Left Atrial Appendage Occlusion (LAAO)
Details
Percutaneous or surgical exclusion of the left atrial appendage to prevent cardioembolic stroke in atrial fibrillation. The LAA is the source of ~90% of AF-related cardiac thrombi, so removing it from the circulation removes the thromboembolic nidus without indefinite systemic anticoagulation. (sources/laao-closureaf-nejm-2026, rating: very high) The organising idea is a trade: a bounded period of procedural and post-implant antithrombotic bleeding risk, bought in exchange for lifelong freedom from anticoagulant bleeding. That trade held for a decade on the strength of trials against warfarin, then failed its first adequate test against contemporary anticoagulation in the highest-risk population. Three randomised trials between 2025 and 2026 reached three different answers in three different risk strata — and the population LAAO was invented for is the one in which it did not work.
Key Facts
- Rationale rests on a single anatomical fact. ~90% of AF-related cardiac thrombi form in the LAA, and occlusion is expected to eliminate that source while leaving systemic haemostasis intact. (sources/laao-closureaf-nejm-2026, rating: very high)
- The benefit is deferred; the harm is immediate. Bleeding is concentrated in the procedural phase and the first six months of post-implant antithrombotic therapy, while stroke prevention accrues only once that therapy is withdrawn — the asymmetry that decided CLOSURE-AF. (sources/laao-closureaf-nejm-2026)
- Three trials, three answers. OPTION: non-inferior with 56% less bleeding after ablation. CHAMPION-AF: non-inferior with bleeding superiority in NOAC-eligible patients. CLOSURE-AF: non-inferiority failed in anticoagulation-unsuitable high-risk patients. (sources/laao-option-nejm-2025; sources/laao-championaf-nejm-2026, rating: very high)
- The divergence is about the population and the post-implant drug, not the device. Watchman FLX was used in all three trials — in CLOSURE-AF alongside several other permitted devices; what differed was baseline risk (CHA₂DS₂-VASc 3.5 versus 5.2) and whether the patient left on a NOAC or on DAPT.
- Surgical closure carries the strongest evidence on this page and percutaneous closure the weakest. LAAOS III gives surgical closure Class I; percutaneous LAAO sits at Class IIa (AHA) and IIb (ESC), both pre-dating every trial above.
- Implantation is reliable; seal and device thrombus are the weak points. Deployment succeeds in ~98–99% of attempts, but complete seal and device-related thrombus rates depend heavily on the imaging modality used to measure them. (sources/laao-option-nejm-2025)
Indications and Guideline Recommendations
| Setting | Recommendation | Class | Source |
|---|---|---|---|
| Percutaneous LAAO, contraindication to long-term OAC | Is reasonable | IIa / B-NR (AHA 2023) | sources/AF-AHA-2023 |
| Percutaneous LAAO, contraindication to all OAC | May be considered | IIb / C (ESC 2024) | sources/AF-ESC-2024 |
| Surgical LAA closure during cardiac surgery | As adjunct to OAC | I / A (AHA 2023); I / B (ESC 2024) | sources/AF-AHA-2023; sources/AF-ESC-2024 |
| Surgical LAA closure during any valve surgery with AF | Recommended | I / B (ESC VHD 2025) | sources/vhd-esc-2025 |
| Surgical LAA closure during endoscopic or hybrid AF ablation | May be considered | IIa / C (ESC 2024) | sources/AF-ESC-2024 |
| Stand-alone epicardial LAA closure, OAC contraindicated | May be considered | IIb / C (ESC 2024) | sources/AF-ESC-2024 |
| Atrial secondary MR — MV surgery plus AF ablation plus LAAO | May be considered | IIa (ESC VHD 2025) | sources/vhd-esc-2025 |
All rows rating: very high.
