A reanalysis of atomic-bomb cataract data, 2000-2002: a threshold analysis

Abstract

To obtain the best statistical model for a previous study of cataract prevalence in atomic-bomb survivors, we tested the fitness of the threshold model in an updated dataset of the study, utilizing re-diagnosis by a single ophthalmologist, use of the DS02 dosimetry system, and separation of the in utero group. The results suggest that, in 730 atomic-bomb survivors, we cannot conclude thresholds are greater than 0 Sv in cortical cataract and posterior sub-capsular opacity since the lower 90% confidence limits of the thresholds were 0 Sv. Threshold dose point estimates were 0.6 Sv (90% CI, <0.0-1.2 Sv) and 0.7 Sv (90% CI, <0.0-2.8 Sv) for cortical cataract and posterior sub-capsular opacity, respectively. Detailed regression analyses with no threshold model showed that nuclear color and nuclear opacity have no dose responses (p > 0.40). Cortical cataract showed a significant dose effect (p = 0.002), with an odds ratio (OR)/Sv of 1.30 (95% CI, 1.10-1.53) and no dose-effect modifiers. Posterior sub-capsular opacity showed a significant dose effect (p < 0.001), with an OR/Sv of 1.44 at age of exposure of 10 y (95% CI, 1.19-1.73). The dose effect decreased significantly with increasing age at exposure (p = 0.022). No dose response was observed for in utero survivors (p > 0.20).