Document 6bbqjLemB92xRnvbLaa70oNZm

Clinical Medicine Asbestos and Gastrointestinal Cancer A Review of the Literature ROBERT W. MORGAN, MD, SMHyg; DONNA E. FOLIART, MD, MPH, and OTTO WONG, ScD, Oakland, California Exposure to asbestos is among several factors cited as possible causes ofesophageal, gastric and colorectal cancer. More than 45 published studies have presented mortality data on asbestos-ex posed workers. For each cohort, we listed the observed and expected rates of deaths from types of gastrointestinal cancer based on the latest published follow-up. Summary standardized mortality ratios (SMRs) were then derived. Finally, we calculated summary SMRs for total gastrointestinal tract cancer for three occupational groups: asbestos factory workers, insulators/shipyard workers and asbestos miners. Statistically significant elevations in summary SMRs were found for esophageal, stomach and total gastrointestinal tract cancer in all asbestos-exposed workers. Esophageal cancer summary SMR remained significantly elevated when data were reanalyzed to include only those cohorts with death certificate diagnoses for cause of observed deaths. However, summary SMRs were not statistically significant for stomach and total gastrointestinal tract cancer after reanalysis. Summary SMRs by occupational group showed a significant elevation for total gastrointestinal cancer in insulators/shipyard workers. The elevation was not significant after reanalysis. Based on the results after reanalysis, the elevations in summary SMRs for stomach and total gastrointestinal tract cancer are of a magnitude that could result from diagnostic and investigator error. We conclude that more studies are required before stomach and colorectal cancers are documented as asbestos-related diseases. (Morgan RW, Foliart DE, Wong O: Asbestos and gastrointestinal cancer--A review of the litera ture. West J Med 1985 Jul; 143:60-65) What causes gastrointestinal cancer is far from clear. Cited as possible causes are factors as diverse as diet, race, blood group, alcohol intake and exposure to asbestos. This review will examine the relationship of occupational asbestos exposure to the risk of developing carcinoma of that portion of the gastrointestinal tract including the esophagus, stomach and colon-rectum. The association of asbestos and peritoneal mesothelioma is well documented, does not involve the gastrointestinal tract and will not be reviewed here. Esophageal cancer is a relatively uncommon cancer for which death rates have remained stable over time: 4.1/ 100,000 in 1955 and4.3/100,000 in 1975, for US white men. Higher death rates have been reported in areas of Africa and Eastern Europe. In the United States, esophageal cancer is found predominantly in men and is associated with cigarette smoking and alcohol intake. Gastric cancer is a slightly more common disease of con siderable geographic variation, with some countries, notably Japan, having rates significantly higher than those of the United States. Dietary factors have been implicated in the pathogenesis of the disease. In the United States both mor tality and incidence of stomach cancer in white men have fallen consistently and dramatically over a 20-year period. For example, the mortality rate dropped from 17.2/100,000 in 1955, to 8.2/100,000 in 1975. Colorectal cancer presents an etiologic puzzle distin guished more by the number of hypotheses than by discovery of causative agents. In addition to asbestos, associations are postulated for intake of fiber, fat and specific foodstuffs. For US white men, death rates for colorectal cancer were 23.4/ 100,000 in 1955 and 24.4/100,000 in 1975. Attempts to study the relationship of asbestos to gastroin- From the Environmental Health Associates. Inc, Oakland, California. Submitted, revised, December 17, 1984. Reprint requests to Robert W. Morgan, MD, Environmental Health Associates, Inc, 520 Third Street, Suite 208, Oakland, CA 94607. 