Associations Between Trace Elemental Composition of PM2.5 and Mortality and Morbidity in Philadelphia

A large number of studies have reported associations between particulate matter (PM) and mortality and morbidity. Since PM is a chemically non-specific index and a mixture of a variety of chemical components from multiple sources, it is possible that use of the total PM mass as an index may cause attenuation of the observed effects, especially if only certain component(s) were responsible for the health effects. Thus, identifying a specific component of PM that is responsible for the health effects would be useful not only for shedding a light for mechanism of PM effects, but also for regulatory purpose to control specific sources.

This study examined thirteen individual trace elemental components of PM22.5 (analyzed by X-ray fluorescence) for their associations with total (non-accidental) mortality, cardiovascular mortality, elderly circulatory hospital admissions, and elderly respiratory hospital admissions in Philadelphia metropolitan statistical area during May 1992 - September 1995 study period. A Poisson generalized additive model was used to regress the health outcomes on each of the trace elements and PM2.5, adjusting for seasonal cycles, day-of-week, major holidays, and weather effects. Distributed lags of 0 through 6 days of the trace elements and PM2.5 were computed, and the sum of the coefficients and standard error (using variance/covariance matrix of lagged coefficients) were computed.

Among the thirteen trace elements, sulfur (S) was most consistently positively associated with the health outcomes, and was more significant than PM2.5. For example, the relative risks per 5th-to-95th percentile increment for S, using the sum of distributed lags, was 1.033 (95CI:0.999-1.069) and 1.079 (95CI: 1.018-1.145) for total mortality and elderly hospital admissions, Respectively. In contrast, the RRs for PM 2.5 for corresponding categories were 1.024(95CI: 0990?1.058) and 1.067 (95CI: 1.006-1.132).

These results suggest that the associations between PM2.5 and health outcomes in this locale may be driven by the secondary aerosols from fossil fuel combustion. This project has been funded by the United State Environmental Protection Agency under cooperative agreement number CR827358 to New York University School of Medicine.


Wilson Jr., W E. Associations Between Trace Elemental Composition of PM2.5 and Mortality and Morbidity in Philadelphia. Presented at Joint International Conference of the ISEE/ISEA, Vancouver, BC, Canada, August 11-15, 2002.