Spatio-temporal variation of effects of short-term exposure to air pollutants on mortality in Seven Major Cities, Korea, 1998-2013

Abstract

Introduction: heterogeneity in effect measure of short-term exposure to air pollution has been one of challenging issues in environmental epidemiology. In this preliminary study, we aim to investigate systematic variation of this effect measures in sixteen years and seven different cities, Korea. Methods: ordinary time-series analysis was conducted for association between PM10, NO2, SO2, O3, CO and mortality in Seoul, Busan, Daegu, Incheon, Gwangju, Daejeon, Ulsan. Temperature, relative humidity, unobserved long-term trend and variation in population size was adjusted. To see systematic variation, for each city, dataset was split to a series of eight years long windows and a regression was serially conducted. Variation of relative risks per unit increment of air pollutants was recorded for each windows and city. Results: we found systematic variation rather than random in spatio-temporality of relative risks. In Incheon, all of relative risks for five air pollutants decreased over time. On the contrary, in Busan, a trend was opposite except for non-trend O3. In Seoul, relative risks for gaseous pollutants decreased over time while that for PM increased. For Daegu, Gwangju, Ulsan, relative risks for several pollutants increased over time while those for the others was stationary. For Daejeon, no trend was found. Conclusion: this preliminary study shows possibility of systematic variation of relative risks for short-term effects of air pollutants on mortality. Possible reasons may be changes in population dynamics such as aging population, potential interaction with meteorological factors, harvesting, measurement errors, and essence of PM (e.g. portion of harmful particles). A main research would shed light on delineation of that systematic variation.