Journal of the American College of Cardiology (JACC) June 30, 2020

Left Atrial Appendage Closure Versus Direct Oral Anticoagulants in High-Risk Patients With Atrial Fibrillation

Pavel Osmancik, Dalibor Herman, Petr Neuzil, et al.

Bottom Line

In high-risk patients with nonvalvular atrial fibrillation, left atrial appendage closure was noninferior to direct oral anticoagulants for the composite of major cardiovascular, neurological, and bleeding events.

Key Findings

1. At a median follow-up of 19.9 months, the annualized rate of the primary composite outcome was 10.99% in the left atrial appendage closure (LAAC) group compared to 13.42% in the direct oral anticoagulant (DOAC) group.

2. LAAC met the criteria for noninferiority compared to DOAC therapy for the primary endpoint (subdistribution hazard ratio [sHR] 0.84; 95% CI, 0.53-1.31; p=0.004 for noninferiority).

3. There was no significant difference in the individual components of the primary endpoint, including all-stroke/TIA (sHR 1.00; 95% CI, 0.40-2.51), clinically significant bleeding (sHR 0.81; 95% CI, 0.44-1.52), and cardiovascular death (sHR 0.75; 95% CI, 0.34-1.62).

4. LAAC implantation was successful in 90.0% of assigned patients (181 of 201), with procedure- or device-related complications occurring in 4.5% of cases.

Study Design

Design

RCT

Open-Label

Sample

402

Patients

Duration

19.9 mo

Median

Setting

Multicenter, Czech Republic

Population Patients with nonvalvular atrial fibrillation at high risk for stroke (mean CHA2DS2-VASc 4.7) and high risk for bleeding (mean HAS-BLED 3.0), with an indication for oral anticoagulation.

Intervention Left atrial appendage closure (LAAC) using predominantly the Amplatzer Amulet or Watchman devices, typically followed by dual antiplatelet therapy for 3 months, then aspirin monotherapy.

Comparator Direct oral anticoagulants (DOACs), predominantly apixaban (used in 95.5% of the DOAC arm).

Outcome Composite of stroke, transient ischemic attack, systemic embolism, cardiovascular death, clinically relevant bleeding (major or nonmajor), or procedure-/device-related complications.

Study Limitations

• The open-label design of the trial may have introduced ascertainment bias, particularly for more subjective clinical endpoints like non-major clinically relevant bleeding.

• The relatively small sample size (N=402) resulted in wide confidence intervals that cross 1.0, meaning small but clinically meaningful differences in efficacy or safety between the two strategies cannot be completely ruled out.

• The follow-up duration (median 19.9 months in the primary report) was relatively short given that stroke prevention in atrial fibrillation is a lifelong need, although subsequent long-term follow-up papers later addressed this.

• The DOAC group heavily favored a single agent (apixaban was used in 95.5% of cases), meaning the results may not perfectly generalize to other DOACs like rivaroxaban or dabigatran.

• The patient population had exceptionally high baseline risk (mean CHA2DS2-VASc 4.7, mean HAS-BLED 3.0), which limits the extrapolability of the findings to younger or lower-risk atrial fibrillation populations.

Clinical Significance

PRAGUE-17 provided the first randomized head-to-head evidence comparing percutaneous left atrial appendage closure to the modern standard of care, DOACs. It established that for atrial fibrillation patients at high risk for both ischemic stroke and major bleeding, structural LAA occlusion is a safe and comparably effective alternative to continuous DOAC therapy. This validates LAAC as a vital option for patients who struggle with long-term oral anticoagulation adherence or tolerance.

Historical Context

The foundational randomized trials for left atrial appendage closure (PROTECT-AF and PREVAIL) compared the WATCHMAN device strictly to warfarin, demonstrating noninferiority and superior bleeding profiles over time. However, following those trials, DOACs revolutionized the management of atrial fibrillation by significantly reducing the risk of intracranial hemorrhage and overall bleeding compared to warfarin. This paradigm shift left a critical gap in the evidence base regarding how device-based closure stacked up against the safer, newer pharmacological standard. PRAGUE-17 was specifically designed to fill this void by comparing LAAC directly against DOACs.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

Anatomically and pathophysiologically, why does left atrial appendage closure (LAAC) serve as an effective alternative to systemic anticoagulation in patients with non-valvular atrial fibrillation?

Key Response

Over 90 percent of thrombi in non-valvular atrial fibrillation originate in the left atrial appendage due to localized blood stasis and endothelial dysfunction. Excluding this structure mechanically mitigates the primary source of thromboembolism without the systemic bleeding risks associated with DOACs.

