Document 5DKXeKQxDMbqxGK9wE9dwRGB0

Mutation Research, 41 (1976) 131--14Z C Elsevier/North-Holland Biomedical Press 131 CARCINOGENIC, MUTAGENIC AND TERATOGENIC RISKS ASSOCIATED WITH VINYL CHLORIDE PETER F. INFANTE, JOSEPH K. WAGONER and RICHARD J. WAXWEILER ;>nision of Surveillance, Hazard Evaluations and Field Studies, National Institute for Occupational Safety and Health, Main Post Office Building, Cincinnati, Ohio 45202 (US.A.) (Received May 13th, 1976) Summary The data presented demonstrate dearly that vinyl chloride (VC) is related to a significant excess of mortality from cancer of the liver, lung and brain among workers occupationally exposed to VC. The risk of dying from cancer of the iymnhatic and hematopoietic system also appears to increase with an increase in latency. These cancer sites could have been predicted by the animal bioassay conducted by Maltoni. With regard to the liver, even the histopathologic type of cancer (angiosarcoma) was observed first in experimental animals. A study of cancer mortality among populations residing proximate to VC polymeriza tion facilities also demonstrated an increased risk of dying from CNS and lymphatic cancer. These latter findings raise cause for concern about out-plant emissions of VC, but without further study these cancers obviously cannot be interpreted as being related to out-plant exposure to VC. Various test systems now have elicited a positive mutagenic response to VC. Thus, our observations of a significant excess of fetal mortality among the *ives of males, who were occupationally exposed to VC, raise public health concern that VC may be mutagenic in humas. With regard to the teratogenicity of VC, observations of a significant excess of children bom with birth defects were reported among populations residing proximate to VC polymerization facilities. Additional epidemiologic study is needed to determine whether a repeated pattern of excessive numbers of children born with birth defects can be observed in other communities with VC Polymerization-facilities. Introduction In 1930, the first adverse health effects of vinyl chloride (VC) were reported (22). Since then, numerous investigators have reported the toxic effects of VC on the central nervous system, the liver, the bones of the fingers and the lungs 132 [2,5,9--12,18,20,28]. More recently, the study of VC toxicity has broadened to assess the spectrum of carcinogenesis [2,15,16,21,26,29,30], mutagenesis [1,3,4,6,13,14,23,24]. and teratogenesis [8]. These efforts were stimulated by the work of Viola et al. [29] who, in 1971, reported the induction of tumors of the skin, lungs and bones in rats exposed by inhalation to 30,000 ppm of VC over a period of twelve months. Widespread concern for the carcinogenic activity of VC, however, did not occur until early 1974, when Creech and Johnson [2] reported four deaths from angiosarcoma of the liver among workers employed in the manufacture of polyvinyl chloride (PVC) resins. It was later learned that Maltoni had previously demonstrated VC-in duced hepatic angiosarcomas, lung adenomas, brain neuroblastomas, lymphomas and various other tumors in mice, rats and hamsters [15,16]. With regard to mutagenicity, several investigators have induced mutations via microbial test systems [1,14,24]. Also, VC metabolites have induced mutations in mammalian cells [6]. In addition, reports from several countries have demonstrated significant excesses of chromosomal aberrations among workers exposed to VC as contrasted with those not exposed [3,4,13,23]. Because of these observations and of the widespread exposure to VC among both workers and the general population, the National Institute for Occupational Safety and Health (NIOSH) in the U.S.A. undertook an epidemiologic program to evaluate the magnitude and spectrum of VC toxicity to humans. This program was 4-fold in nature and sought to evaluate: (1) The site-specific risk of cancer among workers exposed to VC [30]. (2) The risk of some cancers among populations residing proximate to VC polymerization facilities [8]. (3) The site specific risk of congenital anomalies among populations residing proximate to VC polymerization facilities [8]. (4) The risk of fetal wastage among the wives of workers occupationally exposed to VC [7]. Cancer risk among workers exposed to VC To assess the neoplastic risk of workers