Biventricular pacing after AV-node ablation for atrial fibrillation with heart failure: something old and something new, future standard, or just for the few?

This editorial refers to ‘A randomized controlled trial of atrioventricular junction ablation and cardiac resynchronization therapy in patients with permanent atrial fibrillation and narrow QRS’†, by M. Brignole et al., on page 3999. Although cardiovascular mortality and morbidity appear to be on the wane compared with other diseases, heart failure (HF) is still an increasing chronic morbidity in our ageing population. Despite many successful landmark trials with multiple drugs, resynchronization therapy in the case of a wide QRS complex, and other novel mechanical therapies, there remains room for improvement with better patient and therapy selection.1 HF may be accompanied by atrial fibrillation (AF) in up to 50% of patients, and both conditions can either trigger or be the result of the other.1 , 2 In advanced stages of HF, AF usually becomes chronic and adds to the burden of symptoms as it can be difficult to control. In patients with uncontrollable AF, AV-node ablation (AVNA) followed by chronic right ventricular apex pacing (RVAP) was already established as a pragmatic and successful way to reduce AF symptoms decades ago (ESC guideline class IIA).2 Previous studies have also shown that patients with a low left ventricular ejection fraction (LVEF) undergoing cardiac resynchronization therapy (CRT) for HF and AF with uncontrollable heart rates could benefit from AVNA (ESC guideline class IIA).3 There are, however, still some knowledge gaps for specific patient populations with the combinations of HF and AF.4 In the present issue of the European Heart Journal, Brignole et al. describe the results of the APAF-CRT randomized controlled trial (RCT), comparing optimal medical therapy (OMT) with AVNA with CRT pacing in patients with permanent AF and narrow QRS complex, that were hospitalized for HF in the previous 12 months.5 The primary endpoint for the trial was the combination of all-cause mortality, or hospitalization for HF in need of increased medication. As this was a well-organized RCT, all events were monitored by a central adjudication committee. Although follow-up was planned for a maximum of 36 months, the adjudication committee ended the trial prematurely after 16 months at the interim analysis of 50% of the analysed patient data, because of a significant imbalance in outcome between the two treatment arms. They observed that AVNA with CRT pacing was superior to OMT for the combined endpoint [20% vs. 38%, hazard ratio (HR) 0.38, 95% confidence interval (CI) 0.18–0.81, P = 0.013], both for mortality and for prevention of HF hospitalization and/or increasing HF medication. The investigators are to be commended for a well-performed trial with a positive outcome for active therapy. HF and AF are formidable foes by themselves, but when combined can be the cause of severe deterioration of symptoms and usually bode ill for those affected. In patients with an important pre-existing depressed left ventricular function (LVEF), or in those that develop severe tachycardiomyopathy, this may be the harbinger of ventricular arrhythmic death, as well as death due to terminal HF. Any treatment able to diminish that risk is of the upmost importance to patients in such dire situations, and the cardiology community will welcome this improvement in mortality and HF, which appears just as important and of the same magnitude as many HF drug trials performed over the last 25 years (Take home figure). One can assume that this piece of the puzzle will be of value in updating guidelines on the optimal therapy for patients with AF and HF. Outcomes of landmark trials for heart failure and AF. It is of equal importance to look beyond the primary endpoint, and examine the basic hypothesis of this trial, and the study set-up and patient population that was studied, to understand what exactly this trial is teaching us. Looking at the study population, one should consider how many patients may have been recommended for other therapies. Although inclusion criteria stated that patients had permanent AF and were not eligible for other curative therapies, recently the CASTLE-AF trial showed a clear benefit in mortality and HF morbidity by an active atrial ablation strategy even in patients that were in AF for a long time.6 Also, there are studies showing improved outcomes of arrhythmic surgery in patients with non-paroxysmal AF.7 In view of these findings, the ESC strongly recommends trying such therapies before performing an irreversible AVNA which makes patients pacemaker dependent for the rest of their lives.3 The investigators base their study on the premise that data are lacking on the effect of AVNA in the specific situation of patients with HF and AF with narrow QRS complex that undergo AVNA with CRT pacing. Any LVEF was eligible for inclusion, and in fact 42% of patients had an LVEF <35%. The control group consisted of patients that were kept on OMT. This leaves room for a biased patient inclusion, as patients not responding to OMT, and therefore eligible for the trial, will most probably also not benefit from more OMT, and show worse outcomes for HF and mortality. The choice for AVNA with CRT pacing should therefore probably only apply to those patients that show deterioration of symptoms despite OMT, and whether one episode of HF hospitalization would already qualify should be considered on a patient to patient basis. Patients with wide QRS complex were excluded, presumably because such patients were considered to have a potential class IA guideline indication for CRT pacing in the case of HF.1 Those with LVEF <35% were rightly excluded as they should be considered for a CRT-D, while patients with wide QRS and LVEF of 35–45% would be eligible for CRT-P.3 These patients might not even need AVNA, but only in case of insufficient rate control, preventing >90% of CRT pacing.3 In patients with wide QRS and LVEF >45%, there is no clear indication for CRT pacing regardless of QRS width. The indication for pacing in the current trial was therefore in fact physician induced due to the strategy of AVNA, as this leaves patients with insufficient heart rates. In patients with permanent AF, RVAP is the standard after AVNA as no