Meta-analysis of the association between secondhand smoke exposure and stroke

Abstract

Active smoking is a recognized risk factor for stroke. We determined the evidence for an association with secondhand smoke exposure. A systematic review was undertaken according to PRISMA guidelines. Random effects meta-analysis provided a pooled estimate of risk, and heterogeneity quantified using I(2) values. Potential publication and study bias were assessed using a funnel plot and Egger's test. Meta-regression analyses were used to investigate sources of heterogeneity. The 20 eligible studies provided 35 estimates of risk derived from 885 307 participants, of whom 5894 (0.7%) suffered a stroke. The pooled estimate of risk was 1.25 (95% CI: 1.12-1.38) with no evidence of significant publication or small-study bias. There was moderate heterogeneity (I(2) = 54.2%, P < 0.001) but no study characteristics were statistically significant in the meta-regression analysis. There was a non-linear dose relationship. The relative risk increased from 1.16 (95% CI: 1.06-1.27) for exposure to 5 cigarettes/day to 1.56 (95% CI: 1.25-1.96) for exposure to 40 cigarettes/day. There is evidence of a strong, consistent and dose-dependent association between exposure to secondhand smoke and risk of stroke, suggestive of a causal relationship, with disproportionately high risk at low levels of exposure suggesting no safe lower limit of exposure.