exposed to VC, a population con sisting of employes from four VC polymerizing facilitates was selected for cohort mortality study [30], Since occupationally induced cancers often take many years to become manifest, the study cohort was restricted to workers who had achieved five or more years of employment and for whom at least 10 years had lapsed since initial employment. Thus, the exposure period was five or more years and the latency period was 10 or more years. Follow-up for study cohort members was greater than 99%. Table I shows the total mortality experience among the study cohort. The expected numbers are based on United States mortality data applied to NIOSH's modified life-table method. For all malignant neoplasms, there were 35 observed vs 23.5 expected, the SMR was 149. This excess was significant at the P< 0.05 level of confidence. For total mortality, 136 deaths were observed vs 126.3 expected. Selecting a cohort with at least five years work experience and 10 years latency eliminated the "healthy worker effect" [19], and so there were more total deaths than expected and this was mostly at the expense of cancer mortality. Table II shows cancer mortality experience by greater than 10 and 15 year latency periods. At the greater than 10 year latency period, only biliary and 133 table 1 mortality experience among cohort workers exposed TO VINYL CHLORIDE Cause of death 1CD code a Observed Expected SMR b All malignant neoplasms Heart >,on-maliKnant respiratory diseases Cirrhosis Another causes <140--205) <400--443) <470--527) <581) 35 57 6 2 36 23.5 54.7 3.4 4.0 40.7 149 c 104 176 50 88 * ICO. International Classification of Diseases. 7Hi revision. 6 SMR. standardized mortality ratio. r significant at P < 0.05. /Z4' "5 iiver cancer deaths were significantly in excess; however, when the sub-cohort with 15 years of latency since initial employment was used, deaths from three categories of cancer were significantly in excess. The excess in mortality from cancer of the lymphatic and hematopoietic systems was not significant; however, the SME increased from 159 to 176 with an increase in latency. These comparisons show the importance of latency when looking for occupationally-induced cancers. table it CANCER MORTALITY BY INTERVAL SINCE INITIAL EXPOSURE AMONG WORKERS EXPOSED TO VINYL CHLORIDE Site of malignancy 10* yean 15+ yean AU malignant neoplasms Brain arret-CNS cance Respiratory system cancer BUiary and liver cancer l-rmphatic and hematopoietic system cancer Ml `her malignant neoplasms 0b*,, observed. r expected. ,, SN,R. standardized mortality ratio. s,*nifkant at /* < 0.05. s**nifteant at /* < 0.01. Ob*.* Exp. b SMR c Ob*. Exp. SMR Ob*. Exp. SMR Ob*. Exp. SMR Oba. Exp. SMR Ob*. Exp. SMR 35 23.5 149 d 3 0.9 329 12 7.7 156 0.6 1155 2.5 159 31 16.9 184 * 3 0.6 498 ' 11 5.7 194 1 7 0.4 1606 3 1.7 176 7 8.4 83 In laboratory studies, Maltoni has induced hepatic angiosarcomas, lung adenomas, brain neuroblastomas and lymphomas [16], Therefore, the predictive value of animal bioassay studies can be demonstrated, not only by the observation of a significant excess of angiosarcoma of the liver among VC workers, but also by significant excesses of lung and brain tumors. Cancer risks in populations residing near VC polymerization facilities The risk of some cancers among adult populations residing in the only three Ohio communities with VC polymerization facilities was assessed [8]. All three communities had at least one facility in operation by 1954 and Painesville had a second plant in operation by 1967. The community population sizes varied from 12,000--24,000. Between 1960--70 the population had remained stable in Painesville and Ashtabula, but had increased by 30% in Avon Lake. As reported previously, for the three communities taken as a whole, there were no apparent differences in racial origin or family income as compared to the aver- j\ . ~j f age for the State [ 8]. C/Ct-vu ***'***** As a result of previous findings 126], cancers of the central nervous system, -j- .r * QUL&tj lymphatic and hematopoietic systems, were selected for study. Because of possible error in death certificate data in terms of metastases from other primary sites, data for lung and liver cancers were not analyzed. - wwkf/Table III shoivs data for observed versus expected CNS cancer deaths in the J^ / white population