Document DMaJ20VrQvm3oGrRQQ3jyajBO

American Journal of Industrial Medicine 18:19-24 (1990) Liver and Biliary Tract Cancer Among Chemical Workers Gregory G. Bond, phd, mph, Elsie A. McLaren, rn, Fred L. Sabel, ms, Kenneth M. Bodner, msph, Thomas E. Lipps, md, and Ralph R. Cook, md, mph A recent cohort mortality study of male, hourly wage employees of a large Michigan chemical production and research facility had found a greater than expected number of deaths coded to liver and biliary tract cancer. In response, an additional investigation was then undertaken of the 44 liver and biliary tract cancer deaths observed between 1940 and 1982. A random sample (N = 1,888) of subjects was selected from the total cohort (N = 21.437) to serve as referents. Company work history records were used to classify cases and referents by work area assignment and potential for exposure to 11 selected chemical agents which have been shown to produce cancer of the liver or biliary passages in experimental animals. Statistically significant associations in both positive and negative directions were found for several work areas within the facility. A sug gestive association was found for vinyl chloride monomer, based on five cases with presumed exposure. Key words: carbon tetrachloride, epidemiology, methylene chloride, retrospective studies, vinyl chloride INTRODUCTION A recently conducted cohort study of the more than 37,000 male employees from Dow Chemical's Midland/Bay City, Michigan, production, research, and head quarters units showed more deaths than expected from cancer of the liver and biliary passages (SMR=140, 95% Cl = 99-191) among hourly workers [Bond et al., 1987], No data relating to personal or occupational exposures were collected as a part of the original study. This was by design since it was not feasible to collect such comprehensive information on all workers. This paper describes the results of a follow-up investigation which used a hybrid study design involving a modification of the traditional case-control method [Kupper et al., 1975]. Department of Epidemiology, Health and Environmental Sciences, The Dow Chemical Company. Mid land. MI (G.G.B., K.M.B.. R.R.C.). Medical Department. Michigan Division. Dow Chemical U.S.A., Midland. MI (E.A. McL.. T.E.L.). Industrial Hygiene Services. Michigan Division. Dow Chemical U.S.A.. Midland. Ml (F.L.S.). Address reprint requests to Dr. Gregory G. Bond, Department of Epidemiology. Health and Environ mental Sciences. The Dow Chemical Company, 1803 Building, Midland. Ml 48674. Accepted for publication February 7, 1990. 1990 Wiley-Liss, Inc, R&S162766 20 Bond et al. METHODS AND MATERIALS Death certificates of all male, hourly employees who died during 1940-1982 (N = 6.259) were reviewed for a statement of primary liver cancer, cancer of the liver not specified as primary or secondary, or cancer of the gallbladder or bile ducts (ICDA-8 codes 155-156 and 197.8) mentioned anywhere on the document. This identified 44 cases, of which I 1 were coded to primary liver cancer, 14 to cancer of gallbladder or bile ducts, and the remaining 19 to cancer of the liver not specified as primary or secondary. From the cohort of male, hourly employees (N = 21,437) who composed the original study population [Bond et al., 1987], a random sample (N = 1,888) was chosen to serve as referents. Company work history records were used to classify employee exposures by work area assignment [Bond et al.. 1985. 1986] as well as by potential for exposure to 1 1 chemical agents chosen because they had been shown to produce cancer of the liver or biliary passages under experimental conditions in animals [Chu et al.. 1984; Haseman et al., 1984; NTP. 1985; Wilbourn et al.. 1986). Visits to the Medical Department by the cases were reviewed by a nurse co-investigator (E.A.M.) to ascertain a history of excessive alcohol use or hepatitis, as they are established risk factors for primary liver cancer [Falk. 1982]. This exercise was of limited usefulness, however, owing to: I) the unavailability of medical records for those subjects who terminated their employment prior to 1947 (N = 24); and 2) a lack of information concerning any post-employment illnesses among the workers. Risk ratios were computed adjusted for birthyear by using the Mantel-Haenszel (1959) procedure. Analyses which additionally controlled for period of