Left Atrial Appendage Closure or Anticoagulation for Atrial Fibrillation (CHAMPION-AF)

Authors, Journal, Affiliations, Type, DOI

• Doshi SK, Kar S, Nair DG, Waggoner T, Agarwal H, et al.; CHAMPION-AF Investigators; Leon MB, Ellenbogen KA (trial chairs)
• N Engl J Med. Published March 28, 2026 (DOI: 10.1056/NEJMoa2517213)
• Lead sites: Cedars-Sinai Smidt Heart Institute, Pacific Heart Institute (Los Angeles); 141 sites across 16 countries
• Sponsored by Boston Scientific (manufacturer of the Watchman FLX device)
• Ongoing, prospective, international, randomised, 1:1; Kaplan–Meier intention-to-treat analysis; hierarchical testing strategy for multiplicity control
• DOI: 10.1056/NEJMoa2517213

Overview

CHAMPION-AF enrolled 3,000 patients with atrial fibrillation who were suitable candidates for NOAC therapy (CHA₂DS₂-VASc ≥2 men/≥3 women; mean 3.5; mean HAS-BLED 1.3) and compared Watchman FLX–based left atrial appendage closure (LAAO) with physician-selected NOAC. At 3 years, LAAO was noninferior to NOAC for the primary efficacy composite (CV death/stroke/SE: 5.7% vs 4.8%; difference 0.9%; P<0.001 for NI) and superior for non-procedure-related bleeding (10.9% vs 19.0%; HR 0.55; P<0.001 for superiority). Published simultaneously with CLOSURE-AF (which failed NI in the highest-risk AF population), CHAMPION-AF represents the first large RCT demonstrating a significant reduction in clinically relevant bleeding with LAAO vs NOAC — but in a lower-risk, anticoagulation-eligible population, not those for whom LAAO has traditionally been advocated.

Keywords

Left atrial appendage closure, atrial fibrillation, Watchman FLX, NOAC, stroke prevention, bleeding, noninferiority, net clinical benefit, CHA₂DS₂-VASc, HAS-BLED, CHAMPION-AF

Key Takeaways

Background

• NOAC therapy is the standard stroke-prevention strategy for AF with CHA₂DS₂-VASc ≥2 (men)/≥3 (women); uptake and long-term adherence are limited by bleeding risk.
• Percutaneous LAAO (Watchman/FLX) was approved for patients with contraindications to long-term anticoagulation; its role in NOAC-eligible patients had not been established.
• Previous pivotal RCTs (PROTECT AF, PREVAIL) compared Watchman vs warfarin, not vs DOACs, in anticoagulation-unsuitable patients; PRAGUE-17 and OPTION demonstrated NI vs NOACs in moderate-risk or ablation-linked populations.
• CHAMPION-AF is the first large RCT specifically designed to test LAAO as an alternative to NOAC in patients who are eligible for anticoagulation.

Trial Design

• Population: AF with CHA₂DS₂-VASc ≥2 (men) or ≥3 (women); exclusion if MI/stroke/TIA within 30 days, or ISTH major bleeding within 30 days.
• Device group (n=1499): Watchman FLX (Boston Scientific) only; 92.5% (1386) received device; post-implant antithrombotic: 85.0% NOAC, 12.6% DAPT × 3 months, then antiplatelet monotherapy.
• Anticoagulation group (n=1501): Any physician-selected approved NOAC.
• Primary efficacy endpoint: Composite of CV death (including haemorrhage-related and unexplained deaths), stroke (ischaemic or haemorrhagic), or systemic embolism at 3 years — tested for noninferiority (margin 4.8 percentage points absolute; equivalent to relative risk 1.4).
• Primary safety endpoint: Non-procedure-related bleeding (ISTH major + clinically relevant nonmajor) at 3 years — tested for superiority.
• Additional primary efficacy endpoint: Ischaemic stroke or SE at 5 years (ongoing follow-up).
• Hierarchical testing: Both primary endpoints at 3 years must be met to enable testing of secondary endpoints and the 5-year primary.
• 141 sites, 16 countries; 3,000 patients; ongoing trial.

Patient Characteristics

• Mean age 71.7±7.5 years; 31.9% women.
• Mean CHA₂DS₂-VASc 3.5±1.3; mean HAS-BLED 1.3±0.8 — moderate stroke risk, low-moderate bleeding risk (starkly different from CLOSURE-AF: CHA₂DS₂-VASc 5.2, HAS-BLED 3.0).
• 68.9% paroxysmal AF; 47.8% had prior catheter ablation.
• Well-balanced baseline characteristics; 76% device group vs 87% anticoagulation group had confirmed adherence at majority of follow-up visits.

Primary Results — Noninferiority Demonstrated (Efficacy); Superiority Demonstrated (Safety)

Primary Efficacy — Noninferiority Confirmed

• CV death/stroke/SE at 3 years: Device 81 patients (KM 5.7%) vs NOAC 65 patients (KM 4.8%).
• Difference: 0.9 percentage points (95% CI −0.8 to 2.6); P<0.001 for noninferiority. NI demonstrated.
• HR 1.20 (95% CI 0.87–1.66) — point estimate favours NOAC but within NI margin.

