Highlights
Integrated Outcomes
Using a hierarchical win ratio analysis that integrated death, stroke, and major bleeding, anticoagulation with edoxaban showed no clinical advantage over a no-anticoagulation strategy in patients with device-detected atrial fibrillation (DDAF).
Dominance of Undecided Pairs
Approximately 84.9% of the patient comparisons in the study were undecided, reflecting the relatively low event rates and the high number of patients who remained event-free during the 21-month median follow-up.
Bleeding vs. Protection
While anticoagulation is standard for ECG-diagnosed AF, this study reinforces the notion that for shorter, device-detected episodes, the risk of major bleeding may offset any potential reduction in thromboembolic events.
Background: The Dilemma of Device-Detected Atrial Fibrillation
The widespread use of cardiac electronic implantable devices (CIEDs), such as pacemakers and implantable cardioverter-defibrillators (ICDs), has led to the frequent identification of subclinical or device-detected atrial fibrillation (DDAF). These episodes, often asymptomatic and brief, present a significant clinical conundrum. While clinical atrial fibrillation (detected via standard 12-lead ECG) is a well-established risk factor for stroke and requires oral anticoagulation (OAC) in patients with elevated CHA2DS2-VASc scores, the evidence for treating DDAF has been less clear.
Traditional clinical trials often use a composite primary endpoint, typically “time-to-first-event.” However, this approach treats all events—whether a minor stroke, a fatal intracranial hemorrhage, or all-cause death—as having equal weight in the statistical analysis. Clinically, however, these events carry vastly different implications for patient prognosis and quality of life. To address this, researchers have increasingly turned to the “win ratio” method, a hierarchical analysis that prioritizes the most severe clinical outcomes.
Study Design: A New Look at NOAH-AFNET 6
The NOAH-AFNET 6 (Non-vitamin K antagonist Oral anticoagulants in patients with Atrial High-rate episodes) trial was originally designed to assess whether edoxaban could reduce stroke, systemic embolism, or cardiovascular death in patients with DDAF. The primary results previously indicated that edoxaban did not significantly reduce these events but did increase the risk of major bleeding. To provide a more nuanced understanding, researchers performed a post-hoc unmatched win ratio analysis of the trial dataset.
Patient Population and Methodology
The analysis included 2,534 patients with a mean age of 77 years. The population was high-risk, with a median CHA2DS2-VA score of 3. These patients had device-detected atrial high-rate episodes (AHRE) lasting at least 6 minutes but no history of ECG-diagnosed AF. The win ratio analysis utilized a hierarchical order of importance for outcomes:
- 1. All-cause death
- 2. Stroke
- 3. Systemic or pulmonary embolism or myocardial infarction
- 4. Major bleeding
In this framework, every patient in the edoxaban group is compared to every patient in the no-anticoagulation (placebo or standard care) group. A “win” is recorded for the group with the better outcome for the highest-priority event. If the highest-priority event is a tie (e.g., neither patient died), the analysis moves to the next event in the hierarchy. If no event occurs in either patient, the pair is considered “undecided.”
Key Findings: The Results of the Win Ratio Analysis
The analysis involved over 1.6 million win ratio pairs. The results were consistent across several hierarchical models. The primary win ratio for edoxaban versus no anticoagulation was 0.87 (95% CI: 0.68–1.10; P = 0.23). Because a win ratio below 1.0 suggests a trend toward the control group (no anticoagulation), this indicates that edoxaban failed to demonstrate superiority.
Distribution of Wins and Losses
The most striking finding was that 84.9% of all comparisons resulted in an undecided outcome, meaning the vast majority of patients did not experience any of the primary endpoints during the median follow-up of 21 months. In the remaining 15.1% of pairs where a winner could be determined, edoxaban won in 46% of cases, while the no-anticoagulation group won in 54%.
The Impact of Death and Bleeding
All-cause death and major bleeding were the most frequent drivers of the analysis. Because death was the highest-priority event and occurred at similar rates in both groups, it often “neutralized” the potential benefits of stroke reduction. Furthermore, the increase in major bleeding events in the edoxaban arm contributed to “losses” in the hierarchy, further dragging the win ratio below the threshold of significance.
Secondary Sensitivity Analyses
The researchers also conducted analyses replacing all-cause death with cardiovascular death and another including patient-reported outcomes (quality of life). Even with these adjustments, the win odds remained 0.98 (95% CI: 0.94–1.01; P = 0.23), reinforcing the lack of a clear benefit for edoxaban in this population.
Expert Commentary: Interpreting the Data for Clinical Practice
The NOAH-AFNET 6 win ratio analysis provides a critical perspective on the management of subclinical AF. Experts note that the stroke risk in DDAF appears substantially lower than that observed in patients with clinical AF. While the CHA2DS2-VASc score is a powerful predictor in clinical AF, its utility may be attenuated in the context of very brief device-detected episodes.
The Comparison with ARTESiA
It is essential to view these results alongside the ARTESiA trial, which looked at apixaban in a similar DDAF population. ARTESiA did show a reduction in stroke with apixaban, though it also showed an increase in major bleeding. The difference in results between NOAH-AFNET 6 and ARTESiA may stem from differences in patient characteristics, trial design, or the specific anticoagulant used. However, the NOAH-AFNET 6 win ratio analysis suggests that when we weigh the severity of all events—including the potential for fatal or life-altering bleeds—the “net” benefit of anticoagulation in DDAF is marginal at best for the average patient.
Clinical Considerations
For clinicians, these findings suggest a more cautious approach. Rather than automatically initiating OAC for any device-detected AHRE over 6 minutes, a personalized approach is required. This should involve assessing the total burden of AF (duration and frequency), the patient’s overall thromboembolic risk, and their specific bleeding risk. Current European Society of Cardiology (ESC) guidelines have already moved toward a more nuanced recommendation, often suggesting a higher threshold of AF duration (e.g., >24 hours) or very high stroke risk scores before starting OAC in subclinical AF.
Conclusion: A Call for Precision Medicine
The win ratio analysis of the NOAH-AFNET 6 trial underscores the complexity of treating device-detected atrial fibrillation. By integrating multiple clinical outcomes into a single hierarchical framework, the study demonstrates that the current standard of anticoagulation does not offer a clear advantage for these patients. The high proportion of undecided pairs suggests that for many, the risk of any major event is low, making the potential harm of anticoagulation-induced bleeding a primary concern.
Future research should focus on identifying specific subgroups within the DDAF population who might still benefit from therapy—perhaps those with the highest AF burden or those with additional imaging markers of atrial cardiomyopathy. Until then, the “wait and see” approach, coupled with close monitoring for progression to clinical AF, remains a scientifically supported strategy for many patients with device-detected episodes.
References
1. Becher N, Köllner G, Bertaglia E, et al. Effects of anticoagulation in patients with device-detected atrial fibrillation and multiple stroke risk factors: a win ratio analysis of the NOAH-AFNET 6 trial. Eur Heart J Qual Care Clin Outcomes. 2025;11(8):1351-1358. doi:10.1093/ehjqcco/qcaf087.
2. Kirchhof P, Blank BF, Calvert M, et al. Prothrombotic and bleeding events in patients with atrial high-rate episodes: the NOAH-AFNET 6 trial. Eur Heart J. 2023;44(41):4351-4364.
3. Healey JS, Lopes RD, Granger CB, et al. Apixaban for Stroke Prevention in Subclinical Atrial Fibrillation (ARTESiA). N Engl J Med. 2024;390(2):107-117.
4. Hindricks G, Potpara T, Dagres N, et al. 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS). Eur Heart J. 2021;42(5):373-498.