Comparative study of approached cancer incidence in a cohort of permanent nuclear workers with that in a sample of French population, period 2009-2018

While many studies have been conducted on mortality in nuclear workers, morbidity remains less studied in these populations. This study aimed to assess the algorithm-derived incidence of cancer in a population of French nuclear workers, by using medico-administrative data from the National Health Data System (SNDS) as no national cancer registry is available, and to compare it with a general population sample. The study cohort included permanent workers with at least 1 year's work history in the company Electricité de France, monitored for ionizing radiation exposure at least once over the period 1961-2003. The comparison population was a representative random sample of 2% of the French population receiving outpatient healthcare. Cancer incidences were estimated over the 2009-2018 period using data collected from the SNDS and validated algorithms. Standardized incidence ratios and 95% Confidence Interval (SIR [CI95%]) were calculated. The cohort comprised 27,473 workers, while the comparison population included 611,953 persons. Incidences in the cohort were significantly higher for prostate cancer (SIR=1.09 [CI95% 1.01-1.18]), malignant melanoma of the skin (1.33 [1.08-1.60]), and breast cancer in women (1.35 [1.03-1.73]) compared with that of the comparison population. Significant deficits were observed for liver, larynx, trachea-bronchus-lung and bladder cancers. This study is the first to use the SNDS in a cohort of workers to assess cancer incidence. In the absence of linkage with individual radiation dose information, it should not be interpreted as evidence of radiation causality. The next step will be to investigate exposure-risk analyses when data access will be completed.