Clinical Characteristics of Heart Failure Patients Undergoing Atrial Fibrillation Ablation Today in Europe. Data From the Atrial Fibrillation Registries of the European Society of Cardiology and the European Heart Rhythm Association

Over the past three decades, catheter ablation of atrial fibrillation (AF) has evolved from an investigational procedure to its current role as a standard treatment option. Current guidelines recommend catheter ablation for symptomatic AF patients, preferring non-pharmacological therapy or resistant to antiarrhythmic drugs.1 However, which patients with AF and heart failure (HF) are actually referred for AF ablation in clinical practice today is still unclear, and the indication criteria are likely to be reconsidered by cardiologists in the light of the recent published literature. A recent non individual-based meta-analysis of six small randomized controlled trials (RCTs) including 775 HF patients, most with persistent AF, mean age 55–64 years, reduced left ventricular ejection fraction (EF) (mean 28.3%), and a mean follow-up longer than 12 months, showed an improvement of both EF and patient functional capacity associated with a significant reduction of both readmission and death rates.2 Since then, results of two larger RCTs have been reported. The CASTLE-AF trial, including 339 HF patients (EF ≤ 35%) out of 3013 patients assessed for enrolment over a period of 8 years, showed a lower rate of the combined endpoint of death and hospitalization for HF with AF ablation than medical therapy (hazard ratio 0.62, 95% confidence interval 0.43–0.87).3 The CABANA trial comparing AF ablation vs. antiarrhythmic pharmacological therapy included 2204 AF patients, with only 15% having a history of congestive HF (the results have been reported but are yet unpublished).4 Based on the original intention-to-treat trial design, the results were neutral, however, in the on-treatment analysis, due to an (expected) shift of approximately one third of patients randomized to drug treatment to the ablation arm, with a significant benefit on the primary endpoint in the ablation group (7% vs. 10.9%; hazard ratio 0.67, 95% confidence interval 0.50–0.89). An attempt to understand which HF patients cardiologists actually do select as candidates for AF ablation is therefore worthwhile, and prospective registries on AF recently conducted by the European Society of Cardiology (ESC) and the European Heart Rhythm Association may offer such an opportunity. This applies in particular to two published registries, the Atrial Fibrillation Ablation (AFA) Registry aimed at providing an observational picture of contemporary real-world AF ablation strategy and its outcome5 and the Atrial Fibrillation General Pilot (AFG) Registry6 collected in Europe within approximately the same time frame, by partly overlapping centre networks, and conducted by the same ESC management team. In the AFA registry, most patients underwent a first ablation procedure, while 19% had a redo procedure. In the AFG registry, 7.6% of patients had undergone an AF ablation procedure before enrolment in the registry, while ablation was performed or planned at admission in 6.8% of patients. We specifically focused on the HF patients enrolled in both studies: 537 (14.9%) in AFA and 1382 (46.5%) in AFG, respectively. The clinical characteristics of these HF cohorts are summarized in Table 1. AFA patients were a decade younger, had less prevalent cardiovascular risk factors (e.g. diabetes, hypercholesterolaemia, smoking habit) than AFG patients, and a lower risk of both stroke (CHA2DS2-VASc) and bleeding (HAS-BLED). While there were no differences in baseline blood pressure and gender ratio between the two cohorts, mean heart rate was faster in the general AF cohort than in ablated patients. Underlying or concomitant clinical disorders such as coronary and peripheral artery disease, valvular heart disease, previous transient ischaemic attack, chronic kidney and liver disease, and chronic obstructive pulmonary disease were less common in ablated patients than in the general AF population. As expected, the rate of paroxysmal AF was two-fold higher in the AFA HF patients, which were more symptomatic than the AFG HF patients. AFA patients had smaller left atrial size and higher EF at echocardiographic imaging. Most AFA patients (77%) had a preserved EF (≥ 50%, HFpEF), while those with mid-range EF (40–49%, HFmrEF) or reduced EF (< 40%, HFrEF) were a relatively small minority (15% and 8%, respectively). The relative proportion of the EF phenotypes among AFG patients was more balanced (HFpEF 46%, HFmrEF 21%, HFrEF 33%). The clinical profile of each EF subtype of ablated patients was more favourable than that of the corresponding HF patients in the AFG registry (data not reported). The few data reported above are descriptive, and the analysis only underscores a few major differences between the two AF cohorts. In fact, they are different by definition. In both groups, the setting is the same – the cardiology department – but the AFA patients were selected for an atrial ablation, whereas the AFG patients were seeking for clinical care. The differences between them outline the interventional niche reserved to co-morbid HF-AF patients in Europe today, which probably reflects a certain clinical cautiousness. The choice for intervention appears more focused on relieving AF-related symptoms than on an attempt to influence the clinical course of HF. Whether or not the most recent randomized trials3, 4 may prompt a more extensive use of AF ablation in HF patients is open to debate.7, 8 Besides other considerations, a potential limitation of the RCTs conducted so far is the uncertain representativeness of the patients enrolled. Organizational and methodological difficulties encountered in enrolling randomized patients and the consequent methodological adjustments inevitably impact on the representativeness of the study population. For instance, in our open AFA cohort, only 42 patients with EF ≤ 35% were included, and only nine of these had a cardioverter-defibrillator implanted (both entry criteria in the CASTLE-AF trial). Conflict of interest: none declared.