Document yrXXjKvVvG0DvGRk96jwRQEb6

14 III. HAZARDS OF VINYL CHLORIDE Vinyl chloride (chloroethene) Chemical Abstracts Service Registry No. 75015, previously had an OSHA standard of 500 ppm because its known toxic effects were thougnt to be relatively insignificant. Studies reported during the 1960's provided evidence of additional health hazards, prompting the American Conference of Governmental Industrial Hygienists (ACGIH) in 1971 to reduce the threshold limit value (TLV) from 500 ppm celling to 200 ppm time-weighted average. As a result of reports of liver cancer deaths among employees at the B. F. Goodrich Chemical Company, OSHA issued an emergency temporary standard of 50 ppm ceiling on April 5, 1974. To date, 24 cases of angioaarcoma have been reported among vinyl chloride workers, including 15 in the United States. 1, SAFETY HAZARDS Vinyl chloride (VC) is a flammable gas with a boiling point of approximately - 14C. It has been used as a refrigerant and as an extraction solvent for heat-sensitive substances because of its high vapor pressure. However, the substance's high flammability, acute toxicity, and readiness to polymerize made these uses undesirable. Handling of vinyl chloride presents a substantial hazard, since it is normally gaseous at room temperature and can form explosive mixtures with air. The lower explosive limit is 3.6 percent by volume in air (36,000 ppm); the upper explosive limit is 26.4 percent by volume in air (264,000 ppm) # VC is normally stored or transported in liquid form under vvc 000005235 15 pressure, with safety blow-off valves usually provided as a precautionary measure to guard against pressure buildup* 2. ACUTE EFFECTS AND TOXICITY OF VINYL CHLORIDE2*3*5 Most early studies Involving acute exposure to vinyl chloride focused on Its narcotic and anesthetic characteristics. In the 1930's, vinyl chloride was considered suitable for supplementary narcosis because of Its slight toxicity, wide range of effect, low concentration in the blood stream, and rapid elimination. Several animal studies were conducted to determine appropriate narcotic concentrations. The range for anesthetic effect in animals, generally between 7 and 10 percent by volume in air, Is also assumed valid for humans. Dangerous effects appear above 12 percent, and fatalities occur In test animals at concentrations between 20 and 40 percent by volume in air. Tests on human subjects at concentrations at and below 2 percent by volume in air (20,000 ppm) revealed symptoms such as dizziness, lightheadedness, nausea, dulling of vision and hearing, and headaches which often persisted for short periods of time. Recovery from these symptoms was rapid under test conditions. The gas does not provide adequate warning of its presence by any odor or Irritant action. It does, however, produce symptoms of dizziness and disorientation In advance of harmful effects, except when It is present In exceedingly high concentrations. VVC 000005236 L6 Two fatalities from vinyl chloride poisoning were reported in the literature. Few pathologic and histologic studies were performed in the early experiments dealing with acute exposure to VC. In fact, these studies seem to have been somewhat perfunctory. Subsequent investigations established definite evidence of cardiac irregularities, such as sinus arrhythmia (the gradual waxing and waning of the pulse rate)? tachycardia (excessively fast heart beat)7and bradycardia (abnormally slow heart beat). These particular results seem to have led to the conclusion that VC is unsuitable for narcosis and anesthesia in man. Later studies on animals showed evidence of pulmonary congestion and congestion of the liver and kidneys after short periods of exposure to VC. Studies involving longer exposures (3-6 months) validated the previous studies, showing some evidence of pneumonia, lung lesions. Increases in liver to body weight ratios, decreases in spleen to body weight ratios, and mixed changes in kidney weight.^ These studies marked the end of research into the acute effects of VC and the beginning of research into chronic effects. 3. CHRONIC AND LONG TERM EFFECTS OF EXPOSURE TO VINYL CHLORIDE Fifteen years ago, little was known or suspected about either the chronic or the long-term effects of exposure to vinyl chloride. Since that time, VC has come under steadily Increasing scrutiny VVC 000005237 17 as evidence of its deleterious health effects mounted. Research efforts of this period can be divided into three areas: (1) those brought about by the persistent appearance of an effect on liver function and involving the examination of the substance's chronic effect on liver function; (2) those delving into the nature and etiology of acroosteolysis; and (3) those involved with the carcinogenic potential of vinyl chloride. (1) Chronic Toxicity from Exposure to Vinyl Chloride^*^*^ The acute toxicity studies o f exposure to vinyl chloride repeatedly gave evidence of an effect on liver function, but the brevity of exposure and apparent reversibility of damage precluded attaching much significance to these results. When an investigation was finally conducted to determine the effect of longer-term exposure at lower concentrations, mlcropathologic changes and increased liver weight were clearly identified in Q several animal species. In a study designed primarily to Indicate the value of a technique correlating clinical and environmental measurements, statistically significant alterations were found in clinical liver tests using measurements of beta-protein, bromsulphaleln, and the Icterus index among workers exposed for several years to vinyl chloride. These animal and human studies were not accorded much importance, because of the belief that liver functions would return to normal when exposure to VC ceased. Reports of diseases of the liver, skin, and other organs and studies of hypertension. VVC 0000523Q 18 vascular lesions, nervous system injuries, kidney damage, and scleroderma (hardening of the skin) were similarly given little weight. More recently, evidence^ of lung changes and neurologic damage has come