Document zQg7yendbXmOGyeQ7e0zQrYrm

Original Articles 4 Cancer Mortality of a Croup of Canadian Workers Exposed to i Vinyl Chloride Monomer G. Theriault, M.D., Dr.P.H., and P. Allard, M.D., M.Sc. The present study was undertaken to find out whether their first exposure more than 15 years before their death. there was an excess of cancer mortality from causes other While most authors have not been able to document a than angiosarcoma of the liver among a group of workers significant excess of any other cancer in their studies, they heavily exposed to vinyl chloride monomer (VCM). The mortality of 451 workers exposed to VCM for more than have noted that such an excess may exist. In order to better document the cancer mortality asso five years was compared with that of 870 workers from the ciated with exposure to VCM, the present study of the same company who had not been exposed to VCM. The workers of the Shawinigan vinyl chloride polymerization relative risk for digestive cancer was significantly higher plant was undertaken. than 1 (6.25, confidence interval 2.69 to 14.52) in the ex The main objectives of the study were (1) to compare posed group. The standardized mortality ratio (SMR) for mortality among vinyl chloride workers with the mortality digestive cancer was also higher (5MR 259.26 p < 0.01) of workers from a nearby industrial complex; and (2) to than that of the general population. No other cancer was in compare the mortality of the VCM workers with that of excess. Since the exposed workers are known to have had a the general population. The Shawinigan vinyl chloride cigarette smoking experience similar to that of those who plant opened in 1943 as an independent company which * }vere not exposed, it is concluded that the association be produced both VCM and PVC (polyvinyl chloride). In tween lung cancer and VCM exposure, if present, is indeed 1958, it was merged into an industrial complex (to be used rather small. as the comparison group) of which it became the Cana dian Resins Division. In the late 1960s, the VCM produc tion ceased and only the polymerization process con Several studies have demonstrated that vinyl chloride tinued operation. In 1972, this division was sold and again became an independent plant. Over the years, the VCM monomer (VCM) produces angiosarcoma of the liver and PVC productions fluctuated. The average annual among people exposed at work.17 There is, however, work force at the plant was 225 men. In 1965, VCM pro much controversy about the capability of VCM to gener duction was estimated at 90 million pounds and PVC at ate other types of cancer in man. One author,* using a 60 million pounds.'0 In 1977. PVC production was 22.3 proportional mortality ratio technique, described ex million pounds." Since several episodes of unconscious cesses of lung cancer (observed, 13; expected, 7.9; obs/exp ness among workers were reported, the level of VCM in 1.6) and brain cancer (observed, 5; expected, 1.2; obs/exp the air at the workplace is believed to have been high dur 4.2) in a group of 161 workers from two vinyl chloride ing this period. Ten cases of angiosarcoma of the liver plants.* Another author,' who studied the mortality of a have been reported since then.7 group of 1,294 American workers exposed to VCM, re The industrial complex from which the comparison ported excesses of lung cancer (observed, 11 expected, group was drawn comprises several divisions: a carbide 7.5; SMR 194, p < 0.05) and brain cancer (observed, 3; ex division which produces calcium carbide and acetylene pected, 0.6; SMR 498, p < 0.05) among men who had had black; a chemicals division which produces several chem icals derived from acetylene: acetone, chloral, butanol, i From tht Department of Social and Preventive Medicine School of Medicine. acetic acid, acetaldehyde, several acetates, solvents and Laval University, Ste-Foy. Quebec, Canada, ClK 7P4. This research was financed partly under Grant No 780396 from the Conseil resins; a stainless steel division which consisted of a small de la Recherche en Sante du Quebec. foundry and a division formed in 1957, which produced Journal of Occupational Medicine/Vol. 23, No. 10/0ctober 1981 671 | (J E ` p Fig 1. -- Population under study. sulfuric acid, chlorine and cyanide. This complex started The men whose