60 THE WESTERN JOURNAL OF MEDICINE ASBESTOS AND GASTROINTESTINAL CANCER ABBREVIATIONS USED IN TEXT AR=attributable risk SMR = standardized mortality ratio testinal cancer are frustrated by a number of factors. First, for any specific site (for example, stomach), the disease is suffi ciently uncommon that few work forces can be expected to account for enough cases from which to draw conclusions. Perhaps more important, there is considerable opportunity for misdiagnosis or inaccurate certification of death, in that some gastrointestinal tract cancers will be certified as mesothe lioma, and vice versa. This question has been addressed by both Newhouse 1 and Selikoff2 in their examinations of death certificate accuracy in asbestos workers' deaths. In the New house series, half of the deaths certified as gastrointestinal cancer were actually due to mesotheliomas. In SelikofFs se ries, only 24 deaths were originally certified as peritoneal mesotheliomas; after examining further clinical and patho logic evidence from other deaths in the cohort, investigators classified 112 deaths as due to that cause, with a pronounced reduction in number of deaths due to "all other cancer'' and "all other causes.'' Although diagnostic inaccuracy is troubling, it is not easily remedied. Selikoff has used "best evidence'' from nu merous sources, rather than the certified cause of death, to calculate standardized mortality ratios (SMRs). However, "best evidence'' data cannot properly be compared with na tional mortality statistics, which are derived solely from death certificates. Unless one can also correct the statistics from the US reference population, comparisons are invalid because of the underlying principle that both study and reference popula tions should have similar ascertainment and classification of death. Another statistically questionable practice is attrib uting to cancer those deaths previously certified as due to a cause other than cancer, but where cancer was present (Seli koff communication quoted by Miller3). Again, without a similar mechanism for including cancer as a contributing cause of death in the national mortality statistics, such com parisons will either create an apparent excess of cancer in the study population or will exaggerate any excess already there. Most of the evidence concerning asbestos and disease comes from the retrospective cohort mortality studies of as bestos-exposed workers. In all of these studies, a cohort ex posed to asbestos has been identified from employment, union or other existing records and then followed prospectively over time to compare their rate and pattern of mortality with a reference population with presumably no exposure or less exposure. The reference population is frequently that of county, state or country. In general, national statistics are used because of availability and the stability imparted by large numbers. Smaller jurisdictions tend to produce statistics that can be heavily influenced by relatively few deaths, including those occurring in the work force under study. Methods To determine the role of occupational asbestos exposure in the risk of gastrointestinal cancer, we attempted to identify all relevant publications from the indexed literature. From each report, we extracted the details concerning cohort definition, exposure of concern and gastrointestinal cancer deaths (both JULY 1985 143 1 observed and expected). We did not include any judgment concerning scientific merit in selecting papers to be included. A summary of published studies of asbestos-exposed workers is presented in Table 1. Some 45 publications have described about 22 different cohorts, although within them there is some overlap. Deciphering the interwoven cohorts is complicated because in many cases a single data set has been repeatedly published, sometimes for an alternative analysis. In other cases, there appears to be no clear reason for the republication. Further confusion is introduced when papers do not specify that the data, or portions thereof, were previ ously reported. Some of the cohort studies reported do not mention gastrointestinal cancer, let alone specific sites, so contribute little to a quantitative analysis. Table 2 summarizes the results of those studies which are informative as to the possible relationship of asbestos to gas trointestinal cancer. Studies were included in Table 2 when the number of gastrointestinal tract cancer deaths was avail able and the cohort discrete. Because of the problem of data overlap, we have used the latest published follow-up of each cohort rather than numbers from earlier papers. When avail able, observed deaths based on death certificates, rather than "best evidence,'' are presented in Table 2. Several of the studies24 28 38 in Table 2 grouped deaths from peritoneal meso theliomas with gastrointestinal cancers. When the actual numbers of mesotheliomas were given in the text, we deleted these from "observed" deaths. One report33 grouped esopha geal and stomach cancer deaths together for analysis. This report was not included in the calculations for those sites. In order to provide a quantitative summary risk index for gastrointestinal cancers based on the studies reviewed, we have examined a number of statistical procedures. One pro cedure is the combination of relative risks based on the vari ance-weighted logarithms of the relative risk.56"58 In this pro cedure, we can treat each cohort as a 2 x 2 table and the SMR as a special type of relative risk.59 The unexposed group in the 2x2 table would be a sample of the standard population, whose sample size is the same as that of the exposed study group. This procedure, however, presents some difficulties because the number of person-years is not always presented in every study reviewed. Another procedure of combining the risk estimates from these studies is the indirect standardization. We can view each cohort as a separate stratum; within each, certain characteris tics are homogeneous (for example, study design, length of follow-up and exposure). This view is similar to that in calcu lating SMR within a cohort, where the strata are homoge neous with respect to age, race, sex and so forth. In the direct standardization only the observed and expected deaths are needed for the calculation. The summary SMR can be calcu lated as follows: So ummary oSwMnR=-S-u--m---o--f--O--b--s-e--rv--e-d---D--e--a-t-h--s xw 100. Sum of Expected Deaths In essence, the summary SMR is weighted by the number of expected deaths, which is the reciprocal of the variance of the individual SMR.58 Thus, this method is, to some extent, sim ilar to the one based on the logarithm of the relative risks. Needless to say, this application of the indirect standard ization in summarizing SMRs from different studies is subject to the same criticisms that are often raised in summarizing 61 ASBESTOS AND GASTROINTESTINAL CANCER age-sex-race-specific rates in a single study. On the other hand, we know of no way to overcome the bias due to the fact that negative studies were less likely to result in published SMRs or observed and expected values, although even studies with negative findings in gastrointestinal cancer but with posi tive findings in respiratory tract cancer would still have a high likelihood of being published. We have also derived summary SMRs for three major occupations: asbestos factory workers, insulators/shipyard workers and miners. In the follow-up of the New York-New Jersey cohort of insulation workers to 1976,2 only "best evi dence" calculations of observed deaths are presented (Table 3). Due to the limitations of "best evidence" data, summary SMRs were also calculated after deleting observed deaths in this cohort. Results Summary SMRs and confidence intervals for esophageal, stomach, colorectal and total gastrointestinal tract cancer ap pear in Table 3. Significant elevations in summary SMRs TABLE 1 --Summary of Studies of Asbestos-Exposed Workers Authors Published First Follow-up Date Re published Cohort Description and Comments Doll ............................... Knox et al ..................... Peto et al .................... Braun and Truan.......... Dunn et al.................... Mancuso et al............... Elwood et al.................. Selikoff et al.................. Selikoff et al.................. Selikoff et al.................. Selikoff et al.................. Selikoff et al.................. Enterline....................... Kleinfeld et al................ Newhouse et al............. Elmes and Simpson . . . McDonald et al............. Enterline....................... 1955 ( 9)* 1958 (14) 1960 (15) 1963 1967 1964 1964 (17) (18) (19) ( 5) 1972 (20) 1972 (21) 1979 1979 1965 1967 1967 1969 1972 (7) ( 8) (23) (24) (25) ( 1) (26) 1973 (27) 1979 (28) 1971 (29) 1971 (31) 1978 (34) 1972 (35) 1965 (10) 1968 (11) 1977 (12) 1965 (16) 1968 ( 6) 1974 ( 4) 1979 ( 2) 1979 ( 2) 1979 (49) 1977 (30) 1974 (32) 1980 (33) Meurman et al ............. Weiss............................ Rubino et al.................. Gillam et al .................. Kolonel et al.................. Jones et al.................... Hobbs et al .................. Rossiter and Coles .... Hughes and Weill .... Newhouse and Berry . . 1974 1977 1979 1979 1980 1980 1980 1980 1980 1982 (39) (40) (41) (42) (43) (44) (45) (46) (13) (47) Acheson et al............... PMR=proportionate mortality ratio 1982 (53) 'Numbers in parentheses () = reference numbers. 