, most exposures were less than proposed international levels during commensurate time periods. However, Dow Chemical Company responded to early reports of vinyl chloride toxicity in animal studies by creating an in-house standard of 50 ppm (Ott et al., 1974; Paddle, 1986). This was well below the acceptable limit during the 1960's and early 1970's and probably does not represent the average exposure at other vinyl chloride polymerization plants. Dow Chemical Company had not reported any cases of angiosarcoma of the liver to 1985 (Forman et al., 1985). Other exposure estimates, such as those presented by Barnes and summarized in Tables B-3 and B-5 of this appendix, may better describe average exposure for the Waxweiler et al. cohort (Barnes, 1976). Data were not given by Barnes to substantiate his exposure estimates, but rather he qualifies his estimates by stating: "the general consensus of opinion throughout the world, today, is that average atmospheric exposure for polymerization workers between 1940 and 1970 might have been of the following order" (Table B-5) (Barnes, 1976). The Barnes estimates approach the existing standards during these time periods. Additionally, the exposure experience among cohorts with angiosarcoma of the liver should be higher when compared to exposure levels in cohorts reporting no angiosarcomas. Consequently, exposure estimates provided by Barnes are used to quantify cumulative vinyl B-5 CMrt 010500 chloride exposure for the risk assessment since the group studied by Taxweiler et al. (1976) was presumably one of the most heavily exposed cohorts and therefore, unlikely to involve exposures lower than those given bv Barnes. The employment dates of the Waxweiler et al. cohort span 1942-1973. Thus, some work histories fell below the boundary of the first time period provided by Barnes (January 1, 1945). These particular histories were counted separately and assigned an exposure level of 1000 ppm. This practice assumes exposure during early process days (pre-1945) to be the same as that during the 1945-1955 exposure period. Methods For Using Average Cohort Exposure Data Data for individual exposure estimates for the Waxweiler et al. occupational cohort do not exist. However, Appendix C of this document gives a method for incorporating average exposure data into a risk assessment analysis used in the following manner. Deaths due to angiosarcoma of the liver occurred between 1964 and 1973. The latency period between vinyl chloride exposure and deaths attributed to angiosarcoma of the liver ranged from five to twenty years (Smith et al., 1980). More recently, Stafford (1983) has calculated the latency period for angiosarcoma of the liver among vinyl chloride workers worldwide to be 22.1 years. The work histories for the cohort were analyzed to identify the person-time in each calendar year for the subset who had at least five years of employment and who began work (and thus vinyl chloride exposure) prior to 1964. These restrictions were incorporated to correspond to the same restrictions used by Waxweiler and B-6 CMA 010501 co-vorkers (1976) in generating their SMR values. Thus, the SMR values correspond to those workers with at least five years of exposure and at least a ten-year latency period from first exposure. The analysis herein assumes that the average worktime cohort exposure for the period 1942 to 1964 was a time-weighted average of 647 ppm of vinyl chloride. The plant had begun operations in 1942. The year 1964 was chosen as an endpoint since 1969 was the midpoint of diagnosis of liver angiosarcoma cases and since exposures in the last five years preceding diagnosis were probably not relevant to etiology (Smith et al., 1980). It is assumed that this exposure rate was experienced for eight hours per day, five days per week, 46 weeks per year, and that, outside these times, the exposure to vinyl chloride was the same as for the general population. Thus, the