hire showed similar results to those which controlled only for birthyear and are not presented. Likewise, analyses of the primary liver cancers separate from the gall bladder/bile duct cancers did not reveal additional information and are not shown. Ninety-five per cent test-based confidence intervals were calculated about the risk ratio estimates [Miettinen, 1976]. RESULTS Analyses by work area assignments (Table I) showed the cases were signifi cantly less likely to have held assignments in electrochemicals or organic chemical production, but were more likely to have held assignments in office maintenance (RR = 3.6; 95% Cl = 1.4-9.1). A positive association with assignment to polyvinyipolyvinylidene chloride production did not reach statistical significance (RR = 2.0; 95% Cl - 0.9-4.6). There were no statistically significant trends apparent when analyses were made of duration of assignment within work areas: however, the odds ratio for an assign ment of five or more years to polyvinyl-polyvinylidene chloride production was 3.7 (95% Cl = 0.9-14.6) based on the contribution of two cases. Assignment to polyvinyl-polyvinylidene chloride production areas was explored further in Figure 1 via a plot of the "attributable fraction" [Goldsmith et al.. 1980] for each year during the period 1940-1980. The peak "attributable fraction" occurred in 1952. a year in which four of the seven cases assigned to these areas were all employed, no two in the same plant, however. None of the chemical agent comparisons revealed statistically significant dif- TABLE I. A Liver and Bi Major/minor work areas Administratu Manufactunn Agricuitur; Herbictc Inseetici Inorganic t Ammon Brine Bromin; Calcium Caustic Cellulo- Elecmx Metal f. Organic cl Acrvlan Alkyl tv Chlorin. Dves Fabrnj ^ai G! Haf Ion e Latex Olefins Phenol Polvsty Polvvin Unknown m Research Service Animal Boiler Constrc Electric Engine^ Garage Lead si Lubric; Machir Mainte Maten; Office Paint S' Pipe si 05 Plant p s? Power 8 Road a \%i ShopsWaste `WorlRrs n counted :1940-1982(N toer ot' the liver t or bile ducts locument. This 14 to cancer of ot specified as = 21,437) who or sample (N = e exposures by iid for exposure ce cancer of the ru et al.. 1984: to the Medical 'r (E.A.M.) to established risk nted usefulness. subjects who , of information \^^B1-Haenszel I entire showed : n resented, call uiadder/bile Ninety-five per ratio estimates es were signifirganic chemical oamtenance (RR nt to polyvinyllce (RR = 2.0: yses were made o for an assign ation was 3.7 .as was explored ath et al.. 1980] liable fraction" these areas were s cant dif- Chemical Worker Liver/Biliary Tract Cancer 21 TABLE 1. Age-Adjusted Risk Ratios for Analyses by Major and Minor Work Areas: Study of Liver and Biliary Tract Cancer Among Chemical Workers, Midland, Michigan, 1940-1982* Exposure frequencies 1%) Major/minor work areas Administration Manufacturing Agricultural chemicals Herbicides insecticides Inorganic chemicals Ammonia Brine Brominated products Calciurri'maanesium Caustic Cellulose products Electrochemicals Metal tabrication Organic chemicals Acrvlamide Alkvl benzenes Chlorinated benzenes Dves Fabricated plastics Glvcols Halogenated hydrocarbons Ion exchange resins Latex Olefins Phenol derivatives Polvstvrene Polyvinyl-polyvinylidene chlonde Unknown manufacturing Research Service Animal handling Boiler shop Construction Electric shop Engineering Garage Lead shop Lubrication Machine shop Maintenance Material handling Office maintenance Paint shop Pipe shop Plant protection Power Road and vard mtc. Shops-steel & instrument Waste management Well services Cases IN = 44) 0.0 68.2 9.1 68 4.6 47.7 0.0 6.8 4.6 9.1 40 2.3 0.0 n7 36.4 0.0 0.0 0.0 6.8 6.8 0.0 4.6 2.3 2.3 4.6 6.8 6.8 15.9 0.0 15.9 65.9 0.0 4.6 13.6 2.3 2.3 0.0 0.0 0.0 4.6 4.6 11.4 11.4 4.6 9.1 9.1 2.3 4.6 6.9 0.0 0.0 Referents IN = 1.888) 5.8 75.7 26.2 17.1 9.7 40.0 1.0 J.6 5.9 12.5 1.9 76 10 2 16.8 56.8 2.1 3.9 3.6 6.1 16.5 1.8 7.7 2.0 2.7 2.5 13.0 15.0 10.1 1.0 22.8 59.8 0.6 3.4 13.6 2.8 1.6 2.5 0.3 0.4 4.2 8.6 8.1 3.3 7.9 10.3 5.6 7.1 2.0 2.3 1.2 0.1 RR 95% Cl -- 0.0-1.5 <,> 7 0.4-1.3 0.4 0.l-l.0 n s 0.2-1.6 0.6 0.1-2.4 1.1 0,6-2.0 -- 0.0-8.4 0 9 0.3-3.0 0,9 0.2-3.7 il, 7 0 2-1.8 : l 0.5-9.5 ) 4 0.1-2.5 -- 0.0-0.8 1.1 0 6-2.2 0.5 0.3-0 9 -- 0.0-4.0 -- 0.0-2.2 -- 0.0-2.3 0.7 0.2-2.3 0.6 0.2-1.9 -- 0.0-4,7 0.4 0.1-1.6 1.2 0.2-8.7 0.8 0.1-6.2 1.5 0.3-6.7 0.4 0.1-1.4 0.5 0.1-2.0 2 0 0.9-4.6 -- 0.0-8 4. 