Risks of discontinuation of anticoagulation after atrial fibrillation ablation in high-risk patients

This editorial refers to ‘Clinical Outcomes of Continuing Versus Discontinuing Oral Anticoagulation After Atrial Fibrillation Ablation: An Updated Systematic Review and Meta-Analysis of Observational Studies and Randomized Controlled Trials’ by S.S. Nayak et al., https://doi.org/10.1093/europace/euag104 The question of whether oral anticoagulation (OAC) can be safely discontinued after apparently successful catheter ablation for atrial fibrillation (AF) is emerging as one of the most clinically important and challenging issues in contemporary clinical management of AF patients. AF is the most common sustained supraventricular arrhythmia and remains a major cause of thromboembolic events, particularly ischaemic stroke.1 Thromboembolic risk is determined by the cumulative burden of clinical factors that are currently stratified using the CHA2DS2-VA score, which has replaced the CHA2DS2-VASc score in the 2024 European Society of Cardiology (ESC) Guidelines.2–5 OAC, preferably with direct oral anticoagulants (DOACs), is recommended long-term in patients with a CHA2DS2-VA score ≥2 (Class I recommendation) and should be considered in those with a score of 1 (Class IIa recommendation), while it is generally not indicated for long-term use in patients with a score of 0 except for temporary indications, such as the peri-cardioversion or peri-ablation period.2,3,6 Catheter ablation has become a cornerstone of rhythm-control therapy for AF, with pulmonary vein isolation representing the standard procedure. More recently, technological advances such as pulsed-field ablation have simplified procedural workflows and improved safety profiles. Depending on AF type and the intensity of rhythm monitoring employed, long-term freedom from documented AF recurrence can be achieved in approximately 60–90% of patients with paroxysmal AF and 40–65% of those with persistent AF.7–9 As catheter ablation has become increasingly widespread, a clinically relevant question has emerged: Can OAC be safely discontinued in patients who appear to achieve durable rhythm control after a successful procedure? In this context, evidence derived from randomized clinical trials and comprehensive meta-analyses is of paramount importance to guide clinical decision-making. In this issue of Europace, Nayak and colleagues provide an updated meta-analysis that offers a timely synthesis of the available evidence.10 The authors performed a systematic review of major databases and included studies enrolling adult patients undergoing AF catheter ablation that compared outcomes between patients who continued vs. discontinued OAC following a successful AF ablation. The primary outcomes included thromboembolic events [with a focus on stroke and transient ischaemic attack (TIA)], major bleeding, and all-cause mortality. Several predefined subgroup analyses were performed, including stratification according to baseline thromboembolic risk. Among 7560 screened records, 31 studies met inclusion criteria, including 28 observational studies and 3 randomized controlled trials, encompassing more than 50 000 patients overall. Of these, 31 509 continued OAC after ablation, whereas 18 818 discontinued OAC treatment. The overall analysis showed that OAC continuation was not significantly associated with a lower risk of stroke or TIA [odds ratio (OR) 1.48, 95% confidence interval (CI) 0.84–2.60; P = 0.16], although substantial between-study heterogeneity was observed. However, the most clinically relevant finding emerged from the subgroup analysis according to thromboembolic risk. In patients with a CHA2DS2-VASc score ≥2, continuation of OAC was associated with an approximately 50% reduction in the risk of stroke or TIA, with an absent between study heterogeneity (I2 = 0%). Conversely, among patients with CHA2DS2-VASc scores ≤1, OAC continuation was not associated with a significant reduction in thromboembolic events. As expected, anticoagulation was associated with a higher risk of bleeding, while no significant differences were observed for all-cause mortality or major adverse cardiovascular events.10 These findings are broadly consistent with current ESC recommendations2 that explicitly report ‘Continuation of oral anticoagulation is recommended after AF ablation according to the patient’s CHA2DS2-VA score, and not the perceived success of the ablation procedure, to prevent ischaemic stroke and thromboembolism’.2,6 However, despite this class I recommendation, with a level of evidence C, a gap exist between guideline-directed management and real-world practice. As a matter of fact, OAC discontinuation after apparently successful AF ablation remains common, often driven by the assumption that successful rhythm control translates into elimination of thromboembolic risk. Actually, the