You are here:
ASSOCIATIONS BETWEEN PARTICULATE MATTER COMPONENTS AND DAILY MORTALITY AND MORBIDITY IN PHILADELPHIA, PA
Citation:
Ito, K., M S. Landis, W Wilson, AND G. Thurston. ASSOCIATIONS BETWEEN PARTICULATE MATTER COMPONENTS AND DAILY MORTALITY AND MORBIDITY IN PHILADELPHIA, PA. Presented at 2003 AAAR PM Meeting, Pittsburgh, PA, March 31-April 4, 2003.
Impact/Purpose:
The analysis of ambient air samples by X-ray fluorescence is a critical component in the interpretation of complex air quality studies. The data products generated by the XRF laboratory are therefore critical to protecting human health and the environment. Analysis of particulate matter by XRF contributes to the following outcomes:
1) By characterizing emissions from different source types EPA is able to ensure timely attainment of the national ambient air quality regulations (NAAQS).
2) These analyses assist in the development of models for predicting pollutant concentrations from source emissions, apportioning sources from air concentration data, and for predicting exposures from different source types. The goal of the aforementioned models is to accurately understand risk and the ways to reduce risks to human health and the environment.
3) Datasets generated by the X-ray fluorescence laboratory cross-cut several laboratories (NERL, NHEERL, NRMRL, etc.) and represent an important collaboration effort. The XRF laboratory products are therefore crucial in the integration of varied, and sometimes wide ranging, agency goals. It is these cross-cutting relationships that enable ORD to collectively assess pollutants which pose the greatest risk to global, regional, state and local populations.
Description:
In evaluating the health risks from particulate matter (PM), the question remains as to which component(s) of PM are most harmful. We investigated this issue using PM mass, PM constituents, mortality, and the elderly hospital admission data in Philadelphia, PA. Daily paired PM2.5 and PM10 samples were collected at one site in downtown Philadelphia between May 1992 and September 1995. Trace elements as analyzed by energy dispersive X-ray fluorescence from PM2.5 filters (including Br, Ca, Fe, K, Mn, Ni, S, Se, Si, V, and Zn), gaseous pollutants (CO, NO2, SO2, and O3), and coefficient of haze (CoH) were also considered in the health effects analyses. Daily cardiovascular mortality, total (non-accidental) mortality, cardiovascular elderly (age over 65) hospital admissions, and respiratory elderly hospital admissions were aggregated for the Philadelphia Metropolitan Statistical Area. Generalized Linear Poisson regression Model (GLM) was used to estimate the excess health outcomes associated with PM components and gaseous pollutants adjusting for temporal trends, weather, day-of-week, and major holidays. Several alternative weather models, as well as varying extent of seasonal smoothing, were applied to examine the sensitivity of results to model specifications. Among the PM components examined, most consistent associations with health outcomes were observed for PM2.5, PM10 (70% of whose mass, on the average, was PM2.5), and sulfur. The PM2.5 (lag 1 day) relative risk estimate for cardiovascular mortality, for example, ranged from about 2 to 4% per 25 ug/m3 increase, depending on the weather model specification and seasonal trend smoothing (natural splines) between 4 to 12 degrees of freedom. PM10-2.5 was not significantly associated with any of the health outcomes. These results suggest the strongest PM associations are between regionally uniformly distributed secondary aerosol and health outcomes in Philadelphia during this time period.
This work has been wholly funded by the United States Environmental Protection Agency under cooperative agreement number CR827358 to New York University School of Medicine. It has been subjected to Agency Review and approved for publication.