- Both percutaneous recommendations are stale. ESC 2024 (IIb/C) and AHA 2023 (IIa/B-NR) were written before OPTION, CHAMPION-AF and CLOSURE-AF reported; AHA 2023 had already upgraded percutaneous LAAO from IIb to IIa on evidence that predates all three. (sources/AF-AHA-2023, rating: very high)
- The two guidelines disagree on the same question — IIa versus IIb for essentially the same OAC-contraindicated patient — and neither distinguishes NOAC-eligible from NOAC-unsuitable patients, which is now the distinction that matters most.
Devices
| Device | Manufacturer | Design | Trial role |
|---|---|---|---|
| Watchman / Watchman FLX | Boston Scientific | Endovascular plug; most studied device | Only device in CHAMPION-AF; used in OPTION and CLOSURE-AF |
| Amulet | Abbott (St. Jude Medical) | Lobe-and-disc | Used in CLOSURE-AF |
| Amplatzer Cardiac Plug | Abbott (St. Jude Medical) | Lobe-and-disc predecessor of Amulet | Used at selected CLOSURE-AF sites |
| LAmbre | LifeTech | Disc-and-umbrella | Bailout option at selected CLOSURE-AF sites |
| Surgical ligation, stapling or epicardial clip (AtriClip) | — | Surgical exclusion | Concomitant cardiac surgery or stand-alone; driven by LAAOS III |
Device roles as reported in sources/laao-championaf-nejm-2026, sources/laao-closureaf-nejm-2026 and sources/AF-ESC-2024, all rating: very high.
- Only Watchman FLX has been tested against a NOAC in a large trial, so CHAMPION-AF and OPTION say nothing about lobe-and-disc devices or surgical closure. (sources/laao-championaf-nejm-2026, rating: very high)
- Lobe-and-disc devices interact with pulsed field ablation; plug devices do not. One-third of patients with an Amulet in situ had aborted left-sided PFA applications from device proximity, so PFA should precede implantation whenever both are planned. See concepts/PFA-Atrial-Arrhythmias. (sources/pfa-af-ehra-2026, rating: high)
The Pivotal Trial Landscape
| Trial | Population | Comparator | Result | Key numbers |
|---|---|---|---|---|
| PROTECT AF | Moderate-risk, warfarin-eligible | Warfarin | NI demonstrated | First Watchman RCT; no DOAC comparator |
| PREVAIL | As PROTECT AF | Warfarin | NI demonstrated | Confirmed PROTECT AF; no DOAC comparator |
| PRAGUE-17 (2022) | Moderate risk, CHA₂DS₂-VASc ~4.7 | NOAC | NI at 4 years | First trial against a NOAC; small |
| OPTION (2025) | Post-ablation, n=1,600, CHA₂DS₂-VASc 3.5, HAS-BLED 1.2 | OAC (95% NOAC) | NI + bleeding superiority | Bleeding 8.5% vs 18.1%, HR 0.44; death/stroke/SE 5.3% vs 5.8%, HR 0.91 |
| CHAMPION-AF (2026) | NOAC-eligible, n=3,000, CHA₂DS₂-VASc 3.5, HAS-BLED 1.3 | NOAC | NI + bleeding superiority | Efficacy 5.7% vs 4.8%, HR 1.20; bleeding 10.9% vs 19.0%, HR 0.55 |
| CLOSURE-AF (2026) | OAC-unsuitable, high risk, n=888, CHA₂DS₂-VASc 5.2, HAS-BLED 3.0 | DOAC (85.1%) | NI failed (P=0.44) | 16.8 vs 13.3/100 patient-yr; RMST difference −0.36 yr |
OPTION data from sources/laao-option-nejm-2025; CHAMPION-AF from sources/laao-championaf-nejm-2026; CLOSURE-AF from sources/laao-closureaf-nejm-2026; all rating very high. Earlier trials as summarised in those sources.