number of years of work for these operations in 1927; some of its divisions opened at a later companies, as established through occupational histories, date; others closed in the late 1960s or early 1970s. The did not exceed five were removed from the study. The re annual number of male workers averaged 1,000.'J maining population was then subdivided into three groups. The first group (known as the unexposed cohort) \ Methods consisted of men who stated they had not been involved Population Under Study. -- All production workers with the production of VCM or PVC, or both, for a period whose names appeared on the unions' lists or the com greater than five months. The second group (constituting panies payrolls of the VC plant and the industrial complex the exposed cohort) comprised men who declared they between January 1,1948, and December 31,1972, consti had worked on the production of VCM-PVC for more tuted the population under study. Several methods (old than five years. Finally, the third group consisted of the addresses, telephone books, Quebec motor vehicles bureau, inquiries with fellow workers and families, men who had worked on the production of VCM-PVC for a period of six months to five years. This group was ex t searches in homes for retired persons) were used to trace cluded from analysis because it neither fulfilled the these men, 18% of whom were still working for the same definition of the exposed nor the unexposed cohort. companies. Once identified, the worker himself or his Data Sources. -- The vital status of these men was W next of kin (spouse, child or close relative) was questioned established as of December 31, 1977. Causes of death at home or it work by a trained interviewer. The ques were determined from death certificates. They were H* 1 I tions permitted the reconstruction of a detailed occupa coded according to the 8th revision (ICDA) by one of the tional history within and outside the companies under authors, who was unaware at the time of the coding as to 0) (fr ostudy. Questions were also asked regarding smoking which group the man belonged. For cancer cases, a histo- habits. For each deceased person believed to have been exposed to VCM, the occupational history was confirmed pathological confirmation was sought on autopsy, biopsy or cytology reports contained in the medical files at the S' with the help of a fellow worker. hospital. Information on Quebec's population and mor- 672 Mortality of Workers Exposed to Vinyl Chloride Monomer/Th6riault and Allard Table 1. -- Age Distribution of the Men in the VCMExposed and Unexposed Cohorts as of December 31, 1977 (Deceased Excluded). Age 25-29 30-34 35-39 40-44 45-49 50-54 55-59 50-64 65-69 70-74 75 + 7o;ai Exposed No. % 0-- 0-- 1 0.3 28 7.1 75 19.1 96 24.5 102 26.0 53 13.5 30 7.7 6 1.5 1 0.3 392 100.0 Unexposed No. % 1 0.2 3 0.5 6 0.9 45 7.1 55 8.6 84 13.2 112 17.6 110 17.3 113 17.7 65 10.2 43 6.7 637 100.0 tality by cause, age and sex was secured from Statistics Canada for the year 1971. In order to ensure the validity of the results, only the causes of death which appeared on the death certificates were used in the comparisons. Statistical Analysis. -- The results were analyzed using the "man-years at risk" method as described by Hill.'3 The number of person-years of observation was established for each five-year age group and for each five-year period between 1948 and 1977. The mortality in the exposed cohort was compared with that in the unexposed cohort according to the method of relative risk. The tests of significance used were those of Mantel, Haenszel14 ,s and Miettinen.'6 '7 The observed mortality was also compared with the mor tality expected from the population of the Province of Quebec according to the method of the standardized mortality ratio (5MR).'