true benefits are observed for RV outflow tract pacing. Several trials such as BLOCK-HF and COMBAT have already shown that CRT-P is better for HF symptoms than RVAP in those with decreased LVEF. Moreover, active RVAP in patients with LVEF <35% was shown to increase morbidity and mortality in those with LVEF <35% in the DAVID trials. There is also abundant evidence that patients with low LVEF clearly benefit from upgrading RVAP to CRT pacing, with a class IB indication.3 As 42% of the included patients in the current trial had an LVEF <35%, international guidelines would therefore already have recommended CRT pacing rather than RV pacing after AVNA in this group.3 Given all the CRT trials in the past, it should come as no surprise that these patients would benefit from CRT, regardless of whether they had a wide or narrow QRS at baseline. Indeed, the current result reconfirmed a significant improvement only in this group of patients, and not in the 58% that had an LVEF >35%. This leaves us with unanswered questions for this latter group, as they do not seem truly to benefit from AVNA and CRT pacing. All in all, it would have been interesting to have a second control group with RVAP, to see whether the benefit of the therapy is based on CRT pacing. With the revival of His-bundle pacing, one might consider this as an even better physiological option. Alternatively, the positive outcome may also be due to the superior and absolute rate control after the AVNA. While the average heart rate was around 100 bpm, both treatment arms received optimization of drug use in the first 30 days after enrolment; however, no data are provided on what drugs were added or changed, and what average heart rates were during follow-up. Moreover, as the aim was to get patients below an average of 113 bpm3 which was already the case in over half of the patients at baseline, it is unclear what was changed and in how many patients optimal rate control was achieved. It remains to be determined if outcomes would be better in the OMT arm, when drug-induced rate control would have achieved an even lower heart rate. Although lower mortality was a hard and non-negotiable endpoint in favour of AVNA with CRT pacing, the improvement in symptoms may be open to discussion. Chronic symptoms found in patient with systolic and diastolic HF often include shortness of breath, exercise intolerance, and fatigue. However, such symptoms appear to be identical in patients suffering from (non-)paroxysmal AF. The improvement in symptoms could therefore be due to proper rate control reducing AF symptoms, or improved left ventricular function improving HF symptoms. The former could be expected in patients with LVEF >45%, and the latter in patients with LVEF <45%. No data are given on LVEF changes during the trial, so from the current analysis it is unclear how many showed an increase in LVEF. One could also argue that this hardly matters; if improvement is achieved, cardiologists should consider applying this strategy. Indeed, in the APAF-CRT trial, all patients showed significant improvement on their SSS scale regardless of treatment arm. Interestingly, the patients with AVNA and CRT with preserved LVEF >35% benefited most, while the change was equal in the other cohorts, which could signify that LVEF improvement was not the main driver of symptom improvement. Finally, some words of caution seem to be appropriate for a few issues in the study. About 10% of the patients crossed over from the OMT arm to AVNA with CRT pacing. Based on intention to treat, they may have benefited from the active therapy, which may have overestimated results in the OMT arm. The results could therefore have been even more in favour of AVNA with CRT pacing than they already were. The other concern is that the trial was performed in only a small sample of 102 patients, and was ended prematurely because of the large difference in mortality and HF outcomes between the two study arms. Although this may seem appropriate in terms of statistical trial management, it is always better to follow the original design of the trial to reduce the element of chance (especially when numbers are small), and see if time in the end would equalize outcomes. On the other hand, the results seem in line with many trials on AVNA and CRT in AF and HF patients.3 Luckily, the follow-up will continue for mortality, so we may expect to hear more on the fate of these patients, unless even more of them cross over to the AVNA with CRT arm to make a proper analysis impossible In conclusion, the current trial has provided us with evidence that in patients with narrow QRS, permanent AF, and HF symptoms that do not respond to OMT, AVNA followed by CRT pacing appears to be better at reducing HF deterioration and mortality than trying to optimize drug therapy, specifically in patients with an LVEF <35%. Symptom reduction, on the other hand, was greatest in patient with AVNA and CRT pacing with LVEF >35%. Whether it is the absolute rate control or the CRT pacing that does the job remains unclear. We already knew that patients with an LVEF <45% and bradycardia were better off for HF with CRT pacing than RVAP, as well as those with adequate rhythm but wide QRS.3 CARE-HF and COMPANION previously suggested that CRT pacing lowered both HF and mortality even in the absence of a defibrillator function.1 It remains unclear, however, if these outcomes also apply to patients with narrow baseline QRS, AF, and LVEF >35%, in the absence of a control group undergoing AVNA with RVAP. Although in general one could argue that there may be merit to performing pacing as physiologically as possible, that would mean that all patients that require active pacing would require CRT pacing. This may not be without potential for harm as well as increased cost.3 It seems likely that future guidelines will encourage us to be liberal with CRT pacing after AVNA for AF and HF regardless of baseline QRS width. On the other hand, we should probably reserve this for those with lower LVEF, as the hard evidence in patients with preserved ventricular contractility is not clear from the current data. Conflict of interest: L.V.A.B. provides services as a consultant to Medtronic, Boston Scientific, and Abbott. The reimbursement for his time is received by the Cardiology Department at the hospital where he works.