Document 7MYy1ZBL87LOYgyQgNJyEyV98

Differences in occupational mortality from pleural cancer, peritoneal cancer, and asbestosis David Coggon, Hazel Inskip, Paul Winter, Brian Panneu MRC Environmental Epidemiology Unit, University of Southampton, Southampton Ceneral Hospital, Southampton SOI 6 6YD D Coggon H Inskip P Winter B Panneu Correspondence to: Dr David Coggon, MRC Environmental Epidemiology Unit, University of Sourhampton, Southampton General Hospital, Soutluimpron SO 16 6YD. Accepted 2d June 1995 Abstract One starting point is to identify the occupa Objective--To assess whether the tions associated with excess mortality from increased risk of disease related to diseases related to asbestos. During the 1980s, asbestos in occupations from the con a large proportion of mate deaths from struction and engineering industries mesothelioma in England and Wales were applies equally to pleural cancer, peri related to work in the construction and engi toneal cancer, and asbestosis. neering industries.1 It is unclear whether the Methods--Analysis was based on deaths increased risk in these jobs applies to mesothe among men aged 20--74 in England and lioma of the pleura, peritoneum, or both. We Wales during 1979-80 and 1982-90. (n = have analysed data from the same period that 1 656 096). Information about cause of indicate important differences in the relative death and the last lull time occupation of frequency of pleural and peritoneal mesathe- decedents was derived from death certifi lioma by occupation, and which may have cates. Proportional mortality ratios implications for control strategy. (PMRs) by occupation were calculated for each of pleural cancer, peritoneal cancer, and asbestosis. Method Results--Altogether, 2S4S deaths were Our analysis was based on all deaths among attributed to cancer of the pleura, 362 to men aged 20-74 in England and Wales in cancer of the peritoneum, and 281 to 1979-80 and 1982-90. Data for 1981 were asbestosis. When occupations were unreliable because of industrial action in that ranked according to PMRs from these year by registrars of deaths. Information about diseases, striking differences were found. the age, underlying cause of death, and most The category of construction workers recent full time occupation of decedents was which Included loggers had the highest obtained from death certificates. Causes of mortality from peritoneal cancer (PMR death were coded to the ninth revision of die 990, 64 deaths), but a PMR of only 160 (77 international classification of diseases (ICD- deaths) for pleural cancer. In contrast, 9), and occupations were coded initially to the several occupations with much higher Office of Population Censuses and Surveys mortality from pleural tumours had no (OPCS) 1980 classificadon of occupations.' excess of peritoneal cancer. PMRs for The occupational units defined in the OPCS asbestosis related more closely to those classification were then aggregated into 194 for peritoneal than pleural cancer. larger job groups.' . Conclusions--These findings suggest that For each job group we calculated propor the exposure-response relations for dis tional mortality ratios (PMRs) for cancer of eases related to asbestos are not all linear, die pleura (I CD-9 = 163), cancer'of the peri and that risks of pleural mesothelioma toneum (ICD-9 = 158-8 and 158 9), and . may be underestimated by simple extrap asbestosis (ICD-9 = 501), with five-year age olation from observations in cohorts with specific proportions in all occupations com heavy exposure. bined as the standard. Confidence intervals (CIs) for PMRs were based on the Poisson (tkcup Environ Med 1995;52:775-777) distribution. Keywords; asbestos; mortality; mesothelioma Occupational exposure to asbestos is a major cause of mortality from asbestosis, mesothe lioma, and bronchial carcinoma. In Britain, legal controls on exposure to asbestos were first introduced in 1933, and since then have been progressively tightened. Recent analysis indicates that mortality from mesothelioma is rising in men at most ages, and suggests that overall rates will continue to increase through to the next century.1 It is important to estab lish whether this trend results from inadequate enforcement of statutory exposure limits, or whether the limits themselves are not suffi ciently stringent. Results Over the 11 year period of study 1 656 096 deaths were recorded in men with adequately described occupations, including 2848 from cancer of rhe pleura, 362 from cancer of the peritoneum, and 281 from asbestosis. The table lists the job groups with significantly (P < 0-05) raised PMRs for at least one of these diseases. Tire ranking of PMRs for cancers of the pleura and peritoneum was quite different. For example, construction workers not else where classified (nec), a group which includes laggers, had the highest mortality from peri toneal cancer (PMR 990, 64 deaths), but a 77ft ' 'jcp'n, /mktp, H imn, famm Jo* poygj tottA sipiificataly (P < 0 05J raised mortalityfixm cancer of thepleura, cancer ofthe peritoneum, or asbawiis Job group Ouupeaonai units * Deatkt fimt aS ZOAtJCS Cancer ofpleura Deaths PMR (95% Cl) Cancer ofprihmeimt Dtathi PMR (95% Cl) Meu! plate workers Vehicle body builders Plumben and pi fitters Uphdilemr Caipcntm Electricians Electrical plant operators Chemical engineers and scientists Welders Managers in construction Boiler operators Electrical engineers (so described) Production fitters Plasterers Sheet rectal workers Builders and handymen Professional engineers nee Construction workers nec Architects and surveyors Dockers and goods porters Chemical workers Preparatory fibre processors Smiths arid forgemcn Machine tool operators 124 3, 124-2 129-3, 131-1 124-1, 125 101-4, 103 104-1, 105-1 120*2, 121*2 120 3,121-4 024 2, 028-1 5867 2250 16941 2763 27425 20517 2681 2712 124-6, 12a 035 1, 035 2 150 9, 161*1 1213 12467 7404 6212 7930 114 4,117 52566 139-3, 140-3 5198 124-2, 126 1 7961 139*5, 140*5, 140 6 22433 025, 026*1, 026*2, 028-2, 17780 028*3,028*4,028-6 139-12,143-1, 143-2, 26886 146-2 031*1, 031*2, 031*3 7463 156*2, 156*3, 157-2, 157-3 13603 088,069, 159-2, <60-2 15480 086 1,087 1 862 108 3* 109-3 2770 in 4*1124 50854 73 709 556-892 24 649 416-968 134 450 377-533 16 366 209-594 167 362 309-421 127 349 291-415 14 301 165-506 13 274 146-468' 56 247 186-320 32 240 164-340 24 240 153-357 31 227 154-323 192 208 180-240 18 207 122-327 25 196 120-274 65 166 128-212 53 162 121-212 77 160 126-200 21 160 99-245 36 156 109-216 39 144 103--J98 2 139 17-501 6 133 49-289 116 132 109-158 1 78 2-432 4 877 239-2246 11 283 141-506 0 0 0-688 6 i 102 37-221 4 83 23-212 0 0 0 643 0 0 0-597 1 33 1-186 6 349 128 759 2 m 21-616 1 58 1-325 12 103 53-179 3 265 55-773 4 235 64-602 5 98 32-229 2 46 6-168 64 990 762--1265 1 58 1-324 7 242 97-500 5 146 47-341 0 0 0-2049 3 561 116*1639 15 136 76-224 *As dclmtd in the Oi*CS 1960 classification of occupations, net = not clscwllCTC classified. Asbestosis Deaths PMR (95K Cl) 3 5 13 0. 9 | 1 0 292 1302 457 0 196 29 212 0 60-854 423-3039 243-782 0-802 90-372 1-164 5-1180 0-813 3 1 '4 1 142 29-415 77 2-429 366 100-937 74 2410 17 189 110-303 0 0 0-431 4 300 82-766 0 0 0-98 0 0 0-120 71 1592 1243-2009 0 0 0^297 4 170 46-435 5 189 61-441 2 1316 159-4753 0 0 8-765 11 126 63-226 PMR of only 160 (77 deaths) for pleural can cer. In contrast, several of the occupations with the highest mortality from cancer of the pleura--metal plate workers (PMR 709, 73 deaths), upholsterers (PMR 366, 16 deaths), carpenters (PMR 362, 167 deaths) and electri cians (PMR 349, 127 deaths)--had no excess of peritoneal cancer. Mortality from asbestosis was more closely related to that torn peritoneal than pleural cancer, with the highest PMR again in con struction workers nec (PMR 1592, 71 deaths). The Spearman rank correlation coefficient across the 24 job groups in the table was 0-43 for asbestosis with cancer of the peritoneum and 0T5 for asbestosis with cancer of the pleura. Discussion In this analysis we restricted attention to the ' three causes of death that are related most specifically to asbestos. Asbestos also causes bronchial carcinoma, but the effect is less dis cernible in analyses of occupational mortality because the relative risk is smaller and because associations arc confounded by differences in smoking and exposure to other lung carcino gens in the workplace. The analysis was limited by inaccuracies that are known to occur hi information obtained from death certificates. Not all deaths ascribed to cancers of the pleura or peritoneum are mesotheliomas, and some mesothelioma deaths ate classified as cancers of other or unspecified sites.* Also, some deaths may be incorrectly attributed to asbestosis on the basis of pleural thickening or plaques, when no fibrosis is present. Moreover, data were only available on the most recent full time job of decedents, and some subjects will have been exposed to asbestos in earlier employment about which we had no information. In general, however. die effect of such errors should be to attenuate occupational associations, and they would not be expected to have a differential effect on the ranking of occupations by mortality from dif ferent diseases related to asbestos. The use of periodic medical examinations or an unusually high rate of necropsies in certain occupations might boost the detection of peritoneal more than pleural tumours, but it would not explain such targe discrepancies in the relative fre quencies of these diseases as were found. Nor would it account for the high frequency of deaths from asbestosis in some occupations with relatively low mortality from pleursl cancer. Another limitation was die use of propor tional mortality rather than true death rates. Although PMRs may have been somewhat depressed or inflated by differences in the overall death rates of job groups, the effect should be similar for each of the three diseases examined, and again would not explain the contrasting tanking of jobs by PMR. . The occupations with significantly raised PMRs in our analysis were largely the same as those found previously to have high mortality from mesothelioma overall.1 All entail poten tial exposure to asbestos although in some cases the exposure is related to work in specific industries rather than a general