Short-term effects of daily air pollution on mortality

The daily variations of air pollutants in the Klang Valley, Malaysia, which includes Kuala Lumpur were investigated for its association with mortality counts using time series analysis. This study located in the tropic with much less seasonal variation than typically seen in more temperate climates. Data on daily mortality for the Klang Valley (2000–2006), daily mean concentrations of air pollutants of PM10, SO2, CO, NO2, O3, daily maximum O3 and meteorological conditions were obtained from Malaysian Department of Environment. We examined the association between pollutants and daily mortality using Poisson regression while controlling for time trends and meteorological factors. Effects of the pollutants (Relative Risk, RR) on current-day (lag 0) mortality to seven previous days (lag 7) and the effects of the pollutants from the first two days (lag 01) to the first eight days (lag 07) were determined. We found significant associations in the single-pollutant model for PM10 and the daily mean O3 with natural mortality. For the daily mean O3, the highest association was at lag 05 (RR = 1.0215, 95% CI = 1.0013–1.0202). CO was found not significantly associated with natural mortality, however the RR's of CO were found to be consistently higher than PM10. In spite of significant results of PM10, the magnitude of RR's of PM10 was not important for natural mortality in comparison with either daily mean O3 or CO. There is an association between daily mean O3 and natural mortality in a two-pollutants model after adjusting for PM10. Most pollutants except SO2, were significantly associated with respiratory mortality in a single pollutant model. Daily mean O3 is also important for respiratory mortality, with over 10% of mortality associated with every IQR increased. These findings are noteworthy because seasonal confounding is unlikely in this relatively stable climate, by contrast with more temperate regions.

► We model the association of air pollution and mortality using Poisson regression.
► RR for each pollutant was obtained for every IQR increase at different lag times.
► O3, CO and PM10 were the important pollutants associated with natural mortality.
► All pollutants except SO2 were associated with respiratory mortality.
► Those who have respiratory diseases are at higher risk of mortality.