Left Atrial Appendage Closure in Atrial Fibrillation Patients With High Risk of Stroke and Bleeding: An Individual Participant Data Meta-Analysis of Randomized Trials
In brief
Left atrial appendage closure showed no stroke or bleeding advantage over medical therapy
In two randomized trials involving 1,290 high-risk patients with atrial fibrillation, the combined risk of stroke, systemic embolism or cardiovascular death was similar with closure and medical therapy: 21.9% versus 20.7%. Major bleeding was more common with closure at one year, but rates were similar by three years; longer follow-up is needed to clarify later effects.
- Journal
- Journal of the American College of Cardiology (Q1)
- Published
- 29 October 2026
- Study design
- Systematic review / meta-analysis of RCTs
- Evidence level
- Level 1, High (CEBM 1a)
- Authors
- Carsten Skurk, Ingo Eitel, Tharusan Thevathasan, Petr Neuzil, Paulus Kirchhof, Holger Thiele, et al.
- PMID
- 43018508
- DOI
- 10.1016/j.jacc.2026.09.033
Why clinicians should know about it
- Picked for Neurology (clinical) (paper of the day, 15 November 2026): Randomized-trial meta-analysis of stroke prevention in high-risk AF
- Picked for Cardiology and Cardiovascular Medicine (top studies of the week, 15 November 2026): Randomized-trial meta-analysis compares appendage closure with medical therapy
Abstract
BACKGROUND: Percutaneous left atrial appendage closure (LAAC) is increasingly used as an alternative to oral anticoagulation for stroke prevention in atrial fibrillation. Its comparative effectiveness and safety in patients at a high risk of both stroke and bleeding remain uncertain. OBJECTIVES: We evaluated the efficacy and safety of LAAC vs contemporary medical therapy in a high-risk population using individual participant data from randomized trials. METHODS: Bibliographic databases and trial registries through August 5, 2026, were searched for randomized trials comparing percutaneous LAAC with contemporary medical therapy in populations with a trial-level mean CHA2DS2-VASc score ≥4. A 1-stage common-effect analysis used Cox models stratified by trial and adjusted for age, sex, CHA2DS2-VASc score, and HAS-BLED score. Both primary outcomes were time to first stroke, systemic embolism, or cardiovascular/unknown death (efficacy) and time to first major bleeding (safety). A flexible time-varying Cox model was used for major bleeding. The protocol was submitted to PROSPERO (CRD420261335972). RESULTS: CLOSURE-AF (Left Atrial Appendage CLOSURE in Patients With Atrial Fibrillation Compared to Medical Therapy) and PRAGUE-17 (Left Atrial Appendage Closure vs. Novel Anticoagulation Agents in Atrial Fibrillation) contributed 1,290 participants (647 assigned to LAAC and 643 to control). Median age was 77.3 years (Q1-Q3: 71.4-81.9 years), and 809 participants (62.7%) were men. The primary efficacy outcome occurred in 142 LAAC participants (21.9%) and 133 control subjects (20.7%) (adjusted HR: 0.98; 95% CI: 0.77-1.24; P = 0.842). Major bleeding occurred in 86 LAAC participants (13.3%) and 80 control subjects (12.4%) (adjusted HR: 1.03; 95% CI: 0.75-1.39; P = 0.874). The treatment effect on major bleeding varied over time (P for treatment-by-time interaction < 0.001), with a cumulative major bleeding incidence of 9.4% with LAAC vs 6.0% with control (absolute risk difference, 3.4 percentage points; 95% CI: 0.4-6.3 percentage points) at 1 year and 13.2% vs 12.1% (absolute risk difference, 1.1 percentage points; 95% CI: -2.8 to 5.0 percentage points) at 3 years. CONCLUSIONS: In high-risk patients with atrial fibrillation, no statistically significant difference was detected between LAAC and contemporary medical therapy for the primary efficacy outcome or major bleeding. The major bleeding hazard varied over time, with higher early and lower later conditional hazard estimates; however, a cumulative bleeding benefit was not demonstrated by 3 years. Long-term randomized data are needed to improve the precision of later treatment-effect estimates.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.