Environment International · Feb 24, 2026

Racial and ethnic differences in the impact of air pollution on the risk of Alzheimer's disease and related dementias in the Multiethnic Cohort Study

By Wu AH, Wu J, Tseng C et al.

BACKGROUND AND OBJECTIVES: Meta-analysis results, based largely among Whites, suggested that fine particulate matter (PM2.5) exposure increases the risk of clinical dementia. This study investigated the association of air pollution and incidence of Alzheimer's disease and related dementias (ADRD) by race and ethnicity. METHODS: We investigated incidence of AD (n = 4,010) and other dementia (n = 4,971) among 44,954 California Multiethnic Cohort (MEC) participants (28% African American, 14% Japanese American, 44% Latino, 14% White adults) who were enrolled in the fee-for-service component of Medicare (2001-2016). We used Cox proportional hazards regression to examine associations between exposure to PM, airport-related ultrafine particles (aUFP) and gaseous pollutants and incidence of AD, other dementia, and ADRD in a minimally- and fully-adjusted model, considering 12 established ADRD risk factors. We conducted stratified analyses to examine associations by sex, and race/ethnicity. RESULTS: ADRD incidence was associated with PM2.5 (per 2 µg/m³), airport-related UFP (aUFP, per 4400 particles/cm³) and nitrogen dioxide (NO2, per 10 µg/m³) with hazard ratios (HRs, 95%CI), respectively, of 1.04 (1.02-1.06), 1.03 (1.01-1.05) and 1.09 (1.06-1.12). The AD-associations with PM2.5 and NO2, were stronger than the corresponding associations with other dementia (Pheterogeneity ≤ 0.003). Similar patterns of results were observed by sex and across race and ethnicity. Statistically significant findings for ADRD with PM2.5, aUFP and NO₂ were observed among African American (respective HRs 1.03, 1.04, 1.09), and Latino and White participants for NO₂ (HR 1.10, 1.08). Results in all and African American participants remained statistically significant in fully-adjusted models. Although the effect of PM2.5 was diluted in a co-pollutant with NO₂, both PM2.5 and aUFP were significantly associated with ADRD incidence in a co-pollutant model, and NO₂ and aUFP (but not PM2.5) remained associated in a multipollutant model. We did not observe consistent modifying effects for any of the 12 established ADRD risk factors. CONCLUSIONS: In this multiethnic population, incidence of ADRD increased with exposures to PM2.5, aUFP, and NO₂ in all subjects and this pattern was most prominent among African American adults. These results emphasize that ADRD prevention should include not only individual-level factors but also population-wide policies and regulation to curb air pollution.

doi.org/10.1016/j.envint.2026.110169Open accessPubMed 41793956Populations