Coffee consumption and risk of stroke: a meta-analysis of cohort studies

Aims: Several studies have been done to investigate the correlation between coffee consumption and risk of stroke. The results are inconsistent. This meta-analysis is to evaluate the pooled effect of coffee consumption on risk of stroke. Methods: We searched PubMed and Embase for all cohort studies on stroke and coffee consumption in humans with original data. Data from 13 studies involving 12414 stroke cases among 492760 individuals were pooled. The outcome considered was the occurrence of any type of stroke. Fixed effect model was used for the analysis of pooled relative risk. Heterogeneity was tested. If any existed, randomized effect model was used. Results: Comparing with non coffee drinking, all coffee consumption could reduce total stroke by 11% (95% CI: 0.81, 0.97). In women, coffee consumption could reduce stroke risk by 17% (95% CI: 0.79, 0.87). Although there is no statistical difference in men, coffee consumption showed the tendency of decreasing stroke risk. Light, moderate and heavy coffee consumption could reduce the all type stroke risk by 10% (95% CI: 0.85, 0.96), 14% (95% CI: 0.76, 0.97), and 17% (95% CI: 0.78, 0.89) respectively. Conclusions: This meta-analysis is strongly supportive of the hypothesis that coffee consumption at different level, reduces the risk of stroke.