by sex for the period 1958--73. North Ridgeville is the only communj^y shown which does not have a PVC polymerization facility. It was jdp c* tfCuA*. included in the analyses of the cancer data because it is located contiguous to julcJU Avon Lake and because it had a high incidence of children bom with birth defects [8]. If North Ridgeville had not been included in the analyses of cancer mortality, little difference in the results would have been observed. C/(S Expected values are based on the occurrence in the balance of the counties iJy+J t over the same period of time. As shown in Table 111, there was a significant excess of CNS cancer deaths in males. The excess was greatest in Painesville and North Ridgeville. With all communities combined, there were 27 CNS cancer deaths observed in males versus 14.1 expected; the SMR was 191. The difference was significant atP< 0.01, In females, with all groups combined, there was only a slight excess. With sex groups combined, the excess was significant at P< 0.01. It may be note worthy that one father-daughter combination for CNS tumor deaths occurred in Painesville. The daughter died of a papillary ependymoma in 1963 at the age of 16. The father died two years later of a glioblastoma multiforme at the age of 58. It would be difficult to determine whether this observation is the result of genetic or environmental factors. Table III also shows deaths from lymphomas. Although the differences between observed and expected deaths in males were not significant, there was a consistent excess in each community. In females, there were excess lymphoma deaths in two of the four communities. With sex groups combined, a significant excess of lymphoma deaths was observed in Ashtabula. With sex groups and communities combined, the number of observed deaths was 61 versus 48.7 expected. The SMR was 125. This was significant at 135 TABLE ltl OBSERVED AND EXPECTED DEATHS FOR THREE TYPES OF CANCER FOR RESIDENTS 45 years AND OLDER IN THE OHIO COMMUNITIES WITH VINYL CHLORIDE POLYMERIZATION FACILITIES, 1958-73 Expected numbers of cancer deaths based on occurrence over the same period of lime in the balance of me counties in which the communities arc located. Males Females Sexes combined Ashtabula I'airiesviltc \v<si\ Lake N. Ridgeville Communities Combined Ashtabula Pamesville Avon Lake N . Ridgeville Communities Combined Obs./Exp. SMR CNS cancer <X9 1 -- 192)a 7/ 6.1 12/ 3.8 2/ 2.3 6/ 1.9 115 316 d 87 316 b 27/14.1 191 c Obs./Exp. 6/ 3.8 2/ 2.8 1/ 1.8 21 1.5 11/ 9.9 Leukemia and aleukemia (204--207)a 13/J4.7 8/ 7.1 1/ 3.6 2/ 3.2 88 113 28 63 11/ 8.6 7/ 6.8 5/ 2.2 0/ 1.9 24/28.6 84 23/19.5 SMR 158 71 56 133 111 128 103 227 0 118 Obs./Exp. SMR 13/ 9.9 14/ 6.6 3/ 4.1 8/ 3.4 38/24.0 131 212 b 73 235 b 158 c 24/23.3 15/13.9 6/ 5.8 2/ 5.1 47/48.1 103 108 103 39 98 Ashtabula Pamesville Avon Lake N*. Ridgeville Communities Combined Lymphomas (200--203) 18/12.6 12/10.4 4/ 3.3 3/ 2.8 143 115 121 107 37/29.1 127 14/ 5.8 41 8.3 5/ 2.9 1/ 2.6 24/19.6 241 c 48 172 38 122 32/18.4 16/18.7 91 6.2 4/ 5.4 61/48.7 174 c 86 145 74 125 * International Classification of Diseases, 8th revision codes. h / < 0.05. c P < 0.01. dP< 0.002. 0.05 < P < 0.10. Data were then analyzed for leukemia and aleukemia mortali ty. As shown in Table III, the observed versus expected mortality was virtually identical. Birth defects among populations residing near VC polymerization facilities The occurrence of birth defects was studied [8) for the same three Ohio communities with VC polymerization facilities. Birth data for residents of North Ridgeville were not combined with data for the three index communities in the initial analyses (Table IV) because the high incidence of birth defects in North Ridgeville, which lies contiguous to Avon Lake, was identified in subsequent analyses (81- Data for North Ridgeville, however, were included in subsequent analyses (Table V) for specific birth malformations. The observa tions among community residents were compared to both the occurrence in 136 TABLE IV RESIDENT BIRTHS. MALFORMATION RATE PER 1000 LIVE BIRTHS IN OHIO AND IN THREE SELECTED COMMUNITIES AND OBSERVED VERSUS EXPECTED NUMBERS OF MALFORMA TIONS IN EACH CITY. YEARS COMBINED. 1970--73 Malformations, are based on codes 740--759 of the International Classification of Diseases. 8th revision. 