CLOSURE-AF Finds LAA Closure Fails to Meet Noninferiority Versus Medical Therapy

March 18, 2026—Catheter-based left atrial appendage (LAA) closure did not achieve noninferiority compared with physician-directed best medical care in patients with atrial fibrillation (AF) at high risk for both stroke and bleeding, according to results from the randomized CLOSURE-AF trial published by Landmesser et al in The New England Journal of Medicine.¹

CLOSURE-AF is a prospective, multicenter, randomized controlled trial conducted across 42 sites in Germany. A total of 912 patients with AF and elevated risk of stroke and bleeding or contraindications to long-term anticoagulation (AC) were randomized to LAA closure or physician-directed best medical care, the default of which was direct oral anticoagulants (DOACs) for those eligible for AC.

Stroke risk was measured as a CHA₂DS₂-VASc score ≥ 2, with higher scores indicating greater thromboembolic risk. High bleeding risk was defined as a HAS-BLED score ≥ 3 or at least one major bleeding-related condition, including prior intracranial, intraspinal, or intraocular hemorrhage (Bleeding Academic Research Consortium [BARC] type 3c); previous major bleeding such as gastrointestinal or genitourinary events meeting BARC type 3a or 3b criteria; advanced chronic kidney disease (estimated glomerular filtration rate of 15–29 mL/min/1.73 m²), or recurrent bleeding episodes that precluded long-term AC. This approach allowed inclusion of patients with either formally elevated bleeding risk scores or clinically significant bleeding histories that would influence real-world AC decisions.

The primary endpoint was a composite of stroke, systemic embolism, major bleeding (defined as BARC type ≥ 3), or cardiovascular or unexplained death, analyzed in a time-to-event framework with a prespecified noninferiority margin of a hazard ratio of 1.3. After a median follow-up of 3 years (IQR, 1.7-4.7 years), a primary endpoint event occurred in 155 patients in the device group (16.8 events per 100 patient-years) and 127 patients in the medical therapy group (13.3 events per 100 patient-years). The difference in restricted mean survival time was minus; 0.36 years (95% CI, −0.70 to −0.01; P = .44 for noninferiority), indicating that LAA closure did not meet the threshold for noninferiority.

Analysis of individual components showed similar stroke and systemic embolism rates between the device and medical therapy groups (2.6 vs 2.7 events per 100 patient-years and 0.3 vs 0.1 events per 100 patient-years, respectively). However, major bleeding (7.4 vs 6.2 per 100 patient-years) and cardiovascular or unexplained death (9.5 vs 7.7 per 100 patient-years) were numerically higher in the device arm, although the study was not powered to detect differences in individual endpoints.

In terms of procedural outcomes, device implantation was successful in 98.3% of patients. However, periprocedural complications occurred in 5.7% of patients, including major bleeding, pericardial tamponade, and two early deaths. Serious adverse events were reported in 82.5% of the device group compared with 77.4% in the medical therapy group. The investigators noted that although LAA closure may reduce stroke risk relative to no AC, this benefit did not translate into improved net clinical outcomes compared with medical therapy in this high-risk cohort. Early bleeding events—particularly within the first 6 months and potentially related to dual antiplatelet therapy—may have contributed to the lack of benefit.

The researchers also noted that these findings contrast with prior studies in lower-risk populations and underscore the importance of patient selection. The results highlight the evolving role of LAA closure in the era of DOACs, particularly as ongoing trials evaluate alternative antithrombotic strategies and device iterations designed to reduce periprocedural risk.

The authors noted several limitations that may affect interpretation. The trial was not powered to detect differences in individual components of the composite endpoint, and the study population—predominantly White and limited to sites in Germany—may restrict generalizability. Treatment strategies reflected currently available devices and medical therapy and may evolve. Complete blinding was not feasible for procedural events, and differences in follow-up between groups introduce the potential for bias. In addition, sample size adjustments and early termination of recruitment may have influenced statistical power and event estimates.

The results of CLOSURE-AF suggest that LAA closure should not be assumed to provide equivalent overall outcomes to contemporary medical therapy in patients with both high thrombotic and bleeding risk, and careful individualized decision-making remains essential.

1. Landmesser U, Skurk C, Kirchhof P, et al. Left atrial appendage closure or medical therapy in atrial fibrillation. N Engl J Med. Published online March 18, 2026. doi: 10.1056/NEJMoa2513310