PACULit
← September 2026 edition

A05 · PMID:41910347

Left Atrial Appendage Closure or Anticoagulation for Atrial Fibrillation

Open the original article record

Bottom line

In anticoagulation-eligible AF patients, WATCHMAN FLX-based left atrial appendage closure (LAAC) met noninferiority against physician-chosen NOAC therapy for the composite of cardiovascular death, stroke, or systemic embolism at 3 years, and cut non-procedure-related bleeding nearly in half, but the device arm had numerically more ischemic strokes and total primary-endpoint events, so this is supportive-but-contested evidence rather than a clean win.

PICO at a glance

Clinical question and design

LAAC has historically been reserved for patients unsuitable for long-term anticoagulation. Whether it also holds up as a genuine alternative, not just a fallback, in anticoagulation-eligible patients had not been established before CHAMPION-AF (NCT04394546), the first large randomized trial to test a device against oral anticoagulation head-to-head in this population.

Of 3000 patients randomized, 1499 were assigned to the device group and 1501 to the anticoagulation group. This 3-year readout is an interim analysis; a third primary endpoint (5-year ischemic stroke/systemic embolism noninferiority) is still to come.

Intervention and outcomes measured

Patients suitable for anticoagulation were randomized to WATCHMAN FLX device-based LAAC or NOAC therapy chosen at the treating physician's discretion, since no single agent or dose was protocol-mandated. The primary efficacy endpoint was the 3-year composite of cardiovascular death, stroke (ischemic or hemorrhagic), or systemic embolism, tested for noninferiority against a 4.8-percentage-point margin. The primary safety endpoint was non-procedure-related bleeding, tested for superiority.

Results

A primary efficacy-endpoint event occurred in 81 patients (5.7%) in the device group versus 65 patients (4.8%) in the anticoagulation group, a difference of 0.9 percentage points (95% CI, -0.8 to 2.6), meeting the prespecified noninferiority margin (P<0.001). Non-procedure-related bleeding occurred in 154 patients (10.9%) with the device versus 260 patients (19.0%) with anticoagulation, hazard ratio 0.55 (95% CI, 0.45-0.67; P<0.001), meeting superiority, roughly a 45% lower hazard of bleeding with LAAC.

Safety and additional findings

The primary-composite hazard ratio was 1.20 (95% CI, 0.87-1.66). Ischemic stroke or systemic embolism occurred in 45 device-group patients (3.2%) versus 29 in the anticoagulation group (2.2%), numerically more with the device; all-cause stroke was 50 (3.6%) versus 33 (2.5%), hazard ratio 1.46 (95% CI, 0.94-2.27), and hemorrhagic stroke occurred in 5 patients in each arm. Pericardial effusion requiring intervention occurred in 10 device patients (0.7%) within 30 days. Device-related thrombus was assessed in 1320 device patients at 4 months and found in 63 (4.8%); 24 of those (1.8%) were clinically relevant and prompted resumption of oral anticoagulation. Once periprocedural bleeding is folded back in, the prespecified secondary safety endpoint of procedure-related plus non-procedure-related ISTH major bleeding was 83 patients (5.9%) with the device versus 87 (6.4%) with anticoagulation, hazard ratio 0.92 (95% CI, 0.68-1.24), noninferior only, not superior. Net clinical benefit favored the device, 215 events (15.1%) versus 300 (21.8%), hazard ratio 0.66 (95% CI, 0.56-0.79).

Uncertainty and limitations

CHAMPION-AF is industry-funded (Boston Scientific). The primary manuscript confirms the device as the Watchman FLX and confirms that the comparator was any approved NOAC, with no agent or dose mandated by protocol, so this trial cannot be read as a comparison against one specific agent. The device arm had numerically more ischemic strokes and total primary-endpoint events, which some cardiologists flagged as inconsistent with a straightforward "as good or better" reading; the wide 4.8-point margin also drew comment. This is a 3-year interim readout with a 5-year endpoint still to come. Death from any cause was 71 (5.0%) with the device versus 67 (4.9%) with anticoagulation, and quality-of-life and cognitive assessments were similar between groups. The trial defines no functional-outcome or modified Rankin endpoint.

Literature review and evidence synthesis

CHAMPION-AF extends a literature that has mostly evaluated LAAC against warfarin or in anticoagulation-ineligible populations. A 2015 patient-level meta-analysis pooling the PROTECT AF and PREVAIL trials established LAAC as noninferior to warfarin, predating the DOAC era. A 2023 overview and a 2024 network meta-analysis of LAAC versus oral anticoagulants trace how that evidence base evolved as DOACs became the standard comparator. A 2025 pooled analysis of long-term outcomes across four randomized LAAC-versus-anticoagulation trials is the most direct precedent for CHAMPION-AF's own comparative question.

How this fits with the broader evidence

On the headline numbers, LAAC met noninferiority for the hard composite efficacy endpoint and roughly halved non-procedure-related bleeding, supporting LAAC as a genuine alternative for anticoagulation-eligible patients rather than solely a fallback. But the topline verdicts undersell two nuances: more primary-endpoint events and ischemic strokes numerically in the device arm, and a bleeding advantage that narrows once periprocedural bleeding is counted. Post-presentation cardiology commentary split on how to weigh this, some called it validation of LAAC, others argued the wide margin and lower-than-expected event rates made noninferiority easier to meet than the underlying signal justifies. Pharmacists should treat CHAMPION-AF as evidence that LAAC can be discussed with anticoagulation-eligible patients who value lower non-procedure-related bleeding, remembering the comparator was physician-discretion NOAC therapy, not one specific DOAC.

ACPE UAN: 0683-0000-26-036-H01-P

Study source: PMID:41910347