0,8 0.3-1.8 1.1 0.6-2.0 -- 0.0-14.0 1.0 0.2-4.1 0.7 0.3-1.7 0.9 0.1-6.4 1.7 0.2-12,5 -- 0.0-3.4 -- 0.0-28.0 -- 0.0-21.0 0.9 0.2-3.5 0.4 0.1-1.6 1 4 0.5-3.7 3 6 1.4-9.1 0.6 0.1-2.5 0.8 0.3-2.3 1.1 0.4-3.2 0.2 0.0-1.3 0.5-9.6 0.7-7.5 -- 0.0-7.0 -- 0.0-83.8 * Workers may have been assigned to more than one work area during their employment and thus may be counted more than once m the table R&S162768 ^*7 22 Bond et al. O 2 0- F R C 10 TI fI O > 0 " I......................................... ......... ................................................................ 1940 1550 1960 1970 198u CALENDAR YEAR Fig. 1. Studv of liver and biliary tract cancer. Attributable fraction tor assignment to polyvinvl-polyvinylidene chloride areas in each year during 1940-1980. ferences (Table II). The highest risk ratio (RR = 2.2: 95% Cl = 0.9-5.8) was for presumed exposure to vinyl chloride based on the contribution ot five cases. There were no statistically significant trends for analysis by duration of exposure to any ot the 11 agents, although in many instances there were too few exposed cases to permit meaningful conclusions. Likewise, there were no statistically significant trends ap parent when analyses were made by estimated dose of exposure. DISCUSSION The present investigation revealed a statistically significant positive association with assignment to office maintenance, and some marginal associations with poly vinyl-polyvinylidene chloride production. The association with an assignment in office maintenance is curious, since this type of work is not generally considered hazardous. In fact, it was a common practice to assign to that department any workers who were not considered medically fit to work in more strenuous jobs. Company medical records were available to review for four of the five cases assigned to office maintenance, and revealed that two of the men had been hospitalized for chronic alcoholism, suggesting that this finding may have resulted from confounding. High levels of vinyl chloride exposure have produced a specific hazard of angiosarcoma of the liver in some work forces [Doll. 1988]. In the present study, five of the seven cases of liver and biliary tract cancer who held assignments in the polyvinvl-polyvinvlidene work area were considered to have had the opportunity tor exposure to vinyl chloride monomer. There was no evidence that any of the five cases experienced angiosarcoma; however, pathology reports and/or medical records were TABLE II. Study of Li\ Michigan, l1 Agent Carbon tetra Chlorodiben Chloroform Ethylene dtf Ethylene dit Methylene i Perchloroetl Polychlorin, 1,1.2-Tnch 2.4.6-Trich Vinvl chloi 'Workers i counted m< available descripti they the ol^F exposu than 20 three hat 01 factors ' cancer, not expi dence i apparen icals (i. increasi St ity of i produci did occ of the ; ACKN ZD 1 22 Wendy a> Muelle no ato> S^^te 1900 V- .8, as for uses. There re to any of es to permit i trends ap- association with polylignment in considered any workers Company ted to office for chronic ading. c hazard of t study, five tents in the tortunity for te tuMcases eco^^vere Chemical Worker Liver/Biliary Tract Cancer 23 TABLE II. Age-Adjusted Risk Ratios for Analyses by Selected Chemical Agent Exposures: Study of Liver and Biliary Tract Cancer Among Chemical Workers, Midland, Michigan, 1940-1982* Exposure frequencies (%) Agent Cases (N = 44) Referents iN = 1.888) RR 95% Ct Carbon tetrachloride Chiorodibenzo-tpl-dtoxins Chloroform Ethylene dibromide Ethylene dichloride Methylene chloride Perchloroethvlene Polvchlorinated biphenyls 1.1.2-Triehloroettiane 2,4.6-Tnchlorophenol Vinyl chloride 11.4 2.3 4.6 2.3 6.8 4.6 13.6 0.0 0.0 04) 1 1.4 20.0 9.1 6.7 7.3 14.6 8.9 11.3 1.2 0.5 09 '8 0.5 0.2-1.4 o 2 0.4-1.4 0.6 0.2-2.7 0.4 0.1-3.0 0 4 0.1-1.3 0.8 0 2-3.6 1.8 U. 8-4.3 -- 0.0-7.0 -- 0 0-16.8 -- 0.0-9.3 -1 A 1 LA be "'Workers may have been exposed to more than one agent during their employment and thus may be counted more than once in the table. available to confirm the diagnosis for only two of them. Table III presents some descriptive information concerning these cases. By virtue of the time period in which they worked and the nature of their jobs, each of the men was considered to have had the opportunity for high levels of exposure to vinyl chloride monomer, with