Effect of antihypertensive agents on risk of atrial fibrillation: a meta-analysis of large-scale randomized trials.

Aims High blood pressure is known to be associated with future risk of atrial fibrillation. Whether such risks can be reduced with antihypertensive therapy is less clear. We conducted a systematic review and meta-analysis of large-scale randomized trials that have reported the effect of antihypertensive agents on atrial fibrillation. Methods and results MEDLINE was searched for randomized trials published between 1966 and February 2014. Randomized, controlled trials were eligible for inclusion if they tested an antihypertensive agent and reported atrial fibrillation as an outcome. Atrial fibrillation, reported as trial outcome or adverse event, and study characteristics were extracted by investigators. In 27 trials with 214 763 randomized participants and 9929 events of atrial fibrillation, pooled using inverse-variance weighted fixed effects meta-analysis, antihypertensive therapy reduced the risk of atrial fibrillation by 10% [risk ratio (RR) 0.90; 95% confidence interval (CI) 0.86, 0.94]. However, the proportional effects differed significantly between trials (P < 0.001 for heterogeneity). In trials that included patients with no prior heart disease, or patients with coronary artery disease but no heart failure, no significant effects were found (RR 1.02; CI 0.88, 1.18 and RR 0.95; CI 0.89, 1.01, respectively). Conversely, proportional effects were larger in trials that predominantly included patients with heart failure (RR 0.81; CI 0.74, 0.87). When classes of medication were compared against each other, no significant differences in effects on atrial fibrillation were observed. Conclusions Antihypertensive therapy reduces the risk of atrial fibrillation modestly but benefits appear to be larger in patients with heart failure, with no clear evidence of benefit in patients without heart failure. Previous suggestions of class-specific effects could not be confirmed in this more comprehensive analysis.