Trial Design Compared
The three 2025–26 trials differed in nearly every design element, which is why they cannot be pooled mentally.
| Feature | OPTION | CHAMPION-AF | CLOSURE-AF |
|---|---|---|---|
| n | 1,600 | 3,000 | 888 |
| Mean age | 69.6 | 71.7 | 77.9 |
| CHA₂DS₂-VASc | 3.5 | 3.5 | 5.2 |
| HAS-BLED | 1.2 | 1.3 | 3.0 |
| Sites | 106 / 10 countries | 141 / 16 countries | Germany only |
| Primary efficacy endpoint | Death, stroke or SE at 36 months | CV death, stroke or SE at 3 years | Stroke, SE, major bleeding or CV/unexplained death |
| NI margin | 5 percentage points (RR 1.5) | 4.8 percentage points (RR 1.4) | HR 1.3 (RMST primary metric) |
| Device(s) | Watchman FLX | Watchman FLX only | Watchman/FLX, ACP, Amulet, LAmbre |
| Post-implant antithrombotic | OAC + aspirin 90 days → aspirin to 12 months | NOAC 85.0%, DAPT 12.6% × 3 months | DAPT 79.6% |
| Follow-up | 36 months | 3 years; 5-year endpoint ongoing | Median 3.0 years (max 6.7) |
| Funding | Boston Scientific | Boston Scientific | DZHK (non-industry) |
| Result | NI met; safety superior | NI met; safety superior | NI not met |
(sources/laao-option-nejm-2025, sources/laao-championaf-nejm-2026, sources/laao-closureaf-nejm-2026, rating: very high)
Why CHAMPION-AF and CLOSURE-AF Disagree
Two trials run in the same era over overlapping populations reached opposite conclusions from the same procedure. Three differences account for it.
- Baseline risk amplifies early harm. CLOSURE-AF enrolled patients with CHA₂DS₂-VASc 5.2 and HAS-BLED 3.0 — mean age 77.9 — who were anticoagulation-unsuitable by protocol definition (prior BARC ≥3a/3c bleed, eGFR 15–29, or recurrent bleeding). A fixed periprocedural hazard weighs far more heavily when the competing anticoagulation strategy is itself already failing. (sources/laao-closureaf-nejm-2026, rating: very high)
- Post-implant antithrombotic therapy decided the bleeding outcome. 79.6% of CLOSURE-AF device patients received DAPT, and the highest major-bleeding incidence fell in the first six months — exactly the DAPT window. CHAMPION-AF gave 85.0% a NOAC instead, and its bleeding curves diverged only after day 180, favouring the device. See concepts/NOAC-AF-Management.
- The non-inferiority margins are not commensurable. CHAMPION-AF used an absolute 4.8-percentage-point margin set against an anticipated 12% device event rate; the observed rate was 5.7%, so the same absolute margin permitted a relative hazard ratio up to 1.4. CLOSURE-AF fixed the margin in relative terms (HR 1.3) and used restricted mean survival time as the primary metric because proportional hazards did not hold. CLOSURE-AF's medical-therapy event rate also exceeded expectation — 13.3 per 100 patient-years — reflecting a genuinely high-risk population, and the device arm was worse on restricted mean survival throughout, so the test failed despite a nominal margin of HR 1.3.
The practical reading: LAAO is non-inferior to a NOAC in a moderate-risk patient who can take one and is left on one afterwards. It is not a rescue for the patient who cannot.
