Alcohol consumption and risk of atrial fibrillation: a pairwise and network meta-analysis

Despite well-established evidence linking alcohol consumption to an increased risk of atrial fibrillation (AF), a clear threshold for safe intake remains undefined. Most clinical guidelines recommend abstinence or moderation, but lack specificity regarding quantitative limits and evidence-based thresholds for primary prevention of AF.

We performed this systematic review, pairwise meta-analysis, and network meta-analysis to assess the association and dose-response relationship between varying quantities of alcohol consumption and incident AF. The predefined alcohol intake categories included: no intake, very low (<12 g/day), low (12.1-24 g/day), moderate (24.1-48 g/day), high (48.1-60 g/day), and very high (>60 g/day). The pairwise meta-analyses compared each intake category to no intake, and a frequentist network meta-analysis was conducted to integrate direct and indirect comparisons. Pooled risk ratio (RR) for AF incidence were estimated using random-effects models, and a subgroup analysis by sex was performed. Twenty-six studies with nearly 15 million participants were included. As compared to those with no alcohol intake, very high alcohol intake was associated with a 75% increased risk of incident AF (RR 1.75; 95% CI: 1.25-2.44, P = 0.04), intake below 48 g/day was associated with modestly lower AF risk, and the lowest risk was observed at <12 g/day (RR 0.72; 95% CI: 0.63-0.83). The weighted mean follow-up was approximately 6.4 years. Network meta-analysis confirmed these findings and demonstrated a significantly increased AF risk with very high compared to low (RR 1.91), moderate (RR 2.03), and high (RR 2.00) levels. Sex-stratified analyses showed similar results.

Alcohol consumption is associated with a nonlinear, threshold-dependent relationship with incident AF. While low to moderate intake may not increase AF risk, intake beyond 60 g/day significantly increases the AF risk. These findings may suggest that a more liberal approach to alcohol consumption, particularly at low to moderate levels, could be reasonable in the general population, though further prospective studies are needed to confirm causality and refine individual risk stratification.