Lancet August 15, 2009

Percutaneous closure of the left atrial appendage versus warfarin therapy for prevention of stroke in patients with atrial fibrillation: a randomised non-inferiority trial

David R Holmes, Vivek Y Reddy, Zoltan G Turi, Shephal K Doshi, Horst Sievert, Maurice Buchbinder, Christopher M Mullin, Peter Sick (for the PROTECT AF Investigators)

Bottom Line

The PROTECT AF trial demonstrated that percutaneous closure of the left atrial appendage with the Watchman device was non-inferior to warfarin for stroke prevention in non-valvular atrial fibrillation, though it was associated with higher upfront procedural safety events.

Key Findings

1. At 1,065 patient-years of follow-up, the primary efficacy endpoint occurred at a rate of 3.0 per 100 patient-years in the Watchman group compared to 4.9 per 100 patient-years in the warfarin group (Rate Ratio [RR] 0.62; 95% CrI 0.35-1.25).
2. The intervention met the prespecified criteria for non-inferiority with a probability greater than 99.9%.
3. Primary safety events were significantly more frequent in the intervention arm (7.4 vs. 4.4 per 100 patient-years; RR 1.69; 95% CrI 1.01-3.19), driven predominantly by periprocedural complications such as pericardial effusions.
4. Despite the early procedural hazard, long-term efficacy endpoints, including all-cause stroke and all-cause mortality, numerically favored the left atrial appendage closure device.

Study Design

Design
Randomized Non-inferiority Trial
Open-Label
Sample
707
Patients
Duration
18 mo
Median
Setting
Multicenter, US and Europe
Population Adult patients with non-valvular atrial fibrillation eligible for long-term warfarin therapy with at least one additional stroke risk factor (CHADS2 score ≥1).
Intervention Percutaneous left atrial appendage (LAA) closure using the Watchman device, followed by discontinuation of warfarin after 45 days and transition to antiplatelet therapy.
Comparator Long-term systemic anticoagulation with warfarin (target INR 2.0-3.0).
Outcome Composite efficacy endpoint of stroke (ischemic or hemorrhagic), cardiovascular death, and systemic embolism.

Study Limitations

The open-label design introduced potential bias in patient management and event ascertainment, although hard clinical endpoints were evaluated.
The trial experienced a relatively high rate of early periprocedural complications (e.g., pericardial effusions), reflecting the operator learning curve for a novel mechanical device.
The study only included patients who were eligible for warfarin therapy, leaving the efficacy and safety in patients with absolute contraindications to anticoagulation unaddressed.
Patients in the device arm were required to take warfarin for 45 days post-procedure, limiting its applicability for those who cannot tolerate even short-term systemic anticoagulation.
The subsequent rise of direct oral anticoagulants (DOACs) altered the clinical landscape, making the warfarin comparator less representative of the modern standard of care.

Clinical Significance

PROTECT AF was a landmark study that established percutaneous left atrial appendage (LAA) closure as a viable, non-pharmacologic alternative to warfarin for stroke prophylaxis in patients with non-valvular atrial fibrillation. By proving non-inferiority to the standard of care, it introduced a major paradigm shift for patients who are suitable for short-term anticoagulation but desire to avoid the bleeding risks and monitoring burdens of lifelong systemic anticoagulation. The trial also highlighted the critical need for operator training to mitigate early procedural hazards.

Historical Context

Prior to PROTECT AF, the undisputed standard for stroke prevention in non-valvular atrial fibrillation was systemic anticoagulation with warfarin. However, warfarin's narrow therapeutic window, numerous drug-food interactions, and bleeding risks resulted in widespread underutilization. Based on transesophageal echocardiography and autopsy data showing that over 90% of atrial thrombi in non-valvular AF originate in the LAA, researchers hypothesized that mechanically excluding the appendage could prevent stroke. PROTECT AF was the pivotal randomized trial validating this concept, eventually leading to FDA approval of the Watchman device after further safety refinements were demonstrated in the subsequent PREVAIL trial.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

Why is the left atrial appendage specifically targeted for percutaneous closure in patients with non-valvular atrial fibrillation, and how does the pathophysiology differ from valvular atrial fibrillation?

Key Response

In non-valvular atrial fibrillation, over 90 percent of thrombi originate in the left atrial appendage due to blood stasis from the loss of coordinated atrial contraction. In valvular atrial fibrillation, such as with mitral stenosis, stasis occurs throughout the globally dilated left atrium, meaning left atrial appendage closure alone would be insufficient for stroke prevention.