to light. When vinyl chloride was Implicated as a cause of liver angiosarcoma, several clinical studies were initiated to determine the presence of pathologic abnormalities among workers. One study ^0 a comprehensive examination of workers in a polyvinyl chloride plant which had been in operation since 1946. The results of this study provided some evidence of peripheral vascular disease, enlargement of the liver and spleen, increase of alkaline phosphatase levels, and pulmonary disease. So far, the only significance that can be attached to these results is that evidence of liver dysfunction may have represented the first stages of angiosarcoma. Further clinical studies are required before the full significance of such abnormalities can be appraised. (2) Acroosteolysis^* ^ * 10,11,12,13,14,15 In the middle 1960's, reports emerged of a disease which appeared among workers involved in the polymerization of vinyl chloride. This disease, the "hand problem" or occupational acroosteolysls, is marked by soreness and tenderness of the fingertips. In some cases, it is accompanied by Raynaud's phenomenon (blanching and numbness of the fingers) and by skin lesions. In most cases no symptomatic complaints were registered; presence of the disease was found only through X-ray examinations vvc 000005239 19 of the hands, which indicated osteolysis of the fingers (i.e. the destruction of bony tissues). As a result of these Initial findings, an extensive epidemiologic* industrial hygiene, and clinical study of 12 13 1^ occupational acroosteolysis was performed in the late 1960's. Since the disease had been found only among workers engaged in various phases of vinyl chloride and polyvinyl chloride manufacturing, the study focused only on these industries with the purpose of identifying the loci of the disease. This study found that workers who developed acroosteolysis had been involved at one time or another in manually cleaning reactors used in polymerization. The study also suggested that unreacted vinyl chloride was the cause of the disease, although the precise etiology could not be delineated. Generally, it was found that the incidence of acroosteolysis was substantially reduced by the Introduction of more automated cleaning of the reactors. In a later study of acroosteolysis, a more detailed examination of workers was undertaken.^ Although soreness and cold sensitivity in the fingers were generally the only symptoms identified in the earlier clinical studies, the workers involved in this study also indicated general complaints of excessive tiredness, forgetfulness and slowing down, dizziness, nausea. Increased perspiration and upper abdominal pain. Studies of X-ray results were consistent with the earlier findings. Detailed internal VVC 000005240 20 examinations showed many abnormalities, Including light to serious thrombocytopenia (small number of platelets in the circulating blood); enlargement of the spleen; toxic liver damage; ventilation blockage; circulatory obstruction; and skin and bone alteration. Although these symptoms have not yet been assigned to a known syndromenor have their pathogeneses been elaborated, the evidence at this point strongly indicates that vinyl chloride Is the etiologlc agent. (3) Carcinogenicity of Vinyl Chloride 6*9,16,17,18,19,21 Evidence that vinyl chloride has a carcinogenic effect first arose from a study designed to examine more fully the nature of acroosteolysis. The first observations of tumors were reported by Viola in 1970. He exposed Wistar rats to 3 percent (30,000 ppm) concentrations of vinyl chloride for a period of 12 months.^ in these experiments, skin, lung, and bone tumors were observed. While these results were statistically valid, they were not viewed with alarm, since the concentration of VC was near the explosive limit and was not likely to be found in industrial situations* Nonetheless, Viola's results did indicate the need for an extensive investigation into the type and degree of VC*s carcinogenic potential;with special attention to experimental conditions. On January 22, 1974, the National Institute for Occupational Safety and Health (NIOSH) was informed of three deaths from liver angiosarcoma among workers at the B. F. Goodrich Company in VVC 0 000052** 21 Louisville, Kentucky. Since then, a total of 18 cases of liver angiosarcoma among PVC production workers and one case among VC monomer production workers have been Identified Thirteen of these cases occurred in the United Statest seven at the B. F. Goodrich plant in Louisville, Kentucky; three at the Goodyear plant in Niagara Falls, New York; two at the Union Carbide plant in South Charleston, West Virginia; and one at the Firestone plant in Fottstown, Pennsylvania. In addition, three other cases of angiosarcoma have been identified as possibly resulting from vinyl chloride exposure: two workers in PVC fabricating plants and a person who lived within two miles of a fabricating plant developed angiosarcomas. The significance of these last three cases is still somewhat in doubt. 1. Results In Animal Studies Subsequent to Viola's work, studies by Cesare Maltonl and by the Industrial BIO-TEST Laboratories were initiated to define more completely the carcinogenic hazards of vinyl chloride. At OSHA's public hearing on February 15, 1974, Maltonl19 Indicated that liver angiosarcoma had been observed in Sprague-Dawley rats exposed to VC concentrations as low as 250 ppm. On April 15, 1974, OSHA was informed that BIO-TEST had observed two cases of angiosarcoma in mice exposed to 50 ppm. In testimony22*22 presented at the public hearing on June 25, 1974, further results from Maltonl showed angiosarcoma at 50 ppm in rats.Later results from BIO-TEST revealed angiosarcoma in hamsters. In all these animals, the following types of tumors were observed: Zymbal VVC 0000052 22 gland carcinomas, liver angiosarcomas and angiomas, kidney nephroblastomas, skin carcinomas, intra-abdominal angiosarcoma, subcutaneous angloperlcltoma, pulmonary tumors, mammary carcinomas, and liver fibroangiomas. The angiosarcoma observed in the test animals