The aim of this thesis is to contribute to a better understanding of the health effects of chronic external low doses of ionising radiation. This work is based on the French cohort of CEA-AREVA NC nuclear workers. The mains stages of this thesis were (1) conducting a review of epidemiological studies on nuclear workers, (2) completing the database and performing a descriptive analysis of the cohort, (3) quantifying risk by different statistical methods and (4) modelling the exposure-time-risk relationship.
The cohort includes monitored workers employed more than one year between 1950 and 1994 at CEA or AREVA NC companies. Individual annual external exposure, history of work, vital status and causes of death were reconstructed for each worker. Standardized mortality ratios using French national mortality rates as external reference were computed. Exposure-risk analysis was conducted in the cohort using the linear excess relative risk model, based on both Poisson regression and Cox model. Time dependent modifying factors were investigated by adding an interaction term in the model or by using exposure time windows.
The cohort includes 36,769 workers, followed-up until age 60 in average. During the 1968-2004 period, 5,443 deaths, 2,213 cancers, 62 leukaemia and 1,314 cardiovascular diseases were recorded. Among the 57% exposed workers, the mean cumulative dose was 21.5 millisieverts (mSv). A strong Healthy Worker Effect is observed in the cohort. Significant elevated risks of pleura cancer and melanoma deaths were observed in the cohort but not associated with dose. No significant association was observed with solid cancers, lung cancer and cardiovascular diseases. A significant dose-response relationship was observed for leukaemia excluding chronic lymphatic leukaemia, mainly for doses received less than 15 years before and for yearly dose rates higher than 10 mSv.
This PhD work contributes to the evaluation of risks associated to chronic external radiation exposure. Our study finds a risk of leukaemia associated with chronic external exposure and allows taking into account modifying factors of this relation. Additional follow-up allows to improve the precision of the estimated dose-response relationship. A combined analysis including the present cohort and other nuclear cohorts could quantify more precisely the risks associated with low doses at low dose rates, in order to validate current underlying hypotheses of the radiation protection system.