NEW ORLEANS — Two multicenter trials suggest less may be more for stroke prevention in patients with atrial fibrillation (AF).

The larger of the two trials, called OCEAN, which evaluated prolonged antithrombotic therapy to prevent stroke after catheter ablation in patients with AF, was stopped early by the data safety monitoring board (DSMB) for futility. In the second, CLOSURE-AF, left atrial appendage closure (LAAC) fell short when compared with the best medical care for reducing embolic events in older patients with AF and a high risk for stroke and bleeding.

The studies were presented in the same late-breaking clinical science session at the American Heart Association (AHA) Scientific Sessions 2025. The OCEAN trial was simultaneously published in The New England Journal of Medicine.

OCEAN: Risk for Events Low After Ablation

The OCEAN trial compared the direct-acting oral anticoagulant rivaroxaban with aspirin with a composite endpoint that included covert stroke, explained lead investigator Atul Verma, MD, cardiac electrophysiologist affiliated with the Centre for Outcomes Research and Evaluation at McGill University Health Centre in Montreal, Québec, Canada.

In this trial, the problem was not relative efficacy but that the annualized rates of stroke, systemic embolism, and covert stroke proved to be so low in both arms after successful ablation for AF that there was little room to show a difference, according to Verma.

The researchers had randomly assigned 1284 patients who had undergone successful catheter ablation at least 1 year earlier when the DSMB recommended stopping the trial on the basis of a likely lack of benefit from rivaroxaban relative to aspirin, according to Verma. The DSMB also cited an interim analysis showing a higher rate of bleeding in the rivaroxaban arm.

To qualify for inclusion, patients had to have a CHA2DS2-VASc score of at least 1 or at least 2 for women or patients in whom vascular disease was a risk factor. Patients were randomly assigned to 15 mg of rivaroxaban or 70-120 mg of aspirin administered once daily at 56 sites in six countries. They underwent brain MRI at baseline and at 3 years to capture covert strokes. Outcomes were adjudicated by an independent blinded committee.

The primary composite outcome of stroke, systemic embolism, or new covert embolic stroke, defined as an asymptomatic infarct measuring at least 15 mm on MRI, occurred in just five patients in the rivaroxaban group and nine patients in the aspirin group, Verma reported. He noted that 96% of patients had no new infarct of any size over the 3 years of follow-up.

The annualized risks for the primary outcome were 0.31% and 0.66% for rivaroxaban and aspirin, respectively (P = .28). An analysis of the individual components did show a 28% relative risk reduction for all stroke or systemic embolism in favor of rivaroxaban, but this finding did not approach statistical significance (hazard ratio [HR], 0.72; 95% CI, 0.23-2.25; P = .28).

Fatal and major bleeding, the primary safety outcome, did not differ significantly with rivaroxaban compared with aspirin (1.6% vs 0.6%; HR, 2.51; 95% CI, 0.79-7.95), though researchers observed more clinically relevant but non-major bleeding in the rivaroxaban group (5.5% vs 0.6%; HR, 3.51; 95% CI, 1.75-7.03).

In anticipation of low rates of clinical strokes, covert strokes were included in the composite endpoint with the justification that these contribute to cognitive decline in patients with AF, according to Verma. Only two occurred, both in the aspirin arm, with an annualized risk of only 0.08%.

The rate of events was far lower than predicted, thwarting the power calculations that drove the study design. As a result, Verma said one take-home from this study is that the endpoints evaluated in this study are “very low after successful AF ablation.”

CLOSURE-AF: LAAC Falls Short

CLOSURE-AF was a randomized noninferiority trial comparing LAAC with best medical care, including oral anticoagulation with a direct-acting oral anticoagulant if available, in a high-risk population. The composite endpoint included both embolic events and bleeding, but both types of events were more common in the LAAC arm to a degree that the prespecified margin of noninferiority was missed, reported Ulf Landmesser, MD, chairman of the Department of Cardiology, Angiology and Intensive Care Medicine at Charité University Medicine in Berlin, Germany.

For the trial, Landmesser and colleagues enrolled patients with AF and high-risk characteristics. These included a CHA₂DS₂-VASc score ≥ 2 plus at least one of three additional factors — a history of bleeding, reduced renal function, or a HAS-BLED score ≥ 3 — at 42 healthcare sites in Germany.

Researchers included 888 patients in the primary analysis (median age, 79 years). The primary composite endpoint was stroke, systemic embolism, cardiovascular or unexplained death, or major bleeding, defined as Bleeding Academic Research Consortium type 3-5. Patients were followed for a median of 3 years.

In the intention-to-treat analysis, the rates of the primary composite endpoint were 16.83% in the LAAC arm and 13.27% in the physician-directed best medical care arm. The 28% increase in the LAAC arm did not include the line of unity and did not remain within the predefined margin of noninferiority to best medical care (HR, 1.28; 95% CI, 1.01-1.62).

When endpoints were evaluated individually, the rates of stroke were the same, about 2.7 per 100 patient-years, but all other secondary endpoints were numerically more common in the LAAC group, including systemic embolism (0.29 vs 0.1 per 100 patient-years), major bleeding (7.43 vs 6.23 per 100 patient-years), cardiovascular or unexplained death (9.47 vs 7.75 per 100 patient-years), and all-cause death (14.83 vs 13.49 per 100 patient-years).

On the basis of available evidence, the latest US and European guidelines have conferred LAAC with a class 2b recommendation, meaning that evidence is limited even if this intervention might be useful in some patients, Landmesser explained.

According to Jeff S. Healey, MD, director of arrhythmia services, McMaster University in Hamilton, Ontario, Canada, the results of the CLOSURE-AF trial are not likely to alter these recommendations.

While he conceded that the trial results “add randomized controlled data,” Healey said that larger studies are underway. He also expressed concern about the decision to compare LAAC with medical care with a composite endpoint that included both efficacy and bleeding endpoints, noting that weighting one outcome relative to others makes the results difficult to interpret.

Landmesser emphasized that due to the enrollment criteria, these results are only relevant to older patients with AF who have a high risk for both stroke and bleeding. Yet, he acknowledged that the results were surprising and disappointing because this is the population in which LAAC is most often used in his center.

The OCEAN trial reported receiving financial support from Bayer. Verma reported having financial relationships with Abbott, Adagio, Biosense Webster, Medtronic, and Volta Medical. CLOSURE-AF was investigator-initiated without industry funding. Landmesser reported having financial relationships with Abbott, Amgen, Aventis, Menarini, Merck Sharp & Dohme, Novartis, and Sanofi. Healey reported having financial relationships with Boston Scientific, Bristol Myers Squibb, Medtronic, and Pfizer.