was of the same morphological type as that observed in the B. F. Goodrich workers. This provided the first clear evidence that vinyl chloride was probabjy the el 1olop.lt' cause of liver angiosarcomas among vinyl chloride workers. Having established the causative effect of VC, the next stage of the animal studies involved determining the dose-response relationship between vinyl chloride and liver angiosarcoma and the appearance of other tumors. The results of one of Maltoni's experiments revealed the following incidence of angiosarcoma in Sprague-Dawley rats following 52 weeks of exposure and 135 weeks of observation: Dose (ppm) Angiosarcomas Positive/Tota1 0 0/68 50 1/64 250 4/67 2,500 13/74 6,000 14/72 10,000 9/69 These results can be fitted to several models to determine risk level (the level of exposure that would be expected to induce one case) and response rate (the number of cases that would be expected at a particular level of exposure). The use of the VVC 0000052*3 probit and logit models, based on different mt :hematical 23 representations of the data* was demonstrated at the public hearing of June 25* 1974.^ At their present stage of development, however, they are only hypotheses as to the form of the actual doseresponse relationship. The selection of a "most appropriate" model or discovery of a scientifically verifiable safe level must await experimental verification. The results of the applications are given in Tables 1 and 2. Table 1 (risk levels) shows the level of exposure to vinyl chloride in parts per billion (ppb) that would induce liver angiosarcoma in Sprague-Dawley rats at the indicated risk level. Thus, for example, the first entry in the top line indicates that at an exposure of 225 ppb, 1 rat in 100,000 exposed would be expected to contract liver angiosarcoma. Table 2 (response rates) shows the probability that a single Sprague-Dawley rat would exhibit liver angiosarcoma at a given dose level. Thus, the first entry in the top line shows that 0.028 or 2.8 percent of rats exposed to 50 ppm would be expected to contract liver angiosarcoma. The table includes levels below those at which animal experiments have hertofore been performed, such as 25, 10 and lppm. The results of these studies have only slightly advanced the understanding of the carcinogenicity of vinyl chloride. They highlight the many unknowns that need to be clarified before a definitive extrapolation to humans can be made. The results to date, along with the human experience, suggest the existence of a long latency period. In fact, they indicate the possibility WC 000005244 23a TABLE 1 Extrapolated 'Safe* Bose Level (ppb) Using Maltoni's Liver Angiosarcoma Data* (99% Assurance Level) Extrapolation Model Risk Level 10-5 icf6 f7 10-8 Probit (Slope = 1, Mantel-Bryan) Logit (Slope 3.454) Logit (Slope - 2.303) and "One-hit" 225 73 26 10 553 119 26 5.5 21 2.1 0.21 0.021 * Only the data up through dose level 500 ppm, but not higher, were used since the response curve flattened out at higher levels. TABLE 2 Estimated Response Rates: Liver Angiosarcoma (99% Upper Confidence Limit) Model Probit (Slope - 1) Logit (Slope = 3.454) Logit (Slope = 2.303) and "One-hit" Bose(ppm) 50 25 10 0.028 0.013 0.0085 0.003 0.02 0.0071 0.0044 0.0008 0.0018 1 0.00015 0.00002 0.00006 WC 00000524-5 24 that* at low dosage levels, the onset o cancer may not occur during the average lifespan of humans. Many of the experimental animals did not exhibit angiosarcoma until late In their lives, and many died, not as a result of the angiosarcoma, but from other causes. Further, the results of these studies suggest that differences between species have an effect upon the incidence of angiosarcoma. This however does not mean that man is necessarily less sensitive than the animals tested. A conservative view Is that man is probably at least as sensitive as the rat. Although the rat has a shorter lifespan than man,tthe rat's metabolic rate Is more rapid than man's--thus it could be construed that the rat exhibits a proportionally accelerated version of the possible response in man. These Btudies also demonstrate that different organs and sites in the body are susceptible to vinyl chloride carcinogenesis at different rates, with the possibility that the angiosarcoma may not actually be the most significant effect. The relative Importance of vinyl chloride's various effects still needs to be adequately delineated. Another area which demands further study Is the effect of vinyl chloride on unborn children. In Maltonl's experiments, two subcutaneous fibrosing angiosarcomas were observed in offspring of Sprague-Dawley rats exposed to 6,000 and 10,000 ppm. The significance of these findings is as yet unclear. However, this result does Indicate that vinyl chloride appears to have a parentteratogenic effect, a phenomenon which is common to other carcinogens. Further study into this problem is clearly required to determine 00 000 25 whether additional precautions should be taken to protect pregnant women from exposure to vinyl chloride. In regard to the mechanisffl of vinyl chloride once It enters the body, it has been suggested that VC is not the ultimate IS carcinogen. Preliminary results from an ongoing study indicate that up to a level of approximately 200 ppm vinyl chloride is metabolized by alcohol dehydrogenase. Above this level, it is apparently metabolized by microsomal oxidases. It has been suggested that metabolism through the microsomal oxidases route may produce the ultimate carcinogen. However, since this action can occur at levels below 200 ppm, this does not preclude the possibility that VC is carcinogenic at low doses* 2^ Exposure Level for Humana The most definitive way to establish a "safe" dose level for humans is to examine the data on past exposures and relate it to the occurrence of angiosarcoma. However, data is not available, and the available evidence on the possible effects of exposure can be consideredas only tentative and preliminary. In view of the relative scarcity of angiosarcoma in the U.S. population, the relationship of the cases identified to