Longitudinal associations between air pollution and incident dementia as mediated by MRI-measured brain volumes in the UK Biobank

Background: Although there is increasing evidence that environmental exposures are associated with the risk of neurodegenerative conditions, there is still limited mechanistic evidence evaluating potential mediators in human populations.

Methods: UK Biobank is a large long-term study of 500,000 adults enrolled from 2006 to 2010 age 40-69 years. ICD-10 classified reports of dementia cases up to 2022 (Alzheimer's disease, vascular dementia, dementia in other classified diseases, and unspecified dementia) were identified from health record linkage. Estimates of residential air pollution, traffic noise, and greenspace exposure have been modelled. Structural brain MRI was conducted from 2014 to 2022, with brain volumes relevant to dementia identified a priori. Associations between environmental exposures, brain volumes, and dementia cases (diagnosed post-MRI) were tested using linear and logistic regression and adjusted for age, sex, household income, ethnicity, education, smoking, and area-level deprivation. Mediation of exposure-outcome associations by plausible brain volumes (those associated with both environmental exposure and dementia outcomes) were modelled using the quasi-Bayesian Monte Carlo method (N = 34,817-39,772).

Results: Small but significant mediating effects (2%-8% of relationships mediated) were observed between PM_2.5abs exposure and dementia risk by reduced total brain volume, NOx and Alzheimer's disease risk by reduced peripheral cortical grey matter, PM_2.5abs and vascular dementia risk by reduced peripheral cortical grey matter, PM_2.5abs and other dementia risk by reduced total grey matter, and PM₁₀ and other dementia risk by reduced total grey matter. Greenspace and noise were not associated with dementia outcomes in the subset of the cohort providing brain imaging data.

Conclusions: This study adds to existing evidence of associations between environmental exposures and dementia outcomes. Our findings provide novel evidence that differences in brain volume may mediate these relationships. Future research is required to prove this mechanism and establish the other mechanisms through which exposure to air pollution might increase dementia risk.