Evaluating the effect of daily PM10 variation on mortality

Many epidemiological studies have shown the association between daily mortality and daily average of PM(10) level. Daily average of PM(10) has been used for regression type time-series analysis. This article suggests that daily mean mortality is not only a function of daily mean but also a function of daily standard deviation of PM(10). We used generalized additive Poisson regression model with overdispersion parameter to investigate this hypothesis using the data from Seoul, Korea, from 1997 to 2001. One interquartile range (IQR = 42.11 microg/m(3)) increase of daily mean level of PM(10) was found to be associated with 2.1% increase of additional daily mortality (RR = 1.021, 95% CI = 1.009-1.035) after controlling for other confounders. Similarly, we also found that one IQR (11.93 microg/m(3)) increase in daily standard deviation of PM(10) is associated with 2.5% increase of additional risk of death (RR = 1.025, 95% CI = 1.000-1.028) when other covariates remained the same. These findings may provide new insight into the possible explanation of health effect of PM(10) and support the hypothesis that PM(10) deviation is also an important risk factor after controlling for daily mean PM(10) level.