Withdrawal of heart failure therapy after atrial fibrillation rhythm control with ejection fraction normalization: the WITHDRAW-AF trial

Abstract Background and Aims Atrial fibrillation-mediated cardiomyopathy (AFCM) represents an important reversible cause of left ventricular systolic dysfunction. Current clinical practice is indefinite heart failure (HF) pharmacotherapy despite left ventricular ejection fraction (LVEF) normalization. However, whether this is necessary to maintain normal LVEF, in addition to rhythm control, is uncertain. Methods This multi-centre, randomized trial conducted between 2021 and 2024 examined the impact of staged withdrawal of HF therapy following AF rhythm control and LVEF normalization in AFCM. Participants were randomized (1:1) to early withdrawal (Group A) or continued therapy for 6 months followed by delayed withdrawal (Group B), in a crossover design. The primary endpoint was the randomized comparison of cardiac magnetic resonance (CMR) LVEF maintenance ≥50% at 6 months, during which time Group A had withdrawn therapy and Group B remained on treatment. Secondary outcomes included cardiac remodelling, functional status, biomarkers, quality of life, and arrhythmia recurrence on vs off HF therapy. The total follow-up duration was 12 months. Results Between July 2021 and May 2024, 60 patients were enrolled (age 60 55–65 years, previous persistent AF 1 year and maintaining sinus rhythm for minimum 6 months following AF rhythm control catheter ablation in 97%). All participants completed treatment withdrawal and 12-month follow-up. In the initial randomized comparison, LVEF was maintained ≥50% at 6 months in 90% of participants undergoing HF therapy withdrawal (Group A), compared with 100% who continued medical therapy (Group B) (odds ratio OR 1.18, 95% confidence interval CI 0.27–2.82, P = .47). CMR LVEF was similar between randomization groups at the end of the randomization phase (Group A: LVEF 58% 95% CI 54–60 vs Group B: LVEF 59% 95% CI 55–64, P = .236) and across study time points (mixed effects P = .37). Transthoracic echocardiography characteristics, N-terminal pro-B-type natriuretic peptide, functional status, quality of life and AF burden were similar on vs off HF therapy in the overall population. Conclusions Withdrawal of HF therapy following AF rhythm control for prior AFCM and recovered LVEF was not associated with a decline in LVEF for most patients in the following 6 months.