|
An association between ozone and daily mortality is controversal and was reported to depend on the proper controls for temperature. To evaluate the relative importance of ozone as a predictor of daily mortality in Seoul, South Korea, the association between total daily mortality and ozone was investigated for the period from January, 1995 to December, 1998. Poisson regression was used to regress daily death counts on ozone. The model included loess smooth function on the time and meteorological factors to capture the seasonal/long-term time trend and any possible nonlinear relationship between the mortality and meterological factors. Ozone exposure was statistically significant with regard to mortality and meteorological factors. Ozone exposure was statistically significant with regard to mortality in the above models, and graphical analysis showed that the relationship was nonlinear. The risk of all-cause mortality associated with a 20.125ppb(IQR) increment in 1-hr maximum ozone was estimated to increase by 2.4%[95%confidence interval(CI), 1.4-3.1] after adjusting for meteorological factors and nonlinearity, 2.0%(CI, 1.2-2.8) when ozone is beyond 30ppb and 1.4%(CI, 0.2-2.6) in summer. Ozone effect on mortality was not statistically significant in winter.
|