Background

Prior analyses of trials comparing direct oral anticoagulants (DOACs) to warfarin in atrial fibrillation (AF) have not routinely incorporated patient preferences, despite substantial variation in how patients value the trade-off between outcomes such as stroke and bleeding. By applying patient-centered approaches, we aimed to provide intuitive metrics to inform shared decision-making, particularly for frail older adults for whom DOAC benefit remains controversial.

Methods

Individual-level data from 58,634 participants in four randomized controlled trials (RCTs) comparing DOACs to warfarin (A Collaboration Between Multiple Institutions to Better Investigate Non–Vitamin K Antagonist Oral Anticoagulant Use in Atrial Fibrillation; COMBINE-AF) were analyzed using two patient-centered methods. Seven clinical outcomes (death, disabling stroke, major bleeding, moderate-severity stroke, systemic embolism, clinically relevant non-major bleeding, and minor stroke) were weighted based on a prior 1028-patient preference study with all values scaled relative to death. For the weighted composite endpoint (WCE), a survival-based approach incorporated weights of initial and recurrent events to estimate event-free survival. For win statistics, outcomes were hierarchically ranked for pairwise comparisons. The primary estimand was the 2-year difference in weighted death-equivalent events per 100 patients for the WCE. The win ratio was a secondary estimand. A prespecified subgroup analysis was conducted in frail, older patients.

Results

In the overall cohort, compared to warfarin, DOACs were associated with a more favorable outcome (WCE: 11.74 vs. 12.85 events per 100 patients; difference, −1.11 [95% confidence interval (CI): −1.61 to −0.61]; P<0.001; win ratio 1.11 [95% CI: 1.07 to 1.15]). In the prespecified subgroup of 5913 frail participants, the difference in the WCE was +0.50 events [95% CI: −1.39 to 2.40]) with a win ratio of 0.99 [95% CI: 0.90 to 1.08]) in individuals treated with DOAC versus warfarin.

Conclusions

In individuals with atrial fibrillation pooled from four RCTs, DOACs were associated with a favorable net clinical benefit compared to warfarin when evaluated using a patient-weighted composite clinical outcome.