Mortality of workers in the British steel and coke industries. Final report on CEC Contract 7246-24/8/001 and 7246.24.009

This report describes a study of mortality over a 10-year period in 86 548 industrial workers who were employed continuously in the British steel industry for at least four-and-a-half years until July 1967. Men at 43 different establishments of the British Steel Corporation (BSC) were included. The objectives were to compare the mortality of steelworkers with that of other men in Britain, to describe differences in mortality between occupational groups within the industry, and to establish whether or not work on coke ovens in Britain was associated with excess lung cancer mortality. For this latter purpose 2 842 men employed at 14 BSC coke works for at least 19 months until July 1967, and a further 3 925 men employed on 1 January 1967 at 13 works owned by National Smokeless Fuels Limited (NSF) , were studied over 12- and 13-year periods respectively.BSC workers eligible for study who had left the industry prior to data collection in 1971/73 (“”leavers””), and those still employed by BSC (“”current employees””) were identified from works personnel records and regionally-maintained computer files respectively. Government agencies traced 20 875 “”leavers”” and established their vital status. Other men were traced primarily via BSC Pensions Office files. There were 9 298 deaths recorded over the 10-year study period among the 81 253 traced men ( follow-up rateSome BSC works had been omitted during data collection. In particular, foundry and tin-plate workers were under-represented among the men studied.Summary descriptions of occupation when data were collected (BSC Line Number) were available, from personnel computer files, for some “”current employees””. Of these, 32 859 men (58.5% of all the traced men who were still with BSC on 31.12.1973) were grouped according to job activity and their mortality studied cover a three-and-a-half year period. Information on work history,’ . smoking habits and various health characteristics had been recorded, on questionnaire forms, for about 30% of the “”current employees”” during special appointments at works Medical Centres. A sample (10.390 of these data was studied and results are reported.Special efforts were made during 1975/77 to ensure complete and reliable information, including lifetime histories of work in the industry, for men in the two coke studies. Men were then classified to occupational groups, with special reference to ovens work. There were no important gaps in the study groups as identified, and 6 608 coke workers (97.4%) were traced successfully. A control group of 3 668 steel-workers who had never worked on the coke departments was identified and followed-up for 10 years, with 257 (7.0%) untraced.Primary cause of death only was used in the mortality analyses. Comparisons with groups external to the studies were made by indirect standardisation for age, calendar-year of follow-up and, where possible, geographical region. Internal comparisons between subgroups were also made in relation to standards based on the pooled results from the studies themselves.The mortality of the steelworkers was less than that expected in the general population among men of similar age, living in the same geographical regions (SMR = 78). Losses to follow-up suggest that this figure underestimates slightly the true rate. Such favourable mortality, related to the ability to carry out manual work, is usual in studies of industrial groups. The (standardised) proportion of deaths due to lung cancer, as a percentage of that expected, was 114. This was higher than among manual workers generally. Analyses of mortality in relation to occupation showed that the lung cancer excess was most clearly marked among blast-furnacemen, whose general mortality was also relatively unfavourable. Further studies of these men are suggested. In a supplementary exercise the summary occupational data used in classifying the men were compared on a sample basis with more detailed work histories supplied by the men at interview, and were found reliable.The observed fraction of deaths from lung cancers among coke workers was consistently higher than that expected compared with all men in Britain, men in the same Social Classes, and the control group of steelworkers (though the latter comparison may be unreliable). On average the proportion of lung cancer deaths was 1.16 times that expected among semi-skilled manual workers in England and Wales (Social Class IV) and the excess was evident in all age-groups. The numbers of lung cancer deaths were greater than expected among men generally, and similar to other semi-skilled workers. An excess was found in men at ages less than 55 years (SMR = 126 compared with Social Class IV).Deaths from all causes among the coke workers were fewer than expected, irrespective of the comparison group, except among NSF workers at age less than 55 years. An excess of bladder cancer among NSF workers, based on just 12 deaths, was not found among BSC men.Among NSF workers aged less than 45 years when follow-up began, the subsequent lung cancer mortality among those with some (full- or part-time) experience of ovens jobs was higher than among non-ovens men. This difference, while clearcut, was based on only 10 deaths. No such difference was found in the BSC study. But lung cancer mortality among BSC coke workers with longer (more than five years) experience in the ovens jobs was less favourable than among other BSC ovens men. There was no association between time on ovens jobs and deaths from lung cancer among NSF ovens workers. An excess of stomach cancers among NSF non-ovens workers was not found in the BSC group.The results did not show that the lung cancer excess was specific to particular works. Limited data on smoking habits indicated that tobacco consumption among coke workers was not greater than among manual workers generally in Britain.The overall favourable mortality of the coke workers, and the age-specific patterns observed, are consistent with “”healthy worker”” effects among the men studied. Three things follow. First, it is sensible to consider the proportions of deaths as well as actual numbers. We consider that an occupationally-related excess of lung cancer has been shown clearly. Second, the absolute death rates are likely to have been under-estimated, especially among older men. We consider that the excess number of lung cancer deaths among younger workers may reflect the real risks more accurately than results for all men. However, lung cancer mortality was clearly not as severe as that reported for coke workers in the USA and Canada, particularly ovens men. This may be due to differences in employment and work patterns in the countries concerned. Finally, dose-response relationships may have been obscured by selection effects. Results from these studies in relation to occupation within the coke wortcs were not clearcut, but the findings were consistent with a hazard from coke oven fume emissions and it is concluded that continued hygiene measures to protect the workforces from carcinogenic fumes are justified. No other occupationally-related excess mortality was demonstrated. Continuation of the follow-up is suggested, and some issues for further study are identified. “”