* The test of significance used was the one described by Bailar and Ederer." Quality Control Procedures. -- A thorough occupa tional history was developed for workers who were ex posed within the VC plant. The questionnaires presented a detailed profile of the plant while specifying possible jobs within its major units. Before the beginning of the study, it was pre-tested with former employees. Interviewers who contacted the workers had received training by the same person and were supervised during the entire study. The completed questionnaires were re viewed weekly by one of the authors and missing infor mation was sought by telephone. Data entered in the computer were systematically checked against the origi nal questionnaires. Results As shown in Fig 1, the population under study originally comprised 1,659 workers. Of these, 48 men were untrace- able, leaving the number in the study population at 1,611 (or 97.1 % of the original number). After the removal of the 156 men who had worked at the companies for a period not exceeding f ive years and the 134 men who had been exposed to VCM for a perod extending from six months to five years, the exposed cohort comprised 451 men and the unexposed one, 870 men. In the exposed cohort, the proportion of men who responded to the questionnaire for themselves was 84.3%. In the unex posed cohort, this proportion was somewhat lower at 68.2%. The age distributions of the workers in the exposed and unexposed cohorts as of December 31,1977, reveals that workers in the unexposed cohort (Table 1) were somewhat older. Since the industrial complex had been in operation for several years when the VC plant opened, the finding of older people in the latter cohort was predictable. Table 2 shows the distribution of the exposed workers according to their length of exposure and their observa tion period (time between first employment and end of follow-up). Among these, 81 % (365 out of 451) had an ob servation period of 15 years or more. For 23% of them, the time elapsed between first exposure and the end of the follow-up period exceeded 29 years. The mean obser vation period for the exposed cohort was 22.6 years. Table 3 gives the diagnosis at death and the diagnosis after histopathological review of all the cancer cases found in the exposed group. On the death certificates, eight cancers of the liver were reported. The histopatho logical review rejected one of them as being a cancer of the sigmoidal flexure of the colon, but added another as an incorrectly diagnosed cancer of the peritoneal cavity. All eight cancers of the liver turned out to be angiosar coma of the liver. It is noted that two more angiosar comas of the liver were certified as cirrhosis on the death certificates and were considered as such throughout this study. Table 4 compares the mortality of the exposed workers with the mortality of the unexposed workers. The con fidence interval (95%) of the relative risk indicates that there is a significant excess of mortality due to cancers of the digestive system in the exposed group (RR 6.25, con fidence interval 2.69 to 14.52). There is no excess in the other causes of death. Table 5 shows the standardized mortality ratios (SMRs) for the exposed workers in comparison with the entire population of the Province of Quebec. The SMR for all causes is 83.02, thereby showing no excess of death in the exposed cohort Although the SMRs for all cancers, respi ratory system diseases and digestive system diseases ex- 2113S003 Table 2. -- Distribution of Exposed Workers by Length of Exposure and Observation Period. Exposure (Years) 5- 9 10-14 15-19 20-24 25-29 30 + Total Observation Period (Years) 5-9 10-14 15-19 20-24 25-29 30 + 30 28 7 40 2 4 -- 28 26 -- -- 17 ------ 19 17 35 19 22 24 5 8 15 -- -- -- -- 34 33 -- -- -- -- -- 38 30 56 50 116 96 103 Journal of Occupational Medicine/Vol. 23, No. 10/October 1981 Total 111 95 79 61 67 38 451 673 Table 3. -- Age at Death, Diagnosis on Death Certificate and Diagnosis on Histological Review o< the 20 VCM-Exposed Workers Who Died of Cancer. Case Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Death Certificate Cancer of the liver Cancer of the intestine Carcinoma of the trachea Angiosarcoma of the peritoneal cavity Cancer of the lung Cancer of the liver Cancer of the liver Cancer ot the prostate Angiosarcoma of the liver Cancer of the pancreas Adenosarcoma ot the intestine Hepatoma Cancer of the liver Cancer of the pancreas Chondrosarcoma Hepatoma Primary cancer of the liver Cancer of the digestive track Cancer ot the biliary ducts Lymphoid leukemia Histological Review Cancer ot the sigmoid flexure of the colon Metastatic cancer ot the colon Carcinoma of the trachea Angiosarcoma of the liver (no review) Angiosarcoma ot the liver Angiosarcoma of the liver Cancer of the prostate Angiosarcoma ot the