Procedural Safety and Imaging
- Implantation succeeds consistently. OPTION: 98.8% of attempts (753/762). CHAMPION-AF: 98.8% of attempts (1,386/1,403), with 92.5% of randomised patients actually implanted. CLOSURE-AF: 98.3% implantation, 97.6% technical and 96.4% procedural success. (sources/laao-option-nejm-2025, rating: very high)
- Pericardial effusion requiring intervention is uncommon: 0.3% in OPTION's device arm and 0.7% within 30 days in CHAMPION-AF. (sources/laao-championaf-nejm-2026)
- CLOSURE-AF recorded the fullest complication picture: periprocedural complications in 5.7% of implantations within 7 days, 18 procedural major bleeds, two periprocedural deaths, one device embolisation requiring surgical retrieval, 20 patients with a device leak of whom 3 exceeded 5 mm. (sources/laao-closureaf-nejm-2026)
- Seal rates are imaging-dependent, not device-dependent. OPTION reported complete seal in 81.0% at 3 months and 79.7% at 12 months against ~90% with TEE in PINNACLE FLX, but CT was used in 16% of OPTION patients at 12 months and detects smaller peridevice leaks. (sources/laao-option-nejm-2025, rating: very high)
- The same artefact inflates device-related thrombus rates. OPTION reported 1.9% at 12 months. CHAMPION-AF reported 4.8% at 4 months on CT or TEE, but CT was used in 18.3% and detects benign hypoattenuated endothelialisation that TEE does not; clinically relevant thrombus requiring resumed OAC was 1.8%, and two patients had a device thrombus with stroke. (sources/laao-championaf-nejm-2026)
Post-Implant Antithrombotic Strategy
| Strategy | Regimen | Where used | Consequence |
|---|---|---|---|
| Short anticoagulation, then antiplatelet | NOAC × ~3 months, then antiplatelet monotherapy or cessation | CHAMPION-AF (85.0% at discharge); OPTION used OAC + aspirin 90 days → aspirin to 12 months | Low early bleeding; the strategy under which bleeding superiority was won |
| DAPT-based | DAPT × 3–6 months, then single antiplatelet | CLOSURE-AF (79.6%); CHAMPION-AF (12.6%) | Higher early bleeding; negated the stroke benefit in high-risk elderly patients |
- Both successful trials kept anticoagulation through endothelialisation and then stopped it. Only 10.1% of OPTION device patients remained on OAC at 36 months, against 84.8% of the OAC arm who maintained it throughout. (sources/laao-option-nejm-2025, rating: very high)
- The regimen, not the device, is the modifiable variable — which is why trials of periprocedural antithrombotic strategy (ADALA and others) are now the field's open question. (sources/laao-closureaf-nejm-2026)
Surgical LAA Closure
- LAAOS III is the evidentiary anchor: surgical LAA closure during cardiac surgery reduced ischaemic stroke or systemic embolism, HR 0.67 over 3.8 years, and drove the Class I recommendations in both major guidelines. (sources/AF-ESC-2024, rating: very high)
- Concomitant surgical closure is an adjunct to OAC, not a replacement. Only stand-alone epicardial closure is framed as filling an anticoagulation gap, and at Class IIb/C — far weaker than the Class I its concomitant counterpart carries.
- ESC VHD 2025 extended it to any valve surgery in patients with AF, and to atrial secondary MR as part of a combined mitral surgery, AF ablation and LAAO strategy. (sources/vhd-esc-2025, rating: very high)
- Stand-alone epicardial closure is far weaker evidence — Class IIb/C with OAC contraindicated, IIa/C during endoscopic or hybrid ablation. (sources/AF-ESC-2024)
- Surgical and percutaneous evidence are not interchangeable. LAAOS III was performed inside an open chest at the time of another operation; its results do not transfer to catheter-based occlusion, and CHAMPION-AF and OPTION say nothing about surgery.
LAAO and Catheter Ablation
- OPTION enrolled only post-ablation patients: ablation preceded randomisation in all cases, either 90–180 days before (sequential, 59.1%) or within 10 days after (concomitant, 40.9%); radiofrequency 59.4%, cryoablation 33.2%; sinus rhythm was restored in 88.1%. (sources/laao-option-nejm-2025, rating: very high)
- Outcome differences were not explained by rhythm control — atrial arrhythmia recurrence was similar between arms, and stroke rates were low in both, consistent with reduced AF burden after ablation. (sources/laao-option-nejm-2025)
- Left ventricular ejection fraction ≤30% was excluded, so OPTION does not support LAAO in the ablation patient with severe LV dysfunction. See concepts/Catheter-Ablation-AF.