feature of the occupation. For example, the excess mortality among welders occurred mainly in centres of shipbuilding.' Many of the high risk occupa tions are in die construction industry where exposure has occurred from the use of asbestos in lagging and odier building materi als. The high PMR for cancer of die pleura hi carpenters compared with many other build ing trades may be related to work with asbestos board. Notable for an absence of sig nificant risk are motor mechanics, (PMR 46, 12 deaths from pleural cancer; PMR 88, three deaths from peritonea) cancer; PMR 80, two deaths from asbestosis) about whom concerns * >' (4r> rJ . ^...I , f-r - rj. have been raised because of the presence of pictit'ai cancer is lower and diluted to a level 9 .4 asbestos in brake linings. It seems that this below that in occupations such as carpenters exposure has no important effect on mortality where a larger proportion of men are exposed, 1itL/ from mesothelioma or asbestosis, perhaps but at a lower level. j because the asbestos fibres are modified by This hypothesis is consistent with findings i.\ heal. on the ratio of pleural to peritoneal mesothe The difference in the ranking of occupa liomas in cohort studies of asbestos workers, tions by mortality from pleural and peritoneal where peritoneal turnouts have tended to be cancer and asbestosis is striking and cannot relatively more common in cohorts with longer realistically be ascribed to chance. Nor is it and heavier exposures,6 7 although with occa likely to be explained simply by differences in sional exceptions.6 It also accords with the the types of asbestos to which occupations are finding of higher fibre contents in the lungs of exposed. Crocidolite and amosite are more patients dying horn peritoneal compared with potent causes of mesothelioma than pleural cancer.* It is an indication for caution chrysolite,5 and all of the occupations with when extrapolating risk estimates for mesothe high PMRs from peritoneal cancer could have lioma at law exposures to crocidolite and involved exposure to crocidolite. But welders, amosite from observations in cohorts with who had a higher PMR than construction heavy exposure. In particular, the risks of workers nec from pleural cancer, are also likely pleural mesothelioma may be underestimated to have had exposure to crocidolite (especially if a linear exposure-response is assumed. This those employed in shipbuilding), and yet had a should be taken into account when control deficit of peritoneal cancers. limits are reviewed. A more plausible explanation is that the exposure-response relations for mesothelioma and asbestosis are non-linear, with the risk of pleural mesothelioma increasing relatively We thank the Office of Population Ccnsusc* and Surveys for providing us with the mortality data, sod Dr Comda Magnani for his helpful comments on die findings. The analysis was sup* ported by a grant from the Health and Safety Executive. more steeply at low exposures but less steeply at high exposures. Where an occupation entails low exposure to asbestos, excess pleural cancer occurs but there is little effect on peritoneal cancer or asbestosis. With high exposures, the risk of pleural cancer is increased further, but that of peritoneal cancer and asbestosis. goes up much more and becomes dominant. Also, the effect on occupational mortality is diluted according to the proportion of men in the job group who are exposed. For example, con struction workers nec include not only laggers with very high exposure to asbestos, but also other occupations such as floor layers with minimal exposure. As a consequence, the PMRs of the group are reduced for diseases related to asbestos. In the case of peritoneal cancer and asbestosis, the risk in iaggers is so high that die PMR of the job group as a whole remains highest in the ranking. The PMR for 1 Peto Jj Hodgson JT, Matthew* FE> Janes JR. Continuing increase in mesothelioma mortality in Britain. Lanttx I995j344;535-9. 2 Office of Population Censuses and Surveys. Cfafsiftcation of occupation* 19&0, London: HMSO, I960, 3 Office of Papulation Censuses and Surveys/licalth and Safety Executive, Occupaumai -Health D&attnal Sufiptc/rttiiil London: HMSO, 1995. -1 Newhouse ML, Berry G, Wagner JC. Mortality of factory worker^ in Ease London. 1933-BO. Br J fiid Med 1985;42:4-11, 5 Acheson ED, Gardner MJ. Atb&w: dtt conavt Htoit for vibestos. l-oudon: HMSO, 1983 6 Sdikoff Q, Hammond EC, Churg J. Me rudr'ry experiences of asbestos insulation workers (943-68. In: Shapiro HA, ed. Pi\Mt*nocD7\iotiK pryetidingi of the intcnuttuntdl eanfer- entc, Johumsbttrgr Capetown: OUP, 1969:180~6, 7 Browne Kk Smitbar WJ. Asbestos-related mesothelioma: factors discriminating between pleural and peritoneal sites. ItrJfudAUd 1983;48:145-52. 8 Armstrong BK, Musk AW, Baker JE, Hunt JM, Mcwall CC, jlenzdl HR, ti at Epidemiology of malignant inesodiuTioma in Western Australia. Mid J Aust 1984; 141;8G 9 Wagner JC, Newhouse ML, Corrin B, Rossirer CER, Griffiths DM. Correlation between fibre content of the lung and disease in East I-ondon asbestos factory work ers. Br`J ind Med 1988j4S.`305^a. < < 4 <c 4 4 <