date and vinyl chloride exposure shown In Table 3 presents strong evidence that VC is a human carcinogen. This data also shows that the time lapse from the initial VC exposure to diagnosis of angiosarcoma ranged from 11 to 36 years, with an average of 21 years, suggesting the existence of a latency period for the cancer. VVC 000005247 TABLE 3 COUNTRY CASE # CONFIRMED CASES OF LIVER ANGIOSARCOMA AMONG WORKERS EXPOSED TO VINYL CHLORIDE OR POLYVINYL CHLORIDE BIRTH DATE 1st VC or PVC EXPOSURE DX ANGIO SARCOMA AGE AT DX YRS 1st EXP TO DX VC MONOMER PRODUCTION WORKERS Sweden 02 11-27-11 00-00-45 05-15-72 61 27 POLYMERIZATION WORKERS United States United States United States United States United States United States United States United States United States United States United States United States United States W. Germany W. Germany o Great Britain o Norway o Sweden w o VJ\ If United States United States 01 02 03 04 05 06 07 08 09 10 11 12 13 01 02 01 01 01 14 15 00-00-22 00-00-34 00-00-15 00-00-24 00-00-12 00-00-29 05-03-22 05-06-20 00-00-31 08-16-13 05-27-09 11-17-18 12-01-21 07-26-31 06-04-30 00-00-01 12-23-15 06-23-27 00-00-13 00-00-25 12-09-48 11-15-55 11-28-45 07-06-52 06-19-44 01-17-62 08-00-44 10-07-46 05-28-45 06-00-51 10-14-46 09-13-49 08-19-44 10-14-57 10-01-57 00-00-46 03-00-50 08-14-51 03-00-71 05-00-70 12-00-73 08-00-67 04-00-64 02-00-74 00-00-68 08-00-61 03-01-74 05-00-68 03-00-70 05-00-69 05-00-74 00-00-71 00-00-69 12-00-72 12-20-71 02-00-70 49 36 58 43 52 45 45 41 43 55 61 50 53 40 39 71 56 43 SECONDARY MANUFACTURING 08-18-38 06-00-73 60 00-00-00 07-00-72 47 22 14 28 15 20 12 24 15 29 17 23 20 30 14 11 26 22 19 36 00 Note; *00* indicates unknown date DX = Diagnosed TOT YRS EXP 23 16 13 28 15 18 12 18 15 17 17 23 15 30 14 11 20 21 18 00 00 DATE OF DEATH 08-16-72 03-03-73 09-28-71 12-19-73 01-07-68 04-09-64 ALIVE 03-23-68 08-29-61 ALIVE 05-10-68 03-16-70 05-02-69 ALIVE 12-14-71 01-25-69 12-00-72 01-04-72 10-20-70 07-03-73 02-15-73 25a 26 The exact exposure levels of VC .experienced by the workers who were stricken with angiosarcoma are not known. However, evidence presented at the public hearing indicated that, at least for the U.S. cases, exposures during the 1940's and 1950*s were estimated by industry to be extremely high, frequently over 1000 ppm and commonly in the 100 ppm range. This would seem to indicate that VC carcinogenicity for humans might only occur as a result of high exposure levels. On the other hand, two cases of angiosarcoma that occured U.S. workers conflict with this hypothesis. One worker was first exposed in 1962 and since that time worked In areas where the exposure levels averaged below 200 ppm TWA. The other case worked in an extruding operation and was probably never exposed to a level above 50 ppm (with TWA's much lower). These two cases cast doubt on the existence of a "safe" level for humans. At this time, a major epidemiologic study nf the relationship between VC levels and carcinogenicity is the 26 Tabershow/Cooper study of 8384 vinyl chloride workers employed in 35 domestic VCM and FVC plants. Workers, employed for at least one year in a job Involving vinyl chloride exposure, were subjectively classified according to an Index representing their relative exposure level. Results were analyzed for the overall study population according to the estimated levels of exposure, the durations of exposed employment, and the values of the exposure index. VVC 000005249 27 The primary finding of the study vas that those exposed to vinyl chloride did not Buffer e statist! celly . significant mortality rate increase over what would be expected la a cosparable U.8* male population. Other Important findings were that cancers of the liver (primarily angiosarcoma)* the respiratory system* the brain* the lymphatic systea. and unknown primary sites occurred most frequently in those meafcere of the study population who suffered the greatest exposure. Even though these excesses were not statistically significant* the findings warranted furth*rrstudy., The results of this epidemiologic study do not definitively establish the case for or against VC carcinogenicity. The tentative conclusion of an Increased nusber of malignant neoplasms with VC exposures raises the possibility that vinyl chloride has a mere wide-ranging affect then liver angiosarcoma dona. Further study and analysis of this data seem warranted* such as by the further development of worker histories and by the analysis of each plant by exposure levels* months of employee exposure* length of time since exposures cease* and ages of the workers. Such an investigation might Include the development of a population of chemical workers with which to compare the subset of vinyl chloride barkers In addition* an epidemiologic study of workers from oompeiindlng sad fabricating plants might be useful. vvc 000005250 28 In viewing any epidemiologic data* conclusions drawn have to be tempered by the consideration that the full effects of past exposures among workers may not have yet surfaced because of the latency period associated with VC carcinogenesis. Other unknown factors include the effects of single or intermittent exposures at high or low levels and the effects of small differences In exposure level (such as between 15 and 25 ppm) on the onset of cancer. VVC 000005251 NOTES 29 1. C. A. Brighton in N. M. Bikales, ed., Encyclopedia of Polymer Science and Technology. Vol. 14, Interscience Publishers, a division of John Wiley & Sons, Inc., New York, 1971, pp. 305-320. 2. K. B. Lehmann and F. Flury, Toxicology and Hygiene of Industrial Solvents. J. Springer, Berlin, 1938, p. 130 (As translated by E. King and H. F. Smyth, Jr., 1941, pp. 143-144). 3. W. F. Von Oettingen, The Halogenated Aliphatic, Olefinic, Cyclic, Aromatic, and Aliphatic-Aromatic Hydrocarbons Including the Halogenated Insecticides Their Toxicity and Potential Dangers, Public Health Service Publication No. 414, U.S. Department of Health, Education, and Welfare, Washington, D.C., 1955, pp. 195-197. 