The association of air pollution and mortality: examining the case for inference.

An association between air pollution measured as particulate matter, and mortality has been reported in several different locations. These studies have been conducted over a wide range of climates and populations. The time-series studies, which examine the joint occurrence of daily fluctuations in air pollution and mortality, provide the strongest evidence of a true association. However, several criteria, including the consistency of the results, need to be explored before causality is inferred from these studies. A striking consistency in the results was observed, after the different studies were converted into a common metric. The mean effect of an 10 micrograms/m3 change in PM10 implied by these studies varies between 0.64 and 1.49%. The fulfillment of other criteria, including specificity, presence of a dose-response relationship, and coherence of results, lend strong support to the existence of an actual association between particulate matter and mortality. However, the biologic mechanism is not well understood at this time. In addition, the precise measure of the pollutant responsible for the health effect--total suspended particles, PM10, fine particles, sulfates, acidic aerosols, sulfur dioxide, or some as yet unmeasured pollutant--is unclear, based on current available evidence.