High ground level ozone amplifies heat-related illnesses in Japan: a nationally representative analysis

Abstract Background . Heat-related illnesses, ranging from heat cramps to life-threatening heatstrokes, are caused by thermal stress. Several previous studies have examined their interactions with air pollutants, however most of these studies focused in morbidity or mortality, cardiovascular disease or respiratory disease, and not heat-related illnesses. We investigated whether PM 2.5 and O 3 modify the association between temperature and heat-related illness. Methods . Data on heat-related illness cases were obtained from ambulance transport records across 47 prefectures during the summer months of 2014–2019. PM 2.5 values measured in each prefecture were used, whereas for O 3 , model-estimated concentrations weighted by population were employed to account for within-prefecture heterogeneity. We applied quasi-Poisson regression with distributed lag nonlinear models for each prefecture to quantify the association between the daily mean temperature and heat-related illness. The interaction term between temperature and the binary variable air pollution (high ≥ 75th percentile, low < 75th percentile) was included. The ratio of relative risks (RRR) between higher and lower air pollution was also estimated to determine the presence or absence of an effect modification. Prefecture-specific estimates were pooled using a meta-analysis framework with a random-effects model. Results . The association between temperature and heat-related illnesses was non-linear, with the risk increasing steeply at higher temperatures. The risk of heat-related illnesses at the same temperature was higher at higher O 3 concentrations, but not in PM 2.5 . The RRRs of heat-related illness at the 90th percentile temperature with confidence interval (CI) were 0.86 (95% CI: 0.63, 1.17) for PM 2.5 and 1.56 (95% CI: 1.22, 1.99) for O 3 , respectively. Conclusion . This multi-prefecture study highlights the amplifying role of O 3 in the relationship between high ambient temperatures and heat-related illnesses during Japan’s summer months, although there was no effect modification by PM 2.5.