Joint radiation risk from several carefully selected models for leukemia mortality in Japanese A-bomb survivors with mortality data (1950 - 2000)

*Jan Christian Kaiser, Helmholtz Zentrum München, Institute of Radiation Protection 
Linda Walsh, Department Radiation Protection and Health, Federal Office of Radiation Protection 

Keywords: leukemia mortality, radiation risk, A-bomb survivors, non-linear dose response, multi-model inference

A recent analysis of leukemia mortality in Japanese A-bomb survivors has used a number of descriptive models of the excess relative risk (ERR) to derive a joint risk estimate by multi-model inference (MMI). The models have been applied in various earlier studies and use a linear-quadratic dose response with differing dose effect modifiers. Here these models have been submitted to a rigorous statistical selection procedure which fosters the parsimonious deployment of model parameters based on pair-wise likelihood ratio tests. Nested models were consequently excluded from risk assessment. Additionally, models of the excess absolute risk (EAR) and two types of non-standard ERR models with sigmoidal responses or two line spline functions with a changing slope at a kink dose have been included in the selection procedure. Due to higher values of the Akaike Information Criterion none of the EAR models has been selected but two non-standard ERR models qualified for MMI. The preferred ERR model applies a purely quadratic dose response which is damped by an exponential factor at high doses and modified by power function for attained age. Compared to the previous analysis the present study reports similar point estimates and confidence intervals (CI) of the ERR from MMI for doses between 0.5-2.5 Sv. However, at lower doses the point estimates are reduced by factors between two and five, although the reduction was not statistically significant. The 2.5% percentiles of the ERR from the preferred quadratic-exponential model did not fall below zero risk in exposure scenarios for children, adolescents and adults at very low doses down to 10 mSv. However, MMI produced risk estimates with a positive 2.5% percentile only above doses of some three hundred mSv. Compared to CI from a single model of choice, CI from MMI are broadened in cohort strata with low statistical power by a combination of risk extrapolations from several models.