Temporal changes in morality attributed to heat extremes for 57 cities in Northeast Asia.

Recent studies have reported that heat-related mortality decreased by adaptation during decades. However, since the frequency of extreme heat events is increasing, it is difficult to conclude with certainty that the heat mortality burden is decreasing. To examine temporal changes in mortality attributed to heat extremes in Northeast Asia, we collected temperature and mortality data covering the years 1972-2012 from 57 cities of 3 countries (Taiwan, Korea, and Japan) in Northeast Asia. Poisson regression curves were fitted to the data from each city. The temporal changes in heat-mortality association were estimated with a time-varying distributed lag non-linear model. Heat extremes were defined as temperatures greater than the 97.5th percentiles of city-specific average temperatures. Attributable deaths were calculated considering temporal variations in exposure and relative risk. The estimates were then pooled through meta-analysis. The results show that the mortality risk on extreme heat days declined during the study period in all countries. However, as summer temperatures in Japan have shown more heat extremes over time, the mortality risk attributed to heat increased during 2003-2012 (0.32%) compared with 1972-1981 (0.19%). Thus, to assess the total health burden due to heat extremes related to climate change, public health strategies should focus on the temporal variation in heat-mortality association as well as changes in the distribution of heat extremes overtime.