I had a Dream…

This article refers to ‘Atrioventricular junction ablation in patients with atrial fibrillation treated with cardiac resynchronization therapy: positive impact on ventricular arrhythmias, implantable cardioverter-defibrillator therapies and hospitalizations’ by M. Gasparini et al., published in this issue on pages 1472–1481. This night, I had a dream… the completion of a randomized controlled multicentre trial comparing different therapeutic strategies in patients with permanent atrial fibrillation (AF) and an indication for cardiac resynchronization device implantation. The endpoint is the combination of mortality and heart failure hospitalizations with a long follow-up; several years, at least three. This trial will answer the crucial question of the real efficacy of cardiac resynchronization therapy (CRT) in this specific population. Cardiac resynchronization therapy is implemented for patients with New York Heart Association (NYHA) class III and IV in the European guidelines since 2005 after the publication of the Multisite Stimulation in Cardiomyopathies (MUSTIC), Multicenter InSync Randomized Clinical Evaluation (MIRACLE) and Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) trials. The results of subsequent trials led to extend the recommendations of CRT use to less severe patients (NYHA class II). However, the vast majority of these trials excluded patients with permanent AF.1 The lack of data from randomized controlled trials in patients with AF and CRT is surprising or even incomprehensible if we consider the importance of this population in CRT. According to different surveys or studies, about 25% of patients implanted with a CRT device have AF2 and CRT is recommended by the 2016 European Society of Cardiology guidelines on the management and treatment of heart failure for NYHA class III and IV AF patients with left ventricular ejection fraction <35% and QRS duration >130 ms if a strategy to ensure biventricular capture is in place, or if the patient is expected to return to sinus rhythm.1 The recommendations are based on limited evidence of efficacy of CRT in this specific population without dedicated randomized controlled trials. In this issue of the Journal, Gasparini et al.3 report interesting data suggesting that atrioventricular (AV) node ablation in patients with permanent AF receiving a CRT defibrillator (CRT-D) reduces significantly the rate of appropriate and inappropriate therapies, all-cause hospitalizations and heart failure hospitalizations compared to patients treated with rate control drugs. Moreover, the outcome for all these endpoints was surprisingly much better in patients with AV node ablation and AF as compared to patients in sinus rhythm. What do we know about the efficacy and safety of CRT in patients with AF? From randomized controlled trials: very few evidence. The MUSTIC trial included 64 patients in NYHA class III and IV with left ventricular systolic dysfunction, chronic AF, slow ventricular rate requiring permanent ventricular pacing, and wide QRS complex (paced QRS width ≥ 200 ms). The primary endpoint was the 6-min walked distance, a soft endpoint. The efficacy analysis showed that CRT increased significantly the distance compared to right ventricular pacing; no relevant data on morbi-mortality were available in this pilot trial.4 The purpose of the Post-AV Nodal Ablation Evaluation (PAVE) study was to prospectively compare chronic biventricular to right ventricular pacing in patients requiring AV node ablation for the management of AF associated with excessive ventricular rates. This study was different because mean left ventricular ejection fraction was 46 ± 16%, higher than in patients with CRT indication. Borderline significant improvements in left ventricular ejection fraction (46 ± 13% vs. 41 ± 13%; P = 0.03) and the 6-min walked distance (82.9 ± 94.7 m vs. 61.2 ± 90 m; P = 0.04) were reported. However, PAVE did not demonstrate an improvement in quality of life or clinical outcomes.5 Brignole et al.6 compared right ventricular apical pacing and biventricular pacing in 186 patients with permanent AF and AV node ablation. The primary endpoint (death from heart failure, hospitalization due to heart failure, or worsening heart failure) was significantly reduced with biventricular pacing by 73% after a follow-up of 20 months, mainly driven by the reduction in hospitalizations for heart failure and worsening heart failure.6 A sub-analysis of the Resynchronization for Ambulatory Heart Failure Trial (RAFT) assessed the efficacy of CRT in 229 patients in NYHA class II and III with permanent AF. All patients had low left ventricular ejection fraction (<30%) and wide QRS (>120 ms) and were randomized to CRT-D or implantable cardioverter defibrillator (ICD).7 After a follow-up of 40 ± 18 months, no differences were observed between groups for the primary endpoint (death and heart failure hospitalizations) or cardiovascular deaths. There was only a trend for fewer heart failure hospitalizations. It is worth noting that only one patient underwent AV node ablation in the RAFT trial. Although RAFT required patients with permanent AF to have good rate control (resting ventricular rate ≤ 60/min and a ventricular rate ≤ 90/min after 6-min walk test) before randomization, only one third of CRT patients received ≥95% ventricular pacing during the first 6 months. Even this may be an overestimate, because Holter monitoring studies have shown that, when device logs indicate ≥90% ventricular pacing in patients with permanent AF but without AV junction ablation, 53% of these paced beats are actually fusion or pseudofusion.7 Data from registries were published to analyse the efficacy of CRT in patients with permanent AF. Most of them were published by Gasparini and consistently showed a significant benefit of CRT when associated with AV node ablation as compared to pharmacological rate control. The benefit was observed for exercise tolerance, left ventricular reverse remodelling but also survival.8-10 Two meta-analyses confirmed these findings but, due to methodological limitations of registries with limited number of patients and no randomization, the evidence is questionable.11, 12 Gasparini et al.3 provide new interesting data from two combined randomized studies and one Italian registry. The two trials were respectively designed to assess the efficacy of biventricular vs. right ventricular anti-tachycardia pacing and to assess whether using long detection interval to detect ventricular tachycardia/ventricular fibrillation reduces anti-tachycardia and shock delivery. AV node ablation in AF patients treated with CRT significantly reduced the primary endpoint of all-cause ICD shocks (appropriate and inappropriate) and all-cause hospitalizations during a 18-month follow-up as compared to rate control drugs. We know the deleterious impact of appropriate and inappropriate shocks on the outcome of the patients, so a therapy that reduces ICD shocks is associated with better outcomes.13 For appropriate shocks, the reduction can be explained by the fact that AV node ablation suppresses short–long–short sequences but also rapidly conducted AF, which are both triggers for ventricular arrhythmias. For inappropriate shocks, we know that one of the major causes is atrial tachycardia with fast ventricular rate, which does not exist with AV node ablation. For the hospitalization rate, this is not surprising as the percentage of biventricular pacing, a key marker of CRT success, is higher in patients with AV node ablation than in AF patients with drug-rate control. Another explanation is the fact that after successful AV node ablation there is no risk of tachycardiomyopathy. These results are suggested by the authors to reinforce the role of AV node ablation in AF patients with CRT with a non-perfect control rate but also raises the question of the type of device to be used in patients with AF and AV node ablation. With a so low rate of appropriate shocks (8 shocks in 262 patients), is a CRT-D still useful? More surprising is the comparison between AF patients and sinus rhythm patients that are definitively different (age, aetiology, QRS duration, QRS morphology, treatment). To be provocative according to the results of this study, we can suggest that it is better to have a permanent AF with a complete AV block induced by AV node ablation, so two diseases, and being pacemaker-dependent than being in sinus rhythm for CRT patients. Do we have to induce AF and perform AV node ablation in sinus rhythm patients? I don't think so. This comparison shows the limitations of the interpretation of registries including different populations. Finally, for patients with paroxysmal and persistent AF, the management has today to integrate the impressive results of the CASTLE-AF trial.14 So, I still have a dream, to have a robust, randomized controlled trial to prove the efficacy and safety of CRT in AF patients, who represent 25% of the indications for CRT with so weak evidence. Conflict of interest: none declared.