Lung cancer from radon and smoking: a multistage model for the WISMUT uranium miners

Full text: In the world's third-largest uranium-mining province located in areas of Saxony and Thuringia in the former German Democratic Republic, the WISMUT Company conducted extensive uranium mining starting in 1946. Up to 1990, when mining activities were discontinued, most of the 400,000 employees had been exposed to uranium ore dust and radon and its progeny. It is well established that, besides smoking, such exposures are associated with an increased risk of lung cancer. From about 130,000 known miners a huge cohort of 59,000 miners has been formed and in an epidemiological analysis lung cancer risks have been evaluated (Grosche et al., 2006). We will present an alternative approach using a biologically-based multistage carcinogenesis model quantifying the lung-cancer risk related to both the exposure to radon and smoking habits. This mechanistic technique allows for extrapolation to the low exposures that are important for present-day radiation protection purposes and the transfer of risk across populations. The model is applied to a sub-cohort of about 35,000 persons who were employed at WISMUT after 1955, with known annual exposures estimated from the job-exposure matrix (Lehmann et al., 2004). Unfortunately, detailed information on smoking is missing for most miners. However, this information has been retrieved in two case-control studies, one of which was nested in the cohort while the other was not (Brueske-Hohlfeld et al., 2006). For these studies, the relevant smoking parameters are assembled in so-called smoking spectra that are next projected onto the entire cohort using a Monte-Carlo sampling method. Individual smoking habits that are randomly assigned to the cohort members, together with the information on annual exposure to radon, is used as an input for the multistage model. Model parameters related to radon and tobacco exposure are fitted with a maximum-likelihood technique. We will show results of the observed and expected lung-cancer mortality among WISMUT miners and present risk calculations associated with the delicate interplay between radon exposure and smoking habits. Finally, we will estimate the excess relative risk associated with the exposure to residential radon, by extrapolation to low exposures. (author)