The Diesel Exhaust in Miners Study is unquestionably the most suitable data material to date to examine a possible link between diesel engine emissions and lung cancer risk. But the results do not appear to be consistent in themselves. The crucial methodological problem in this study, however, has yet to be discovered, to which the lack of any description of age related information (year of birth, year of hire, year of first exposure, year of death) for the cohort as well as for the cases might have contributed. This information is important to understand the flaws in the analysis. It turns out that the year of birth is associated with the exposure, i.e. with the chance to be exposed over a certain period of time as well as with the chance to be an ever-smoker. A further important issue for the interpretation of the results is the validity of the data on smoking, which are mainly obtained from next of kin for decedents up to 50 years after death. Taking all these aspects into account, it can be concluded that only the SMR-analysis can be considered from all published results.

This article is published in the "European Journal of Epidemiology", Volume 33, Issue 12, pp. 1251-1254.