4. E. Mastromatteo, A. M. Fisher, H. Christie, and H. Danzlger, "Acute Inhalation Toxicity of Vinyl Chloride to Laboratory Animals," American Industrial Hygiene Association Journal. Volume 21, No. 5, October 1960, pp. 394-398. 5. D. Lester, L. A. Greenberg, andWW. Robert Adams, "Effects of Single and Repeated Exposures of Humans and Rats to Vinyl Chloride," American Industrial Hygiene Association Journal, May-June 1963, pp. 265-275. 6. Statement of Dr. Irving J. Selikoff, Mt. Sinai School of Medicine, at OSHA hearing on February 15, 1974. 7. T. R. Torkelson, F. Oyen, and V. K. Rowe, "The Toxicity of Vinyl Chloride as Determined by Repeated Exposure of Laboratory Animals, "American Industrial Hygiene Association Journal. Volume 22, No.*5, October 1961, pp. 354-361. 8. C. G. Kramer, and J. E. Mutchler, "The Correlation of Clinical and Environmental Measurements for Workers Exposed to Vinyl Chloride," American Industrial Hygiene Association Journal. Volume 33, No. 1, pp. 19-30. 9. Statement of Anton Vittone, Jr., B. F. Goodrich Chemical Company, at OSHA hearing on February 15, 1974. ooooo^2 30 10 Statement of Peter Bomnarlto, President, United Rubber, Cork, Linoleum, and Plastic Workers of America, at OSHA hearing on February 15, 1974. 11. R. H. Wilson, W. E. McCormick, C. F. Tatum, and J. L. Creech, "Occupational Acroosteolysls, Report of 31 Cases," The Journal of the American Medical Association. Volume 201, Ho, 8, August 21, 1967, pp. 83-87 12. B, D. DInman, W. A. Cook, W. M. Whltehouse, H. J. Magmuson, and T. Ditcheck, "Occupational Acrossteolysls: I. An Epidemiological Study," Archives of Envir^n^M! Health. Volume 22, pp. 61-73, January 1991. 13. W. A. Cook, P. M. Glever, B. D. DInman, and H. J. Magmuson, "Occupational Acroosteolysls: II. An Industrial Hygiene Study, "Archives of Environmental Health, Volume 22, pp. 74-82, January 1971. 14. V. N. Dodson, B. D. DInman, W. M. Whltehouse, A. M. M. Naer, and H. J. Magnuson, "Occupational Acroosteolysls: III. A Clinical Study, "Archives of Environmental Health. Volume 22, pp. 83-91, January 1971. 15. S. Juhe, C. E. Lange, G. Stein, and G. Veltman, "On the So-called Vinyl Chloride Disease," Deutsche MedizlnBche Wockenachrlft. 98 (43), pp, 2034-2037, English translation, 1973. 16. P. L. Viola, "Pathology of Vinyl Chloride," Medclnla del Lavoro. Volume 61, No. 3, pp. 174-179, March 1970, 17. P. L. Viola, "Carcinogenic Effect of Vinyl Chloride," Presented at the Tenth International Cancer Congress, Houston, Texas, May 22-29, 1970. 18. P. L. Viola, A. Blgottl, and A. Caputo, "Oncogenic Response of Rat Skin, Lungs, and Bones to Vinyl Chloride." Cancer Research, Volume 31, pp. 516-522, May 1971. 19. Statement of Cesare Maltoni, Instltuto dl Oncologla, Bologna, Italy, at OSHA hearing on February 15, 1974. 20. Statement of Irving J. Selikoff, Mount Sinai School of Medicine, at OSHA hearing of June 25, 1974. VVC 000005253 30A 21. Statement of Marcus M, Key, National Institute for Occupational Safety and Health, at OSHA hearing of June 25, 1974. 22. Statement of Daniel P. Boyd, Occupational Safety and Health Administration, at OSHA hearing of June 25, 1974. 23. Statement of M. L. Keplinger, Industrial BIO-TEST Laboratories, at OSHA hearing of June 25, 1974. 24. Statement of Marvin A. Schnelderman, National Cancer Institute, at OSHA hearing of June 25, 1974. 25. Statement of Perry Gehrlng, Dow Chemical Company, at OSHA hearing of June 25, 1974. 26. "Epidemiological Study of Vinyl Chloride Workers," Tabershaw/Cooper Associates, Inc., Berkeley, California, May 3, 1974 (Submitted to the Manufacturing Chemists Association, Washington, D.C.) (Reproduced as Exhibit 20J of the OSHA hearing of June 25, 1974). "VC 0052S4 31 IV. OCCUPATIONAL EXPOSURE TO VINYL CBLOltllg An assessment of the health hazard from exposure to vinyl chloride must Include an estimate of the number of workers exposed, a determination of their past and present exposure levels, and an appraisal of the number of workers who may be expected to contract liver angiosarcoma or other forms of cancer as a result of these exposures. A full characterization of employee exposure to vinyl chloride, however, cannot be accomplished at this point since very little information is available concerning the levels to which workers have been exposed in the past. A reliable model to predict the expected number of cases cannot be formulated because of the lack of data on past exposures and of the mechanism of vinyl chloride carcinogenesis. Consequently, this chapter only attempts to describe the primary potential sources of occupational exposure to vinyl chloride. Vinyl chloride (VC) is used primarily inthe production of pAlyvinyl chloride (PVC), a thermoplastic resin which has application in a host of products for consumers and Industry. The conversion of the VC monomer into a polymer or copolymer is an incomplete process, i.e., not all of the monomer is reacted. Furthermore, despite the fact that the process of stripping removes most of this monomer from the polymer resin, it only partially accomplishes this function. Consequently, residual monomer is visually entrapped In polymer and copolymer formations, and often escapes -- thus posing a workplace hazard in both VVC 000005255 PVC and VC plants. The VC content and form of the polymer determine the rate at which VC separates, and the extent of exposure that will occur. Even after the polymer has been fused Into a product which will undergo no further heating in bulk (at least to melting temperatures); escape of the vinyl chloride may still occur because of migration to the surface. The OSHA proposed regulation excludes companies using fabricated products since no facts are available to show measurable emissions from fabricated products. Manufacturing processes involving VC or PVC (containing entrapped monomer) fall into five distinct categories. Figure 1 presents a schematic which shows the interrelationship of these process stages. These are: o Production of vinyl chloride monomer; o Production of polyvinyl chloride homopolymer or copolymer; o Compounding of PVC homopolymers or copolymers with plasticizers, stabilizers, lubricants, fillers, pigments, and other additives; o Processing of compounded PVC into finished or semi-flnlshed products; and o Fabrication of finished products from intermediate stock. The fourth stage may yield finished products or semi-finished products used in the fifth stage. Since companies In the fifth stage will not be regulated, this report will examine only the first four stages to estimate the number of workers exposed and VVC 000005256 33 Figure 1. Manufacturing Processes Involving Vinyl Chloride or Polyvinyl Chloride 00005257 VVC 0 34 che level of their exposure to vinyl chloride. 