SAN FRANCISCO — Thirteen months after percutaneous left atrial appendage closure (LAAC), outcomes did not differ whether closure was performed using the Amulet (Abbott Cardiovascular) or the Watchman FLX (Boston Scientific) devices.
The SWISS-APERO trial randomly assigned patients undergoing clinically indicated LAAC whose LAA anatomy was deemed suitable to receive either device.
Rates of residual LAA patency and its subtypes at 13-month coronary computed tomography angiography (CCTA); rates device-related thrombus (DRT) at 13-month CCTA; and rates of clinical outcomes at 13-months did not differ between the randomized device groups, reported Roberto Galea, MD, Bern University Hospital, Bern, Switzerland at the Transcatheter Cardiovascular Therapeutics 2023 congress. The paper was also published online simultaneously in Circulation.
Watchman FLX and Amulet are the two most frequently used devices for LAAC worldwide, and the SWISS-APERO trial is the first randomized controlled trial comparing them, he said. LAAC is meant to reduce thrombus formation and stroke risk in patients with non-valvular atrial fibrillation.
A previous report from the SWISS-APERO trial with a 45-day time frame showed similar results. The current report extends the earlier findings by including prespecified subanalyses of the trial at 13 months.
Patients in the SWISS-APERO trial (N = 221) had clinical indications for LAAC. Eligibility included a high bleeding risk, LAA morphology suitable for either device, and no LAA thrombus. They were randomly assigned to the Amulet device (n = 111) or to Watchman FLX (n = 110).
During the procedure, they underwent transesophageal echocardiography (TEE) and intracardiac echocardiography (ICE). Forty-five days post-procedure they were evaluated with TEE and CCTA and then with CCTA at 13 months to evaluate long-term LAA patency; 84 Amulet patients and 80 Watchman FLX patients were evaluable. Aspirin plus clopidogrel or a direct oral anticoagulant were prescribed for 45 days post-procedure and evaluated at that time.
Patients had a mean age of 76.9 years, 70.6% were male, 39.3% had a prior cerebrovascular event, and 87.8% had prior relevant bleeding (cohort CHA2DS2-VASc score 4.3, HAS-BLED score 3.1).
The primary endpoint at 13 months was residual LAA patency and related subtypes, DRT by CCTA, and clinical outcomes — namely, all-cause or cardiovascular death, all stroke, systemic or pulmonary embolism, and bleeding according to Bleeding Academic Research Consortium (BARC) classification. At discharge, at 45 days, and at 13 months, the Amulet and WATCHMAN FLX groups were equivalent in terms of proportions on each type of antithrombotic therapy or none.
No Difference in LAA Patency at 13 Months
At 13 months, the Amulet and Watchman FLX groups did not differ in LAA patency or the subtypes of leaks. Comparing these results with the previously published 45-day results, Galea said, "One thing that was interesting at that time was that the leak rates as identified by CT were about 70% in both groups, plus minus. And now we see that they have improved at the 1-year time point to about 50%, and relatively similar between the groups, still relatively high at 50%."
Overall patency for Amulet was 53.6% vs 48.8% for Watchman FLX (P =.537). The most frequent leak subtype was peri-device or mixed (28.6% vs 27.5%, respectively), followed by intradevice (23.8% vs 17.5%), or patent appendage with no visible leak (7.1% vs 12.5%) (all P >.20).
DRT rates were 1.2% vs 1.3%, respectively (P =.972), and the rates of definite or possible DRT also did not differ (2.4% vs 3.8%; P =.610).
Of patent appendages at 45 days, 30.5% resolved by 13 months. Galea reported that only 13.2% of appendages became patent after 45 days.
There were also no significant differences in clinical outcomes at 13 months. The composite outcome of cardiovascular death, stroke, or systemic embolism was 9.5% in the Amulet group and 10.2% for Watchman FLX (P = .829) and, similarly, no differences in any of the components of the composite outcome.
Major bleeding (BARC 3-5) occurred in 16.5% of patients in the Amulet group and 9.3% of the Watchman group (P = .122)
Galea noted several limitations of the study. First, it was open label, since the two devices can easily be distinguished during CCTA assessment. Second, the trial was not powered to show differences in clinical endpoints. Third, several patients received Watchman 2.5 devices, a version preceding Watchman FLX. And finally, because the study involved a high-risk population, a "relevant percentage" died before the 13-month assessment.
