Document 15DLJedZ6oKrLNJNxk1v11Mpq

Ow^cAdsuJU ^JU Original Articles Cancer Mortality of a Group of Canadian Workers Exposed to 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 there was an excess of cancer mortality from causes other than angiosarcoma of the liver among a group of workers heavily exposed to vinyl chloride monomer (VCM). The mortality of 451 workers exposed to VCM for more than five years was compared with that of 870 workers from the same company who had not been exposed to VCM. The relative risk for digestive cancer was.s/gn)7/can(/y_ higher than 1 (6.25, confidence interval 2.69 to 14.52) in the expost*! group, the slnntbrdi/cr! mortality ratio (SMR) lor digestive cancer was also higher (SMR 259.26 p < 0.01) than that of the general population. No other cancer was in excess. Since the exposed workers are known to have had a cigarette smoking experience similar to that of those who pere not exposed, it is concluded that the association betweeQ.lung cancer and. VCM exposure,. ifjireseptf is. indeed tather. small. Several studies have demonstrated that vinyl chloride monomer (VCM) produces angiosarcoma of the liver among people exposed at work.1-7 There is, however, much COntrover<v about the rxnahilify of VCM to nenerate otht: 4ypes ot cancer in man. One author,1 using a proportional mortality ratio technique, described ex* cesses of lung cancer (observed, 13; expected, 7.9; obs/exp 1.6) and brain cancer (observed. 5; expected, 1.2; obs/exp 4.2) in a group of 161 workers from two vinyl chloride plants.' Another author,* who studied the mortality of a group of 1,294 American workers exposed to VCM, re ported excesses of lung cancer (observed. 11 expected. *7.5; SMR 194, p < 0.05) and brain cancer (observed, 3; ex pected, 0.6; SMR 498, p < 0.05) among men who had had From the Department of Social and Preventive Medicine School of Medicine. Laval University. Ste-Foy. Quebec. Canada. G1K 7P4 This research was financed partly under Grant No 790398 from die Cornell de la Recherche en Sant* du Quebec their first exposure more than 15 years before their death. While most authors have not been able to document a significant excess of arty other cancer in their studies, they have noted that such an excess may exist. In order to better document the cancer mortality asso ciated with exposure to VCM, the present study of the workers of the Shawinigan vinyl chloride polymerization plant was undertaken. The main objectives of the study were (1) to compare mortality among vinyl cl4uride workers with Ihe mortality of workers from a nearby industrial complex; and (2) to compare the mortality of the VCM workers with that of the general population. The Shawinigan vinyl chloride plant opened in 1943 as an independent company which produced both VCM and PVC (polyvinyl chloride). In 1958, it was merged into an industrial complex (to be used 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 corv tinued operation. In 1972, this division was sold and again became an independent plant Over the years, the VCM and PVC productions fluctuated. The average annual ' work force at the plant was 225 men. In 1965, VCM production was estimate"! at 90 million pounds and PVC at 60 million pounds. * i." IT7'. **vC production was 22.3 million pounu*.'" Since several episodes of unconscious ness among workers were reported, the level of VCM in the air at the workplace is believed to have been high dur ing this period. Ten cases of angiosarcoma of the liver have been reported since then/ The industrial complex from which the comparison group was drawn comprises several divisions: a carbide division which produces calcium carbide and acetylene black; a chemicals division which produces several chem icals derived from acetylene; acetone, chloral, butanol, acetic acid, acetaldehyde, several acetates, solvents and resins; a stainless steel division which consisted of a small foundry and a division formed in 1957, which produced Journal ol Occupational Medicine/Vol. 23, No. 10/0ctober 1981 671 URL 19394 -'{ j j ~ > . - . :\b rtf. l t i \ URL 19395 Fig 1. -- Population undar study. sulfuric acid, chlorine and cyanide. This complex started operations in 1927; some of its divisions opened at a later date; others closed in the late 1960s or early 1970s. The annuaL number of male workers averaged I.OOO.'1 .Methods Population Under Study. -- All production workers whose names appeared on the unions' lists