1979 (38) 1983 (48) 1979 ( 2) 1981 (52) 1979 (50) 1981 (52) 1980 (22) 1981 (51) 1973 (36) 1973 (37) Asbestos textile factory workers Followed through 1961 Followed through 1966 Followed through 1974 Canadian asbestos miners followed 1950-55 Multiple occupational groups studied 1954-62 Follow-up through 1962 Factory workers followed 1938-60 Same cohort? Asbestos sheeting factory workers followed 1936-62 New York-New Jersey insulators followed through 1962 Followed through 1964 Followed through 1971 Followed through 1976 Asbestos insulation workers, including NY-NJ cohort survivors, followed through 1971 Followed to 1976 Amosite factory workers through 1971 Followed through 1977 Shipyard workers from 1967-76, a subcohort of 1972a study Chrysotile miners from Quebec Asbestos products workers 1948-51, followed through June 1963 Asbestos products workers to 6/30/63, some from previous cohort PMR study of asbestos workers Asbestos factory men Asbestos factory women employed 1936-1942, followed through 1968 Follow-up of above 2 cohorts to 1970 Follow-up to 1975 170 Shipyard workers through 1966 Followed through 1975 Quebec miners followed to 11/1/66 Followed through 1969 with different analysis Followed through 1975 199 Exposed to crocidolite traced through 1975 Retirees from asbestos mfg followed through 1969 Retirees presumed identical to the above Reanalysis of previous two papers Retirees followed through 1973 Case-control study of anthophyllite miners Chrysotile factory workers 1935-54 followed through 1974 Miners followed through 1974 Gold miners exposed to asbestiform rock followed through 1973 Shipyard workers followed through 1969 Gas mask workers followed through 1978 Crocidolite miners followed through 1978 Dockyard workers 1947-78 Factory workers followed through 1974 Asbestos (mainly chrysotile) friction material workers, 1942-79 Followed through 1980 Female gas mask workers, 40-year follow-up 62 THE WESTERN JOURNAL OF MEDICINE ASBESTOS AND GASTROINTESTINAL CANCER were found for esophageal (SMR=214) and gastric (SMR=183) cancer. Although summary SMRs for these sites were based on data from four studies, numbers were small and the 95 % confidence intervals were wide. The slight excess in SMR for colorectal cancer (SMR=113) was not statistically significant. The SMR for total gastrointestinal tract cancer (SMR = 108) was minimally but significantly ele vated. Table 4 presents site-specific summary SMRs after "best evidence" data are deleted. Esophageal cancer remains ele vated (SMR=238,,P< .05). Total gastrointestinal tract cancer deaths and summary SMRs for three exposure categories are presented in Table 5. Of the three occupational groups, only insulators/shipyard workers had a significant elevation in total gastrointestinal tract cancer (SMR= 132). Total gastrointestinal tract cancer was not significantly elevated in insulators when "best evi dence" data were removed. Discussion The strongest evidence linking asbestos to esophageal and gastric cancer is a series of studies of the New York-New Jersey (NY-NJ) and US-Canadian insulation workers.2*4-6 The summary SMR for esophageal cancer was significantly TABLE 2 --Summary of SMRs from Various Cohort Studies Author (Date)/Comments Cohort Person- (Ref. No.) Size Years Standard Population Esophageal Cancer Stomach Cancer SMR Obs Exp SMR Obs Exp Colorectal Cancer SMR Obs Exp Total Gl Cancer SMR Obs Exp Peto et al (1977) .................. (12) 1,106 16,072 Asbestos factory workers Brit Natl 102 16 15.7 Selikoff et al (1979)............... NY-NJ insulators to 1976 ("best evidence" used) ( 2) 632 13,925 US 71 V 1.4 352 19* 5.4 277 23* 8.3 285 43* 15.1 Selikoff et al (1979)............... ( 2) 17,800 166,853 US 253 18 7.1 127 18 14.2 152 58 38.1 158 94 59.4 US-Canada insulators Selikoff et al (1980)............... (22) Amosite factory workers 582 6,311 New Jersey 125 1 0.8 200 4 2.0 212 11 5.2 200 16 8.0 Selikoff et al (1979)............... ( 7) Shipyard workers 389 US 97 3 3.1 Nicholson et al (1979).......... Chrysotile mining and milling ( 8) 544 7,408 Canada Natl 105 10 9.5 McDonald et al (1980)............ Chrysotile miners, 20 or more years after 1st em ployment; stomach/ esophageal deaths analyzed together (33) 11,379 Quebec 127 130* 102.0 127 130* 102.0 78 79 101.0 103 209 202.9 Newhouse and Berry (1979) . . (28) Rates exclude mesotheliomas 5,522 Brit 136 60* 44.2 Enterline and Kendrick (1967) . (24) Various exposures (bldg, friction/textile products), rate includes malignant mesotheliomas 21,755 US 106 83 78.2 Henderson and Enterline (1979) (38) Rates include mesotheliomas 1,348 US 138 110 79.7 Meurman et al (1974)............. (39) 1,092 Anthophyllite miners Finland 47 7 14.9 Weiss (1979)............................ (40) Chrysotile miners; stomach & colorectal cancers combined to calculate Gl cancer SMR 264 US 105 4 3.8 Rubino et al (1979).................. (41) Chrysotile miners 952 21,459 Italy 98 19 19.4 Jones et al (1980) .................. (44) Gas mask workers 951 Brit 49 10 20.4 Rossiter and Coles (1980) . . . (46) 6,292 Shipyard workers Engl/Wales 83 63 75.9 Hughes and Weill (1980) . . . . (13) 5,645 Factory workers US 50 25 50.0 Mancuso and Coulter (1963) . . (17) 1,495 23,257 US 200 1 0.5 107 2 1.9 145 5 3.5 135 8 5.9 Factory workers Berry and Newhouse (1983) . . (48) Asbestos friction materials (mainly chrysotile) workers 11,182 163,009 Engl/Wales 98 132 134.6 Gl=gastrointestinal. SMR=standardized mortality ratio *See comment in Author (Date)/Comments column JULY 1985 143 1 63 ASBESTOS AND GASTROINTESTINAL CANCER elevated even after "best evidence" data on observed deaths in the New York-New Jersey cohort were removed. Gastric cancer in the NY-NJ cohort contributed 19 observed deaths of 43 deaths in the calculation of the summary SMR. For the NY-NJ cohort, unfortunately, causes of death used in analysis were obtained not only from death certificates but from "best evidence." The authors did not present the number of ob served deaths derived solely from death certificate informa tion. The gastric cancer summary SMR was not significantly elevated when "best evidence" data were deleted. The same researchers found no excess of gastrointestinal cancer in ei ther shipyard workers7 or chrysotile miners.8 One of the bestperformed and longest-term cohort studies9"12 has found no increase of gastrointestinal cancers of any type after more than 50 years of meticulous follow-up from first exposure. Most of the evidence favoring an association between as bestos and gastrointestinal cancer comes from a series of studies by one group of investigators; these results must with stand critical review if a causal relationship is to be accepted. The fallacy of using "best evidence" for the cause of death has been discussed above, as a bias that inflates SMRs. In addition, repeated publishing of the same results may have created the illusion of more positive evidence than actually exists. Four cohorts have resulted in at least 14 publications (see Table 1). The original table describing the original New York-New Jersey cohort, first published in 1964, was repub lished in 1979 without incorporating the follow-up that had been accomplished and published in the interim. The 1976 follow-up was published in both 1979 and 1981. The 1976 follow-up of the US-Canadian cohort was published twice in 1979 and the amosite cohort follow-up of 1977 was published in 1979,1980 and 1981. The negative studies7 8 have not been republished. In an earlier review, Miller3 came to the conclusion that asbestos and gastrointestinal malignant conditions were caus ally related, although he did not attempt to calculate a sum mary risk estimate. Most of his conclusions were based upon the work of Selikoff and the review did not have the later negative studies coming from Peto,12 Hughes13 and Selikoff's associates.7,8 Miller also failed to appreciate the error in volved in the practice of using "best evidence" rather than the certificated cause of death for comparison with national statis tics. Finally, although the review paper acknowledges the work by Newhouse in showing the extent of cause-of-death 64 THE WESTERN JOURNAL OF MEDICINE ASBESTOS AND GASTROINTESTINAL CANCER misclassification, there was little attention paid to its potential impact on study results. Conclusion Malignant lesions of the lung, pleura and peritoneum have been clearly linked with exposure to asbestos. Some studies of cohorts occupationally exposed to asbestos have also found elevations in incidence of gastrointestinal cancer, while other reports have not found excesses after years of follow-up. In an effort to summarize the available studies of gastroin testinal cancers in asbestos-exposed workers, we have calcu lated summary SMRs for esophageal, gastric, colorectal and total gastrointestinal tract cancer. In