Interrupted vs uninterrupted direct oral anticoagulants compared with uninterrupted vitamin K antagonists for atrial fibrillation ablation - A meta-analysis of safety

Abstract Background At the time of atrial fibrillation (AF) ablation, direct oral anticoagulants (DOACs) are either uninterrupted (U) or briefly interrupted (I). A meta-analysis was performed to compare the safety of these 2 approaches, in comparison to the prior gold standard of uninterrupted vitamin K antagonists (VKA). Objective To compare thromboembolic and bleeding complications following AF ablation using one of the following 3 strategies: U DOACs, I DOACs and U VKA (INR > 2). Methods A literature search identified studies that evaluated these arms. Procedural thromboembolic complications included stroke/TIA or systemic embolism; bleeding complications included GI bleeding, pericardial effusion, access site bleeding hematoma, or hemoptysis. DOACs were interrupted 12-96 h prior to ablation. The Mantel-Haenszel random effect model for pooling the study results was used and a random effects model was used for heterogeneous samples/results. The RevMan 5.3 software [Review Manager, Copenhagen] was used for analysis. Results 6739 patients (pts) on I DOACs were compared with 8370 pts on U VKA; 5027 pts on U DOACs were compared with 5011 pts on U VKA. Thromboembolic complications were similar (0.30% I DOACs vs 0.26% U VKA; 0.28% U DOACs vs 0.32% U VKA) (Figure A, B). However, bleeding complications were lower in the I DOACs vs U VKA groups (4.4% vs 6.5%) but similar in U DOACs vs U VKA groups (7.8% vs 8.3%) (Figure C, D) Conclusions This meta-analysis demonstrates that a strategy of I DOACs during AF ablation is associated with the lowest bleeding rate and not associated with an excess in thromboembolic complications. These data support this strategy, especially since randomized clinical trials to compare I vs U DOACs in this setting are highly unlikely to occur.