1. PRODUCTION OF VINYL CHLORIDE Vinyl chloride first attained commercial significance during the 1930*8 after the discovery that PVC could be processed and converted Into a rubber-like product. The ready availability of natural rubber limited development of VC manufacture prior to the outbreak of World War II. When the war cut off supplies of natural rubber, production of vinyl chloride for PVC, as a rubber substitute, was initiated on a large scale. Developments since that time have made vinyl chloride one of the most Important starting materials of the plastics industry. (1) Process Descriptionsi23,5,6,7 Vinyl chloride is presently produced by 10 companies In 15 plants In the United States (Including Puerto Rico), as shown In Table 1. The substance was Initially produced by the caustic hydrolysis of ethylene dichlorlde, subsequently by the pyrolysis of ethylene dichlorlde to vinyl chloride and hydrogen chloride, and still later by the hydrochlorination of acetylene.19 Most recently, a newer process Involving the oxychlorlnation of ethylene has become increasingly prominent In VC production, so that presently 13 plants, with almost 92 percent of the capacity, are using the ethylene route. Hydrochlorinatlon of Acetylene - In this process, also known as the balanced process, vinyl chloride Is produced by the hydrochlorinatlon of acetylene, usually with mercuric chloride as the catalyst. A flow diagram for this process Is VVC 000005258 Company Allied Chemical Corporation Azaeri can Chemical, Inc. B*F.Goodrich Chemical Company Continental Oil Company Dow Chemical Company Ethyl Corporation Monochem. Inc. PPG Industries Shell Chemical Co. Textnuco Chemicals, Inc. Table 1 Vinyl Chloride Monomer Producers Location Nominal Capacity (MM lb.) Baton Rouge, La. Watson, Calif. Calvert City, Ky. Lake Charles, La. Freeport, Texas Oyster Creek, Texas Plaquemine, La. Baton Rouge, La. Houston, Texas Gelsmar, La. Lake Charles, La. Guayanilla, P. R. Deer Park, Texas Norco, La. Houston, Texas 300 170 1,000 625 180 *700 340 270 150 350 300 500 840 700 255 6,680 Process Direct and oxychlorinatlon t It ft 11 1! M Direct chlorination ?i u i, it Direct U and oxychlorinatlon f1 Acet' Icrie Direct and oxychlorinatlon ir M It Acetylene II Sources: See Bibliography, references 1, 2, 3, 4. VVC 0 0 0 0 0 5 2 5 9 Lua> J 3 63 6 e E J 36 & B t g J txS II o VVC 000005Z60 37 shown in Figure 2. Acetylene and hydrogen chloride are first passed through a drying tower and then mixed in a chamber packed with activated charcoal. This mixture is passed through several hundred vertical tubes in which the catalyst is packed. The temperature of these tubes is limited to ranges which permit the reaction to take place. The product gases are then fractionally condensed, and the liquid condensate fractionally distilled to obtain vinyl chloride. The VC is finally purified by caustic scrubbing. Most of the variations in this process arise from the use of different catalysts or mixtures of catalysts. Use of this route in the United States has been declining due to the reduced availability of acetylene feedstocks. Oxychlorination of Ethylene - Ethylene can be directly chlorinated to produce vinyl chloride; however, difficulties with this technique have led to VC production from ethylene through the intermediate formation of 1,2-dichloroethane (also called ethylene dichloride) a process used by several plants in the United States, Japan, and Europe. A schematic of this process is shown in Figure 3. There are several variations of this process which have demonstrated commercial feasibility. Differences among these generally arise in the amount of recycling or in the materials used as catalysts. The EDO comes from two sources: chlorine combined with ethylene in a liquid phase reaction with a dissolved catalyst WC 000005261 FIGURE 3.. MANUFACTURING OF VCM BX DIRECT CHLORINATION AND OXZCHLQRINATION Direct Chlorination? f Cl2 --CH2C1-CH2C1 CEDC> Oxychlorinations C2H4 + 2HC1 + 02 --------* CH2C1-CH2C1 (EDC) + H20 Cracking of EDC: CH2C1-CH2C1 (EDC)--------- * CH2-CHC1 (VCM) + HC1 < < o VCM STORAGE LO 00 000005262 39 (in a chlorinator) and ethylene combined with hydrogen chloride and oxygen in a vapor phase reaction with a catalyst (in an oxychlorination reactor). The crude EDC from these two units Is then purified by distillation. The final operation involves the thermal cracking (pyrolysis) of this EDC to form vinyl chloride and anhydrous hydrogen chloride. The resulting effluent is then separated by fractionation. The vinyl chloride is stored! the hydrogen chloride is passed back to the oxychlorination unit, and unreacted EDC is recycled. (2) Worker Exposure3*8*9*10*11 Since vinyl chloride is produced in outdoor plants, the build-up of the substance in the workplace is less than that observed during the subsequent process stages Involving the monomer or polymer. The job categories of workers at VC production plants are shown in Table 2. No data is currently available showing the number of people nationwide who are employed in each category. At the OSHA hearing on February 15, 1974, the Manufacturing Chemists Association estimated that 1,500 people work In monomer production. However, since monomer plants differ significantly in age, type, and degree of automation, it is not possible to validate this number. For example, PPG Industries states that it employs between two and four workers per million tons of stated capacity. Based on this figure, the number of workers in monomer plants would be estimated at 1,700 to 