Fine Particle Sources and Cardiorespiratory Morbidity: An Application of Chemical Mass Balance and Factor Analytical Source-Apportionment Methods

Abstract

Abstract Background Interest in the health effects of particulate matter (PM) has focused on identifying sources of PM, including biomass burning, power plants, and gasoline and diesel emissions that may be associated with adverse health risks. Few epidemiologic studies, however, have included source-apportionment estimates in their examinations of PM health effects. We analyzed a time-series of chemically speciated PM measurements in Atlanta, Georgia, and conducted an epidemiologic analysis using data from three distinct source-apportionment methods. Objective The key objective of this analysis was to compare epidemiologic findings generated using both factor analysis and mass balance source-apportionment methods. Methods We analyzed data collected between November 1998 and December 2002 using positive-matrix factorization (PMF), modified chemical mass balance (CMB-LGO), and a tracer approach. Emergency department (ED) visits for a combined cardiovascular (CVD) and respiratory disease (RD) group were assessed as end points. We estimated the risk ratio (RR) associated with same day PM concentrations using Poisson generalized linear models. Results There were significant, positive associations between same-day PM2.5 (PM with aero-dynamic diameter ≤ 2.5 μm) concentrations attributed to mobile sources (RR range, 1.018–1.025) and biomass combustion, primarily prescribed forest burning and residential wood combustion, (RR range, 1.024–1.033) source categories and CVD-related ED visits. Associations between the source categories and RD visits were not significant for all models except sulfate-rich secondary PM2.5 (RR range, 1.012–1.020). Generally, the epidemiologic results were robust to the selection of source-apportionment method, with strong agreement between the RR estimates from the PMF and CMB-LGO models, as well as with results from models using single-species tracers as surrogates of the source-apportioned PM2.5 values. Conclusions Despite differences among the source-apportionment methods, these findings suggest that modeled source-apportioned data can produce robust estimates of acute health risk. In Atlanta, there were consistent associations across methods between PM2.5 from mobile sources and biomass burning with both cardiovascular and respiratory ED visits, and between sulfate-rich secondary PM2.5 with respiratory visits.