Abstract Body: Introduction. Ambient air pollution is linked to increased stroke risk, but it is unclear if associations apply to long-term exposures, equally to different stroke subtypes (ischemic versus hemorrhagic) and differ by pollutants. Of particular interest is exposure to ultrafine particles (PM0.1) for which there is sparse literature on its association with stroke.
Hypothesis. Long-term exposures to particulate matter (PM0 2.5, PM10, PM0.1), ozone (O₃), and nitrogen dioxide (NO₂) are associated with stroke risk.
Methods. In a prospective cohort study of 133,477 women enrolled in the California Teachers Study in 1995-96, we assessed 110,120 participants who resided in California from 2000-2018 for residential exposure to 5 air pollutants (O₃, NO₂, PM2.5, PM10, and PM0.1) based on a chemical transport model (4-km grid) applied to geocoded residential histories across follow-up (median follow-up=19 years). Strokes were identified with ICD-9 and ICD-10 codes via linkage of cohort participants to California state hospitalization records (Department of Health Care Access and Information). Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CI) for associations between the air pollutants (average exposure across follow-up) with total stroke (n=4,348) and stroke subtypes (ischemic stroke n=3,596; hemorrhagic stroke n=752), adjusted for stroke risk factors.
Results. Consistent with state-wide trends, median levels of NO₂ and PM masses assessed among participants declined and plateaued in recent years (2000-2018), whereas median annual O₃ levels increased. The exposures observed among study participants also reflected that of the State; annual median PM2.5=8.23 ug/m³ (range: 0.84-85.6). We observed increased risks for total, ischemic, and hemorrhagic strokes per ug/m3 increase of PM2.5, PM10.0, and PM0.1 (overall stroke: HR_PM0.1=1.43, 95% CI=1.29-1.58; HR_PM2.5=1.45, 95% CI=1.30-1.63; HR_PM10=1.36, 95% CI=1.23-1.51); the risk magnitudes were higher for hemorrhagic stroke (HR_PM0.1=1.57, 95% CI=1.22-2.02; HR_PM2.5=1.84, 95% CI=1.40-2.42; HR_PM10=1.62, 95% CI=1.27-2.08). Associations observed between NO₂ and stroke were also elevated (HR=1.12 per ppm, 95% CI=1.07-1.18).
Conclusion. All PM masses, including ultrafine particles, were associated with overall stroke risk and for both ischemic and hemorrhagic stroke. Further delineation of independent versus combined effect of pollutants is warranted and will be presented.