peak exposures probably exceeding 1,000 ppm. Although two of the cases had spent more than 20 years each in areas of potential vinyl chloride monomer exposure, the other three had very brief exposures on the order of one to several months. Of interest, none of the other chemical agents suspected a priori as potential risk factors was shown to be statistically significantly associated with liver or biliary tract cancer. It may be that some of the agents are weak carcinogens, but this study could not explore that hypothesis. An examination of the upper limits of the 95% confi dence intervals about the risk ratios presented in Table II shows that the study apparently had sufficient power to rule-out modest increases in risk for several chem icals (i.e.. carbon tetrachloride and ethylene dichloride), but only relatively large increases in risk could be excluded for others. Some misclassification of exposure may have occurred; however, the availabil ity of industrial hygiene data from the distant past for many of the plants at this production site should have reduced this possibility considerably. If misclassification did occur, and was non-differential, the tendency would be to diminish the strength of the associations. ACKNOWLEDGMENTS The authors gratefully acknowledge technical assistance from Tom Bradley, Wendy Dahar, Mike Dupuis, Stan Gorgacz. Harold Hoyle, Bruce Houtman, Carl Mueller. Gordon Roush, Dan Schmidt. Belinda Scortichini. Tricia Stirrett. and Mike Stricter. They also thank Diana Diamond for help with preparation of the manuscript. H&St 62770 24 Bond et al. TABLE III. Descriptive Data on Five Cases Presumed to Have Been Exposed to Vinyl Chloride Monomer: Study of Liver and Biliary Tract Cancer Amon . Chemical Workers, Midland, Michigan. 1940-1982 Case no. Year first exposed Year last exposed Duration ot` exposure ( vr: mo> Year of death Job classification Best diagnosis (source )J i 1943 1943 0; <1 I960 Assistant Adenocarcinoma operator of common *> 1943 1943 0: 3 1964 Janitor bile duct (DC) Carcinoma of liver (DC) 3 1950 1970 20: 0 1976 Operator Hepatocellular carcinoma (PR) 4 1946 1967 20: 8 1980 Operator Intrahepatic bile duct cancer (MR and PR) 5 1952 1952 0: 2 1972 Handymatv Cancer of liver packager (DC) `DC = death certificate. MR = medical record, PR = pathology report. REFERENCES Bond GG. Flores GG. Shellenberger RJ, Cartmill JB, Fishbeck WA. Cook RR (1986): Nested casecontrol study of lung cancer among chemical workers. Am J Epidemiol 124; 1:53-66. Bond GG. McLaren EA. Cartmill JB. Wymer KT. Sobel W. Lipps TE. Cook RR (1987): Cause-specific mortality among male chemical workers. Am J Ind Med 12:353-383. Bond GG. Shellenberger RJ. Flores GH. Cook RR. Fishbeck WA (1985): A case-control study of renal cancer mortality at a Texas chemical plant. Am J Ind Med 7:123-139. Chu KC. Cueto C. Ward JM (1984): Factors in the evaluation of 200 National Cancer Institute Bioassays. J Toxicol Environ Health 14:621-639. Doll SR (1988): Effects ot exposure to vinyl chloride--An assessment of the evidence. Scand J Work Environ Health 14:61-78. Falk H (1982): Liver. In Schottenfeld D. Fraumeni JF (eds): "Cancer Epidemiology and Prevention." Philadelphia: W.B. Saunders Company, pp 668-682. Goldsmith OF. Smith AH. McMichaei AJ (1980): A case-control study of prostate cancer within a cohort of rubber and lire workers. J Occup Med 22:533-541, Haseman JK. Crawford DD. Huff JE. Boorman GA. McConnell EE (1984): Results from 86 two-year carcinogenicity studies conducted by the National Toxicology Program. J Toxicol Environ Health 14:621-639. Kupper LL. McMichaei AJ. Spirtas R (1975): A hybrid epidemiologic study design useful in estimating relative risk. J Am Statistical Assoc 70: 351:524-528. Miettinen OS (1976): Estimability and estimation in case referent studies. Am J Epidemiol 103:226-235. National Toxicology Program (NTP) (1985): NTP Technical Report on the toxicology and carcinogenesis studies of dichloromethane in F-344/N rats and B6C3FI mice (Inhalation studies) NTP-TR-306 (board draft). Wilhourn J. Haroun L, Haseltine E, Kaldor J, Partenskv C. Vainio H (1986 Response of experimental animals to human carcinogens: An analysis based upon the IARC Monographs programme. Car cinogenesis 7:1853-1863. *v Cart Amc Olaf A Per N; David PhD, K Key w INTR neurt Depar istry ( Bone: 33 Natioi 0 SanS>o Direct Addrc I. No 5>