Several studies have investigated the role of smoking in incident atrial fibrillation (AF) but have reported contradictory results. Identifying the quantitative association between smoking and AF risk is important for AF management and prevention; therefore, we aimed to estimate the association of smoking with incident AF.

We systematically retrieved relevant studies reporting on the association between smoking and AF using the Cochrane Library, PubMed, and Embase databases. The data were extracted from applicable articles, and we used a random effects model to pool the effect estimates.

Sixteen prospective studies with an overall number of 286,217 participants and 11,878 AF cases met the inclusion criteria. A higher prevalence of AF was confirmed among smokers (risk ratio [RR] = 1.23, 95% confidence interval [CI] 1.08–1.39; P = 0.001). These results were stable in the sensitivity analysis. The pooled RRs showed consistent positive associations in most subgroups. Specifically, 8 articles compared both current smokers (RR = 1.39, 95% CI 1.11–1.75) and former smokers (RR = 1.16, 95% CI: 1.00–1.36) with never smokers. Four articles compared ever smokers (pooled RR = 1.21, 95% CI 0.93–1.57) with never smokers, and 7 articles compared current smokers (pooled RR = 1.21, 95% CI 1.03–1.42) with non-current smokers. Additionally, we estimated that 6.7% of the total risk of AF in men and 1.4% of the risk in women were attributable to smoking worldwide.

Based on the published literature, smoking is associated with a modest increased risk of incident AF. Smoking cessation seemed to reduce but not entirely eliminate the excess risk of AF.