Abstract P213: Habitual Coffee Intake, Blood Pressure and Risk of Stroke: a Meta-analysis of Prospective Studies

Background: Prospective cohort studies estimated the predictive role of exposure to coffee consumption and risk of hypertension (HP) or cerebrovascular disease, with contrasting results. Therefore, we carried out a meta-analysis to assess the relation between habitual coffee intake and risk of HP and stroke in the general population. Methods: We performed a systematic search for prospective studies published without language restrictions between 1966 and July 2013. Inclusion criteria were: original article, prospective design, adult general population study, assessment of coffee intake as baseline exposure, assessment of HP or stroke as outcome, and follow-up of at least 2 years. To evaluate risks of HP and stroke, relative risks and 95% confidence intervals for “low-moderate” (2-3 cups/d), “high”(4-6) and “very high”(>6) compared with “reference” (0-1) coffee intake categories were extracted and pooled using a random effect model. Heterogeneity, publication bias, subgroup and meta-regression analyses were performed. Results: Four studies for HP (171,460 participants and 39,851 HP diagnoses) and 10 for stroke (969,656 participants with 12,990 events) met the inclusion criteria. Coffee intake was assessed by dietary questionnaire. In pooled analyses, “low-moderate” and “high” coffee intake were both inversely associated with the risk of stroke (RR=0.84, 95% C.I. 0.78 to 0.91 and 0.83, 0.72 to 0.96, respectively) but not with risk of HP. “Very high” coffee intake was not significantly associated with either risk of stroke or of HP. There was significant heterogeneity between studies, while some evidence of publication bias was detected only in the analysis of HP risk. Sub-group analysis suggested an influence of gender and of geographic location, and meta-regression analysis detected age at baseline and length of follow-up as further sources of heterogeneity. Decaffeinated coffee consumption was not related to risk of stroke or HP. Conclusions: This meta-analysis of the available prospective studies showed that habitual coffee intake is not associated with higher risk of HP in the general population and that moderate-high coffee consumption is inversely associated with risk of stroke. A protective effect of coffee consumption towards cerebrovascular disease might be exerted through mechanisms independent of changes in blood pressure.