Abstract Rationale Outdoor fine particulate air pollution (PM2.5) affects health, but most studies emphasize total mass, not composition. Spatial differences in PM2.5 components may bias health risk estimates for total PM2.5. Associations between PM2.5 composition and other social determinants of health further complicate the issue. We examined whether specific components (black carbon, BC; organic carbon, OC; nitrate, NO3; sulfate, SO4) are differentially associated with pediatric asthma emergency department (ED) visits. Methods We analyzed census tract-level pediatric (age 18) asthma-related ED visits (2016-2017) in Dallas, Houston, San Antonio, Austin, and El Paso. Exposures were annual averages (IQR-scaled) for PM2.5 and components. Incidence risk ratios (IRR) were estimated using single-pollutant quasi-Poisson models including city fixed effects. Sensitivity models adjusted for percentage of the census tract living in poverty and racial/ethnic composition (share Black or Latinx) of the census tract. We also summarized pairwise correlations between each pollutant and tract-level SES and demographic characteristics. Results Adjusting for city, we found strong positive associations for all exposures, with individual components showing higher risk per IQR than bulk mass: BC = 2.94 2.38-3.64, OC = 1.37 1.28-1.46, NO3 = 4.42 3.16-6.20, SO4 = 3.54 2.69-4.66, PM2.5 IRR = 1.20 95% CI:1.16-1.24. Adjustment for poverty alone attenuated estimates (BC = 1.00 0.83-1.20, OC = 1.02 0.97-1.08, NO3 = 1.24 0.93-1.66, SO4 = 1.34 1.06-1.69, PM2.5 = 1.02 0.99-1.05). Full adjustment for poverty and racial/ethnic composition further reduced associations to near-null (PM2.5 = 1.01 0.98-1.04, OC = 0.99 0.94-1.04, BC = 0.95 0.80-1.15, SO4 = 1.03 0.82-1.29, NO3 = 1.20 0.91-1.59). Pollutant-SES correlations were modest overall, with a PM2.5-poverty r = 0.24 and component correlations ranging from r = -0.02 (SO4) to 0.23 (BC), as were pollutant-race/ethnicity correlations. However, correlation varied significantly across cities (e.g., SO4-poverty range: 0.05 - 0.54; OC-share Black range: -0.10 - 0.13), reinforcing both the potential for confounding and city-specific exposure-context coupling. Conclusions Component-specific heterogeneity was evident before adjustment, but attenuation of component-specific estimates after SES and racial/ethnic covariate control indicates strong confounding by neighborhood context. Studies of PM2.5 composition and pediatric asthma should account for these structural factors, which reflect environmental injustice and differential siting of traffic and industrial sources. This abstract is funded by: NIH
Chambliss et al. (Fri,) studied this question.