results of this meta-analysis suggest that, in patients at high thromboembolic risk, discontinuation of anticoagulation may be unjustified and potentially not supported by current evidence. How should these results be interpreted considering the two contemporary randomized trials that were recently published, ALONE-AF and OCEAN?1,11.1 Several considerations are warranted in evaluating the impact of these trials, whose characteristics and results are summarized in Table 1. Characteristics of RCT investigating discontinuation of OAC therapy after successful AF ablation AF, atrial fibrillation; IQR, interquartile range; LVEF, left ventricular ejection fraction; n, number; SD, standard deviation; TIA, transient ischaemic attack. First, both trials enrolled predominantly low-risk populations. Only approximately one-third of participants had a CHA2DS2-VASc score ≥3, thereby limiting the number of expected thromboembolic events (Table 1). Consequently, both studies were underpowered to detect meaningful differences in stroke outcomes. In ALONE-AF, only nine primary outcome events occurred over two years of follow-up, with one event in the OAC discontinuation group and eight in the OAC continuation group (0.3% vs. 2.2%; absolute difference −1.9%, 95% CI −3.5 to −0.3; P = 0.020).11 Importantly, this endpoint was a broad composite that included both thromboembolic and bleeding events. Similarly, OCEAN was terminated prematurely for futility, with the consequence that the statistical power was reduced.12 Moreover, in the OCEAN, the investigational treatment was a reduced-dose rivaroxaban (15 mg daily) which is not the recommended dose for all patients (e.g. 20 mg for patients with a creatinine clearance > 50 mL/min). Furthermore, the study was designed to demonstrate superiority of rivaroxaban over aspirin, despite the established efficacy of anticoagulation already being reflected in contemporary guideline recommendations. Notably, although the results did not reach statistical significance, the direction of effect provides some evidence in favour of continuation of anticoagulation (RR 0.56, 95% CI 0.19–1.65; P = 0.280). Of note, a superiority trial that does not achieve significance cannot be interpreted in terms of equivalence between the tested treatments/strategies. Planning a non-inferiority trial comparing OAC discontinuation vs. maintenance after successful ablation, applying the margin of non-inferiority used in the studies that validated DOACs vs. vitamin K antagonist will be particularly challenging, since if focused on patients with a CHA2DS2VA ≥2 would result in enrolling at least 7500 patients. How should the available data inform clinical decision-making? Taken together, the available evidence suggests that OAC discontinuation after successful AF ablation may be a reasonable strategy in carefully selected low-risk patients, particularly those with CHA2DS2-VA scores of 1. The meta-analysis by Nayak et al.10 strengthens the evidence supporting this individualized approach. However, the same data also stress the risks of extrapolating these findings to patients at higher thromboembolic risk, i.e. with a CHA2DS2VA ≥2. Furthermore, the dynamic changes in CHA2DS2-VA should be considered and monitored, since an increase of the CHA2DS2-VA in 50% of patients after a median follow-up of 1.7 years was recently reported for patients with an initial CHA2DS2-VA scores of 0–1.13 An additional direction for future research may lie in moving beyond the traditional binary classification of post-ablation AF recurrence. Increasing evidence suggests that residual AF burden after ablation may provide a more accurate estimate of stroke risk than the simple presence or absence of documented arrhythmia and that AF burden is associated with adverse outcomes in a dose-dependent manner.14 We even do not know, at the present time, if a ‘pill-in-the-pocket’ management of AF risk may be safe and protective against thromboembolic complications in patients at risk. Any approach based on improved detection and quantification of AF burden should be coupled with a more detailed assessment of the extent of derangements in AF-associated atrial substrate (‘atrial cardiomyopathy’)15 and in its potential reversal/improvement after successful ablation and this may allow clinical advice for individualized decision-making. Until new clinical evidence becomes available, we believe that the message is clear: catheter ablation is an effective rhythm-control strategy, but it does not abolish the underlying substrate that confers thromboembolic risk. A successful AF ablation modifies rhythm; it does not necessarily eliminate risk. Therefore, in patients with a CHA2DS2-VA score ≥2, oral anticoagulation should remain a cornerstone of long-term management, regardless of apparent procedural success, in line with the recommendations of 2024 ESC guidelines.2