Resident

When evaluating a patient with atrial fibrillation and a high risk of both stroke (high CHA2DS2-VASc) and bleeding (high HAS-BLED), how does the PRAGUE-17 trial influence your decision to recommend LAAC over a DOAC?

Key Response

PRAGUE-17 demonstrated that LAAC is non-inferior to DOACs for the composite outcome of cardiovascular, neurological, and bleeding events in high-risk patients. For patients who have a particularly high risk of non-procedural bleeding, LAAC becomes an attractive, evidence-backed option to avoid lifelong systemic anticoagulation.

Fellow

The PRAGUE-17 trial utilized both Watchman and Amplatzer Amulet devices. How do the differences in these device designs and post-procedural antithrombotic requirements complicate the interpretation of the non-inferiority margin against DOACs?

Key Response

Watchman typically requires short-term post-procedural OAC or DAPT, whereas the Amulet device often allows for DAPT or SAPT immediately post-implant. Mixing devices with different post-procedural drug regimens introduces heterogeneity in the early bleeding hazard phase, complicating the direct comparison of bleeding events against the continuous risk profile of DOACs.

Attending

Given that the PRAGUE-17 trial composite endpoint compares the procedural complications of LAAC against the continuous bleeding risk of DOACs, how should we counsel patients regarding the time-varying risk profile of these two strategies?

Key Response

The risk profiles cross over time. LAAC carries an upfront procedural risk but offers long-term protection from major bleeding. DOACs lack upfront procedural risk but carry a cumulative, lifelong bleeding risk. Counseling must frame this as trading early, procedure-related risks for long-term bleeding reduction.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

PRAGUE-17 utilized a non-inferiority design for a broad composite endpoint (cardiovascular, neurological, and bleeding events). What are the statistical limitations of using such a broad composite in non-inferiority trials, and how might opposing treatment effects on individual components drive the overall result?

Key Response

In broad composite endpoints, a treatment might be substantially inferior for one component but superior for another, artificially driving the composite toward the non-inferiority margin. This dilution effect requires strict evaluation of the individual event rates to ensure clinical acceptability of the trade-offs.

Journal Editor

As a peer reviewer assessing the PRAGUE-17 manuscript, what concerns would you raise regarding the trial's open-label design and its potential impact on the reporting and adjudication of subjective or less severe bleeding endpoints?

Key Response

In an open-label trial, both patients and investigators know the treatment assignment, introducing detection bias, particularly for endpoints like clinically relevant non-major bleeding. Reviewers must scrutinize the blinding of the clinical events committee (PROBE design) and look for discrepancies between objective endpoints and subjective ones.

Guideline Committee

Current ACC/AHA/HRS guidelines give LAAC a Class IIb recommendation for patients with AFib at increased stroke risk who have contraindications to long-term anticoagulation. Based on the PRAGUE-17 findings comparing LAAC directly to DOACs, should this recommendation be upgraded, and what level of evidence does this trial provide?

Key Response

PRAGUE-17 provides Level of Evidence B-R supporting LAAC as a non-inferior alternative to DOACs in high-risk patients. This could prompt committees to upgrade LAAC to a Class IIa recommendation for patients who are candidates for DOACs but prefer to avoid long-term therapy or have high bleeding risk, moving it beyond just those with absolute contraindications.

Clinical Landscape

Noteworthy Related Trials

2009

PROTECT AF Trial

n = 707 · Lancet

Tested

Percutaneous LAA closure (Watchman device)

Population

Non-valvular AF patients eligible for warfarin

Comparator

Warfarin

Endpoint

Composite of stroke, cardiovascular death, and systemic embolism

Key result: LAA closure was non-inferior to warfarin for the primary efficacy endpoint.

2011

ARISTOTLE Trial

n = 18,201 · NEJM

Tested

Apixaban 5 mg twice daily

Population

Patients with non-valvular atrial fibrillation and at least one stroke risk factor

Comparator

Warfarin

Endpoint

Stroke or systemic embolism

Key result: Apixaban was superior to warfarin in preventing stroke or systemic embolism, caused less bleeding, and resulted in lower mortality.

2014

PREVAIL Trial

n = 407 · JACC

Tested

Percutaneous LAA closure (Watchman device)

Population

Non-valvular AF patients with elevated stroke risk

Comparator

Warfarin

Endpoint

Composite of stroke, systemic embolism, and cardiovascular/unexplained death

Key result: LAA closure achieved similar stroke prevention rates to warfarin with improved procedural safety compared to earlier trials.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis