Radon Exposure and Incident Stroke Risk in the Women's Health Initiative

Abstract

Background and objectives: Little is known about the role of radon in the epidemiology of stroke among women. We therefore examined the association between home radon exposure and risk of stroke among middle-aged and older women in the United States.

Methods: We conducted a prospective cohort study of postmenopausal women aged 50-79 years at baseline (1993-1998) in the Women's Health Initiative. We measured exposures as 2-day, indoor, lowest living-level average radon concentrations in picocuries per liter (pCi/L) as estimated in 1993 by the US Geological Survey and reviewed by the Association of American State Geologists under the Indoor Radon Abatement Act. We used Cox proportional hazards models to estimate risk of incident, neurologist-adjudicated stroke during follow-up through 2020 as a hazard ratio and 95% CI, adjusting for study design and participant demographic, social, behavioral, and clinical characteristics.

Results: Among 158,910 women without stroke at baseline (mean age 63.2 years; 83% white), 6,979 incident strokes were identified over follow-up (mean 13.4 years). Incidence rates were 333, 343, and 349 strokes per 100,000 woman-years at radon concentrations of <2, 2-4, and >4 pCi/L, respectively. Compared with women living at concentrations <2 pCi/L, those at 2-4 and >4 pCi/L had higher covariate-adjusted risks of incident stroke: hazard ratio (95% CI) 1.06 (0.99-1.13) and 1.14 (1.05-1.22). Using nonlinear spline functions to model radon, stroke risk was significantly elevated at concentrations ranging from 2 to 4 pCi/L (p = 0.0004), that is, below the United States Environmental Protection Agency Radon Action Level for mitigation (4 pCi/L). Associations were slightly stronger for ischemic (especially cardioembolic, small vessel occlusive, and large artery atherosclerotic) than hemorrhagic stroke, but otherwise robust in sensitivity analyses.

Discussion: Radon exposure is associated with moderately increased stroke risk among middle-aged and older women in the United States, suggesting that promulgation of a lower Radon Action Level may help reduce the domestic impact of cerebrovascular disease on public health.

Figure 1. WHI Participant Distribution and Stroke Hazard

(A) Distribution of Participants at the baseline visit, by EPA radon zone and US Census Region (WHI, 1993–1998). Of the 158,906 participants without missing zone (n = 4), 52,712 (33%), 62,029 (39%), and 44,165 (28%) lived in zones 3, 2, and 1, respectively. (B) Fully adjusted stroke hazard ratio (95% CI) as a spline function of the USGS radon index, centered at its median (9). p-log likelihood ratio = 0.0004. EPA = US Environmental Protection Agency; pCi/L = picocuries per liter; USGS = US Geological Survey; WHI = Women's Health Initiative.