Associations of Temperature Variation and Mortality in 47 Japanese Prefectures

Abstract

Background / Aim: Few studies have investigated the mortality effects of temperature variation (TV) in Japan. In this study, we employed the newly proposed indicators of inter- and intra-day TVs computed separately using daily mean temperatures above or below the minimum mortality temperature to assess the TV-mortality associations and their relative contribution in comparison to daily mean temperature across Japan.Methods: We collected daily data on temperature and mortality during 1972-2012 from 47 Japanese prefectures. We performed a quasi-Poisson regression analysis incorporating a distributed-lag non-linear model to estimate associations of daily mean temperature, inter- and intra-day TVs with mortality in each prefecture.Results: Most prefectures showed a significant increase in mortality risk associated with daily mean temperature, with relative risks (RRs) reaching 1.236 (95% confidence interval (CI): 1.163, 1.313) for heat in Aomori, and 1.676 (95% CI: 1.313, 2.138) for cold in Wakayama. Inconsistent results for inter- / intra-day TVs were found, except for some protective associations between intra-day TV and mortality in Kochi (RR 0.979; 95% CI: 0.965, 0.993) and Gifu (RR 0.988; 95% CI: 0.98, 0.996) on cold days.Conclusion: We found that the association between mortality and TV was generally small compared with daily mean temperature in Japan.Acknowledgement: this research was supported by the Environmental Research and Technology Development Fund (S-10 & S-14) of the Ministry of the Environment, Japan.