Total effect of heat on mortality considering heat-mediated air pollution and interaction effect under demographic,...

Health impact assessments of high ambient temperature (HAT)-related mortality have typically been assessed through direct thermal effects alone, overlooking critical pathways. Here, we conceptualize HAT impacts as a multifaceted process that includes both direct effects and indirect effects mediated through HAT-induced increases in ambient air pollution, notably ozone, and their interactions. We developed a unified epidemiologic framework of mediation, interaction, and exposure measurement error, and estimated the total effect (TE) of HAT on mortality. Using national mortality and environmental data from South Korea, we projected HAT-related mortality from 2025 to 2100 under demographic and mitigation/adaptation scenarios, which includes the direct effect of HAT and indirect effect of HAT (through HAT-mediated ozone) and the interaction between HAT and HAT-mediated ozone. Across all scenarios, the interaction between HAT and HAT-mediated ozone contributed to a substantial fraction of HAT-related mortality. Under a low mitigation/adaptation scenario, SSP5-RCP8.5, HAT is projected to cause 855,202 deaths (95% CI: 518,447-1,224,610), of which 481,214 deaths (126,493-884,528) are attributable to the interaction between HAT and HAT-mediated ozone, while 335,363 deaths (186,224-522,933) are attributable to the direct effect of HAT alone. The total deaths could be reduced to 389,446 deaths (229,335-563,693) under a high mitigation/adaptation scenario, SSP1-RCP2.6, of which 221,025 deaths (58,078-406,881) are attributable to the interaction, while 163,744 deaths (85,012-241,821) are attributable to the direct effect of HAT. Our findings imply that heat-related mortality may be higher than previously reported in the literature, highlighting the need for effective mitigation and adaptation strategies to reduce the TE of HAT.