Carson Mowrer, Eric Whitsel, Jason M Collins, Shu Chen, Ayesha Ekanayaka, JoAnn Manson, Jaymie Meliker, Alexander P Reiner, Gary G Schwartz, Aladdin H Shadyab, Cuiran Shi, Richard Smith, James D Stewart, Hari H Venkatachalam, Matthew Bozigar
Higher radon concentrations were consistently associated with elevated malignant CNS cancer hazard, with the clearest concentration-response observed in continuous models. Residential radon exposure may contribute meaningfully to the overall burden of CNS cancers in this largest cohort study of the radon-CNS cancer association to date.
BACKGROUND: Central nervous system (CNS) cancers, including brain cancer, are rare but highly lethal, and risk factors remain poorly understood. Radon, a known lung carcinogen, may affect non-pulmonary sites, including the CNS.
METHODS: We estimated the radon-related hazard of incident malignant CNS cancer in a large United States (U.S.) cohort of postmenopausal women from the Women's Health Initiative. Incident CNS cancers were identified through physician adjudication or death certificates. Radon concentration estimates at residential addresses included U.S. Geological Survey (USGS) Radon Index and Zones, U.S. Environmental Protection Agency (EPA) Radon Zones, and models from the Lawrence Berkeley National Laboratory (LBL) and Li et al. (2025). Cox proportional hazards models were used to estimate hazard ratios, including study design, sociodemographic, behavioral, and environmental covariates. Corresponding population attributable fractions were calculated.
RESULTS: Over follow-up (mean=17.7 years), 451 incident malignant CNS cancers were identified among 159,391 participants. Absolute risks of malignant CNS cancer increased with increasing radon concentration across all metrics. Non-linear spline models for USGS Radon Index, LBL estimates, and Li et al. estimates showed increasing CNS hazard even at low radon concentrations (global test of spline terms p=0.022 for Li et al.). In categorical analyses, the hazard ratio comparing the highest vs lowest Li et al. tertile was 1.49 (95% CI: 1.12-1.99), and the population attributable fraction was 20.1% (4.8%-33.9%).
CONCLUSION: Higher radon concentrations were consistently associated with elevated malignant CNS cancer hazard, with the clearest concentration-response observed in continuous models. Residential radon exposure may contribute meaningfully to the overall burden of CNS cancers in this largest cohort study of the radon-CNS cancer association to date.