[Effects of chronic uranium internal exposure on mortality: results of a pilot study among French nuclear workers]. / Effets de la contamination chronique à l'uranium sur la mortalité : bilan d'une étude-pilote chez les travailleurs de l'industrie nucléaire en France.

BACKGROUND: This article presents the mortality data compiled among a cohort of workers at risk of internal uranium exposure and discusses the extent to which this exposure might differentiate them from other nuclear workers. METHODS: The cohort consisted of 2897 Areva-NC-Pierrelatte plant workers, followed from 1st January 1968 through 31st December 2006 (79,892 person-years). Mortality was compared with that of the French population, by calculating Standardized Mortality Ratios (SMR) and 95% confidence intervals (CI95%). External radiation exposure was reconstructed using external dosimetry archives. Internal uranium exposure was assessed using a plant-specific job-exposure-matrix, considering six types of uranium compounds according to their nature (natural and reprocessed uranium [RPU] and solubility [fast-F, moderate-M, and slow-S]). Exposure-effect analyses were performed for causes of death known to be related to external radiation exposure (all cancers and circulatory system diseases) and cancer of uranium target-organs (lung and hematopoietic and lymphatic tissues, HLT). RESULTS: A significant deficit of mortality from all causes (SMR=0.58; CI95% [0.53-0.63]), all cancers (SMR=0.72; CI95% [0.63-0.82]) and smoking related cancers was observed. Non-significant 30%-higher increase of mortality was observed for cancer of pleura (SMR=2.32; CI95 % [0.75-5.41]), rectum and HLT, notably non-Hodgkin's lymphoma (SMR=1.38; CI95 % [0.63-2.61]) and chronic lymphoid leukemia (SMR=2.36; CI95% [0.64-6.03]). No exposure-effect relationship was found with external radiation cumulative dose. A significant exposure-effect relationship was observed for slowly soluble uranium, particularly RPU, which was associated with an increase in mortality risk reaching 8 to 16% per unit of cumulative exposure score and 10 to 15% per year of exposure duration. CONCLUSION: The Areva-NC-Pierrelatte workers cohort presents a non-significant over-mortality from HLT cancers, notably of lymphoid origin, unrelated to external radiation exposure. The pilot study suggests an association between mortality from the HLT and lung cancers and exposure to slowly soluble RPU compounds. The results of this study should be investigated further in more powerful studies, with a dose-response analysis based on individual assessment of uranium absorbed dose to uranium-target organs.