The biological rationale for the procedure rests on the peculiar anatomy of the left atrial appendage, a finger-like, trabeculated pouch in the upper left chamber of the heart. In non-valvular atrial fibrillation, the quivering atrium fails to contract properly, blood stagnates, and clots form overwhelmingly in this appendage. The review’s authors describe a broader concept of atrial myopathy, in which abnormal contractility combines with endothelial dysfunction, inflammation, reduced nitric oxide availability, and activation of the renin-angiotensin-aldosterone system to create a prothrombotic environment. Appendage morphology itself matters: autopsy studies have linked a greater number of lobes to thrombus formation, and radiographic classifications such as cactus, chicken-wing, cauliflower, and windsock shapes carry different risks, with non-chicken-wing geometries most closely associated with clot and embolic stroke, possibly because of lower peak-flow velocities within the pouch.

Yet the authors emphasize a crucial mechanistic caveat that any patient considering the procedure should understand. Oral anticoagulation acts systemically, protecting against clots wherever they form, whereas appendage occlusion acts locally. Mechanical exclusion of the appendage cannot prevent thromboembolism arising elsewhere in the heart, nor can it address non-cardioembolic stroke mechanisms such as large-artery atherosclerosis and small-vessel disease, which are common in the elderly, vascular population that develops atrial fibrillation. In the classic Stroke Prevention in Atrial Fibrillation trials, 52 percent of ischaemic strokes were classified as probably cardioembolic, 24 percent as non-cardioembolic, and 24 percent as uncertain; strikingly, among patients on warfarin, 56 percent of breakthrough strokes were non-cardioembolic. Occluding the appendage removes the dominant source of cardioembolism but does not eliminate all sources of stroke.

The technical execution of the procedure has matured considerably since the first percutaneous occlusion in 2001 with the PLAATO device. Contemporary implants follow three design principles: the plug principle, exemplified by the Watchman FLX, in which a lobe obstructs the appendage neck and relies on endothelialization for sealing; the pacifier principle, exemplified by the Amulet, which adds a proximal disc to seal the ostium; and ligation, achieved with the LARIAT device through a combined endocardial and epicardial approach. Implantation is performed through a femoral vein with a transseptal puncture, guided by fluoroscopy alone, transoesophageal echocardiography, or intracardiac echocardiography, each with distinct trade-offs in anaesthesia requirements, radiation exposure, and resource use. Real-world data from 3,822 consecutive procedures in the United States immediately after FDA approval showed a 95.6 percent implant success rate, a median procedure time of 50 minutes, a cardiac tamponade rate of 1.02 percent, and a procedure-related stroke rate of just 0.078 percent, even though 71 percent of operators had not participated in the clinical trials.

The evidence base changed dramatically over the past two years with the reporting of three randomized trials comparing occlusion with direct oral anticoagulants. OPTION randomized 1,600 patients undergoing catheter ablation to appendage closure with the Watchman FLX or continued anticoagulation, finding superior reduction in non-procedure-related bleeding (8.5 versus 18.1 percent) and non-inferior efficacy for death, stroke, or systemic embolism. CHAMPION-AF, the largest trial to date, enrolled 3,000 patients eligible for anticoagulation and demonstrated non-inferiority for the composite of cardiovascular death, stroke, or systemic embolism (5.7 versus 4.8 percent), along with a substantial reduction in non-procedure-related bleeding (10.9 versus 19.0 percent). But the nuances matter: observed event rates were lower than projected, making the non-inferiority margin permissive; ischaemic stroke was numerically more frequent in the device arm (3.2 versus 2.2 percent); and device-related thrombus appeared in 4.8 percent of recipients on imaging.

CLOSURE-AF, by contrast, addressed the population where occlusion is most often performed in practice: 912 older patients at high risk of both stroke and bleeding, or with contraindications to anticoagulation, mean age 77.9 years. After a median follow-up of three years, the primary composite endpoint occurred at 16.8 versus 13.3 events per 100 patient-years, and the trial failed to meet its non-inferiority criterion, with an adjusted hazard ratio of 1.28. The authors argue this should not be read as invalidating the therapy, since individual endpoint components occurred at similar rates and the trial was not powered for them, but it does raise a sobering question about whether the upfront procedural risk is worthwhile in frail patients whose remaining life expectancy may be too short for benefits to accumulate.

