Climate Futures and Projected Mortality Due To Non-Optimal Temperature From 2020 To 2100: A Global Burden of Disease Forecasting Study

Abstract

Background: A recent study estimated that 1·69 million (95% uncertainty interval [UI] 1·52–1·83) deaths were attributable to non-optimal temperature in 2019. Globally, most of this burden was attributed to low temperatures, but high heat-attributable mortality rates were observed in warmer regions such as south Asia, north Africa and the Middle East, and sub-Saharan Africa. The impact of increases in global temperatures and demographic trends on future temperature-related mortality is unclear. Methods: We linked cause-specific temperature-mortality risk curves to daily temperature forecasts from the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset. The NEX-GDDP dataset comprises 21 individual General Circulation Models (GCM) that were spatially downscaled for two Representative Concentration Pathways (RCPs): RCP4.5 and RCP8.5. For each pixel, day, and cause, we estimated the population attributable fraction (PAF) based on the risk of cause-specific mortality at a given temperature relative to the theoretical minimum risk exposure level (TMREL); we then estimated temperature-attributable deaths by multiplying these PAFs by location and cause-specific mortality estimates for the 1990 to 2020 reference period and by forecasted mortality estimates for the period from 2020 to 2100. Other upstream drivers in the mortality forecast model, including the Socio-demographic Index and air pollution levels, were held constant. We additionally estimated PAFs for a temperature scenario based on the 2015 Paris Agreement, assuming that countries meet their nationally determined contributions and continue to reduce emissions at the same rate after 2030. Findings: Globally, we estimate increases in the number and rate of deaths attributable to non-optimal temperature, with 2·26 million (1·66–2·94), 2·54 million (1·86–3·31) and 3·17 million (2·11–4·35) annual deaths by the end of the century, for the Paris scenario, and the RCP4.5 and RCP8.5, respectively. These increases are driven largely by heat-attributable burden, with smaller and more variable changes in cold-attributable deaths. Warm, low-latitude regions will be disproportionally affected, with the largest increases in temperature-attributable mortality expected in sub-Saharan Africa, south Asia, and Latin America and the Caribbean. Heat-attributable mortality was reduced by half for the Paris scenario relative to the RCP8.5 with 1·02 million (0·64–1·48) deaths versus 2·08 million (1·13–3·13) deaths in 2100. As our study accounts for adaptation and does not extrapolate relative risks beyond the exposures currently experienced, these estimates are likely conservative. Interpretation: Globally increasing temperatures are projected to substantially increase the mortality burden attributable to non-optimal temperature. Higher global temperatures only lead to minor decreases in cold-related mortality which do not offset increases in heat-related mortality. Funding: Bill & Melinda Gates Foundation Declaration of Interest: The authors declare that there are no conflicts of interest.