AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Atrial Fibrillation Pattern and Competing Outcomes After Left Atrial Appendage Occlusion: Persistent Disease Marks Mortality Risk
By Singh A, Bandaru RR, Singh V et al.
Introduction: Atrial fibrillation (AF) pattern is often treated as a single axis of disease severity, with more sustained AF assumed to confer uniformly higher risk. Whether paroxysmal, persistent, and permanent AF differentially shape the competing risks of bleeding, thromboembolism, cardiac complications, and death after left atrial appendage occlusion (LAAO) remains poorly defined. Methods: We studied 2,750 patients undergoing LAAO at 8 centers participating in the National Cardiovascular Data Registry LAAO Registry. We fit a multistate survival model with four outcome states (major bleeding, stroke/transient ischemic attack/systemic embolism, cardiac complications, and death), using flexible parametric models on the log cumulative hazard scale with spline-based baseline functions to estimate 2-year risks according to AF pattern (paroxysmal, persistent, and permanent). To enable fair comparisons, we applied regression standardization, calculating each pattern’s 2-year risk as if all patients shared the same distribution of CHA ₂ DS ₂ -VASc scores and other clinical covariates observed in the full cohort, so that residual differences reflect AF pattern rather than case mix. We then estimated absolute risk differences relative to paroxysmal AF, with corresponding 95% confidence intervals. Results: AF pattern differentiated patients only with respect to mortality, and only for persistent AF. Patients with persistent AF had higher 2-year mortality than those with paroxysmal AF (9.01%, 95% CI 6.76-11.95 vs 5.72%, 95% CI 4.79-6.82; risk difference +3.29 percentage points, 95% CI +0.53 to +6.05), while risks of major bleeding, stroke/systemic embolism, and cardiac complications did not differ significantly (all confidence intervals included zero). Permanent AF was not significantly associated with any outcome, including mortality (2-year mortality 6.05%; risk difference vs paroxysmal AF +0.33 percentage points, 95% CI –2.40 to +3.06), indicating that the excess mortality signal was confined to persistent AF. Conclusions: After LAAO, AF pattern does not behave as a simple ordinal severity gradient. Persistent AF specifically marks higher 2-year mortality without measurable increases in stroke, bleeding, or cardiac complications, whereas permanent AF is not associated with excess risk across outcomes. These findings support persistent AF as a distinct prognostic phenotype that may warrant closer postprocedural surveillance and follow-up.