Pooling the four randomized trials that compared occlusion with direct oral anticoagulants, a 2026 meta-analysis of 5,890 patients with median follow-up of 36 to 42 months delivers the most informative summary available. The device significantly reduced non-procedure-related major bleeding (relative risk 0.79) and non-major bleeding (relative risk 0.47), but showed no significant difference in all-cause major bleeding, intracranial haemorrhage, all-cause mortality, or cardiovascular death. Most provocatively, the risk of ischaemic stroke was numerically higher with the device (relative risk 1.36), reaching statistical significance once the OPTION trial was excluded (relative risk 1.45). The reviewers argue that observational studies claiming a mortality advantage for the procedure suffer from healthy-candidate bias, since patients accepted for elective structural procedures are systematically fitter than similarly scored patients managed medically, and that survival claims should be limited to what randomized data support, which is currently equivalence.

Device-specific hazards add further layers of complexity. Peri-device leaks greater than 5 millimetres are considered significant, but even smaller leaks have been associated with increased thromboembolic risk at one year, making complete sealing during implantation a priority that modern fully recapturable devices are better positioned to achieve. Device-related thrombus, detected in an international registry of 711 patients at rates approaching 25 percent within 45 days, more than doubles the risk of death, stroke, or systemic embolism, and its predictors include hypercoagulability, renal insufficiency, deep implantation, and non-paroxysmal atrial fibrillation. Head-to-head comparisons suggest the Amulet achieves more complete sealing with fewer device thrombi, while the Watchman carries a lower risk of pericardial effusion, including a network meta-analysis finding a 2.12-fold higher effusion risk with the Amulet, though no device has demonstrated superiority in stroke or mortality outcomes.

The question of what antithrombotic therapy to give after implantation remains among the least resolved aspects of the procedure. Two randomized trials now favour short-term anticoagulation over dual antiplatelet therapy: the ANDES trial of 510 patients found the composite of death, stroke, bleeding, or device-related thrombus significantly lower with a direct oral anticoagulant (22.5 versus 34.9 percent), primarily due to less bleeding. For patients who cannot tolerate any anticoagulation, single antiplatelet therapy or even complete cessation may be considered, with one large series of 431 high-bleeding-risk patients showing a 1-year ischaemic stroke rate of 2.2 percent despite 41.7 percent receiving no antithrombotic therapy at six months, but the supporting evidence remains observational and the authors recommend reserving complete withdrawal for patients in whom every antithrombotic agent poses prohibitive bleeding risk, following documented exclusion of device thrombus and significant leak.

The bottom line, according to the review, is that appendage occlusion is an established option for patients with contraindications to or intolerance of long-term anticoagulation, and OPTION now supports it as a reasonable alternative in patients undergoing ablation. For the broader population of patients who tolerate direct oral anticoagulants well, the procedure has become a randomized-tested alternative, but the evidence supports an individualized, shared decision rather than a universal offer. Ongoing trials such as CATALYST, which compares the Amulet with anticoagulation in 2,650 patients, and studies enrolling patients after intracranial haemorrhage, will help define the boundaries further. Until then, the authors contend, honest counselling about upfront procedural risks, residual non-appendage stroke mechanisms, and the true magnitude of the bleeding benefit should guide a therapy that has been evocatively described as a mechanical vaccination, provided that metaphor never obscures what the procedure cannot do.

Subject of Research: Left atrial appendage occlusion as a mechanical alternative to oral anticoagulation for stroke prevention in atrial fibrillation

Article Title: Left atrial appendage occlusion: an alternative to anticoagulation in atrial fibrillation

Article References: Left atrial appendage occlusion: an alternative to anticoagulation in atrial fibrillation. (n.d.). https://doi.org/10.3389/fcvm.2026.1889918

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1889918

Keywords: atrial fibrillation, left atrial appendage occlusion, stroke prevention, oral anticoagulation, direct oral anticoagulants, Watchman FLX, Amulet, device-related thrombus, peri-device leak, randomized controlled trials, meta-analysis, interventional cardiology