addition, we have de rived total gastrointestinal tract cancer summary SMRs for workers exposed to asbestos in three specific capacities: fac tory work, shipyard/insulation work and mining. Although there are statistically significant elevations of SMR for esophageal and gastric cancer for all groups and for total gastrointestinal tract cancer for insulators and shipyard workers, we are not convinced of the relationship to asbestos exposure, especially for stomach cancer, where the finding is inconsistent and possibly erroneous in the largest positive study. For colorectal cancer, at this time there is no discern ible relationship to asbestos exposure. We hope that the occu pationally exposed cohorts now under study will provide more conclusive evidence soon--refuting or supporting such an association. REFERENCES 1. Newhouse ML, Wagner JC: Validation of death certificates in asbestos workers. Br J Ind Med 1969; 26:302-307 2. Selikoff IJ, Hammond EC, Seidman H: Mortality experience of insulation workers in the United States and Canada, 1943-1976. Ann NY Acad Sci 1979; 330:91-116 3. Miller AB: Asbestos fibre dust and gastrointestinal malignancies--Review of literature with regard to a cause/effect relationship. J Chronic Dis 1978; 31:23-33 4. Selikoff U: Epidemiology of gastrointestinal cancer. Environ Health Perspect 1974;9:299-305 5. Selikoff U, Churg J, Hammond EC: Asbestos exposure and neoplasia. JAMA 1964; 188:22-26 6. Selikoff U, Hammond EC, Churg J: Asbestos exposure, smoking and neoplasia. JAMA 1968; 204:106-112 7. Selikoff, U, Lilis R, Nicholson WJ: Asbestos disease in United States shipyards. Ann NY Acad Sci 1979;330:295-311 8. Nicholson WJ, Selikoff U\ Seidman H, et al: Long-term mortality experience of chrysotile miners and millers in Thetford Mines, Quebec. Ann NY Acad Sci 1979; 330:11-21 9. Doll R: Mortality from lung cancer in asbestos workers. Br J Ind Med 1955; 12:81-86 10. Knox JF, Doll RS, Hill ID: Cohort analysis of changes in incidence of bronchial carcinoma in a textile asbestos factory. Ann NY Acad Sci 1965; 132:526-535 11. Knox JF, Holmes S, Doll R, et al: Mortality from lung cancer and other causes among workers in an asbestos textile factory. Br J Ind Med 1968; 25:293-303 12. Peto J, Doll R, Howard SV, et al: A mortality study among workers in an English asbestos factory. Br J Ind Med 1977; 34:169-173 13. Hughes J, Weill H: Lung cancer risk associated with manufacture of asbestoscement products. In: Biological effects of mineral fibres. Lyon, France, IARC Sci Publ 1980; 2(30): 627-635 14. Braun DC, Truan TD: An epidemiological study of lung cancer in asbestos miners. Arch Ind Health 1958; 17:634-653 15. Dunn JE, Linden G, Breslow L: Lung cancer mortality experience of men in certain occupations in California. Am J Public Health 1960; 50:1475-1487 16. Dunn JE, Weir JM: Cancer experience of several occupational groups followed prospectively. Am J Public Health 1965; 55:1367-1375 17. Maneuso TF, Coulter EJ: Methodology in industrial health studies. Arch En viron Health 1963; 6:210-226 18. Maneuso TF, El-Attar AA: Mortality patterns in a cohort of asbestos workers. J OccupMed 1967;9:147-162 19. Elwood PC, Cochrane AL, Benjamin IT, et al: A follow-up study of workers from an asbestos factory. Br J Med 1964; 21:304-307 20. Selikoff U, Hammond EC, Churg J: Mortality experience of asbestos insulation workers, 1912-1971. Read before the Fourth International Pneumoconiosis Conference, Bucharest, 1972, pp 224-231 21. Selikoff U, Hammond EC, Churg J: Carcinogenicity of amosite asbestos. Arch Environ Health 1972;25:183-186 22. Selikoff U, Seidman H, Hammond EC: Mortality effects of cigarette smoking among amosite asbestos factory workers. JNCI1980; 65(3):507-513 23. Enterline PE: Mortality among asbestos products workers in the United States. Ann NY Acad Sci 1965; 138:156-165 24. Enterline PE, Kendrick MA: Asbestos dust exposures at various levels and mortality. Arch Environ Health 1967; 15:181 -186 25. Kleinfeld M, Massite J, Kooyman O: Mortality experience in a group of as bestos workers. Arch Environ Health 1967; 15:177-180 26. Newhouse ML, Berry G, Wagner JC, et al: A study of the mortality of female asbestos workers. BrJ Ind Med 1972; 29:134-141 27. Newhouse ML: Asbestos in the work place and the community. Ann Occup Hyg 1972; 16:97-102 28. Newhouse ML, Berry G: Patterns of mortality in asbestos factory workers in London. Ann NY Acad Sci 1979; 330:53-60 29. Elmes PC, Simpson