Registry Data Reassure on Safety With Next-Generation Watchman Device

Procedural success and safety outcomes in real-world populations with nonvalvular atrial fibrillation (AF) undergoing left atrial appendage occlusion (LAAO) with the Watchman FLX device (Boston Scientific) align with those of patients in the pivotal trial, new data suggest.

Despite having higher thromboembolic and bleeding risk than patients enrolled in the PINNACLE FLX trial and being slightly older, the registry patients had an implantation success rate of 98% and 45-day all-cause mortality rate of 0.82%. Additionally, the key safety endpoint occurred in less than 0.5%.

Reporting the results in Circulation: Cardiovascular Interventions, Samir R. Kapadia, MD, (Cleveland Clinic, OH), and colleagues say the findings suggest “that patients treated in routine clinical practice are achieving [comparable] short-term outcomes as those observed in the pivotal PINNACLE FLX approval study.” In addition, the researchers say the similarities between the registry and trial patients highlight “that appropriate patients are being treated with proper expertise after the commercialization of the procedure.”

Commenting on the analysis for TCTMD, Mohamad Alkhouli, MD (Mayo Clinic, Rochester, MN), who wrote an accompanying editorial, was a little more cautious. He said the study reassures on one front: despite the high rates of uptake of the device in recent years, implant success and safety profiles are excellent and any potential indication creep is being appropriately confined to high-risk patients.

“It speaks to how many high-risk patients are out there and how this can evolve, but we still need to solve the efficacy data problem,” he said.

As he elaborates in his editorial, Alkhouli pointed out that efficacy data with this device remain scarce and said the current analysis cannot be said to help fill that gap.

“These safety data are great and tell us how much engineering refinements and technology improvement have lowered the complication rate,” he said. “But they still don't address the fact that we only have three randomized trials done several years ago, and of modest size, addressing the efficacy of the approach. That the device is safe is clear, and the safety has improved over time. It's just that the messaging that ‘this is showing us more evidence that the device prevents strokes’ is incorrect.”

Registry Results

The 97,185 patients (mean age 76.4 years; 41% women) with nonvalvular AF in the National Cardiovascular Data Registry’s LAAO postapproval registry were treated between August 2020 and September 2022 at 772 hospitals in the United States. The average CHA₂DS₂-VASc score was 4.8, and the mean HAS-BLED score was 2.8. Approximately 20% of the cohort had a prior history of stroke and 66% had a prior history of bleeding. Paroxysmal AF was present in 62.3%, persistent AF in 20.7%, permanent AF in 11.6%, and long-standing persistent AF in 5.4%.

The implantation success rate was 97.5%, with unsuitable anatomy accounting for most of the unsuccessful implants. In approximately 32% of patients, the device size used was 27 mm, while the 24-mm device was used in 31%.

The key safety endpoint—defined as all-cause death, ischemic stroke, systemic embolism, or device or procedure-related events requiring open cardiac surgery or major endovascular intervention between device implantation hospital discharge—occurred in 0.45% of patients. Of these, 0.22% were events requiring intervention and 0.16% were deaths.

The most common in-hospital adverse event was major bleeding, which occurred in 1.01%, followed by pericardial effusion requiring intervention in 0.39%. On postimplant imaging, the rate of peridevice leak > 5 mm was 0.02%. No peridevice leaks were noted at the conclusion of the procedure in 96.4%.

The most common medication regimen at discharge was direct oral anticoagulant (DOAC) plus aspirin in 47.5%, followed by DOAC alone in 24%.

At 45 days, the rate of all-cause death was 0.81%, with an ischemic stroke rate of 0.23%, major bleeding rate of 3.11% and a rate of pericardial effusion requiring intervention of 0.50%. Device-related thrombus occurred in 0.44%. More than 80% had no leak, and 99.6% had effective device closure. Of the leaks that did occur, the majority were in patients implanted with devices ≤ 3 mm in size.

At 1 year, the rate of all-cause death was 8.2%, with any stroke in 1.5% and ischemic stroke in 1.2%. Major bleeding occurred in 6.4%.

We still have to be vigilant that we don’t have a ton of efficacy data and we’re still waiting for it. Mohamad Alkhouli

In multivariate analysis, predictors of increased risk for early safety events included age, congestive heart failure, abnormal renal function, prior thromboembolic event, and prior clinically relevant bleeding.

Factors associated with an increased risk of death at 45 days and 1 year were age, baseline LV dysfunction, prior clinically relevant bleeding, CAD, diabetes, increased fall risk, chronic lung disease, congestive heart failure, and abnormal renal function.

Prior stroke, age, hypertension, diabetes, and prior transient ischemic attack or thromboembolic event were associated with increased risk of stroke events at 45 days. Prior stroke remained a predictor of that event at 1 year, while added risk also was seen among patients with cardiomyopathy, chronic lung disease, and increased fall risk.

At 45 days and 1 year, age, diabetes, chronic lung disease, congestive heart failure, abnormal renal or liver function, and prior bleeding were associated with increased bleeding risk.

Compared with the PINNACLE FLX results, the key safety endpoint in the registry was similar, with lower rates of ischemic stroke at both 45 days and 1 year as well as a lower rate of major bleeding at 1 year. However, the registry patients had higher rates of all-cause death at both time points.

Efficacy Data on the Horizon

Kapadia and colleagues note that since the LAAO registry relies on site-reported data, there may be over- or underreporting of patient, hospital, or physician data. Additionally, with only 75% of registry patients having had transesophageal echocardiography or CT at the 45-day follow-up, there may be underreporting of device-related thrombus.

In his editorial, Alkhouli says this does appear to be the case, noting that, for example, the 0.44% rate of device-related thrombus at 45 days is significantly lower than historical data and “could be due to missing data (only 75% had imaging data at 45 days), underreporting of device-related thrombus, follow-up imaging obtained too soon, or an actual improvement in the device-related thrombus rate over time.”

To TCTMD, he said ongoing trials hope to fill the evidence gaps, particularly in understanding the efficacy of LAA closure devices in contemporary clinical practice.

The bottom line, Alkhouli added, is that “we still have to be vigilant that we don’t have a ton of efficacy data and we’re still waiting for it.”