Exposure to air pollution and incident dementia and modification by APOE gene in the UK Biobank

Abstract

Background/Aims Air pollution may cause inflammatory and oxidative stress damage to the brain, although the association between air pollution and specific types of dementia, and possible genetic modification by APOE status, is relatively underexplored. We examined the association between air pollution (PM2.5, PM10, PM2.5-10, PM2.5abs, NO2, NOX) and incident dementias (Alzheimer’s disease (AD), fronto-temporal dementia (FTD), and vascular dementia (VAD)) and modification by APOE status.Methods The UK Biobank repository followed >500,000 participants from enrollment through 2017. After restricting to individuals aged > 60 years, with no dementia diagnosis prior to 2010, our sample included 187,165 individuals (N= 680 incident AD, N=incident VAD, N= 63 incident FTD). Annual averages of air pollutant estimates were estimated from ESCAPE land use regression methods for 2010 and were scaled to interquartile ranges (IQR). Dementia was identified from NHS, hospital, and survey records. Time to incident dementia, and follow-up time was reported from baseline (January 1, 2010) to last censor event (death, last hospitalization, or loss to follow-up). Cox proportional hazard ratios (HR) were calculated to estimate the association of air particulates with risk for dementia.Results We observed strong and consistent associations of PM2.5 exposure with all measures of dementia except FTD. These include any incident dementia (HR for 1 IQR increase=1.20, 95%CI: 1.13, 1.27), incident AD (HR=1.20, 95% CI 1.09, 1.32), and with incident VAD (HR=1.15, 95% CI: 1.00, 1.31). NO2 was also associated with any incident dementia (HR=1.21, 95% CI 1.13, 1.29), AD (HR=1.19; 95% CI 1.07, 1.32) and VAD (HR=1.21; 95% CI 1.06, 1.39). No associations were observed for any air pollutants with FTD. APOE did not modify the association between any air particulates and dementia risk.Conclusions PM2.5 and NO2 may be associated with several types of dementia, and these associations are not modified by APOE in this cohort.