Temporal effect modifiers of the radon-associated excess relative rate of lung cancer in the Pooled Uranium Miners Analysis (PUMA) study

Abstract

Aim: The Pooled Uranium Miners Analysis (PUMA) is the largest pooled study of uranium miners to date. Here we report initial observations on temporal patterns of variation in the association between cumulative exposure to radon decay products and lung cancer mortality in this large multinational pooled study. Methods: PUMA combines cohorts of uranium miners from Canada, the Czech Republic, France, Germany, and the United States. PUMA includes approximately 120,000 males followed for lung cancer deaths between 1946 and 2014. Associations between occupational radon progeny exposure and lung cancer mortality were estimated using internal Poisson regression with background stratification by cohort, attained age, calendar period, and race. This analysis focuses on windows of age at exposure and time since exposure modeled simultaneously and separately as modifiers. Results: In this study with 7,774 lung cancer deaths and 4.4 million person-years, there was statistically significant evidence of a positive association between cumulative exposure to radon decay products, under a 5-year lag assumption, and lung cancer mortality. There was statistically significant evidence of variation by age at exposure in the association between cumulative exposure to radon decay products, under a 5-year lag assumption, and lung cancer mortality. There also was statistically significant evidence of variation in the association between cumulative exposure to radon decay products and lung cancer mortality with time since exposure; the estimated excess relative rate (ERR) per working level month (WLM) tended to diminish across windows defined by increasing time since exposure; however, estimated ERR per WLM remained above the null even 35+ years after exposure. Conclusion: Preliminary analyses yield estimates with substantial statistical precision of the radon-lung cancer association and its variation with age at exposure and time since exposure. Further analyses will investigate additional effect modification by exposure rate and heterogeneity between individual studies.