Bayesian prediction of childhood leukemia risks potentially dye to radon, terrestrial gamma and cosmic rays in France

摘要

Background: Previous Quantitative Risk Assessments (QRA) have suggested that natural background radiation (NBR) may be a cause of childhood leukemia but have often neglected important sources of uncertainty when radiation-related cancer risks were extrapolated from the Japanese atomic bomb survivors (LSS) to current populations. Aims: To explore the potential impacts of uncertainty in risk models, parameters and minimal latency period between exposure to NBR and the expression of risk, when predicting the excess risk of childhood leukemia in France related to 3 components of NBR: radon, cosmic rays, and terrestrial gamma rays (TGR). Methods: A Bayesian approach was used to fit 6 recently published leukemia risk models to the LSS dataset. The resulting uncertainty distributions in the risk parameters were used to derive predictive distributions of childhood leukemia excess risks due to NBR for French children (0-14 years). Results: Detailed predictions results based on the excess relative (ERR) and excess absolute (EAR) risk models proposed by the United Nations Scientific Committee on the Effects of Atomic Radiation provide a synthetic range for variation of results according to model choice. The posterior predictive medians of the percentage of cases potentially attributable to the 3 components of NBR grouped together are 20% (95% credible interval (CI): 0-68%) and 4% (95% CI: 0-11%) under these ERR and EAR risk models respectively if a 2 years latency period is assumed, 16% (95% CI: 0-66%) and 3.5% (95% CI: 0-9%) if uncertainty in the minimum latency period is addressed following recent recommendations from the US National Cancer Institute. Conclusions: A sizeable proportion of childhood leukemia might be attributable to radon, TGR and cosmic rays in France. However, the large CIs andthe different predictions provided by the 6 models show that the LSS has important limitations to help predicting leukemia due to NBR as far as populations of children are concerned.