Proctor DM, Suh M, Tachovsky JA, Abraham L, Hixon JG, Brorby GP, Campleman SL. 2014. Cumulative risk assessment of urban air toxics: A pilot study in San Antonio, Texas. Presented at the Society of Toxicology’s 53rd Annual Meeting, Phoenix, AZ, March 23-27, 2014.
A pilot cumulative risk assessment project was conducted to test a modeling strategy for estimating the community health burden potentially associated with air toxics while simultaneously accounting for background air toxics, criteria air pollutants, and non-chemical stressors in an urban population of San Antonio, TX. Generalized additive models (GAMs) were used to quantitatively evaluate the potential association between airborne exposure to certain air toxics (lead, arsenic, antimony and chloromethane) generated by two coal-fired electricity generating units (EGUs), background levels of these air toxics, PM2.5, ozone, and social stressors on all cancer and heart disease mortality for 34 census tracts within the 10-mile radius of the EGUs. Four demographic and socioeconomic indicators (i.e., percent African American, percent female, age, and poverty) were significantly associated with all cancer and heart disease mortality, accounting for 70.6% and 79.1% of the variance, respectively. Accounting for these four demographic and socioeconomic indicators in the GAMs, concentrations of the four EGU-specific air toxics were not significantly associated with either all cancer or heart disease mortality in the study area. However, arsenic from non-EGU related on-road and non-road sources was significantly associated with all cancer mortality (88.8% of the residualized variance). PM2.5, lead from non-point sources, and ozone were significantly associated with heart disease mortality (88.1% of the residualized variance). With 17.9 and 21.27 degrees of freedom for cancer and heart disease, respectively, the GAMs were highly complex and non-linear. Although this pilot study has several limitations, it utilizes an innovative approach to quantitatively assess cumulative risk from non-chemical and chemical stressors and may serve as a preliminary model for future analysis.