The Mortality of Steel Foundry Workers: A Cohort Study

Abstract

Patterns of mortality were analysed among 11069 steel foundry workers: all male manual workers who started work in ten British steel foundries between 1946 and 1965 and were employed at least one year. They were followed up for mortality until 1983. Person-years were summed and expected deaths calculated based on England and Wales age and calendar-year specific death rates. Correction factors based on regional differences in mortality were applied. Standardised Mortality Ratios (SMRs) were calculated as ratios of observed to expected deaths and the excess risk was analysed by modelling its relationship with duration of employment using GLIM. A group of 623 inmigrants had high loss to follow up and very few deaths and were analysed separately. The vital status of the remaining 10446 workers was established for 96.8% of the cohort. The SMR for all causes is 111 based on 2789 deaths (p<0.001). The excess mortality is entirely accounted for by significant excesses of lung cancer, stamach cancer and non-malignant respiratory disease mortality. The excess lung cancer mortality is apparent in the foundry subgroup (SMR = 159, observed/expected = 164/103.3, p<0.001) and the fettling shop subgroup (SMR = 172, obs/exp = 86/49.9, p<0.001) but not those unexposed to these environments (SMR = 119, obs/exp = 77/64.6). There is no consistent pattern with duration of employment, but the excess is considered as unlikely to be attributable to smoking and, given the consistent findings in other studies is largely attributable to foundry employment. The role of specific exposures in the foundry environment could not be evaluated and the level of risk does not appear to be changing over calendar period. Mortality from non-malignant respiratory disease is significantly raised among the exposed (SMR = 144, obs/exp = 317/219.9, p<0.001), being highest for the fettling shop workers (SMR = 172). The risk has been falling over time and no risk is evident for workers who started after 1960, although in this latter group a small excess (obs/exp = 8/5.1) persists for bronchitis, emphysema and asthma. For stomach cancer the excess mortality is restricted to the foundry subgroup (SMR = 192, obs/exp = 52/27.1, p<0.001). Correction for regional mortality has little impact on any of these excesses although it removes a significant excess of circulatory disease mortality (the SMR falls from 111 to 98).