Mortality risk and long-term exposure to ultrafine particles and primary fine particle components in a national U.S. Cohort.

Abstract

The objective of this study was to estimate all-cause, cardiopulmonary, and cancer mortality associations for long-term exposure to ultrafine particles (UFP) and primary PM2.5 components. We utilized high-resolution, national-scale exposure estimates for UFP (measured as particle number concentration; PNC) and three primary PM2.5 components, namely black carbon (BC), traffic-emitted organic PM2.5 (hereafter, hydrocarbon-like organic aerosols; HOA), and cooking-emitted organic PM2.5 (cooking organic aerosols; COA). Two analytic cohorts were constructed from a nationally representative U.S. health survey. The larger cohort consisted of 617,997 adults with information on a broad set of individual-level risk factors; the smaller cohort was further restricted to those with information on physical activity (n=396,470). In single-pollutant models, PNC was significantly associated with all-cause (larger cohort HR=1.03, 95% CI [1.02, 1.04]; smaller cohort HR=1.02, 95% CI [1.00, 1.04]) and cancer mortality (larger cohort HR=1.05, 95% CI [1.02, 1.08]; smaller cohort HR=1.06, 95% CI [1.02, 1.10]). In two-pollutant models, mortality associations varied based on co-pollutant adjustment; PNC mortality associations were generally robust to controlling for PM10-2.5 and SO2, but not PM2.5. In contrast, we found some evidence that the HOA and COA mortality associations are independent of total PM2.5 mass exposure. Nevertheless, PM2.5 mass was the most robust predictor of air pollution related mortality, providing some support for current regulatory policies.