Introduction: While several studies report estimates of heat or cold-mortality relationships, few have directly compared the burden attributable to different temperature components. Methods: Two-stage time series analysis of 384 locations in 13 countries (Australia, Brazil, Canada, China, Italy, Japan, South Korea, Spain, Sweden, Taiwan, Thailand, UK, USA) in 1985-2012. Temperature-mortality relationships were estimated with a distributed lag non-linear model with 21 days of lag, and then pooled in a multivariate meta-regression. Attributable deaths were calculated for heat and cold, defined as temperatures above and below the minimum mortality temperature, and for moderate and extreme temperatures, defined using cut-offs at the 2.5th and 97.5th percentiles. Results: Cold is accountable for a much higher mortality impact if compared to heat, with estimates of attributable fraction of 7•29% (95%CI: 7•03-7•49%) and 0•42% (0•39-0•44%) overall, respectively. Results display a substantial inter-country variation: cold-related deaths are lower in tropical countries such as Thailand (2.61%) or Brazil (2.83%) and surprisingly in cold temperate countries such as Canada (4.46%), while heat-attributable deaths do not show a clear pattern. Only 0.63% (0.61-0.64%) and 0.23% (0.22-0.24%) of deaths are attributable to extreme cold and heat, respectively. Conclusions: The analysis of the largest dataset ever collected for assessing temperature-mortality associations, including over 74 million deaths, indicates that cold is consistently accountable for a much higher proportion of deaths. In addition, most of the attributable deaths are actually due to moderate temperatures. These findings are important for the development of strategies to reduce the health impact, and may have implications in estimating the burden of climate change. Reference Mortality risk attributable to high and low ambient temperature: a multi-country study. The Lancet. 2015 (accepted for publication).
Read full abstract