Outdoor Fine Particles and Nonfatal Strokes: Systematic Review and Meta-analysis

BACKGROUND: Epidemiologic studies find that long- and short-term exposure to fine particles (PM(2.5)) is associated with adverse cardiovascular outcomes, including ischemic and hemorrhagic strokes. However, few systematic reviews or meta-analyses have synthesized these results. METHODS: We reviewed epidemiologic studies that estimated the risks of nonfatal strokes attributable to ambient PM(2.5). To pool risks among studies we used a random-effects model and 2 Bayesian approaches. The first Bayesian approach assumes a normal prior that allows risks to be zero, positive or negative. The second assumes a gamma prior, where risks can only be positive. This second approach is proposed when the number of studies pooled is small, and there is toxicological or clinical literature to support a causal relation. RESULTS: We identified 20 studies suitable for quantitative meta-analysis. Evidence for publication bias is limited. The frequentist meta-analysis produced pooled risk ratios of 1.06 (95% confidence interval = 1.00–1.13) and 1.007 (1.003–1.010) for long- and short-term effects, respectively. The Bayesian meta-analysis found a posterior mean risk ratio of 1.08 (95% posterior interval = 0.96–1.26) and 1.008 (1.003–1.013) from a normal prior, and of 1.05 (1.02–1.10) and 1.008 (1.004–1.013) from a gamma prior, for long- and short-term effects, respectively, per 10 μg/m(3) PM(2.5). CONCLUSIONS: Sufficient evidence exists to develop a concentration-response relation for short- and long-term exposures to PM(2.5) and stroke incidence. Long-term exposures to PM(2.5) result in a higher risk ratio than short-term exposures, regardless of the pooling method. The evidence for short-term PM(2.5)-related ischemic stroke is especially strong.