MJC: Insulation workers in Belfast--3. Mortality 1940-66. BrJ Ind Med 1971; 28:226-236 30. Elmes PC, Simpson MJC: Insulation workers in Belfast--A further study of mortality due to asbestos exposure (1940-75). Br J Ind Med 1977; 34:174-180 31. McDonald JC, McDonald AD, Gibbs GW, et al: Mortality in the chrysotile asbestos mines and mills of Quebec. Arch Environ Health 1971; 22:677-686 32. McDonald JC, Becklake MR, Gibbs GW, et al: The health of chrysotile as bestos mine and mill workers of Quebec. Arch Environ Health 1974; 28:61-68 33. McDonald JC, Liddell FDK, Gibbs GW, et al: Dust exposure and mortality in chrysotile mining, 1910-1975. BrJ Ind Med 1980;37:11-24 34. McDonald AD, McDonald JC: Mesothelioma after crocidolite exposure during gas mask manufacture. Environ Res 1978; 17:340-346 35. Enterline PE, DeCoufle P, Henderson V: Mortality in relation to occupational exposure in asbestos industry. J Occup Med 1972; 14:897-903 36. Enterline P, DeCoufle P, Henderson V: Respiratory cancer in relation to occu pational exposures among retired asbestos workers. BrJ Ind Med 1973; 30:162-166 37. Enterline PE, Henderson V: Type of asbestos and respiratory cancer in the asbestos industry. Arch Environ Health 1973; 27:312-317 38. Henderson VL, Enterline PE: Asbestos exposure: Factors associated with ex cess cancer and respiratory disease mortality. Ann NY Acad Sci 1979; 330:117-126 39. Meurman LO, Kiviluoto R, Hakama M: Mortality and morbidity among the working population of anthophyllite asbestos miners in Finland. Br J Ind Med 1974; 31:105-112 40. Weiss W: Mortality of a cohort exposed to chrysotile asbestos. J Occup Med 1977; 19:737-740 41. Rubino GF, Piolatto G, Newhouse ML, et al: Mortality of chrysotile workers at the Balangero mine. Northern Italy. BrJ Ind Med 1979; 36:187-194 42. Gillam JD, Dement JM, Lemen RA, et al: Mortality patterns among hard rock gold miners exposed to an asbestiform mineral. Ann NY Acad Sci 1976; 271:336-344 43. Kolonel LN, Hirohata T, Chappell BV, et al: Cancer mortality in a cohort of naval shipyard workers in Hawaii: Early findings. JNCI 1980; 64:739-743 44. Jones JSP, Smith PG, Pooley FD, et al: The consequences of exposure to asbestos dust in a wartime gas-mask factory. In: Biological effects of mineral fibres. Lyon, France, IARC Sci Publ 1980; 2(30):637-650 45. Hobbs MST, Woodward SD, Murphy B, et al: The incidence of pneumoconi osis, mesothelioma and other respiratory cancer in men engaged in mining and milling crocidolite in Western Australia, In: Biological effects of mineral fibres. Lyon, France, IARC Sci Publ 1980; 2(30):615-625 46. Rossiter CE, Coles RM: H.M. Dockyard Devonport: 1947 mortality study, In: Biological effects of mineral fibres. Lyon, France, IARC Sci Publ 1980; 2(30):713-721 47. Newhouse ML, Berry G, Skidmore JW: A mortality study of workers manufac turing friction materials with chrysotile asbestos. Ann Occup Hyg 1982; 26:899-909 48. Berry G, Newhouse ML: Mortality of workers manufacturing friction materials using asbestos. BrJ Ind Med 1983; 40:1-7 49. Seidman H, Selikoff IJ, Hammond EC: Short-term asbestos work exposure and long-term observation. Ann NY Acad Sci 1979; 330:61-89 50. Hammond EC, Selikoff U, Seidman H: Asbestos exposure, cigarette smoking and death rates. Ann NY Acad Sci 1979; 330:473-490 51. Selikoff U, Seidman H: Health effects of amosite asbestos exposure. J Environ Biol 1981;2:63-78 52. Selikoff U: Constraints in estimating occupational contributions to current cancer mortality in the U.S.--Banbury Report, Quantification of Occupational Cancer. Cold Springs Harbor Laboratory 1981, pp 3-19 53. Acheson E, Gardner MJ, Pippard EC, et al: Mortality of two groups of women who manufactured gas masks from chrysotile and crocidolite asbestos; A 40-year follow-up. BrJ Ind Med 1982; 39:344-348 54. MacMahon B, Pugh TF: Epidemiology: Principles and Methods. Boston, Little, Brown, 1970 55. Bailar JC, Ederer F: Significant factors for the ratio of a Poisson variable to its expectation. Biometrics 1964; 20:639-643 56. Cornfield J: A statistical property arising from retrospective studies. Proc Third Berkeley Symp Math Stat Prob 1956; 4:135-148 57. GartJ: On the combination of relative risks. Biometrics 1962; 18:601-610 58. Armitage P: Statistical Methods in Medical Research. New York, John Wiley and Sons, A Halsted Press Book, 1971 59. Enterline PE: SMR interpretation explained (Letter). J Occup Med 1978 May; 20:314 JULY 1985 143 1 65