or the com panies payrolls of the VC plant and the industrial complex between January 1,1946, and December 31,1972, consti tuted the population under study. Several methods (old addresses, telephone books, Quebec motor vehicles bureau, inquiries with fellow workers and families, searches in homes for retired persons) were used to trace these men, 18% of whom were still working for the same companies. Once identified, the worker himself or his next of kin (spouse, child or close relative) was questioned at home or at work by a trained interviewer. The ques tions permitted the reconstruction of a detailed occupa tional history within and outside the companies under study. Questions were also asked regarding smoking habits. For each deceased person believed to have been exposed to VCM, the occupational history was confirmed with the help of a fellow worker. The men whose number of years of work for these companies, as established through occupational histories, did not exceed five were removed from the study. The re maining population was then subdivided into three groups. The first group (known as the unexposed cohort) consisted of men who stated they had not been involved with the production of VCM or PVC, or both, for a period greater than five months. The second group (constituting the exposed cohort) comprised men who declared they had worked on the production of VCM-PVC for more than five years. Finally, the third group consisted of die men who had worked on the production of VCM-PVC for a period of six months to five years. This group was ex cluded from analysis because it neither fulfilled the definition of the exposed nor the unexposed cohort. Data Sources. -- The vital status of these men was established as of December 31, 1977. Causes of death were determined from death certificates. They were coded according to the 8th revision (ICDA) by one of die authors, who was unaware at the time of the coding as to which.group the man belonged. For cancer cases, a histopathological confirmation was sought on autopsy, biopsy or cytology reports contained in the medical files at the hospital. Information on Quebec's population and mor- 672 Mortality of Workers Exposed to Vinyl Chloride Monomer/ThSriault and Allard "k'v* ' TaW* 1. -- Ago Distribution of tha Man In tha VCMExpossd and Unaxpasad Coharts as al Dacamtwr 31, 1977 (Dacoasad Exchidad). Ago 25*29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75 + Total Expiud No. % 0 0 _ -- 1 0.3 26 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 Unaxpasad Na. H 1 0.2 3 0.5 6 0.9 45 7.1 55 6.6 64 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 cor...;::;Lons. Statistics/ Analysis. -- The results were analyzed using the "man-years at risk" method as described by Hill.'* 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, Haenszel'4 H and Miettinen.'* 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 (SMK).M The tosl of significance used was the one described by Uoilar and fcderer." 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 fobs 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 Figl, the population under study originally comprised 1.659 workers. Of these. 48 men were untraceable, 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 five 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 alt causes is 83.02, therebv showing no excess of death in the exposto conort Aiunvv-tr e ^<or all cancers, respi ratory sysiw..; dLouses digestive system diseases ex- ExpatHr* (Yaars) 5- 9 10-14 15-19 20-24 25-29 30-1Total Tibia 2. -- Distribution of Expasad Warkars by Langth at Exposure and Gbsarvattea Parted. Obtanratte* parted (Yaars) 5-1 10-14 19-19 29-24 29-29 39 4- 30 26 7 40 2 -- 26 26 19 17 4 5 -- 17 as 19 8 -- -- -- _ -- -- t 22 24 34 15 33 _ -- -- _ -- 38 30 56 50 116 96 103 TaW 111 95 73 61 67 38 451 Journal of Occupational Medicine/Vol. 23, No. iO/October 1981 673 URL 19396 i .p: : i1 r!' it i! . t- ' Table 3. -- Aga at Death, Diagnosis on Death Certificate and Diagnosis tin Histological Review of the 20 VCM-Exposed Workers Who DM ef Cancer. Case .Humber Death Certificate Histeiogicel Review 1 : *2 '3 4 5 6 7 8 9 10 11 12 13. 14 15 16 1? 