liver Epithelioma of the head of the pancreas Adenosarcoma ot the intestine Angiosarcoma of the liver Angiosarcoma ot the liver Anaplastic epithelioma ot the pancreas Chondrosarcoma Angiosarcoma of the liver Angiosarcoma of the liver Cancer of the cecum Cancer of the head of the pancreas Lymphoid leukemia Age at Death (Years) 73 56 50 41 54 42 57 64 62 40 53 53 53 59 43 48 53 52 54 48 Total Exposure (Years) 15 16 15 11 16 19 17 26 18 8 25 24 22 29 7 22 5 26 12 21 Latency Period (Years) 28 17 15 11 16 19 23 29 23 13 28 24 23 30 10 22 12 26 12 21 ceed TOO, they are not statistically significant. On the a group of workers with the mortality of a whole popula other hand, deaths by accident and violence are signifi tion are well documented. They are termed: "healthy cantly lower (p < 0.01). As for specific cancer deaths, a population selection effect," "survivor population significant excess is noted for cancer of the digestive^ effect,"20 "ethnic/cultural/social characteristics." They are system only. This excess is accounted for exclusively by j believed to underestimate the actual mortality of a group cancer of the liver (8 cases observed versus 0.14 expected, T of workers and therefore to underscore the effects of the SMR 5,714.29). No other cancers are in excess. i risks to which these individuals are being exposed at The cancer mortality among the workers who were still 4 work. There are no easy ways to avoid such difficulties. alive 15 years after their first exposure shows the same One method would be to compare a group of workers results as the mortality in the entire group, namely that with another group of workers with the same characteris only cancers of the digestive system are in excess in the tics except for the exposure to the risk under study. The exposed cohort (Table 6). problems encountered in so doing are no easier. Although the group most similar to industrial workers is probably Discussion another group of industrial workers from the same town, Difficulties encountered in comparing the mortality of it is almost impossible to think of a group of industrial "3 1 i V* .4 Table 4. -- Deaths Observed in the VCM-Exposed and Unexposed Cohorts, Age-Adjusted Relative Risks and Confidence Intervals of the Relative Risk. Cause of Death No. ot Cases Exposed Unexposed Relative Risk 95% Confidence Interval of the Relative Risk I All cancers Mouth and pharynx Digestive (140-209) ; (140-149) (150-159) 20 52 1.48 0 2-- 14 9 6.25 0.84- 2.61 -- 2.69-14.52' Respiratory (160-163) 2 20 0.36 0.08- 1.58 Bone, skin, connective (170-173) 2 2 1.67 0.21-12.99 Urogenital Eye. CNS (185-189) (190-192) 1 7 0.83 0.10- 7.10 0 2-- -- ' Leukemia Other cancers (200-209) 1 4 1.15 0.07-17.80 0 6-- -- Cardiovascular (390-458) 25 124 0.80 0.47- 1.36 Respiratory (460-519) 6 15 0.81 0.44- 9.35 Digestive (520-577) 4 12 1.44 0.49- 4.25 Accidents and violence (800-899) 2 15 0.51 0.21- 1.24 Others Total 2t 15 0.57 59 233 1.07 0.14- 2.39 0.79- 1.45 'Significant according to Miettinen's test17 fOne cause unknown J 674 Mortality of Workers Exposed to Vinyl Chloride Monomer/Thdriault and Allard Table 5. -- Observed and Expected* Numbers of Deaths in the VCM-Exposed Cohort and Standardized Mortality Ratios (SMRs). Causa ol Death All cancers Mouth and pharynx Digestive Respiratory Bone, skin, connective Urogenital Eye, CNS Leukemia Other cancers Cardiovascular Respiratory Digestive Accidents and violence Others Total (140-209) (140-149) (150-159) (160-163) (170-173) (185-189) (190-192) (200-209) (390-458) (460-519) (520-577) (800-999) Observed 20 0 14 2 2 1 0 1 0 25 6 4 2 2* 59 Expected 16.37 0.64 5.40 5.78 0.38 1.33 0.60 1.67 0.54 31.67 3.21 3.85 10.58 5.40 71.07 SMR 122.17 -- 259.261 34.60 526.32 75.19 -- 59.88 -- 7B.94 186.91 103.90 18.901 37.03 83.02 'Expected numbers obtained from 1971 males Quebec death rates applied to person-years in each age group tp<0.01 tOne cause unknown workers not exposed to some kind of carcinogenic risks. In the present study, the authors decided to compare the exposed workers with both the general population and a group of unexposed workers. It must be noted that among the unexposed workers, no causes of death were found to be in excess when their mortality was compared with that of the general population. The comparison of the mortality of the exposed