BACKGROUND: Climate change is suspected to cause adverse health effects, and increased ozone concentration is one of the proposed pathways. We examined the mediation of ozone on the association between temperature and daily mortality and estimated excess mortality due to climate change.

METHODS: Daily mean temperature, 8-hour maximum ozone concentration, and daily number of non-accidental deaths from 7 metropolitan cities in Korea (Seoul, Busan, Daegu, Incheon, Daejeon, Gwangju, and Ulsan) between January 1, 2006 and December 31, 2019 were analyzed. A mediation analysis using a linear regression model for temperature and ozone and a Poisson regression model for temperature and mortality adjusting for ozone was conducted on days with temperature higher than or lower than city specific minimum mortality temperature. We calculated excess mortality due to direct and indirect effects of daily temperature exceeding average daily temperature from 1960 to 1990.

RESULTS: The daily mean temperature from 2006 to the end of 2019 was 1.15 ± 2.94 °C higher than the average daily temperature from 1960 to 1990. The pooled relative risk (for a 1 °C increment) of indirect effects through increased ozone were 1.0002 [95% confidence interval (CI): 0.9999, 1.0004] and 1.0003 (95% CI: 1.0002, 1.0005) in days with higher than or lower than minimum mortality temperature, respectively. The numbers of excess deaths during the study period were 2072.5 (95% CI: 1957.1, 2186.5) due to direct effects in days with higher than minimal mortality temperature, and 94.6 (95% CI: 84.3, 101.7) and 268.5 (95% CI: 258.4, 289.1) due to indirect effects in days with higher than and lower than minimal mortality temperature, respectively.

CONCLUSION: We observed a mediating effect of ozone between temperature and daily mortality. There has been excess deaths due direct effect of temperature and indirect effects through ozone.