Effects of solar radiation exposure on ischemic heart disease mortality: country-level spatial regression models.

BACKGROUND: Previous studies have analyzed the association between sun exposure and ischemic heart disease (IHD). However, the association has not been assessed globally and may differ when adjusting for spatial dependency. This study aimed to clarify whether this global association remains even while incorporating spatial adjustment. METHODS: The most recent age-adjusted IHD mortality data (per 100,000) by country (1987-2022) were obtained from the World Health Organization (WHO) database as the dependent variable. As the independent variable, global solar radiation (GSR) data (MJ/m2/day; mean of 1994-2018) were retrieved from the Global Solar Atlas, with values clipped to each capital's location. Covariates included smoking prevalence, alcohol consumption, salt intake, gross domestic product, and health expenditure. To assess associations by sex, ordinary least squares (OLS) regression and three spatial regression models (spatial lag model, spatial error model, and spatial Durbin model) were applied. Additionally, an income-level stratified analysis was conducted. All analyses were performed with R version 4.5.0. RESULTS: After listwise deletion of missing values, 94 countries remained. The mean (SD) IHD mortality rates for males and females were 96.5 (80.4) and 52.4 (48.0), respectively. The mean (SD) GSR was 15.9 (3.7). In the OLS model, GSR showed a significant negative association with IHD mortality (males: β =  - 8.82, p = 0.002; females: β =  - 6.31, p < 0.001). The spatial lag model was the best fit for both sexes, and the association persisted (males: β =  - 4.78, p = 0.041; females: β =  - 3.86, p = 0.005). Stratified analysis largely supported these findings. However, coefficients substantially decreased after spatial adjustment. CONCLUSIONS: Sun exposure retained a significant inverse association with ischemic heart disease mortality after adjusting for spatial dependency, although adjustment markedly reduced the strength of association. However, our results require careful interpretation due to several limitations in the study.