The effect modification of PM2.5 and ozone on the short-term associations between temperature and mortality across the urban areas of Japan

Abstract

Background: The acute effects of temperature and air pollution on mortality are well-known environmental factors that have been receiving more recognition lately. However, the health effects resulting from the interaction of air pollution and temperature remain uncertain, particularly in cities with low levels of pollution. This study aims to examine the modification effects of particulate matter with a diameter of 2.5 µm or less (PM_2.5) and ozone (O₃) on the association between temperature and mortality.

Methods: We collected the daily number of all-cause, cardiovascular, and respiratory mortality from 20 major cities in Japan from 2012-2018. We obtained meteorological data from the Japan Meteorological Agency and air pollution data from the National Institute for Environmental Studies. We conducted analyses using a quasi-Poisson regression model with a distributed lag non-linear model for temperature in each city and subsequently performed a random-effects meta-analysis to derive average estimates.

Results: We found that high levels of O₃ might positively modify the mortality risk of heat exposure, especially for cardiovascular diseases. Subgroups such as the elderly and females were susceptible. We did not observe consistent evidence of effect modification by PM_2.5, including effect modification on cold by both pollutants.

Conclusion: PM_2.5 and O₃ may positively modify the short-term association between heat and mortality in the urban areas of Japan. These results highlight the need for public health policies and interventions to address the collective impacts of both temperature and air pollution.

Keywords: Air pollution; Effect modification; Mortality; Ozone; PM2.5; Temperature.

Fig. 1

Study locations.

Fig. 2

Overall pollution modified relative risk (RR) of temperature percentile and all-cause mortality. Left panel: effect modification from PM_2.5; Right panel: effect modification from ozone; Blue color: days with low concentration; Red color: days with high concentration; Shaded region: 95% confidence interval; Dotted lines: 1^st, 10^th, 90^th, and 99^th percentile of the temperature distribution.