average excess exposure per day is the average worktime exposure multiplied by (1/3) x (5/7). x (46/52) which equals 0.211. The average overall exposure rate above background for vinyl chloride workers was estimated to be 0.211 x 647 ppm or 136 ppm of vinyl chloride. Lifetime Exposure Estimations and Risk of Liver Cancer The above estimates relate to workplace exposures occurring over a relatively short period of time. The SMR for liver cancer was 1155. The average duration of employment for this cohort between 1942 and 1964 was 11.6 years and the average age at first exposure was 29.7 years. Based on the assumptions stated in Appendix C, the adjustment used to extrapolate to lifetime exposure is duration of employment/(age + duration of employment), CMA 010502 B-7 1 \j f\ ni \ -- or (29.- 11,6). Ihererore, the average lifetime exposure estimated zo resui a liver cancer SMR of 1155 is [(136 ppm) (0.281)] or 38.3 ppm. 3etween 1960 and 1979, 67,782 deaths from liver cancer occurred among white males in the United States (International Classification of Diseases (ICD) codes 155,156). The total number of deaths among the same group was approximately 18,297,297, Thus, one in approximately 270 deaths was from liver cancer. Since the workers under discussion in the Waxweiler et al. cohort had an SMR of 1155, their lifetime risk of dying from liver cancer would increase above the background rate of one in 270. An SMR of 1155 is equivalent to a relative risk of 11.55 (Symons and Taulbee, 1981). The lifetime risk of dying from liver cancer calculated herein for workers exposed to vinyl chloride is thus the relative risk multiplied by the background lifetime risk or 11.55 x 1/270. The added lifetime risk would be the excess relative risk times the background lifetime risk, or 10.55 x 1/270 - 1/25.6. Thus, the added lifetime risk of liver cancer attributable to an exposure rate causing an SMR of 1155 would be one in 25.6 or that lifetime exposure to 38.3 ppm of vinyl chloride would cause a 39,062 in a million increase in the lifetime risk of death from liver cancer. Cancer Risk Scenarios Including Brain__ and__ Lvng Cancer AS Well,,33 Live? The same approach can be adopted for brain cancer and lung cancer. Evidence for their relationship to vinyl chloride exposure is discussed in the next B-8 ^ Ol0503 Ctt* 01 sections. In the absence of evidence to the contrary, it is assumed that the latency period is the same as for angiosarcoma of the liver. The number of deaths from brain cancer (ICD codes 191 & 192) between 1960 and 1979 in the United States was 79,847 (1/229 of deaths), and for lung cancer (ICD codes 160-163, 165) 978,504 (1/18.7 of deaths). Applying the same procedure indicated above, the added brain cancer lifetime risk was one in 100 and, for lung cancer, was one in 33.4, assuming the average exposure rate of 38.3 ppm. Brain and Lung Cancer and VinvLChloride Exposure While it is clear that exposure to vinyl chloride causes angiosarcoma of the liver, the causal relationship to brain and lung cancer is not so welldefined. One review suggested that there was a consistent relationship with brain cancer in occupational, studies, but not with lung cancer (Beaumont and Breslow, 1981). However, it would seem appropriate to consider lung cancer in the risk assessment along with liver and brain cancer, since this is consistent with a conservative approach and the relationship with lung cancer cannot be rejected out of hand. In fact, the report referenced above focused on statistical power independent of the degree of exposure experienced by the various cohorts reviewed. The lung cancer findings become more consistent when considered in conjunction with the liver cancer excess experienced by each cohort. Since excesses of liver cancer can be used as a surrogate indicator of exposure, this suggests that some studies not finding an excess of lung cancer may have been a result of relatively low exposures. B-9 