PREVAIL Published: Watchman, Warfarin Show Comparable Stroke Protection in A-fib

Download this article's Factoid (PDF & PPT for Gold Subscribers)

In patients with nonvalvular atrial fibrillation (A-fib), occlusion of the left atrial appendage (LAA) with the transcatheter Watchman device is safe and provides ischemic stroke protection similar to warfarin therapy. The findings from the PREVAIL trial, which largely confirm those of its randomized predecessor study, PROTECT AF, were published in the July 8, 2014, issue of the Journal of the American College of Cardiology.

For PREVAIL, investigators led by David R. Holmes Jr, MD, of the Mayo Clinic (Rochester, MN), randomized patients with nonvalvular A-fib and either a CHADS2 score of at least 2 or a score of 1 plus another high-risk characteristic, in a 2:1 ratio, to chronic warfarin (n = 138) or LAA closure with the Watchman device plus warfarin for at least 45 days postimplantation (n = 269). Warfarin was discontinued when transesophageal echocardiography (TEE) documented either complete LAA closure or residual peridevice flow of less than 5 mm and no visible large thrombus on the device.
Patients’ mean age was about 74 years; baseline characteristics were similar between the treatment and control arms.

Warfarin Discontinued Per Protocol in Most Device Patients

The device was successfully implanted in 95.1% of patients in whom it was attempted (252 of 265). In this group, 92.2% were able to discontinue warfarin at 45 days and 99.3% by 12 months.

At 18 months, rates of the first primary efficacy endpoint (stroke, systemic embolism, and cardiovascular/unexplained death) were similar between the device and control groups, but the prespecified margin of noninferiority was not achieved. However, the rate of the late ischemic primary efficacy endpoint (stroke or systemic embolism more than 7 days after randomization) did meet the criterion for noninferiority (table 1).

Table 1. Coprimary Efficacy Endpoints at 18 Months


(n = 269) 

(n = 138)  


  Rate Ratio 
(95% CrI)


Main Efficacy Endpoint




Late Ischemic Endpoint 




CrI = credible interval

a 95% CrI upper bound < 1.75 for noninferiority
b 95% CrI upper bound < 0.0275 for noninferiority

Early safety events (all-cause death, ischemic stroke, systemic embolism, or device-/procedure-related events requiring open cardiovascular surgery or endovascular intervention within 7 days) occurred in 2.2% of the device arm, meeting the prespecified performance goal of 2.67%. Specifically, there were 2 cases of device embolization and 1 each of ateriovenous fistula, cardiac perforation, pericardial effusion with cardiac tamponade, and major bleeding requiring transfusion.

Importantly, implantation success and overall early safety were improved in PREVAIL and the CAP (Continued ACcess PROTECT AF) registry compared with PROTECT AF despite study populations in the former that were older, at higher risk, and had higher CHADS2 scores (table 2).

Table 2. Comparison of Implantation Success, Earlya Safety



CAP Registry


P Valueb

Implant Success





All Complications





Pericardial Effusion
Requiring Surgery





Pericardial Effusion with Pericardiocentesis





Procedure-Related Stroke





Device Embolization





a Within 7 days.
b For CAP registry and PREVAIL vs PROTECT AF.

By design in PREVAIL, at least one-quarter of device patients were treated by new operators, yet similar rates of successful implantation and complications were seen between new and experienced operators.

The authors say the failure of the Watchman to achieve noninferiority to warfarin for the first primary efficacy endpoint is due to wide confidence intervals. One potential explanation for this, they say, is “the substantially lower-than-expected number of events, particularly in the control group.”

“Taken in totality, the Watchman studies are the first to demonstrate, in a prospective randomized fashion, the utility of exclusion of the LAA in preventing stroke in patients with [nonvalvular A-fib] compared with warfarin,” writes Randall J. Lee, MD, PhD, of the University of California-San Francisco (San Francisco, CA), in an accompanying editorial.

He notes that the efficacy data from PROTECT AF, mortality data of the 5-year follow up, and improved safety results of the CAP registry and PREVAIL all contributed to an FDA panel voting 13 to 1 to approve the Watchman device in December 2013.

Not a Universal Replacement for Anticoagulation  

Though these studies provide proof of principle of the Watchman’s efficacy, Dr. Lee cautions, “LAA device implants should not be considered universally as a substitute for oral anticoagulation therapy.” The most vulnerable A-fib patients—those who are older, have had a previous stroke, or have renal dysfunction and thus are at increased risk of both ischemic and bleeding events—must be carefully evaluated, he observes. Until more data become available, the decision to treat these patients with LAA occlusion instead of oral anticoagulation “should be made on the basis of [their] overall health, ability to tolerate oral anticoagulation therapy, procedural risks, and the operator’s experience,” he concludes.

One important patient subset for which the proper strategy remains uncertain is those with contraindications to oral anticoagulants, Dr. Lee points out. Though results of the ASAP trial suggest the Watchman can be safely implanted in the presence of aspirin and clopidogrel rather than transitional warfarin, the study was small and observational, he notes.

Moreover, complications such as thrombus formation on the Watchman device suggest the need for ongoing surveillance, Dr. Lee says, recommending TEE at 45-days postprocedure and longer term if there is a leak or partial uncovered LAA lobe.

Finally, Dr. Lee cautions, if the Watchman is approved in the United States, “[i]dentifying operators with the necessary skillset to implant the… device will be paramount for the long-term success of LAA exclusion as a treatment for stroke [prevention] in [A-fib] patients,” he concludes.

Novel Oral Anticoagulants the Default Therapy?

In a telephone interview with TCTMD, Christopher Granger, MD, of Duke University (Durham, NC), said that while PREVAIL adds to the evidence that LAA occlusion with the Watchman is an effective strategy for preventing stroke in A-fib patients, the dataset and number of events remains relatively small, undermining confidence in its noninferiority.

Echoing the editorial, Dr. Granger also observed that the Watchman was tested against an inferior anticoagulant. “The newer oral anticoagulants have about a 10% lower death rate and up to a 35% lower stroke rate than warfarin, with less bleeding,” he said. Nonetheless, the Watchman trials remain relevant because warfarin will continue to be used for some time, he acknowledged. Moreover, a trial comparing the Watchman with a novel oral anticoagulant is very unlikely since it would require enrolling about 50,000 patients, he added.

“We’re left with a situation where if you want to have the most confidence that you’re preventing stroke effectively, you would use a novel oral anticoagulant,” Dr. Granger contended. The main exceptions, he said, are patients who either are unable or refuse to take long-term anticoagulation. LAA occlusion would also be a good choice for those with refractory bleeding, he suggested, although that has not been studied.

“I think the occluder is an important option,” Dr. Granger said, adding that he expected it to be approved “with some caveats and labeling.” But given the current limited database, he cautioned, “it should be reserved for selected populations.”


1. Holmes DR, Kar S, Price MJ, et al. Prospective randomized evaluation of the Watchman left atrial appendage closure device in patients with atrial fibrillation versus long-term warfarin therapy: the PREVAIL trial. J Am Coll Cardiol. 2014;64:1-12.

2.Lee RJ. Evolution of stroke prevention in nonvalvular atrial fibrillation patients. J Am Coll Cardiol. 2014;64:13-15.


Related Stories:



  • The PREVAIL trial was sponsored by Atritech/Boston Scientific.
  • Dr. Holmes reports having a financial interest in study-related technology, which has been licensed to Atritech.
  • Dr. Lee reports holding equity in and serving as a consultant to SentreHEART.
  • Dr. Granger reports receiving research funding from and serving as a consultant to multiple pharmaceutical companies.