The noninferiority trial of patients at risk for both bleeding and stroke, in a surprise turn, found standard care to be superior.
NEW ORLEANS, LA—For patients with atrial fibrillation (AF) who are at high risk of both stroke and bleeding, left atrial appendage occlusion (LAAO) isn’t noninferior to standard medical care including anticoagulants, results from the CLOSURE-AF trial show.
The study, as it turned out, went in a different direction than the trialists anticipated.
“I think the result was indeed somewhat unexpected for us, because this is exactly the patient population where we are now doing left atrial appendage closure,” said Ulf Landmesser, MD (Deutsches Herzzentrum Charité, Berlin, Germany), who presented the trial yesterday at the American Heart Association (AHA) 2025 Scientific Sessions. To TCTMD, he predicted that he will change his practice on the basis of CLOSURE-AF, given that it suggests “patients at high risk are probably not optimal candidates for this procedure.”
The current US guidelines for AF consider percutaneous LAAO to have a class 2b, level of evidence B, indication in this patient population, while the latest European guidelines give the therapy a class IIb, level of evidence C, recommendation.
“You can appreciate that the recommendation is not very strong, and clearly more evidence is needed,” he said at an AHA press conference, noting that this knowledge gap inspired the CLOSURE-AF trial.
More data are on the way in populations with different risk profiles, Landmesser highlighted. Trials of interest include CHAMPION-AF and CATALYST in lower-risk patients, as well as LAAOS-4 in very-high-risk patients.
Jeff Healey, MD (McMaster University, Hamilton, Canada), the discussant for CLOSURE-AF in the late-breaking session, noted that LAAO is widely used for AF patients, because while “oral anticoagulation is very effective, . . . discontinuation is common and often [occurs] in the setting of bleeding.” Thus far, however, there has been only “a modest amount of RCT data to guide practice,” he said, especially given that patients with a history of bleeding are often excluded from major trials.
With CLOSURE-AF, “the punch line here is that we actually saw an increase of the primary outcome of stroke, systemic embolism, [cardiovascular or unexplained mortality], and major bleeds in the occlusion arm,” said Healey, noting that this diverges from what might have been expected.
He cautioned that with this composite, “you don’t really have statistical power to understand” nuances related to the individual endpoints, though he pointed out that a quarter of the bleeding was periprocedural and most of the overall difference between groups was related to CV death.
The study “does add to the body of evidence . . . , but I think we do need the forthcoming trials, which are much larger,” Healey added.
CLOSURE-AF
CLOSURE-AF enrolled 912 patients (median age 79.1 years; 38.6% women) at 42 centers in Germany. All had a high risk of stroke (CHA2DS2-VASc score ≥ 2) and increased bleeding risk (HAS-BLED score ≥ 3, history of bleeding, or chronic kidney disease based on an eGFR of 15-29 mL/min/1.73m2). Mean scores were 5.2 for CHA2DS2-VASc and 3.0 for HAS-BLED.
Researchers randomized participants, stratified by center, to receive either LAAO (mostly with Abbott’s Amplatzer Amulet or Boston Scientific’s Watchman FLX) or physician-directed standard care (including anticoagulation, if eligible). The vast majority of patients assigned to standard care took a direct oral anticoagulant (DOAC). Initially, most patients in the LAAO arm were on dual antiplatelet therapy, but most switched to a single antiplatelet after 3 months. Follow-up lasted for a median of 3 years.
By 6 years, LAAO did not meet noninferiority criteria for the estimated cumulative incidence of the primary endpoint—a composite of stroke, systemic embolism, cardiovascular/unexplained death, or major bleeding (BARC ≥ 3)—compared with standard care (P for noninferiority = 0.44). Incidence was 16.83 per 100 person-years for LAAO and 13.27 per 100 person-years for standard care in an intention-to-treat analysis—amounting to an adjusted hazard ratio of 1.28 (95% CI 1.01-1.62) that significantly favored standard care.
Each of the individual components occurred at similar rates in the two groups, though the trial was not powered for those endpoints.
The benefit in terms of preventing stroke and preventing bleeding has to outweigh the risk of procedural complications. Andrew Goldsweig
Andrew Goldsweig, MD (University of Massachusetts-Baystate, Springfield), commenting on the new data for TCTMD, said the high-risk population enrolled in CLOSURE-AF captures what’s commonly seen in practice and mirrors the profile of patients who took part in contemporary trials like Amulet IDE, PRAGUE-17, and the ongoing CONFORM pivotal trial of the Claas closure device (Conformal Medical).
“It’s the population for whom left atrial appendage occlusion is approved and indicated,” said Goldsweig, who served as writing committee chair for the LAAO practice guidelines recently released by the Heart Rhythm Society and the Society for Cardiovascular Angiography and Interventions.
CLOSURE-AF isn’t a signal that there should be concern over LAAO, he noted. Rather, the issue is that the trial doesn’t reflect current practice—enrollment started in 2018 at a time when operators were less experienced and over the next 6 years used a mix of devices, including some that are no longer on the market. “If you were to do the trial with current devices today, you would have very different results,” Goldsweig predicted.
That’s not to say there are no messages for today’s practice, he specified. What it highlights is that when considering LAAO for an individual patient, “you’re trying to balance risks and benefits,” said Goldsweig. “The benefit in terms of preventing stroke and preventing bleeding has to outweigh the risk of procedural complications.”
For Goldsweig, longer follow-up would be revealing as well, given that bleeding events would be expected to continue to accrue in patients who stay on anticoagulation, putting standard care at a disadvantage compared with LAAO.