CMA 010504 However. a more recent large study presents evidence against a relationsnip between lung cancer and vinyl chloride exposure (EHA, 1985). This study considered deaths between 1942 and 1982 inclusive for a cohort of 10,173 men who had worked for at least one year in jobs involving exposure to vinyl chloride. The SMR for liver cancer was 641, for brain cancer was 180, but for lung cancer was only 95.8. Most of the liver and brain cancer excess was in two of the 37 plants forming the cohort. Unfortunately, lung cancer SMRs were not presented for these two plants. In spite of this, the study provides evidence against a vinyl chloride-lung cancer association. However, without lung cancer data for the two plants with the highest liver and brain cancer excesses, it would seem inadvisable to completely exclude lung cancer from the risk assessment. Confidence.. Limits for the Lifetime Risk Estimates Confidence limits for the risk estimates are calculated by combining the risks for tumor development for each site (by summing observed and expected values for each site) and then calculating the 95% confidence limits of that single point estimate assuming a Poisson distribution. The 95% confidence interval for the liver cancer SMR is (467-2404). To estimate the upper 95% confidence limit for the excess risk estimate, the upper limit of excess risk (24.04 - 1 - 23.04) is multiplied by the lifetime risk for the average person of dying from liver cancer (1/270). Therefore, the upper 95% confidence limit for the added risk due to vinyl chloride exposure is 23.04 x 1/270 - 0.085 or 1/11.7. B-10 CHA 010505 The same calculation is used to estimate the upper 95% confidence limit based on liver cancer and brain cancer combined. In this case, the combined observed and expected values for liver and brain cancer (7+3)/(0.6+0.9) results in 95% confidence interval for the SMR of (319 - 1226). Thus, the upper 95% confidence limit for the excess risk estimate is (12.261)(1/270+1/229), or 1/11.0. For liver cancer, brain cancer, and lung cancer combined, the observed to expected ratio is 22/9.2 and the 95% confidence interval for the SMR is (149 362). Using the same strategy, the upper 95% confidence limit for the estimate of added risk is (3.62 - 1)(1/27Q + 1/229 + 1/18.7), or 1/6.20. Extrapolation of Risk to Low Dose Exposure There are many models for extrapolating risks to low dose exposure. The method of analysis employed here gives only one exposure point and therefore limits the models that may be used. A linear extrapolation of excess risk was chosen as the most appropriate for this analysis. This approach is very close to a one-hit model extrapolation. In turn, the one-hit model extrapolation is very close to a multistage extrapolation with linearization as recommended by the U.S. Environmental Protection Agency (EPA) Carcinogen Assessment Group for use with animal data. Thus, a simple linear extrapolation would provide similar results to the more complex multistage model approach that could have been used with more extensive data. As indicated in the above section entitled "Lifetime Exposure Estimations and Risk of Liver Cancer", lifetime exposure to 38.3 ppm of vinyl chloride B-11 CMA 010506 Lb < A r T is calculated to increase lifetime liver cancer risk by one in 25.6, or 39,062 in a million. The corresponding unit risk estimate for one in a million added risk is 0.98 ppb. Table B-4 gives corresponding results for the addition of brain cancer and lung cancer and the 95% confidence limits on unit risk exposures. Assumptions and Uncertainties The confidence limits that were calculated for the risk estimates measure only the uncertainty related to the SMR statistics for workers and do not measure the uncertainty of the risk assessment process overall. This risk assessment is based on specific assumptions which, if incorrect, affect the assessment by either overstating or understating the true risk. These assumptions are listed below. 