3,300. However, it is not possible to characterize such plants for the purpose of categorizing workers VVC 000005263 40 Table 2 Job Categories of Workers Vinyl Chloride Plants Control Operator Distillation Operator EDC Operator Electrician Instrument Technician Laboratory Technician Laborer Loading Operator Maintenance Worker Material Controller Millwright 0X7 Operator Pipefitter Quality Control Specialist Sample Man Shift Supervisor Tank Farm Operator VCM Operator VVC o 41 according to exposure levels. As a result, exposure data can only be addressed in general terms, until a comprehensive survey of the industry can be conducted. An estimate of the number of workers in each job classification is necessary for any characterization of the total hazard of vinyl chloride production, since data presented to OSHA has indicated different peak and time-weighted-average concentrations for the various categories. The major sources of exposure under normal conditions are judged to be: o Sampling and analysis of vinyl chloride for quality control; o Loading of vinyl chloride for shipping; o Entry of vinyl chloride containing vessels for maintenance and repair work; and o Leaks of vinyl chloride in the process area. Some sampling data indicate peak concentrations over 200 ppm for laboratory technicians, over 150 ppm for loading operators, and near 50 ppm for quality control specialists. Time-weightedaverages generally fall well below 50 ppm. Since these results were obtained prior to the emergency temporary standard, they possibly do not represent current levels of exposure to vinyl chloride. Thus, this information should be viewed only as Indicative of the categories of workers where high exposures to VC may occur. Wc 00005265 42 2. PRODUCTION OF VINYL CHLORIDE POLYMER AND COPOLYMER RESINS Virtually all vinyl chloride produced in the United States and not exported is used to manufacture either vinyl chloride homopolymers or vinyl chloride copolymerized with such monomers as vinyl acetate, vinylidene chloride, and acrylates. Generically, these homopolymers and copolymers are known as polyvinyl chloride (PVC) resins. They are generally found in the form of powders which, after being compounded with a variety of additives, are processed to produce a large variety of plastic or resinous end products. Table 3 lists the 22 companies and 37 plants In the United States who produce the homopolymer resin, as well as the capacity of each plant. 19 Additional plant openings and the expansion of present plants that are currently planned are estimated to add to the capacity by approximately 2,000 million pounds. 19 Most commercial resins have molecular weights between 30,000 and 120,000, the weight being a determinant of the resin's processing and performance characteristics. These characteristics are also affected by the specific additives with which the resins are compounded. Since theoretical results do not definitively Indicate what mixtures will give rise to what properties, formulation and processing is largely a matter of experience. 2.66 000* ^0 VVC 0 0 0 0 0 5 2 6 7 Table 3 Vinyl Chloride Polymer and Copolymer Reain Producers Company Air Products & Chemicals, Inc. 13 American Chemical Corporation * ^ Atlantic Tubing & Rubber Company B.A.S.F. tfyandote Corporation* B. F. Goodrich Chemical Company *** 1,3,4 Borden, Inc. 1 Continental Oil Company Diamond Shamrock Corporation*1 3 Dow Chemical Ethyl Corporation1 Firestone Tire & Rubber Company ' 2 General Tire & Rubber Company Goodyear Tire & Rubber Company Great American Chemical Corn. Hooker Chemical Corporation1** m *x Keysor Century Corporation1*3 Monsanto Company Morton-Norwich Products, Inc. . _ , National Starch & Chemical Corp. * * Olin Corporation 1*3 Pantasote Company 1*3 Location Calvert City, Ky. Pensacola, Fla. Long Beach, Calif. Cranston, R.I. South Kearney, N.J. Avon Lake, Ohio Henry, 111. Long Beach, Calif. Louisville, Ky. Pedricktown, N.J. Illiopolls, 111. Leominster, Mass. Aberdeen, Miss. Oklahoma City, Okla. Delaware City, Del. Deer Park, Texas Midland, Michigan Baton Rouge, La. Perryville, Md. Pottstewn, Pa. Ashtabula, Ohio Point Pleasant, W. Va. Niagara Falls, N.Y. Plaquemlne, La. Fitchburg, Mass. Burlington, N.J. Hlcksvllle, N.Y. Saugus, Calif. Springfield, Mass. Rlngwood, 111. Meredosia, 111. Assonet, Mass Passaic, N.J. Nameplate Capacity (MM lbs,) 150 50 150 Unk. Unk. 140 140 140 340 170 140 180 285 240 100 270 Unk. 180 230 270 125 50 100 100 40 180 15 35 70 Unk. 10 150 60 Table 3 (cont'd) Vinyl Chloride Polymer and Copolymer ReSln Producers Company Roblntech, Inc.1 - S.C.M. Corporation ^ Stauffer Chemical Company Tenneco Chemicals, Inc. Union Carbide Corporation1 * 4 1 Uniroyal, Inc. 3 W. R. Grace and Company Location Falnsvllle, Ohio Huron, Ohio Delaware City, Del. Burlington, N.J. Fleolngton, N.J. South Charleston, W. Va. Texas City, Texas Falnsvllle, Ohio Owensboro, Ky. South Acton, Mass. 77 Homopolymer. 2/ Vinyl chloride - propylene. 3/ Vinyl chloride - vinyl acetate copolymer. 4/ Vinyl chloride - vlnylldene chloride copolymer. Sources: See bibliography references 2, 12, 13, 14. Nameplate Capacity (MM lbs 250 Unk. 175 165 70 160 240 140 Unk. Unk. < < o 42b 00000526S 43 Four basic processes are employed In polymerization or copolymerization: bulk, suspension, emulsion, and solution* The bulk process is a dry process requiring no water, suspending agents, or emulsifiers. The other three processes require water, solvents, or other liquids, but they are similar enough to use the same equipment, although each plant usually limits itself to only one process. Polymeriza tion of monomers is initiated by free radicals produced by thermal decomposition from peroxides, azo compounds, and persulfates. The rate of polymerization and the molecular weight distribution of the resin are largely determined by the temperature of the reaction and the concentration of initiators. General schematics for the bulk, suspension and emulsion processes are shown in Figures 4 and 5. (1) Process Descriptions1*12*13*3'5 In the United States, approximately 78 percent of PVC resins is produced