After Galea's presentation, Vivek Reddy, MD , Mount Sinai Heart Health System, New York City, provided expert commentary, noting first that "high-quality trials" like SWISS-APERO are needed because of a lack of randomized data concerning LAAC.
By way of background, Reddy said that in LAAC leaks "matter quite a bit," and high-quality studies have shown that "leaks approximately double the rate of ischemic stroke. And conversely, [if] you don't have a leak, your annualized ischemic stroke rate is on the order of only 1.2, or 1.3%. It's quite remarkable." Therefore, trying to minimize and eliminate leaks is very important, even for small leaks, he said.
The SWISS-APERO patients had high CHA2DS2-VASc scores and high previous bleeding rates, constituting a high-risk population. In the current trial, Reddy said 78% of the Watchman patients had received the updated FLX device, and they experienced a "still relatively high" but improved 50% leak rate, similar to the Amulet group, compared with the 70% rate seen in the 45-day results.
Reddy pointed to the trial's initial 45-day results showing leaks not just on the primary phase of CT imaging but also leaks on the secondary phase (ie, on the venous phase of imaging), which were apparent about 90% of the time by CT, suggesting almost universal leaks. He asked Galea if he had done venous phase imaging with CT at the 1-year point, to which Galea replied that they did, "and still the percentage is quite high."
He said the investigators are discussing the clinical importance of this finding in the venous phase compared with the leak observed at the arterial phase. They intend to follow the patients for 5 years.
Moderator David Cohen, MD, of St. Francis Hospital in Roslyn, New York, referred to other data showing leaks are associated with excess thromboembolic events, and those data were shown using TEE, a less sensitive imaging method that shows lower rates of leak than CT does. "Do you think [CT] is too sensitive, perhaps, because we're finding leaks that don't matter clinically?" he asked Galea.
"This is a very good question," Galea replied. "I personally think that with a CT, we see too much, and we are seeing some difficulties, difficulties to interpret what we see. With a TEE, we have clear definitions of failure device leak, and there is consensus about that. Unfortunately, we don't have still clear definitions of device with a thrombus and leak as evaluated by CT."
He suggested development of a consensus document about the definitions to use and then a prospective study properly sized to assess clinical events.
Acknowledging that the study was under-powered for hard clinical endpoints, panelist Brian Whisenant, MD, of Intermountain Medical Center Heart Institute in Murray, Utah, asked if there was any association between leaks and stroke or DRT and if that analysis was done.
Galea answered that the investigators, in comparing the leaks at 45 days with 1-year outcomes, did not find any association. "But let me add that we had high mortality at 1 year, so probably we lost a bit of power…So probably we're going to test the 1-year leak as predictors of thromboembolic events during the follow up."
Reddy had the final word, addressing the issue of what is a good endpoint when it comes to leaks. Referring to some published abstract data showing that any amount of leak increases the risk of stroke, he suggested considering leak "not just as the overt leak that you see flow, but even leak because there's an echo-free space behind the device, which obviously means there's a covert leak."
He predicted that if one defines leak on TEE as any atrophy space, then the incidence of leak will be similar to CT, "and they're both bad. So, in some ways, I guess you can look at that as a little disappointing/depressing because that means we have a lot of leaks. On the other hand, the flip side is that also means that if we can occlude the appendage without having any leaks, we're going to improve our outcomes even better than we have."
The trial was investigator-initiated. The study sponsor, Insel Gruppe AG, Universitätsklinik für Kardiologie, Bern, Switzerland conducted the study, supported by local funding and a research grant from Abbott, which was not involved in any study processes.
Galea reports no relevant financial relationships. Reddy had the following disclosures: equity/stock(s)/options in Endobar Solutions LLC; grant support/research contract from Abbott Vascular and Boston Scientific Corporation; and is on the scientific advisory board of AtaCor Medical. Cohen has received grant support/research contract from Boston Scientific Corporation, Edwards Lifesciences, Abbott, Corvia, Ancora Heart, Cardiac Dimensions, Philips, Zoll, and Irhythm; and consultant fee/honoraria/speaker's bureau fees from Medtronic, Boston Scientific Corporation, and Abbott. Whisenant has received consultant fee/honoraria/speaker's bureau fee from Boston Scientific Corporation, Abbott, Edwards Lifesciences, Biosense Webster, and Medtronic.