1. Assumptions are made concerning the exposure estimates. This can affect the accuracy of the risk estimates in either direction. 2. The relationship between excess relative risk and lifetime average exposure rate is assumed to be linear. If the relationship is better described by a supralinear curve, then a linear assumption will understate the risk. Conversely, if the relationship is better described by a sublinear curve, then a linear assumption will overstate the risk. 3. It was assumed that cancer risks were dependent on cumulative exposure and not on exposure rate. A given cumulative exposure achieved as an B-12 010507 adult, is assumed, to carry the cancer risk equal to the same cumulative exposure starting at birth. 4. It was assumed that relative risk was dependent only on cumulative exposure and not on age. 5. Based on the pattern of excess exposure for this cohort (Smith et al., 1980), it was assumed that the dose accumulated five years prior to death was not relevant to causation of cancer. 6. The SMRs used were calculated using United States general population cancer rates. If national cancer rates were higher than local rates, the value of the SMR is underestimated, and vice versa. 7. It is assumed that lung cancer and brain cancer are causally associated with vinyl chloride exposure and that the dose accumulated in the five years immediately prior to death was not relevant to causation of cancer. If these cancers are not associated with exposure to vinyl chloride, then the true risk is overstated by including them in the analysis. 8. It is assumed that the effect of a given cumulative exposure is the same in men and women. Conclusions This risk assessment analysis suggests that an average exposure to 38.3 ppm of vinyl chloride may result in an added lifetime cancer risk of 1/25.6 for DR, liver cancer, 1/100 for brain cancer, and 1/33.4 for lung cancer, assuming each cancer is related to vinyl chloride exposure. If one adopts a linear extrapolation approach, one would conclude that an exposure to 0.485 ppb for a lifetime might result in a cancer risk of one in a million, if all these cancers are related to exposure. In the most likely case that only liver and brain cancer are related to exposure, an exposure to 0.781 ppb might result in a one in a million lifetime cancer risk. B-14 CMA 010509 TABLE B1 COHORT CHARACTERISTICS Of SELECTED VIHYL CHLWUOE STUDIES Author i Byren et at., 1976 Cooper, 1981*+ Duck et al., 1975 fok et al,, 1977 Keldaac et al., 1984 Nonson et al., 1975** Cohort Site 750 10173 Total Ho. of Deatha 38 707 Nuaiber of Cancer Deatha 11 139 Range of Expoeure Duration Minima Maxima >0 yr 15Z > lOyre > 1 yr 23 yra 2120 136 35 >0 yr > 15 yre 7)61 *54 3)3 50 161 115 >0 yr 23X > 10 yra. OX > 20 yra 23 > 1 yr 35X > 5 yTs 41 Wicholson et at., 1975 Tabershaw et al., 1974*+ 2)7 030* 24 352 9 ------------------(not given)---------------------- 79 > 1 yr >30 yra Length of Longeat Pol low-up 1940's 1974 19401s - 1972 No tec on Fu 1 low-up 97t 95X 1948 - 1974 99.6X 1940 - 1974 99. IX 1953 - 1979 1946 - 1974 1947 - 1973 1930's - 1972 99. JX 851 lakaaura, 1983 4)2* 209 37 > 1 yr >15 yra 1950 - 1975 Place of Company 1nvolved Sweden 3T Plants in United St.ites South Uiilrt Great Bi i t.i i n Louisvi lie,, Kentucky New Yoi k 33 Plants In United States Japan Theriault et at., 1981 451 59 Uaxweller et at., 1976*+ Weber et at., 1981 1207 7021 136 *14 Wong et al., 1986++ 10173 1336 Proportional mortality study Overlapping cohorts of Goodrich ccmpany workers 20 > 5 yra > 30 yra 35 > 5 yra 27 yra 94 >0 yr > 10 yra 359 >1 yr 30 yr , v 1943 1974 1940*s - 1973 1940's 1974 1942 - 1982 SIX > 15 yra, Canada 23X > 29 yra 99. SX 4 Plants in United States 90X West GCf m,ny 92X Uni led Si ii llKAl 1 1I yo CMA 010510 --n H TABLE E 2 Author Byren et al., 1976 Cooper, 1981* Duck et al., 197S Fox et al., 1977 STANDARD MORTALITY RATIOS (AND 90S CONFIDENCE INTERVALS) FOR SELECTED VINYL CHLORIDE STUDIES Liver 0 E StiK 90% C.l. liver end pancreas 4 0.97 41) (140.), 941.9) Number of Liver Angiosarcomas 2 Lung 90% C 0 K SKR ) 1.78 166 (45.5. 