Primary Safety — Superiority Confirmed

• Non-procedure-related bleeding at 3 years: Device 154 (KM 10.9%) vs NOAC 260 (KM 19.0%).
• Difference: −8.1 percentage points (95% CI −10.8 to −5.5); HR 0.55 (95% CI 0.45–0.67); P<0.001 for superiority. Superiority demonstrated — LAAO significantly reduces non-procedure-related bleeding vs NOAC.
• ISTH major bleeding (non-procedure-related): 5.1% vs 6.4%.
• Clinically relevant nonmajor bleeding: 7.0% vs 14.2%.
• All-bleeding (post-hoc, procedure + non-procedure combined): 12.8% vs 19.0% — device still lower.
• Kaplan–Meier curves diverge after day 180; anticoagulation group had more events after first 6 months and curves continued to diverge at 3 years.

Individual Component Results (Secondary)

• CV death: 38 (KM 2.7%) vs 36 (KM 2.7%) — equal.
• Stroke (any): 50 (KM 3.6%) vs 33 (KM 2.5%) — numerically higher with device.
• Ischaemic stroke or SE: 45 (KM 3.2%) vs 29 (KM 2.2%) — numerically higher with device (difference 1.0% over 3 years; ~0.3%/yr).
• Haemorrhagic stroke: 5 vs 5 (0.4% both) — equal.
• Systemic embolism: 0 vs 2 (KM 0.1%) — numerically lower with device.
• All-cause death: 71 (KM 5.0%) vs 67 (KM 4.9%) — equal.
• Procedure + non-procedure ISTH major bleeding (secondary safety): 83 (KM 5.9%) vs 87 (KM 6.4%); difference −0.6 pp (95% CI −2.4 to 1.2); HR 0.92 (95% CI 0.68–1.24); P<0.001 for NI — NI demonstrated.
• Net clinical benefit (CV death/stroke/SE + non-procedure bleeding): 215 (KM 15.1%) vs 300 (KM 21.8%); difference −6.7 pp (95% CI −9.6 to −3.8); HR 0.66 (95% CI 0.56–0.79); P<0.001 for NI — device significantly better overall.
• Quality of life and cognitive assessment: similar between groups.

Procedural Data

• Implantation success: 1386/1499 (92.5%) received device; 1386/1403 attempts (98.8%) successful deployment.
• Effective closure (residual leak ≤3 mm) at 4 months: 998/1012 (98.6%).
• Pericardial effusion requiring intervention within 30 days: 10 patients (0.7%).
• Device-related thrombus at 4 months (CT or TEE): 63/1320 (4.8%); clinically relevant thrombus (resumed OAC): 24 (1.8%); device thrombus → stroke: 2 patients.
• Note: COVID-19 pandemic increased CT use (18.3%) vs TEE (81.7%); CT is more sensitive for subtle hypoattenuated thickening (benign endothelialization), likely inflating the 4.8% device-related thrombus rate.
• 206 anticoagulation-group patients crossed over to device group (119 before primary endpoint event) — intent-to-treat analysis maintained.

Contextual Comparison — CHAMPION-AF vs CLOSURE-AF

The divergent conclusions are mechanistically coherent: in higher-risk elderly patients (CLOSURE-AF), periprocedural and DAPT-driven early bleeding negates the stroke-prevention benefit. In lower-risk patients on NOAC post-implant (CHAMPION-AF), the dominant effect is elimination of long-term NOAC-driven bleeding.

Limitations of the Document

• NI margin is absolute (4.8 percentage points), not relative. Given lower-than-expected event rates (5.7% vs anticipated 12% in device group), this allows a potentially higher relative risk — LAAO had HR 1.20 for efficacy endpoint without meeting statistical superiority.
• Only Watchman FLX tested — results cannot be generalised to other LAAO devices, surgical LAA closure, or different antithrombotic regimens post-implant.
• Highest-risk patients underrepresented: Majority had CHA₂DS₂-VASc 3–4; those with score ≥5 underrepresented — results may not apply to the highest-risk stratum.
• Industry-sponsored (Boston Scientific, the device manufacturer) — potential for design and publication bias; steering committee included industry representatives at data access level.
• 206 crossovers from anticoagulation to device group (119 before primary endpoint) — ITT analysis maintained but crossovers can attenuate between-group bleeding difference.
• COVID-19 pandemic altered imaging protocol (increased CT use), likely overestimating device-related thrombus rate.
• 5-year data pending — the additional primary efficacy endpoint (ischaemic stroke/SE at 5 years) is not yet reported; the observed ischaemic stroke difference (3.2% vs 2.2%) may widen or narrow.
• Open-label design inherent to device trial.

Key Concepts Mentioned

• concepts/Left-Atrial-Appendage-Closure — NI vs NOAC in NOAC-eligible AF; bleeding superiority; contrasted with CLOSURE-AF in highest-risk population
• concepts/Anticoagulation-in-AF — NOAC comparator; LAAO as alternative in eligible patients

Key Entities Mentioned

• entities/Atrial-Fibrillation — pLAAO section updated; CHAMPION-AF result added alongside CLOSURE-AF; new contradiction documented

Wiki Pages Updated

• wiki/sources/laao-championaf-nejm-2026.md — created
• wiki/entities/Atrial-Fibrillation.md — pLAAO section updated with CHAMPION-AF result; contradictions section updated; source_count 24→25
• wiki/sourceindex.md — new entry added
• wiki/wikiindex.md — Atrial-Fibrillation description updated