Ambient air pollution and stroke in two population-based cohorts: roles of systemic inflammation and polygenic risk

Long-term exposure to ambient air pollution has been associated with an increased risk of stroke, but whether this association differs according to inflammatory burden and genetic susceptibility remains unclear.

We conducted separate analyses in two population-based cohorts. In the 2015 cross-sectional wave of the China Health and Retirement Longitudinal Study (CHARLS; n = 21,095), multivariable logistic regression was used to examine associations between ambient air pollution and prevalent self-reported physician-diagnosed stroke, with additional analyses evaluating effect modification by high-sensitivity C-reactive protein. In the prospective UK Biobank cohort (n = 368,644), recruited between 2006 and 2010, Cox proportional hazards models were used to examine associations between residential air pollution and incident ischemic stroke identified from linked health records using ICD-10 code I63, with additional analyses according to polygenic risk score categories.

In CHARLS, increased exposure to PM10 and PM2.5 was associated with higher odds of prevalent stroke (OR 1.31, 95% CI 1.08-1.59 for PM10; OR 1.26, 95% CI 1.04-1.53 for PM2.5). Significant positive multiplicative interaction terms between hs-CRP and several pollutants were observed in relation to prevalent stroke. In UK Biobank, every 5 µg/m³ increment in PM2.5 was associated with a higher hazard of incident ischemic stroke (HR 1.29, 95% CI 1.15-1.45). Participants with both high genetic susceptibility and high PM2.5 exposure had the highest hazard ratio for incident ischemic stroke (HR 1.44, 95% CI 1.32-1.57).

Ambient air pollution was associated with prevalent stroke in the cross-sectional CHARLS analysis and with incident ischemic stroke in the prospective UK Biobank analysis. Higher inflammatory burden in CHARLS and joint exposure to higher polygenic risk and air pollution in UK Biobank identified potentially susceptible population groups, although the PRS analyses did not formally evaluate interaction. Because the cohorts differed in calendar period, study design, exposure assessment, and stroke ascertainment, the cohort-specific findings should be interpreted as complementary rather than directly quantitatively comparable evidence.