Resident
Resident

Given the initial PROTECT AF findings of higher upfront procedural complications with the Watchman device compared to warfarin, how should you counsel a patient with non-valvular atrial fibrillation and a high HAS-BLED score about this intervention?

Key Response

Residents must weigh procedural risks, such as pericardial effusion and procedural stroke, against the long-term bleeding risks of anticoagulation. Counseling involves explaining that while the device offers freedom from long-term anticoagulation, it carries an acute procedural risk, requiring shared decision-making based on the patient's individual bleeding risk and ability to tolerate a short-term post-procedural antithrombotic regimen.

Fellow
Fellow

The PROTECT AF protocol required patients to remain on warfarin for 45 days post-implantation. How does this requirement complicate the use of the Watchman device in patients who have absolute contraindications to systemic anticoagulation?

Key Response

Fellows should understand the clinical paradox of left atrial appendage closure: the procedure is often most desired for patients who cannot tolerate anticoagulation, yet the trial protocol required 45 days of warfarin to prevent device thrombosis until endothelialization occurred. This necessitates knowledge of alternative off-label antiplatelet regimens or newer trial data for patients with absolute contraindications.

Attending
Attending

How do we interpret the non-inferiority of left atrial appendage closure established in PROTECT AF in the context of the modern era where direct oral anticoagulants are the standard of care rather than warfarin?

Key Response

Attendings must integrate legacy trial data with current practice. PROTECT AF compared the Watchman to warfarin. Today, direct oral anticoagulants are generally safer and more effective than warfarin. Therefore, the modern risk-benefit calculus for left atrial appendage closure must implicitly account for the improved safety profile of direct oral anticoagulants, raising the threshold to recommend an invasive procedure.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

PROTECT AF used a non-inferiority margin of 2.0 for the rate ratio of the primary efficacy endpoint. In the context of a composite endpoint including both ischemic and hemorrhagic stroke, how does this margin affect the interpretation of the device's true efficacy in preventing ischemic events?

Key Response

A non-inferiority trial with a wide margin on a composite endpoint can mask worse performance in one component, like ischemic stroke, if offset by better performance in another, like hemorrhagic stroke, which inherently decreases off warfarin. Methodologists must evaluate if the non-inferiority margin is clinically acceptable for each individual component.

Journal Editor
Journal Editor

As a reviewer, what concerns would you raise regarding the unblinded nature of the PROTECT AF trial and the potential for detection bias or variations in the control arm's Time in Therapeutic Range?

Key Response

Unblinded trials with procedural interventions are highly susceptible to performance and detection bias. A critical reviewer would scrutinize whether the warfarin arm achieved adequate Time in Therapeutic Range, as poor warfarin control could falsely inflate the relative efficacy and safety of the device arm.

Guideline Committee
Guideline Committee

Based on the PROTECT AF findings and subsequent registry data, what Class of Recommendation should be assigned to percutaneous left atrial appendage closure, and how should guidelines position it relative to direct oral anticoagulant therapy?

Key Response

Current AHA/ACC/HRS guidelines give percutaneous left atrial appendage closure a Class 2b recommendation for patients with non-valvular atrial fibrillation at increased risk of stroke who have contraindications to long-term anticoagulation. The committee must balance the long-term non-inferiority shown in PROTECT AF against the lack of head-to-head randomized data against direct oral anticoagulants, positioning it as a second-line option.

Clinical Landscape

Noteworthy Related Trials

2014

PREVAIL Trial

n = 407 · JACC

Tested

Watchman LAA closure device

Population

Patients with non-valvular atrial fibrillation

Comparator

Warfarin therapy

Endpoint

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

Key result: Watchman was non-inferior to warfarin for the second primary efficacy endpoint, with significantly improved procedural safety compared to PROTECT-AF.
2020

PRAGUE-17 Trial

n = 402 · JACC

Tested

Percutaneous LAA closure

Population

Patients with non-valvular atrial fibrillation at high risk for stroke and bleeding

Comparator

Direct oral anticoagulants (DOACs)

Endpoint

Composite of stroke, TIA, systemic embolism, CV death, major bleeding, or procedure complications

Key result: LAA closure was non-inferior to DOAC therapy for the composite endpoint of major cardiovascular and neurological events.
2021

LAAOS III Trial

n = 4,811 · NEJM

Tested

Surgical occlusion of the left atrial appendage

Population

Patients with atrial fibrillation undergoing cardiac surgery for another reason

Comparator

No surgical LAA occlusion

Endpoint

Ischemic stroke or systemic embolism

Key result: Surgical LAA occlusion resulted in a 33% lower risk of ischemic stroke or systemic embolism compared to no occlusion.

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