1968. Area Births Malformations Rate/101 Number observed Number expected a 2 X Entire state of Ohio Ashtabula city Painesville city Avon Lake city All three communities combined 719,287 1900 1381 738 4019 10.1 17.4 18.1 20.3 18.2 7295 33 25 15 73 _ 19.3 14.0 7.5 40.8 -- 9.78 10.29 7.56 27.13 * Expected numbers are based on state rate per 1000 live births. b P < 0.01. c P < 0.001. TABLE V OBSERVED. EXPECTED AND RELATIVE RISK FOR SPECIFIC CONGENITAL ANOMALIES IN INDEX AREAS INCLUDING N. RIDGEVILLE. 1970--73 * * Excludes skin, hair and nails (757). b RR. relative risk (observed/expected). c International Classification of Disease codes. 8th revision, are shown in parentheses. 137 located, the differences remained significant. Therefore, whether you compare the occurrence in the communities to the expected, based on the average for the State, or.for the balance of the counties, the excess of children with congenital anomalies in the communities appears to be significant. Table V shows observed versus expected numbers and the relative risk for total and selected malformations. Although an increase in most organ systems was observed, the greatest excess of severe defects included malformations of the CNS, cleft lip and palate, club foot and genital organs. These observations have been reported previously in more detail [8], Fetal mortality among wives of workers exposed to VC Since VC had elicited a positive mutagenic response via microbial test systems [1,6,14,24] and also had been associated with significant excesses of chromosomal aberrations in the lymphocytes of workers occupationally exposed to VC [3,4,13,23], concern was expressed that VC may induce germinal mutations. Thus, a questionnaire-interview survey was conducted to ascertain the incidence of fetal loss (defined as any product of conception not born alive) among wives of workers exposed to VC [7a]. All current VC polymerization and polyvinyl chloride (PVC) fabrication workers were included for study together with a similar number of current rubber workers selected from work areas relatively free from known toxic materials and tsv\ 7X4 d matched as a group to the VC workers by age. No interviews were conducted with workers' wives and no data were obtained concerning maternal age, except indirectly through paternal age. Group participation rates ranged from 62--77 percent. Data for the wives of VC polymerization workers (study group) were juaWwuSl AjL i/c ot contrasted with data for the wives of PVC fabrication and rubber workers ("controls"), who were known to have had very low or no VC exposure, U'ruXrfl respectively. The data in Table VI show the paternal age distribution for fetal deaths according to the husband's exposure. Prior to husband's exposure, the crude fetal death rates for the control and study group were 6.9 and 10.1%, respectively. Subsequent to husband's exposure, the crude rates were 8.8 and 16.5%, respectively. As can be seen in Table VI, the excess in fetal mortality subsequent to husband's exposure, was associated with younger-aged husbands. For husbands 30 years of age and older, the rates for the control group 17/142 (12.0%) and primary VC exposure group 9/69 (13%) were about the same; whereas, for husbands less than 30 years of age, the rates for the control group 7/131 (5.3%) and primary VC exposure group 14/70 (20.0%) were significantly different, P < 0.001, xJ = 10.52, with one degree of freedom (df). The excess of fetal mortality among wives of younger-aged husbands may be a reflection of a practice of placing newly hired personnel because of little or no seniority, in jobs where occupational exposures to VC may have been worse. This hypo thesis, however, needs further assessment in other working populations. Since parental age was positively correlated with fetal mortality in this population and in a previous study [25], fetal mortality rates for the study group were adjusted to the age-distribution of the control group. 