Safety of discontinuing oral anticoagulation after atrial fibrillation ablation: an updated systematic review and meta-analysis

Résumé fourni par la source

Oral anticoagulant (OAC) therapy after atrial fibrillation (AF) ablation is recommended according to thromboembolic risk rather than procedural success alone. However, the clinical balance between residual thromboembolic risk and bleeding risk after OAC discontinuation remains uncertain because randomized controlled trials (RCTs) and observational studies have reported variable estimates. We conducted an updated systematic review and meta-analysis comparing OAC discontinuation with OAC continuation after AF ablation. Online databases were searched from inception to May 2026 for RCTs and observational studies reporting thromboembolic events, major bleeding, or all-cause mortality after post-ablation OAC discontinuation versus continuation. Random-effects Mantel-Haenszel models were used, and estimates were reported as risk ratios (RRs) with corresponding 95% confidence intervals (CIs). Thirty-two studies comprising 271,808 patients were included, with 88,513 patients discontinuing OAC after ablation. OAC discontinuation, compared with continuation, was not associated with a statistically significant difference in thromboembolic events (RR=0.93; 95%CI=0.72 to 1.20; P=0.58; I²=53%). In contrast, OAC discontinuation significantly reduced major bleeding (RR=0.37; 95%CI=0.27 to 0.52; P<0.00001; I²=60%). No significant effect was observed for all-cause mortality (RR=0.87; 95%CI=0.70 to 1.08; P=0.22; I²=0%). In this updated systematic review and meta-analysis, OAC discontinuation after AF ablation was associated with lower major bleeding without a statistically significant difference in thromboembolic events or all-cause mortality. These findings support individualized, risk-based anticoagulation decisions among selected post-ablation patients rather than routine discontinuation. However, the very low certainty of evidence for thromboembolic outcomes, coupled with a CI that does not exclude potential harm (RR up to 1.20), underscores that these findings should not be interpreted as evidence of safety.