BACKGROUND: Ninety percent of left atrial thrombi in nonvalvular atrial fibrillation arise in the left atrial appendage, which can be occluded percutaneously (left atrial appendage closure [LAAC]). Two 2026 trials (CHAMPION-AF and CLOSURE-AF) more than doubled the available randomized evidence comparing LAAC with oral anticoagulation. We evaluated pooled effects of LAAC on stroke, bleeding, and mortality by bleeding risk, anticoagulation eligibility, and comparator.METHODS: In this systematic review and meta-analysis, we searched PubMed, Scopus, and Embase from inception to April 15, 2026, for randomized controlled trials comparing percutaneous LAAC with oral anticoagulation in nonvalvular atrial fibrillation. Risk of bias was assessed with the Cochrane Risk of Bias 2 tool. Cumulative-incidence outcomes were pooled as odds ratios using random-effects models, with prediction intervals, prespecified subgroups, and trial sequential analysis.RESULTS: Six randomized controlled trials (2014-2026; 358 sites in North America, Europe, and Asia; 7028 patients) compared LAAC with warfarin (2 trials) or a direct oral anticoagulant (4 trials); populations were predominantly anticoagulation-eligible with low prior-stroke prevalence, plus 2 high-bleeding-risk and 1 postablation trial. Overall stroke rates were similar between arms (odds ratio, 1.10 [95% CI, 0.78-1.56]). Ischemic stroke was more frequent with LAAC (odds ratio, 1.41 [95% CI, 1.07-1.86]) but did not reach the required information size. Nonprocedural bleeding was lower with LAAC (odds ratio, 0.58 [95% CI, 0.35-0.96]), the only outcome reaching a sufficient information size, in standard-risk and postablation but not high-bleeding-risk patients.CONCLUSIONS: LAAC is best positioned to reduce nonprocedural bleeding in anticoagulation-eligible patients at standard-to-moderate bleeding risk, rather than as a stroke-equivalent substitute for oral anticoagulation.REGISTRATION: URL: https://www.crd.york.ac.uk/PROSPERO/; Unique identifier: CRD420261370546.