138 TABLE V] paternal ace distribution iqr petal deaths according to husband's vc expo SURE Paternal age group (years) <20 20-24 25-29 30--34 >35 All ages crude rate Mean paternal age at conception (year) Age-adjusted * rate "Controls" Pregnan cies Fetal deaths (N) (N) <%) Prior to husband's exposure 31 80 38 6 4 159 23.0 2 6.5 4 5.0 4 10.5 1 16.7 0 0.0 11 6.9 (6.9) Primary VC Exposure Pregnancies Fetal deaths (N) (N) (%) 7 44 56 27 14 148 26.4 0 o.o 2 4.5 7 12.5 & 18.5 1 7.1 15 to.i (6.1) >20 20--24 25--29 30-34 >35 AU ages crude rate Mean paternal age at conception (year) Age- adjusted " rate Subsequent to husband's exposure 1 45 87 87 55 273 30.4 o 0.0 4 9.3 3 3.4 7 8.0 10 18.2 24 8.8 (8.8) 0 22 48 36 33 139 30.2 O 0.0 3 13.6 11 22.9 3 8.3 6 18.2 23 16.5 (15.8) * Fetal mortality rates for primary VC exposure group are direct age-adjusted to the paternal age-distribu tion of the pregnancies In the control group (shown an parentheses). TABLE VII MEAN PATERNAL, NUMBER OF PREGNANCIES AND AGE-ADJUSTED FETAL DEATH RATES ACCORDING TO HUSBAND'S VC EXPOSURE "Controls" a Primary VC exposure b Number of families Mean paternal age at conception (years) Number of fetal deaths among wives Number ol pregnancies Age-adjusted fetal deaths,TOO preg. c Number of families Mean paternal age at conception (years) Number of fetal deaths among wives Number of pregnancies Age-adjusted fetal deaths/100 preg. c Prior to husband's exposure 95 23.0 11 159 6.9 70 26.4 15 148 6.1 Subsequent to husband's exposure 113 30.4 62 30.2 24 273 8.8 23 139 15.8 d r u * Rubber and PVC fabrication workers, b VC polymerization workers. c Rates age-adjusted to ``control" group paternal age distribution. d Subsequent to husbands' exposure, the frequency of fetal deaths among wives was significantly greater in the primary VC exposure group as compared to the "controls" IP < 0.05) or to the frequency in the study group prior to husband's exposure (P < 0.02) bv age-adjusted Chi-square testing. / 139 The data in Table VII show fetal death rates per 100 pregnancies for the wives of male workers in the control and study groups. Prior to their husband's exposures, the rates were 6.9% and 6.1% for the control and study groups, respectively. Subsequent to husband's exposure, however, the rates were S.8% for the control group versus 15.8% for the study group. This difference was significant at the P<0.05 level by Mantel-Haenszel Chi-square testing (171 (x2= 4.84, df = 1). Further, the before and after exposure comparisons indicated changes in rates from 6.9 to 8.8% for the control group as compared to a change from 6.1 to 15.8% for the study group. The rates for before and after husband's exposure in the study group were significantly different iP < 0.025, x3 = 5.78, df = 1). For the study group, it may be noted that before exposure age-adjustment reduced the crude rate from 10.1 to 6.1%; whereas, after exposure the crude rate was reduced from 16.5 to 15.8%. The greater reduction in rates for the before exposure age-adjustment resulted from a difference in the age distribution of the husbands in the two groups. As can be seen in Table VII, prior to husband's exposure, the mean paternal age in the study group was 26.4 years as compared to only 23.0 years for the control group, whereas, subsequent to husband's exposure, the mean ages of the two groups were virtually the same, i.e., 30.4 years versus 30.2 years. In both situations mean age was a good measure of central tendency. To determine whether women who had chronically experienced abortions might have weighted the results in favor of a higher fetal mortality rate in the primary VC exposure group, data for the pregnancies of women who had two, three or four or more spontaneous abortions were eliminated. The data were then recalculated to determine whether or not the trend of a greater mis* carriage rate could be maintained. As shown in Table VIII, the trend was maintained for each analysis. TABLE VIII NUMBER OF PREGNANCIES AND AGE-ADJUSTED FETAL DEATH RATES ACCORDING TO HUS BAND'S VC EXPOSURE EXCLUDING PREGNANCIES IN WOMEN WITH >2. 3 OR 4 FETAL DEATHS Halts for lhe primary VC exposure group are age-adjusted to the Control group Control* Number of PKfDUein Fetal death rate Primary VC exposure Number of pregnancies Fetal death rate Before husband's exposure Alter husband's exposure Seiore husband's exposure Alter husband's exposure ^2 Fetal deaths excluded 155 2&S 5.8% 4.7% >3 Fetal deaths excluded 159 265 6.9% 6.8% ' 126 111 141 120 1.7% 6.2% 3.1% 10.8% he lore husband's exposure a fter husband's exposure Fetal deaths excluded 159 265 6.9% 6.6% 142 127 5.8% 11.8% 140 As reported previously [7a], additional analyses suggested that thesignifican excess in fetal mortality after the husband's exposure would not seem to be tin result of bias from interviewers nor from respondents. Because of the highh volatile nature of vinyl chloride [27], carry-home exposure to the wife wouk seem unlikely. Therefore, the leading possibility for the mechanism involvec would seem to be germ-cell damage in the male through direct VC exposure. References 1 Baitsch, H., C. Malavielle and H. Montesano. Human, rat and mouse liver-mediated mutagenicity of vinyl chloride in 5. lyphimurium strains. Int, J. Cancer. IS (1975) 429--437. 