Mortality Risk of Hot Nights: A Nationwide Population-Based Retrospective Study in Japan

The health effects of heat are well documented; however, limited information is available regarding the health risks of hot nights. Hot nights have become more common, increasing at a faster rate than hot days, making it urgent to understand the characteristics of the hot night risk.

We estimated the effects of hot nights on the cause- and location-specific mortality in a nationwide assessment over 43 y (1973-2015) using a unified analytical framework in the 47 prefectures of Japan.

Hot nights were defined as days with a) minimum temperature $$ ($$) and b) minimum temperature $$ percentile ($$) for the prefecture. We conducted a time-series analysis using a two-stage approach during the hot night occurrence season (April-November). For each prefecture, we estimated associations between hot nights and mortality controlling for potential confounders including daily mean temperature. We then used a random-effects meta-analytic model to estimate the pooled cumulative association.

Overall, 24,721,226 deaths were included in this study. Nationally, all-cause mortality increased by 9%-10% [$$ relative risk $$, 95% confidence interval (CI): 1.08, 1.10; $$ $$, 95% CI: 1.09, 1.11] during hot nights in comparison with nonhot nights. All 11 cause-specific mortalities were strongly associated with hot nights, and the corresponding associations appeared to be acute and lasted a few weeks, depending on the cause of death. The strength of the association between hot nights and mortality varied among prefectures. We found a higher mortality risk from hot nights in early summer in comparison with the late summer in all regions.

Our findings support the evidence of mortality impacts from hot nights in excess of that explicable by daily mean temperature and have implications useful for establishing public health policy and research efforts estimating the health effects of climate change. https://doi.org/10.1289/EHP11444.