Workers of the Commissariat for Atomic Energy and Alternative Energy (CEA) may be potentially exposed to tritium over long periods. We aimed to assess the effect of tritium exposure on mortality in a cohort of employees followed by radiotoxicological monitoring. A total of 1,746 employees who worked for at least six months at one of three CEA centers were included between 1962 and 2011 (median follow-up 29.6 years). The cumulative dose of tritium was based on the quantification of tritium present from urinary excretion monitoring data from the beginning of occupational exposure to the end of such exposure or December 2011. Mortality was first compared to that in the French population using the standardized mortality ratio (SMR). Then, mortality risk ratios (RRs) per category of cumulative dose of tritium were estimated using categorical Poisson models adjusted for age at the onset of exposure, age, calendar period, sex, smoking, employment status, CEA center, and taking into account the number of person-years. The main causes of mortality were tumors (48%) and cardiovascular diseases (20%). The comparison of mortality within the cohort to that in the French population highlighted a lower rate for all-cause mortality and that due to cancer, related to the healthy worker effect bias. The regression model showed no effect of cumulative dose on all-cause mortality. The risk of death for most malignancies was positive, but not significant for the higher classes of doses relative to the reference class. The highest risk (not significant) was present for the class of higher doses for tumors of the larynx, trachea, bronchi and lung. The risk was significant for the higher doses for tumors of the pancreas and bladder (based on a limited number of cases: five and six deaths, respectively). Significantly more smokers died from tumors of the respiratory system than non-smokers, as expected. We were unable to show an effect of cumulative tritium dose due to the small size of the cohort and the low exposure level. However, our study underlines the need to continue following tritium-exposed workers and conducting multicenter studies.