435.1) digesttva 29 40.0 7) (50.8, 96.9) respiratory a 25 23.9 107 (72.7. 146.1) digestive II 11.09 99 (55.6, 164.2) 0 16 15.5 103 (64.7, 156.8) Brain (CNS) 0 E SHR 90% C.J. 2 0.33 612 (104.6, 1904.5) 12 5.9 203 (117.3, 329.5) -- - - 4 2 46 51.23 89.8 (69.2, 114.8) 2 3.66 54.6 (9.4, 171.7) Heldaas et al., 1964 Monson et al., 1975* 17 biliary and liver 0.7 1100 (568.3, 2061.5) 1 5 5 2.84 180 (69.2, 370.0) - - - - 13 7.9 160 (97.3, 261.6) 5 1.2 420 (163.8, 675.6) Nicholson et al., 1975 Tabershaw et at., 1974* ) 0.12 2 500 digestive 19 21.67 94 (675.2, 6454.2) (57.4, 128.6) 3 6 0 1.1 0 respiratory 25 23.93 112 (72.6, 145.9) 1 0.1 1000 (39.5, 4728.4) -- -- Takaaura, 1981 Theriault et al., 1961 Uaxweiler et al., 1976* 6 2.54 2)6 digestive 14 5.4 259 biliary 6 liver 7 0.6 1155 (102.7, 466.0) (156.7, 405.3) (547.0, 2190.6) 1 8 11 2 2.3) 86 (14.8. 269.7) rcaplratory 2 5.78 34.6 ( 6.0, 108.7) rcaplratory 12 7.7 156 (#9.9, 252.5) Ueber et al., 1901 12 0.79 1523 (876.3, 2460.7) 4 -- - - Nong et al., 1966* liver end billlary 37 5.77 641.2 (478.2, 843.6) 115] 115 122 94.2 (80.3, 110) `Overlapping cohorts of Goodrich company workers. ( All confirmed cases, irrespective of meeting the five-year exposure and ten-year latency criteria of this cohort study 90% C.l. - 990^%^coonfidence interval CMrt 010511 -- -- 0 0.6 0 3 0.9 329 (90.0, 860.6) (28.1, 511.0>C 2 1.23 162 3 23 12.76 180 ( 123.2, 255. if M TABLE B 3 HISTORIC EXPOSURE IEVELS Cppa): VINTi CHLORIDE hull- l>w - Hen me HuIm ttlvMkt* Cminntln ACCII lairlcii Owlinm of CmriHit Industrial Hyglaalata OSU OccaRKlooat Safety w4 Health Adalalatratloa m tlaa walghted avaraga (la ppo) keaad oa aaeight-hour day ppm m part par all Hon range TWA blgbaat and lovaat TWA reported t TWA average TUA for ghat period peaka htgbaat content rat toe reported (ppa). Ttiere la little documentation regarding frequency or duration of peaks reactor area work area eaeuad aed within tha veaael where aolynerltatlon took place scraper' hands = after polynerization; workers entered vessel to scrape any buildup of polymer off the reactor vessel walls; measurements represented this exposure p[eclpitor, centrifus* and drying ov.ns - post polymarlration work srsas CHA 010512 la id ppa zu TABLE 3-4 EXPOSURE LEVELS ASSOCIATED WITH ONE IN A MILLION LIFETIME CANCER RISKS Cancer Site Liver Liver and Brain Liver, Lung and Brain Mas.t Likelv 0.981 ppb 0.781 ppb 0.485 ppb Upper 95% __ Limit 2.82 ppb 2.17 ppb 1.27 ppb Lower 95% __ Limit 0.448 ppb 0.421 ppb 0.238 ppb B-18 OlOSl TABLE B-5 BARNES' EXPOSURE ESTIMATES1 Time Period 1945-1955 1955-1960 1960-1970 Mid 1973 1975 Exposure Estimate 1000 400-500 300-400 150 5 1 Barnes, 1976. B-19 CHA 010514 TABLE 3-6' HISTORICAL EVOLUTION OF OCCUPATIONAL EXPOSURE LIMITS I, 2 Year 1954 1962 1971 1972 1974 Authoritv MAC 3 ACGIH3 OSHA3 OSHA OSHA Late 1974 OSHA April 1975 OSHA five for 15 minutes Vinyl Chloride Limit ('ppm') 500 500 500 200 50 - temporary emergency standard over an eight hour period Proposed non-detectable limit 1 - averaged over eight hour period with a maximum of 3From Paddle Correspondence (1986) 2 As a point of interest, Table F presents a summary of historical occupational standards for vinyl chloride. 