Thesis: The relation between lung cancer risk and radon exposure has been clearly established, especially from the studies on uranium miner cohorts. But the association between radon exposure and extrapulmonary cancers and non-cancer diseases remains not well known. Moreover, the health risks associated with the other mining-related ionizing radiation exposures are still under consideration. The aim of this thesis is to contribute to the estimation of the radio-induced health risks at low-doses through the analysis of the kidney cancer and Circulatory System Disease (CSD) mortality risks among uranium miners.Kidney cancer mortality risk analyses were performed from the French cohort of uranium miners (n=5086; follow-up period: 1946-2007), the post-55 cohort (n=3,377; follow-up period: 1957-2007) and the German cohort of the Wismut (n=58,986; follow-up period: 1946-2003) which included 24, 11 and 174 deaths from kidney cancer, respectively. The exposures to radon and its short-lived progeny (expressed in Working Level Month WLM), to uranium ore dust (kBqh.m-3) and to external gamma rays (mSv) were estimated for each miners and the equivalent kidney dose was calculated. The dose-response relation was refined considering two responses: the instantaneous risk of kidney cancer mortality (corresponding to the classical analysis, Cause-specific Hazard Ratio (CSHR) estimated with the Cox model) and its occurrence probability during the follow-up (Subdistribution Hazard Ratio (SHR) estimated with the Fine & Gray model). An excess of kidney cancer mortality was observed only in the French cohort (SMR = 1.62 CI95%[1.04; 2.41]). In the Wismut cohort, a decrease of the kidney cancer mortality was observed (0.89 [0.78; 0.99]). For these three cohorts, the occupational radiological exposures (or the equivalent kidney dose) were significantly associated neither with the risk of kidney cancer mortality (e.g. CSHRWismut_radon/100WLM=1.023 [0.993; 1.053]), nor with its occurrence probability during the follow-up (e.g. SHRWismut_radon /100WLM=1.012 [0.983; 1.042]).CSD mortality risk analyses in the French cohort showed a significant increase of the risks of mortality from CSD (n=442, CSHR/100WLM=1.11 [1.01; 1.22]) and from CerebroVascular Disease (MCeV, n=105, CSHR/100WLM=1.25 [1.09; 1.43]) with radon exposure. A case-control study nested in the French cohort was set up to collect the information related to CSD risk factors (overweight, hypertension, diabetes…) from the medical records of 313 miners (76 deaths from CSD (including 26 from Ischemic Heart Disease (IHD) and 16 from MCeV) and 237 controls). For the three radiological exposures, the exposure-risk relation was analyzed in a pseudo-cohort (n=1,644 pseudo-individuals, obtained from the weighting of the observations by their inverse selection probability) with the Cox model, adjusted for the CSD risk factors. The association between the radiological exposure and the risk of mortality from CSD, IHD or MCeV was not significant (e.g. CSHRCSD_radon/100WLM=1.43 [0.71; 2.87]). The adjustment for CSD risk factors did not substantially change the exposure-risk relation.The lack of a significant dose-response relation suggests that the excess of kidney cancer mortality among the French uranium miners may be induced by other risk factors, unavailable for this study. The small change of the coefficients observed after adjustment for CSD risk factors in the nested case-control study supports the assumption of the existence of the MCeV mortality risk increase associated with radon exposure in the French cohort of uranium miners. Future analyses based on further follow-up updates should allow to confirm or not these results.