18 19 20 Cancer of Ihe liver Cancer ol (he intesline Carcinoma of the trachea Angiosarcoma of the peritoneal cavity Cancer ol the lung Cancer ot Ihe liver Cancer ol ihe liver Cancer oi Ihe prostate Angiosarcoma ot the liver Cancer o( the pancreas Adenosarcoma of the intestine Hepatoma Cancer of the liver Cancer ot the pancreas Chondrosarcoma Hepatoma Primary cancer 0t the liver Cancer ol the digestive track Cancer of the biliary ducts Lymphoid leukemia Cancer of the sigmoid flexure of the colon Metastatic cancer ol the colon Carcinoma of the trachea Angiosarcoma ot the liver (no review) Angiosarcoma of the liver Angiosarcoma of the liver Cancer ot the prostate Angiosarcoma of Ihe liver Epithelioma ot the head ot the pancreas Adenosarcoma of the intestine Angiosarcoma ot the liver Angiosarcoma of the liver Anaplastic epithelioma of the pancreas Chondrosarcoma Angiosarcoma ot ihe liver Angiosarcoma ol the liver Cancer of the cecum Cancer of the head of the pancreas Lymphoid leukemia Age si Death (Years) 73 56 50 41 54 42 57 64 62 40 S3 53 53 59 43 48 53 52 54 48 Total Exposure (tears) 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 100, they are not statistically significant On the other hand deaths by accident and violence are signify cantly lower (p < 0.01). As for specific cancer deaths, a significant excess is noted for cancer of the digestive system only. This excess is accounted for exclusively by cancer of the liver W cases observe* vAri|<0i4**oec.t*d SMR 5.714.,,*'>. ino otner cancers are in excess. The rancor mortality among the workers who were still alive 15 years after their first exposure shows the same results as the mortality in the entire group, namely that only cancers of the digestive system are in excess in the exposed cohort (Table 6). Discussion Difficulties encountered in comparing the mortality of a group of workers with the mortality of a whole popula tion are well documented. They are termed: "healthy population selection effect," "survivor population effect,"1* "ethnic/cultural/social characteristics." They are believed to underestimate the actual mortality of a group of workers and therefore vo underscore the effects of the risl'S to which tnese dv'-* are being cxp'xed at work, fhcit* mo easy ways to avoid such difficulties. One method would be to compare a group of workers with another group of workers with the same characteris tics except for the exposure to the risk under study. The problems encountered in so doing are no easier. Although (Ik.> group most similar to industrial workers is probably another group of industrial workers from the same town, it is almost impossible to think of a group of industrial Table 4. -- Deaths Observed In the VCM-Exposed and Unoxpesed Cohorts. Ag-Ad)usM Rolattva Risks and Confidence Intervals of the Relative Risk. Causa of Death All cancers (140-209) Mouth and pharynx (140-149) Digestive (150-159) Respiratory (160-163) Bone. skin, connective (170-173) Urogenital (185-189) Eye. CNS (190-192) Leukemia (200-209) Other cancers Cardiovascular (390-458) Respiratory (460-519) Digestive (520-577) Accidents and violence (800-899) Others Total Ne. of Cases Exposed Unoxpesed 20 52 02 14 9 2 20 22 17 02 14 06 25 124 6 15 4 12 2 15 2t . 15 59 *233 Rotative Risk 1.48 -- 6 25 0.36 1.67 0.83 _ 1.15 -- 0.80 0.81 1.44 0.51 0.57 1.07 95* Confidence Interval ef the Rotative Risk 0.84- 2.61 -- 2.69-14.52* 0.08- 1.58 0.21-12.99 0.10- 7.10 -- 0.07*17.80 -- 0.47- 1.36 0.44- 9.35 0.49- 4.25 0.21- 1.24 0.14- 2.39 0.79- 1.45 674 `Significant according to Miattlnen's tost'7 tOne cause unknown Mortality of Workers Exposed to Vinyl Chloride Monomer/ThOriault and Allard URL 19397 > `\r * Table 5. - Observed and Expected* Members el Deaths in the VCM-Expesed Cehert and Standardized Mortality Reties (SMRs). Causa at Death Ail cancers Mouth and pharynx Oigesttve Respiratory Bone, skin, connective Urogenital Eye. CNS Leukemia Other cancers Cardiovascular Respiratory titgwiwe Accidents and violence Others Total (140-209) (140-149) (150-159) (160-163) (170-173) (185-169) (190-192) (200-209) (390-458) (520-577) (BOO-999) Observed 20 0 14 2 2 1 0 t 0 2f5t 4 . 