Health Impact Assessment of PM2.5 in Japan

Abstract

P-077 Introduction: Among industrialized countries, Japan still maintains an old set of guidelines for particulate matter (PM), which were established in 1973. Although Japanese agency has set a guideline for daily average concentrations of Suspended Particulate Matter (particles<10μm), it has not yet set guidelines for PM2.5 (particles<2.5μm) and for annual average concentrations of both particles. Therefore, we estimated the attributable number of cases caused by PM2.5 short- and long-term exposure above the defined reference level in 2002. Methods: Subjects were 7,825,745 men and women older than 30 years old (national statistics, 2000), in one metropolitan area, where population density is the highest in Japan. Our model assumed a linear relationship between exposure and health effects, and its slope was an exposure-response function. In the model, five data components were required: definitions of health outcome, exposure-response function, reference level, population exposure distribution and outcome frequency. Primary health outcomes were all causes of mortality. In the case of PM2.5 long-term exposure, we added cardiopulmonary and lung cancer deaths. As an exposure-response function, we used relative risk per 10μg/m3, which we extrapolated from the WHO meta-analysis for PM2.5 short-term, and from US large cohort study for PM2.5 long-term exposure. We used two reference levels: US EPA's guideline for PM2.5 (65μg/m3 for 24-hour concentration, 15μg/m3 for annual concentration) and the reference level (30μg/m3 for 24-hour concentration, 12μg/m3 for annual concentration) which was established by multiplying the PM10 targeted value in EU in 2010 by 0.6. We constructed the population exposure distribution for PM2.5 short- and long-term exposure from daily average concentrations in 2002, which the environmental public agency published. For final component, outcome frequency, we extracted data from vital statistics in 2002. Results: In 2002, 83,798 people had died, including 36,042 cardio-pulmonary and 4899 lung cancer deaths. Annual PM2.5 concentration was 26.4μg/m3. If we used the EPA's guideline as the reference level, the attributable number of cases was 9 (7–12) due to short-term exposure and 5364 (1788–9835) deaths, 3354 (1118–5962) cardio-pulmonary deaths, and 674 (193–1108) lung cancer deaths due to long-term exposure. Furthermore, if we used another established reference level, the same attributable number of cases was 441 (323–559), 6664 (2221–12,217), 4136 (1379–7353), and 822 (235–1350), respectively. Conclusion: Even at the recent PM2.5 exposure level, the number of deaths will occur after both short- and long-term exposure. This assessment shows that guidelines for PM2.5, especially for long-term exposure, should be recommended in Japan.