- Sequencing matters with pulsed field ablation: perform PFA first and implant afterwards, because a lobe-and-disc occluder in situ distorts the field. (sources/pfa-af-ehra-2026, rating: high)
Contradictions / Open Questions
- LAAO failed in the population it was designed for. CLOSURE-AF enrolled exactly the anticoagulation-unsuitable patient the guidelines name — CHA₂DS₂-VASc 5.2, HAS-BLED 3.0, mean age 77.9 — and non-inferiority was not demonstrated, with numerically more major bleeding and cardiovascular death. The trial succeeded instead in the NOAC-eligible moderate-risk patient it was not originally indicated for. This inverts the indication logic of a decade. (sources/laao-closureaf-nejm-2026, sources/laao-championaf-nejm-2026, rating: very high)
- CHAMPION-AF's non-inferiority is margin-dependent. The margin is absolute (4.8 percentage points, HR up to 1.4) and was set against an anticipated 12% device event rate that proved to be 5.7%. The HR of 1.20 for the primary efficacy endpoint points the wrong way, so a clinically meaningful stroke excess would still have passed the test. (sources/laao-championaf-nejm-2026)
- There is an unresolved ischaemic stroke signal. CHAMPION-AF recorded 3.2% versus 2.2% for ischaemic stroke or systemic embolism with the device — about 0.3% per year absolute excess — while haemorrhagic stroke was identical at 0.4%. The prespecified 5-year ischaemic stroke endpoint is unreported and could widen this. (sources/laao-championaf-nejm-2026)
- Post-implant antithrombotic therapy is undefined and confounds every cross-trial comparison. Three trials used three protocols — NOAC, OAC plus aspirin bridged to aspirin, and DAPT — so their bleeding rates cannot be compared directly, and no trial has randomised the regimen itself. (sources/laao-option-nejm-2025; sources/laao-closureaf-nejm-2026)
- Seal and device-thrombus rates are measurement artefacts as much as device properties. CT detects benign endothelialisation and sub-millimetre leaks that TEE misses, producing a 1.9% versus 4.8% thrombus rate and 79.7% versus 90% seal rate across trials, with unknown clinical meaning for leaks ≤3 mm. (sources/laao-option-nejm-2025; sources/laao-championaf-nejm-2026)
- Generalisability is narrower than the headlines suggest. CHAMPION-AF tested Watchman FLX only, was industry-sponsored by its manufacturer, and under-represented patients with CHA₂DS₂-VASc ≥5; CLOSURE-AF was conducted entirely in Germany. No trial informs lobe-and-disc devices, surgical closure as an alternative to OAC, or non-White populations. (sources/laao-championaf-nejm-2026; sources/laao-closureaf-nejm-2026)
- Guideline reclassification is pending on both sides. AHA 2023's IIa for OAC-contraindicated patients is directly challenged by CLOSURE-AF and expected to fall; CHAMPION-AF would justify a recommendation in NOAC-eligible patients that no guideline currently makes. The two guidelines already disagree (IIa versus IIb) on the same question. (sources/AF-AHA-2023; sources/AF-ESC-2024, rating: very high)
- Whether LAAO can replace indefinite anticoagulation after ablation is not settled by OPTION alone. OPTION showed non-inferiority to 36 months in a population with low absolute stroke rates (1.2% versus 1.3%), which limits the precision of the comparison, and durability of seal beyond 12 months is unknown. (sources/laao-option-nejm-2025, rating: very high)
- The mechanism that should make LAAO work is confirmed; the strategy that delivers it is not. Stroke reduction matched anticoagulation in OPTION and CLOSURE-AF, and LAAOS III demonstrated that anatomical exclusion prevents stroke, so the failures are periprocedural and antithrombotic harms layered on a working principle — but no trial has yet optimised those layers before testing the principle again.
Connections
- Primary indication and frameworks: entities/Atrial-Fibrillation · concepts/AF-CARE (Avoid stroke pillar) · concepts/CHA2DS2-VA
- Comparator therapy: concepts/NOAC-AF-Management
- Procedural context: concepts/Catheter-Ablation-AF (OPTION) · concepts/PFA-Atrial-Arrhythmias (device–energy interaction and sequencing)
- Special populations: concepts/AF-Special-Populations
- Surgical and structural context: concepts/Valvular-Heart-Disease · concepts/Secondary-Mitral-Regurgitation