and the unexposed workers gave results similar to the one reached by the comparison with the general population. The number of cancer of the digestive system was signifi cantly higher in the exposed cohort. No other cancer was in excess and there even seemed to be a deficit (not statis tically significant) in respiratory cancers. Considering the fact that animals experimentally ex posed to VCM have developed not only angiosarcomas of the liver but also malignant tumors of several organs (skin, lung, bone, Zimbal glands, kidneys, mammary glands)2' 2225 and since several epidemiological studies on VCM-exposed workers have found an excess of lung cancer,24 the results of this study of heavily exposed men are surprising. Three questions can be raised. One con cerns the ascertainment of the population. In retrospec tive studies conducted from old company or union rec ords it sometimes happens that records of people who die or leave get lost. These losses could account for the low cancer rates observed. But this is a remote possibility, since the approach followed was a historical-prospective retrospective one, which began with old lists of workers composed in 1948 and 1952 and followed almost every one through time, adding new workers as they first ap peared on more recent lists. This was done without knowl edge of exposure to VCM. A second reason for low lung cancer rates could be a bias in the classification of persons as exposed or not ex posed. The classification was made essentially on the basis of personal recollection for living persons and by family members with confirmation by co-workers for deceased persons. To minimize such bias, much attention had been given to the occupational history (with indica tion of the buildings the subject had worked in) and workers with VCM exposure shorter than five years were rejected. There remains the possibility that some of the men who died of liver cancer, had they lived longer, might have developed lung cancer. The small size of the population under study increased the probability of not detecting an excess that was actual ly present. This was a case where "the number of person/ years at risk may be too small to detect an increased risk of cancer."25 It was particularly true for rare cancers such as cancer of the brain and cancer of the lymphatic system. Inasmuch as cancer of the lung was concerned. Table 6. -- Relative Risks and SMRs for Cancer Deaths Among Men Having an Observation Period of 15 Years and More. Cancers All cancers Mouth and pharynx Digestive Respiratory Bone. skin, connective Urogenital Eye. CNS Leukemia Other cancers (140-209) (140-149) (150-159) (160-163) (170-173) (185-189) (190-192) (200-209) No. 15 0 11 2 0 1 0 1 0 p < 0.01'9 Journal of Occupational Medicine/Vol. 23, No. 10/0ctober 1981 Relative Risk 1.30 -- 5:68 0.35 -- 0.82 -- 0.59 -- 95% Confidence Interval ot the Relative Risk 0.97- 1.74 -- 2.72-11.58 0.25- 4.73 -- -- -- -- -- SMR 129 -- 281* 47 -- -- -- -- -- 675 the situation was somewhat different. When one applies the Quebec lung cancer death rate after adjustment for age to the exposed cohort, the probability of finding an excess twice as high as in Quebec was 73%. This proba bility would have been 99% had the excess been three times as great in the exposed group." In view of the fact that the smoking experience of the exposed workers was similar to that of the unexposed and considering that the subjects in this study are believed to have been heavily exposed to VCM, it seems reasonable to conclude that the excess of lung cancer associated with VCM exposure, if present, is indeed rather small. The authors wish to express their gratitude to the following who contributed to this study, workers, union representatives and management personnel of the companies Canadian Resins and Chemicals. Shawinigan Chemicals. B.F. Goodrich and Gulf Canada. Drs H Hurra and j. Fabia. epidemiologists, and Dr. F. Delorme, pathologist, for their help and advice; the several hospitals con tacted. and the several persons who assisted with the collection of information and the writing of this article. References 1. Creech |L |r and Johnson MN: Angiosarcoma of liver in the man ufacture of polyvinyl chloride. I Occup Med 16:150-151, 1974. 2 Tabershaw IR and Caffey WR: Mortality study of workers in the manufacture of vinyl chloride and its polymers. 