3 MAC Maximum Allowable Concentration; ACGIH - American Conference of Government Industrial Hygienists; OSHA - Occupational Safety and Health Administration B-20 CMA 010515 REFERENCES DRA 3arnes AW (1976) Vinyl chloride and the production of PVC. Proc Roy Soc Med 69:278-281. Seaumont JJ and Breslow NE (1981) Power considerations in epidemiologic studies of vinyl chloride workers. Am J Epidemiol 114:723-34. Cooper WC (1981) Epidemiologic study of vinyl chloride workers: mortality through December 31, 1972. Environ Health Perspect 41:101-106. Environmental Health Associates (1986) An update of an epidemiological study of vinyl chloride workers, 1942-1982. Report for Chemical Manufacturers Association, pp. 1-37. Forman D, Bennett B, Stafford J and Doll R (1985) Exposure to vinyl chloride and angiosarcoma of the liver: a report of the register of cases. Br J Ind Med 42:750-3. Jones J (1974) Worker exposure to vinyl chloride during production and fabrication of vinyl chloride and polyvinyl chloride. NIQSH Report Contract CDC-99-74-59, pp. 2-161. Monson RR, Peters JM and Johnson MN (1974) Proportional mortality among vinyl chloride workers. Lancet pp.397-398. Ott MG, Langner RR and Holder BB (1975) Vinyl chloride exposure in a controlled industrial environment. Arch Environ Health 30:333-339. Paddle G (1986) Personal Communication to A.H. Smith. Smith AH, Waxweiler RJ and Tyroler HA (1980) Epidemiologic investigation of occupational carcinogenesis using a serially additive expected dose model. Am J Epidemiol 112:787-797. Symons MJ and Taulbee JD (1981) Practical considerations for approximating relative risk by the standard mortality ratio. J Occup Med 23:413-416. Tabershaw IR and Gaffey WR (1974) Mortality study of workers in the manufacture of vinyl chloride and its polymers. J Occup Med 16:509-518. Waxweiler RJ, Stringer W, Wagoner JKC, Falk H and Jones J (1976) Neoplastic risk among workers exposed to vinyl chloride. Ann NY Acad Sci 271:40-48. 8-21 CMA 010516 DRAFfr CHA 010517 DRAFT Agpew^a-S Using The Average Exposure Race of a Cohort In Risk Assessment Analysis C-l CHft 010518 Intr^duc cion DRAS Individual exposure data that could be used for health risk assessment are frequently not recorded in occupational health cohort studies. However, a reasonable estimate of health risk may be found in the average exposure that the cohort experienced in certain time periods. This appendix presents a method for using such data for risk assessment by employing certain assumptions. Objectives in Risk Assessment The primary objective behind health risk assessment is to contribute to risk _ management decision-making, including regulating permissible exposure levels. Workplace exposures and general environmental exposures are two major areas considered. In the case of general environmental exposures, the underlying goal is to establish levels of exposure, which, if experienced over a lifetime, would result in levels of risk acceptable for the public. The ideal human study would thus involve lifetime exposure measurements at a series of rates of exposure, the establishment of a dose-response relationship, and the extrapolation of the data to low levels of exposure. Such human studies for vinyl chloride are not available. Animal studies of this agent may produce relevant information, but the extrapolation to humans may include large potential errors. The purpose of this paper is to present one approach to establishing exposure levels by using relevant human C-2 CMA 0X0519 information which, in the case of chemical exposures, usually involves occupational settings. The Model Consider a cohort of a number (N) of individuals who are exposed at constant rates to a particular chemical. The rates of exposure differ among the individuals, but each individual has a constant lifetime rate of exposure. The model assumes a linear relationship between excess relative risk and the exposure rate: RR - 1 - bE where RR refers to relative risk, b is a constant, and E is an exposure rate (lifetime). Strictly speaking, E is the difference in exposure between the study population and the comparison population used in calculating relative risk. However, the exposure of the comparison population is usually negligible when contrasted to that of the exposed study population. The exposed study population also usually experiences the background population exposure. In the case of occupational exposures, estimation of workplace exposures effectively gives an estimate of the difference in the worker