by suspension polymerization, 13 percent by emulsion polymerization, 6 percent by bulk polymerization, and 3 percent by the solution method. The bulk polymerization process is relatively new and has been used for an increasing proportion of the production of PVC resins in recent years. VVC 000005269 FIGURE 4 MANUFACTURING OF PVC BY BULK POLYMERIZATION 4> VVC 0 0 0 0 0 5 2 7 0 FIGURE 5 . MANUFACTURING OF PVC BY SUSPENSION AND EMULSION POLYMERIZATION 45 VVC 000005271 46 Suspension Polymerization - This process is essentially a batch process generally carried out in glass-lined reactors having a capacity of 2,000 to 6,000 U.S. gallons. Monomer soluble initiators, such as lauroyl, decanoyl, and benzoyl peroxides, are dissolved In liquid monomer dispersed in water to form a suspension. The reactor is first charged with the required amount of deionized water. Then the dispersant, a buffer, and the initiator are added. Oxygen is evacuated from the reactor, and the vinyl chloride and any comonomer are fed in. Agitation (to Increase the degree of dispersion) is begun, and the contents of the reactor are brought up to the polymerization temperature (45-60C). After the ten to sixteen hours of the polymerization process, the contents of the reactor (the PVC slurry), are dumped into a large vessel and excess vinyl chloride is recovered for distillation and reuse. The slurry is then blended with other batches in another vessel to reduce small variations and dehydrated by a continuous solid bowl centrifuge to yield a polymer cake containing about 20 percent moisture. The polymer is finally dried in a flash dryer; the resulting particles are sorted out by passage through a cyclone separator. The product is then screened and sent to storage. 0OOOO52 47 Emulsion Polymerization - This process is very similar to the suspension process, and is primarily used to produce exceptionally fine particles. There are two primary differences: (1) The suspension process uses mechanical mixing, while the emulsion process uses surfactants and agitation to convert the monomer to a stable emulsion containing very small monomer particles and; (2) In the emulsion process, the polymer is separated from the water by spray drying, which the nature of the polymer necessitates. While this process is more costly, it is important to users who need PVC in liquid form. In addition, this type of polymer has the ability to form a paste when combined with a plasticizer. In some cases, the slurry is not dried, but Instead is produced in a latex form. Bulk Polymerization - In this process, vinyl chloride is polymerized without the addition of other liquids and in the presence of a free-radical initiator. A portion of a day's supply of monomer is initially pumped into a prepolymerlzer. The mixture from the prepolymerlzer, together with an equal amount of fresh monomer, is transferred into an autoclave equipped with slowly rotating agitator blades where the reaction is Induced by high pressures. The reaction is autocatalytlc and accelerates with increasing conversion of the monomer. After the monomer has been converted, unreacted monomer is removed by a vacuum and recovered in a recycle condenser. The resin is transferred to a VVC 000005273 48 resin receiver. Fines are collected in a dust separator and removed through a manhole. The output of bulk processing plants is said to be more than twice that of good suspension plants of comparable size. 12 This factor will probably lead to a greater use of bulk polymerization during the coming years. Solution Polymerization - This process is used exclusively for the production of copolymers of vinyl chloride and vinyl acetate. The mixed monomers are dissolved in an organic solvent Inside a stirred autoclave heated to 40C. The reaction in the autoclave has the same autocatalytlc characteristic as that occurring in bulk polymerization. In this case, however, the phase separation, arising from the formation of a concentrated polymer base, leads to a decrease in the rate of polymerization. The contents of the autoclave are passed through a filter press to retain the polymer, and the filtrate is recycled. The filtered polymer (in the filter press) is then washed, during which a certain amount of unpolymerized vinyl chloride is recovered by fractional distillation. (2) Worker Exposure It has been estimated that the total number of workers engaged in all aspects of the production of PVC resins is about 5,000. The validity of this figure is suspect, however, since an industrial hygiene survey conducted in 1969 examined over 5,000 PVC-related workers at a time when production was approximately half the current level. Only a small proportion of these workers VVC 000005274- 49 were engaged in operations outside the actual resin production. The productivity of these workers has probably increased since 1969 in view of some significant technological changes such as the introduction of the bulk polymerization process, but, on balance, the total number of workers has probably increased beyond 5,000, although less than 10,000. Some of the job classifications of workers involved in resin production with potential exposure to VC are shown in Table 4. No evaluation of the total industry exposure to vinyl chloride can be made at present, because no estimates of the personnel distribution in these categories are available for the entire industry, and because exposure varies according to job category. The information available does tend to indicate that exposure to VC is greater in resin producing plants than in those synthesizing the monomer. The areas of greatest exposure in a PVC plant are judged to be the following: Unloading of incoming vinyl chloride; * Reactor cleaning; Entry of vinyl chloride containing vessels for maintenance and repair work; Entry of PVC storage silos; Shipping or packaging of PVC; and Leaks of vinyl chloride in the process area. Levels of VC exposure in these plants have been decreasing rapidly since January, 1974. For the entire industry, the average TWA is probably well below 25 ppm. Exposures differ VVC 000005275