1 Occup Med 16:509-518. 1974. 3. Nicholson W|. Hammond EC. Seidman H. et al: Mortality experi ence of a cohort of vinyl chloride polyvinyl workers. Ann NY Acad Sci 246:225-230. 1975. 4 Duck BW, Carter |T. and Coombes E|: Mortality study of workers in a polyvinyl-chloride production plant. Lancet 2:1197-1199. 1975. 5 Byren D. Engholm C. Englund A. and Westerholm P: Mortality and cancer morbidity in a group of Swedish VCM and PVC production workers. Environ Health Perspect 17:167-170. 1976. 6 Fox AJ and Collier PF: Mortality experience of workers exposed to vinyl chloride monomer in the manufacture of polyvinyl chloride in Great Britain. Br ) Ind Med 34:1-10. 1977. 7. Delorme F and Theriault C: Ten cases of angiosarcoma of the liver in Shawinigan. Quebec / Occup Med 20:338-340, 1978. 8. Monson R. Peters ). and Johnson M: Proportional mortality among vinyl chloride workers. Lancet 2:397-398, 1974. 9. Waxweiler JR. Stringer W, and Wagner JJ: Neoplastic risk among workers exposed to vinyl chloride. Ann NY Acad Sci 270:40-48,1976. 10. Special Report by the Editors: PVC aims for a bigger share in plastics. Can Chem Process 49:41-44. 1965. 11. Conseil des Sciences du Canada: Vue d'ensemble des dangers de la contamination par le chlorure de vinyle au Canada. Quebec Science (Suppl) :1-24. 1977. 12. LaRochelle F: Shawinigan depui$75 ans. Ateliers de I'imprimerie Publicite Paquet Enr.. 1976, pp 497-525. 13. Hill ID: Computing man years at risk. Br / Prev Med 26:132-134. 1972. 14. Mantel N and Haenszel W: Statistical aspects of the analysis of data from retrospective studies of disease. ) NatI Cancer Inst 22:710-748. 1959. 15. Mantel N: Evaluation of survival data and two rank order statis tics arising in its consideration. Cancer Chemother Rep 50:163-170. 1966. 16. Miettinen OS: Standardization of risk ratios. Am j Epidemiol 96:383-388. 1972. 17. Miettinen OS: Simple interval-estimation of risk ratio. Am I Epidemiol 100:515-516. 1974. 18. Caffey WR: A critique of the standardized mortality ratio. / Oc cup Med 3:157-160. 1976. 19. Bailar JC and Ederer F: Significance factors for the ratio of a Poisson variable to its expectation. Biometrics 20:639-642, 1964. 20. Fox A| and Collier PF: Low mortality rates in industrial cohort studies due to selection for work and survival in the industry. Br) Prev Soc Med 30:225-230.1976. 21. Viola PL: Cancerogenic E ffect of Vinyl Chloride. VoI. 29, X Inter national Cancer Congress. Houston, 1970. 22. Viola PL. Bigotti A. and Caputo A: Oncogenic response of rat skin, lungs and bones to vinyl chloride. Cancer Res 31:516-519. 1971. 23. Maltoni C and Lefemine C: Cancerogenicity bio assays of vinyl chloride: Current results. Ann NY Acad Sci 246:195-218. 1975. 24. Wagoner )K: Vinyl chloride and pulmonary cancer. / Env Pathol Toxicol 1:361-362. 1978. 25. Pell S. O'Berg MT. and Karrh B: Cancer epidemiologic surveil lance in the DuPont Company. ) Occup Med 20:725-740,1978. * f Reducing Financial Risks There is too little evidence" that the nuclear industry, which was never very good at calculating its own interest, is seriously contemplating the effect of another Three Mile Island on its future -- or its fortunes. The industry has already returned to compfaints that the safety bureaucracy is nitpicking it to death, running up construction costs and delaying licenses. The latest complaint is that, in the aftermath of the TMI accident, NRC diverted safety reviewers to deal with operating reactors and fell behind in ap proving new licenses. There is a kind of tunnel vision in the failure to see the connection between safety and the special financial risk involved in nuclear investment. To use the current energy crisis to avoid safety requirements is to ignore the fact that safety is essential to protec ting the heavy investment in nuclear power. Some hope lies in the fact that, although the industry thinks safety requirements are a pain in the neck, they may be forced to look at them as an investment in financial public relations to alleviate the bankers' skepticism. -- Victor Gilinsky. Commissioner, the Nuclear Regulatory Commission, as quoted in "Notable and Quotable." in The VVa// 5ffeef Journal. May 21. 1981 , I 21136000 676 Mortality of Workers Exposed to Vinyl Chloride Monomer/Th6riault and Allard