Long-term Total and Source-specific PM2.5 in Relation to Incident Dementia in the U.S.

Abstract

Keywords: Air Pollution, Source-specific PM2.5, Dementia Background and Aims: Emerging, but sometimes inconsistent, associations have been observed between PM2.5 and incident dementia. One possible explanation for differing associations across studies could be differing toxicity by source. We investigated the associations of long-term exposures to total and source-specific PM2.5 with incident dementia in a national, population-based cohort. Methods We used biennial data between 1998 and 2016 from participants >50 years in the Health and Retirement Study. Incident dementia was identified using a validated algorithm based on cognitive testing and informant information. We predicted 10-year average total PM2.5 at participant residences before each interview using spatiotemporal modeling. Then, we derived source-specific PM2.5 by leveraging spatially refined fractions of PM2.5 from agriculture, traffic, energy, industry, open-fire and windblown dust as estimated by GEOS-Chem. Associations with incident dementia were estimated with Cox models, adjusting for individual demographics, area-level characteristics, time, and geographic trends. Results Among 27,857 participants, 4,105 (15%) developed dementia during follow-up. Mean PM2.5 levels were 11.5+3.7 μg/m3 overall, with the largest contributing individual sectors of agriculture (1.6+0.6) and the smallest of windblown dust (0.2+0.4). In single-source models, we observed a hazard ratio of 1.16 (95% CI: 1.05 to 1.29), 1.20 (1.01 to 1.41), 1,65 (0.98 to 2.77), 1.07 (0.94 to 1.22) and 1.07 (1.02 to 1.13) for incident dementia per 1-μg/m³ greater PM2.5 from agriculture, road traffic, non-road traffic, energy coal, and open fires, respectively. These associations remained after adjustment for the sum of PM2.5 from other sources. Associations corresponding to PM2.5 from other sources were inconsistent with adverse effects. Conclusions We found evidence that specific emission sources may be more important in the relationship between PM2.5 and dementia risk, suggesting the possibility of targeted interventions to reduce the burden of dementia in the future.