2t 59 ExpeeM 16.37 0.64 5.40 5.78 0.38 1.33 0.60 1.67 0.54 31.87 a.'t ^35 10.58 5.40 71.07 4MR 122.17 -- 259.26f 34 60 526.32 7_5.19 5_9.88 78 94 186.91 103.90 16.90f 37.03 83.02 --------- 1 `Expected numbers obtained Irom 1971 mates Quebec dealt) rates applied to person-years In each age group tp <0.01'9 $0ne 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 thejr_rnortajity 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 nuYnber of cancer of the digestive system was signifi cantly higher in the exposed cohort. No other cancer was . ip excess and there even schemed to bo a deficit (not stntis(ically 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)1' 1111 and since several epidemiological studies on VCM-exposed workers have found an excess of lung cancer,1* 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 1946 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 tire 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."11 It was particularly true for rare cancers such as cancer of the brain and cancer of tire lymphatic system. Inasmuch as cancer of the lung was concerned. URL 19398 Table 6. -- Ratafee Riska and SMRe fer Cancer Deaths Amen* Men Having an Observation Parted el 19 Years and Mere. Cancers Ail cancers Mouth and pharynx Digestive Respiratory Bone, skin, connective Urogenital Eye, CNS leukemia Other cancers (140-209) (140-149) (150-159) (180-163) (170-173) (165-189) (190-192) (200-209) Mebve Me. Risk 15 1.30 0-- 11 5.68 2 0.35 0-- 1 0.82 0 t1 0.59 0 n% CenMence Interval af the Rotative Rick 0.97- 1.74 -- 2.72-11.58 0.25- 4.73 -- -- -- -- p<0.01* Journal J Occupational Medlcine/Vol. 23, No. tO/October 1981 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 author* with lo exprett their gratitude to the followinu who rontributed lo ih\ study workers, union representatives and management personnel ol the <onip.mit*s (..io.mIi,in Resins and Chemicals. Shavrinigan Chemicals, rtf. C.oodrit h and (.nil Canada. Drs H Iturra and |. Fabia. epidemiologists, and Or. I Delorme, pathologist, lor their Help and advice; the several hospitals con tacted. and the several persons who assisted with the collection ol information and the writing ol this article References 1 Creech |L Jr and Johnson MN: Angiosarcoma ol liver in the man ufacture of polyvinyl chloride. I Occup Med 16:150-151, 1974. 2 Tabershaw IR and Gaffey WR: Mortality study of workers in the manufacture of vinyl chloride and its polymers. / Occup Med 16509-Sift. 1974 3 Nicholson WJ. 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 UW, Carter JT, and Coombes El: Mortality study of workers in a polyvinyl-chloride production plant. Lancet 2:1197-1199, 1975. 5. Syren D, Engholm C, Englund A, and Westerholm P; Mortality and cancer morbidity tn a group of Swedish VCM and PVC production workers Environ Health Petspect 17167-170. 1976. < fox AI <ind Collier PF Mortality ex|Mrnncr of workers exposed to vinyl < hloritk- monomer m the monufat lure of |M>lyvinyl chloride in Great Britain. Hr J Ind Med 34:1-10, 1977 7. Delorme F and Theriault C: Ten eases of angiosarcoma of the liver in Shawinigan, Quebec. / Occup Med 20:338-340, 1978. 8. Monson R, Peters I. and lohmon M: Proportional mortality among vinyl chloride workers. Lancet 2:3S7-]Mk, 3974. 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 dus Sciences du Canada: Vue d'vfVMimblt* dm dangers de la contamination par le chloruru de vtnyle au Canada. Quebec Science (Suppi) :1-24, 1977. 12. LaRochelle F: Shawinigan depuis 75ans. Ateliers de I'imprimerie Publicity Piquet Enr., 1976, pp 497-52S. 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. / Nad 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 l 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. / Oocup Med 3:157-160, 1976. 19. Bailar |C and Ederer F: Significance factors for the ratio of Poisson variable to its expectation. Biometrics 20.639-642.1964. 20. Fox AJ and Collier PF: Low mortality rates m industrial cohort studies due to selection for work and survival In the industry. Br/ Pree Soc Med 30:225-230, 1976. 21. Viola PL: Cancerogenic Effect of Vinyl Chloride, Vo). 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 31.316-519,1971. 23. Maltoni C and lefemine C: Cancerogenicity bio assays of vinyl chloride: Current results. Ann NY Acad Scr 246:195-218.1975. 24. Wagoner )K Vinyl chloride and pulmonary cancer. / nv Pathoi Toxicol 1:161-162, 1978. 25. Pell S. O'Uerg Ml, and Karrh B: Cancer epidemiologic surveil lance in (he DuPont Company. / Occup Med 20-.72S-740,1978. URL 19399 4 1 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 <K.jo\w see the connecting .,*tween safety and the special financial risk involved in nucleat investment To use >he ..uticr.t 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 Cilintky. Commiisioner. the Nuclear Regulatory Commisiion. as Quoted in "Notable and Quotable." in The Wall Street Journal. May 21,1981. k 676 Mortality ol Workers Exposed to Vinyl Chloride Monomer/Thferiault and Allard