Diesel motor exhaust and lung cancer mortality: reanalysis of a cohort study in potash miners

The aim of the reanalysis is to reassess lung cancer risk associated with occupational exposure to diesel motor exhaust in potash miners, while controlling for potential confounders such as smoking and previous occupational history. Our investigation is based on a cohort study of nearly 6,000 German potash miners, who were followed up from 1970 to 2001. The reanalysis also takes into account the employment periods before potash mining, in particular uranium mining. Different approaches (nested case–control study and Cox model) were used to adjust for confounding. The exposure estimates were recalculated, lagging the exposure by 5 years. Exposure groups were defined by tertiles of cumulative respirable elemental carbon (REC) exposure estimates and occupational categories, where exposure was estimated originally by representative measurements of total carbon for different occupations. The highest REC concentration was measured for production workers, about twice as much as for other occupations. The reanalysis revealed that while about 4 % of all study subjects had worked earlier in uranium mines, 10.3 % of later lung cancer cases did so. Although their absolute number was small, the corresponding relative risk estimator was significantly elevated. Our analysis did not show any notable association between cumulative REC exposure and lung cancer risk. Introducing cumulative REC exposure as a continuous variable into the conditional logistic regression model yielded an odds ratio of OR = 1.04 [0.70–1.53]95 % adjusted for smoking and previous employment. The study results give no evidence for an association between REC exposure and lung cancer risk. Only for very high cumulative dose, corresponding to at least 20 years of exposure in the production area, some weak hints for a possible risk increase could be detected. The study underlines the importance of assessing the entire occupational history in occupational studies, especially if the supposed dose–response-relationship is weak.