Étude de mortalité des vétérans ayant bénéficié d’une surveillance dosimétrique lors des essais nucléaires français dans le Pacifique

Abstract Background A mortality study was carried out in a cohort of veterans present on the sites of the French nuclear experiments center in the Pacific (CEP) from 1966 to 1996, and for whom external dosimeter monitoring recordings were available. Methods The cohort included 32,550 veterans having had at least one dosimetery recording. Current vital status was collected from the National Register of Identification of Physical People and causes of death data from the national causes of death database. Total mortality and mortality by cause were compared with mortality of the French population using standardized mortality ratios (SMR). To test the effect of a dosimeter recording higher than the threshold (0.2 mSv), i.e., no null dosimetry, the mortality of veterans was compared inside the cohort, using standardized ratios and Poisson regression models. Results The mortality analysis was performed among 26,524 men, of whom 8% had had at least one non-null dosimeter. Five thousand four hundred and ninety-two (21%) veterans died before December 31, 2008 and causes were available for nearly 97% of these deaths. Comparing the mortality between the cohort and the French population highlighted a deficit of mortality, for all causes, by cancer and for radiation-induced pathologies; these results were related to the “healthy worker effect”. The data showed that all causes mortality and cancer mortality of the cohort of veterans with no null dosimeter were not different from those of other veterans, but also showed an excess of hematological malignancies in this sub-population: this excess was significant in the regression model (RR = 1.82; CI 95% [1.6–2.0]). Conclusion Among veterans with an external dosimeter monitoring recording, presence on the sites of CEP from 1966 to1996 does not constitute a factor of increased mortality compared with the national population. However, an increased risk was observed for mortality by hematological malignancies among veterans with no null dosimetry. This result is in line with studies on veterans present during nuclear experiments abroad.