cohort exposure and that of the comparison population. CHA 010520 C-3 As sum-prion Concerning Age The model assumes chat a constant rate of exposure to the chemical over a lifetime results in a relative risk for a particular disease outcome that does not vary by age. This implies that exposure would have a multiplicative effect on background incidence rates. Such an assumption would follow from a multistage model situation in which the exposure of concern affected particular stages while the background exposures predominantly affected other stages. Variation in Exposure Rates The model considers a cohort with a variety of exposure rates, but in which all individuals are of the same age. It includes a subcohort "i" within it that experiences exposure rate , composed of individuals of the same age. For this subcohort: RRi, - 1 - bE1.. If one uses an indirect standardization measure of relative risk such as the SMR, then: RRl. u. 1 where o^ is the observed number of cases with the disease in the subcohort and is the expected number of cases based on some external "standard" C-4 CMA 010521 population. Hence, npa I I it T v 3 *- * cT1 - 1 - bE1.. Now for any one particular age group, e. i - n,i e where e is the expected number of deaths per person in the subcohort and n L is the number of persons in subcohort i. Hence, o,l-in.e(l + bEL,). Summing overall exposure levels, So. - Zn.e(1 + bE.) ii i - Ne - Znl.ebEi.. where N is the total number in the cohort for the particular age involved. - Ne + ebZni.E.i C-5 CMA 010522 However, DRA Hence, So. - Ne + ebNE 1 - Ne(l + bE) Now, Ne - Se.i Therefore, So. __ - 1 + bE Se. x and hence, E gives an estimate of b. Combining Age Groups In each age group, one can obtain an estimate of b as shown in the previous section involving the average exposure rate. If one expands the summation over all age groups, we get the familiar standardized mortality ratio. (SMR): So, SMR - Se, Assuming that the SMR is not age dependent, then the following equation, (SMR-1) b- E C-6 CMA 010523 # l\ will provide an estimate of b, where E is now the average exposure rate for everyone in the cohort. This estimate is based on the assumption that each individual experiences a constant exposure rate. It also implies that the constant exposure rate has been experienced for a lifetime, since no distinction has been made according to age. Variation in Exposure Rates. Duration, and Age at Exposure Most chemical risk assessments are based on occupational studies in which workers experience variation in exposure rates, variable exposure duration, and different ages at exposure. In the case of cancer, the risk assessment is also complicated by cancer latency. To deal with the variable exposure rate, it will be assumed that the effect of exposure is related to the average exposure rate during relevant etiologic time periods. (This assumption is analogous to the assumption that risk is related to cumulative exposure independent of the pattern of exposure, for example the peaks or troughs of the exposure rate.) To deal with latency, the assessment of exposure rate will focus on relevant time periods in relation to dates of diagnosis. For example, the assessment of average exposure rate could be assessed up to a point in time 5 years before the mid-year of diagnosis of the cases. C-7 CMA 010524 Finally, DRAi to deal with lifetime exposure risks, an adjustment will be mace for the age at first exposure. A linear relationship will be assumed between risk and duration of exposure. If the average duration of exposure up to 5 years before the mid-year of diagnosis is d, and if the average age at first exposure is a, then the calculated average rate of exposure will be corrected by multiplying by d/(d+a). (It should be noted that most carcinogen animal bioassays do not commence dosing until the animals are fully grown, so this correction is a conservative element in using human data). C-8 CMA 01052-j