Associations between exhaust and non-exhaust particulate matter and stroke incidence by stroke subtype in South London

Background: Airborne particulate matter (PM) consists of particles from diverse sources, including vehicle exhausts. Associations between short-term PM changes and stroke incidence have been shown. Cumulative exposures over several months, or years, arc less well studied; few studies examined ischaemic subtypes or PM source. Aims: This study combines a high resolution urban air quality model with a population-based stroke register to explore associations between long-term exposure to PM and stroke incidence. Method: Data from the South London Stroke Register from 2005-2012 were included. Poisson regression explored association between stroke incidence and long-term (averaged across the study period) exposure to PM2.5(PM < 2.5 mu m diameter) and PM w(PM < 10 mu m), nitric oxide, nitrogen dioxide, nitrogen oxides and ozone, at the output area level (average population = 309). Estimates were standardised for age and sex and adjusted for socio-economic deprivation. Models were stratified for ischaemic and haemorrhagic strokes and further broken down by Oxford Community Stroke Project classification and Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification. Results: 1800 strokes were recorded (incidence = 42.6/100,000 person-years). No associations were observed between PM and overall ischaemic or haemorrhagic incidence. For an interquartile range increase in PM-2.5, there was a 23% increase in incidence (Incidence rate ratio 1.23 (95%CI: 1.03-1.44)) of total anterior circulation infarcts (TACI) and 20% increase for PM2.5 from exhausts (120(1.01-141)). There were similar associations with PM10, overall (1.21(1.01-1.44)) and from exhausts (1.20(1.01-1.41)). TACI incidence was not associated with non-exhaust sources. There were no associations with other stroke subtypes or pollutants. Conclusion: Outdoor air pollution, particularly that arising from vehicle exhausts, may increase risk of TAO but not other stroke subtypes. (C) 2016 Elsevier B.V. All rights reserved.