Radiation dose dependent risk of liver cancer mortality in the German uranium miners cohort 1946-2003.

An increased risk of mortality from primary liver cancers among uranium miners has been observed in various studies. An analysis of the data from a German uranium miner cohort (the 'Wismut cohort') was used to assess the relationship with ionising radiation. To that end the absorbed organ dose due to high and low linear energy transfer radiation was calculated for 58 987 miners with complete information on radiation exposure from a detailed job-exposure matrix. 159 deaths from liver cancer were observed in the follow-up period from 1946 to 2003. Relative risk models with either linear or categorical dependence on high and low linear energy transfer radiation liver doses were fitted by Poisson regression, stratified on age and calendar year. The linear trend of excess relative risk in a model with both low and high linear transfer radiation is -0.8 (95% confidence interval (CI): -3.7, 2.1) Gy(-1) and 48.3 (95% CI: -32.0, 128.6) Gy(-1) for low and high linear energy transfer radiation, respectively, and thus not statistically significant for either dose. The increase of excess relative risk with equivalent liver dose is 0.57 (95% CI: -0.69, 1.82) Sv(-1). Adjustment for arsenic only had a negligible effect on the radiation risk. In conclusion, there is only weak evidence for an increase of liver cancer mortality with increasing radiation dose in the German uranium miners cohort considered. However, both a lack of statistical power and potential misclassification of primary liver cancer are issues.