2 Creech, J.L., Jr. and M.N. Johnson, Angiosarcoma of liver in the manufacture of polyvinyl chloride. J. Occup. Med., IS (1974) 150--151. 3 Ducatman, A., K. Hirschhorn and i.J. SelLkoff, Vinyl chloride exposure and human chromosome aberrations. Mutation Res., 31 (1975) 163--168. 4 Funes-Cravioto, F., B. Lambert and J. Lindsten, Chromosome aberrations in workers exposed to vinyl chloride. Lancet 1 (1975) 459. 5 Gedrigk, P., R. Muller and H. Bechtelsheimer, Morphology of liver damage among polyvinyl chloride workers, Ann. NY Acad. Sci.. 246 (1976) 278--285. 6 Huberman, E., H. Bartsch and L. Sachs. Mutation induction in Chinese hamster V79 cells by inn vinyl chloride metabolites, chlorocthylene oxide and 2-chloroaeelaldehyde, Int. J. Cancer. 16 (1975) 639--644. 7a Infante, P.F., J.K. Wagoner and AJ. McMichael, Genetic risks of vinyl chloride; Lancet. 1 (1976) 734--735, 7b Infante, P.F., J,K. Wagoner, A .J. McMichael, R.J. Wax writer and H. Falk, Genetic risles of vinyl Chloride, Lancet. 1 (1976) 1289-1290. 8 Infante, P.F., Oncogenic and mutagenic risks in communities with polyvinyl chloride production facilities. Ann. NY Acad. Sci.. 271 (1976) 49--57. 9 Kramer, C.G. and J.E. Mutchler, The correlation of clinical and environmental measurements for workers exposed to vinyl chloride, J. Amtt. lnd. Hyg. Assoc.. 33 (1972) 19--30. 10 Lange, C.E., S. Juhe, G. Stein and G. VeJtman. Further results in polyvinyl chloride production workers, Ann. NY Acad. Sci.. 246 (1975) 18--21. 11 Lester. D., L.A. Greenberg and W.R. Adams, Effects of single and repeated exposures of humans and rats to vinyl chloride, J. Amer. lnd. Hyg. Assoc.. 24 (1963) 265--275. 12 Lilts. R., H. Anderson. WJ. Nicholson, S. Daum. A.S. Fischbein and l.J. Sclikoff. Prevalence of disease among vinyl chloride and polyvinyl chloride workers. Ann. NY A-ad Sci , 246 (1975) 22--41 13 Loken, E. and E. Thiis-Evensen, Preliminary report on the medical examination of 288 employees at the PVC plant, Norsk Hydro a.s., Porsgrunn Fabrikker, PorsgTunn, Norway, unpublished. 14 Loprieno, N., R. Barale and S. Baroncelii. Evaluation of the genetic effects induced by vinyl chloride monomer (VCM) under mammalian metabolic activation: studies in vitro and in vivo. Mutation Res.. 40 (1976) 85--95. 15 Maltoni. C.. Preliminary report on the carcinogenicity bioassays of vinyl chloride, U.S. Dept. Labor Informal Fact-Finding Hearing on Possible Hazards o! Vinyl Chloride Manufacture and Use. Wash ington. D.C., February 1974. 16 Maltoni. C. and G. Lefemine. Carcinogenicity bioassays of vinyl chloride. Current results. Ann. NY Acad. Sci.. 246 (1975) 195--218. 17 Mantel, N. and W. Haenszel, Statistical aspects of the analysis of data from retrospective studies of disease. J. Natl. Cancer. Inst.. 22 (1959) 719--748. 18 Marsteller. H.J., W.K. Lelbach. R. Muller and P. Gedigk. Unusual splenomegaJic liver disease as evi denced by peritoneoscopy and guided liver biopsy among polyvinyl chloride production workers. Ann. NY Acad. Sci., 246 <1975)95--134, 19 McMichael, A.J., S.G. Haynes and H.A. Tyroler. Observations on the evaluation of occupational mortality data. J. Occup. Med., 17 (1975) 128--131. 20 Miller, A., A.S. Teimein, M. Chuang, I.J. Selikoff and R. Warshaw. Changes in pulmonary function in workers exposed to vinyl chloride and polyvinyl chloride. Ann. NY Acad. Set., 246 (1975) 42--5-- 21 Nicholson, W.J., E.C. Hammond, H. Seidman and I.J. Selikoff, Mortality experience of a cohort of vinyl chloride-polyvinyl chloride workers, Ann. NY Acad. Sci.. 246 (1975) 225--230- 22 Patty. F.A.. W.P. Yant and C.P. Waite. Acute response of guinea pigs to vapors of some new com mercial organic compounds, Public Health Reports, 45 (1930) 1963--1971. 23 Purchase, L.F.H., C.R. Richardson and D. Anderson, Chromosomal and dominant lethal effects of vinyl chloride. Lancet, 11 (1975) 410--411. d 3D