Temporal change in minimum mortality temperature from 1969-2018: a multi-country multi-city study

Abstract

Background: Minimum mortality temperature (MMT) is an important indicator to assess the temperature-mortality association, indicating adaptation to climate. Limited evidence is available about the temporal change of the MMT and its determinants.Methods: We collected data for temperature and mortality from 478 cities in 20 countries from 1969-2018. We estimated time-varying temperature-mortality associations for each community, using a generalized linear model with Poisson distribution. We pooled the time-varying associations by country to derive the country-specific time-varying MMT using multivariate meta-regression and Monte Carlo simulation. We assessed the temporal trend of MMT and the effects of community-level indicators on the trend using random-effects meta-regression.Results: The temporal change of MMT varied by country, with some countries showing increasing trend (South Korea, Japan, Taiwan, UK, Netherlands, US) while others showed decreasing trend (Spain, Portugal, Romania, Czech, Estonia, Canada, Australia) or roughly constant trend (German, Swiss, Sweden, Norway, Finland, Ireland, Kuwait) over the last decades. Globally, the MMT has increased (p-value<0.01 for a linear trend). The country-specific MMT estimate ranged from 13.4 (95%CI: 7.9, 19.3) to 30.5 (12.2, 41.0) oC in 2000 while it ranged from 17.3 (13.2, 23.4) to 35.5 (25.4, 43.7) in 2010. Lower latitude (i.e., closer to the equator) and higher annual average temperature were associated with larger positive slope of linear trend of MMT (p-values<0.01). Conclusions: The temporal change of MMT varied by country. Population adaptation under climate change depends on country-specific factors and climate conditions. Future prediction of MMT should account for the country-specific and climate-dependent adaptation.