Atrial fibrillation and cardiac resynchronization therapy: how should rate be controlled?

Abstract

This editorial refers to ‘Cardiac resynchronization therapy in patients with permanent atrial fibrillation. Is it mandatory to ablate the atrioventricular junction to obtain a good response?’, by J.M. Tolosana et al. doi.1093/eurjhf/hfs024 Atrial fibrillation (AF) and heart failure (HF) are common cardiac conditions; the lifetime risk of AF is 1 in 4, and of HF 1 in 5 persons. Both conditions often occur concomitantly, and the prevalence of AF increases with the severity of HF.1,2 In Cardiology practices in Europe, the prevalence of HF in AF patients was found to be 34%, and the prevalence of AF in HF patients 42%.3,4 Reasons for the co-existence of HF and AF are diverse. HF and AF share risk factors such as hypertension, valve disease, coronary artery disease, and diabetes mellitus. However, the haemodynamic consequences, neurohormonal activation, and myocardial structural and electrical remodelling that occur in both AF and HF may predispose to one or the other. Both HF and AF separately increase the mortality risk, and, when both conditions occur at the same time, the mortality risk is even higher.2 In addition to medical therapy of HF, cardiac resynchronization therapy (CRT) has proven to be beneficial for HF patients by improving symptoms, functional capacity, and left ventricular remodelling, and reducing hospitalizations for HF and mortality. Based on large randomized trials, the 2010 Focused Update of European Society Cardiology guidelines on device therapy in HF5 recommend CRT (Class I, level of evidence A) for HF patients, optimal medically treated, with left ventricular ejection fraction (LVEF) ≤35%, in New York Heart Association (NYHA) functional class III/IV, and with QRS ≥120 ms, or in NYHA functional class II with QRS ≥150 ms, all in sinus rhythm. Despite the co-existence of AF and HF, the evidence for CRT in HF patients with AF is less robust.6 As a result, the most recent guidelines5 state that CRT should be considered to reduce morbidity (mortality benefit has not been proven; Class IIa, level of evidence B) for optimally medically treated HF patients with AF, with LVEF ≤35%, in NYHA functional class III/IV, and with QRS ≥130 ms. Both atrioventricular node ablation and pharmacological rate control, in order to ensure maximal biventricular pacing, are options, although the majority of patients have undergone atrioventricular node ablation. To date, randomized studies of CRT have been almost exclusively restricted to patients in sinus rhythm; this in great contrast to daily clinical practice where CRT is routinely used for patients with AF.5 The European CRT survey indicated that ~1 in 5 patients receiving CRT has permanent AF.7 A prerequisite for effective and most successful CRT is biventricular capture of all heart beats, at rest and during exercise. However, ensuring complete ventricular capture is a challenge for patients with AF, because of the uncontrolled ventricular rates, in contrast to the majority of patients with sinus rhythm. Several observational studies have reported that AF-induced uncontrolled ventricular rates lead to suboptimal and even failure of CRT.8–10 Data from 1812 CRT patients included in two registries (CRT RENEWAL and REFLEx)9 showed that the greatest magnitude of reduction in HF hospitalization and all-cause mortality was observed with a biventricular pacing >92% of the time. Patients paced ≤92% had worse outcomes and were less likely to improve in terms of NYHA functional class. Boriani et al.11 last year published the results of an observational study of 1404 patients with CRT, of which 443 (32%) had AF at some point during follow-up. Uncontrolled ventricular rates were related to more HF hospitalizations and all-cause mortality. AF was the main predictor of uncontrolled ventricular rates leading to biventricular pacing of ≤95% of all heart beats. It is important to note that most studies looking at the percentage of biventricular pacing, including the above, are based on device interrogation data, and may structurally overestimate the percentage of biventricular pacing since they do not account for fusion and pseudo-fusion between intrinsic (not paced) and paced beats, which may also lead to suboptimal CRT.12 Two approaches are available to control the ventricular rate in patients with AF, namely pharmacological rate control or atrioventricular node ablation. Institution of negatively chronotropic drugs, such as beta-blockers, digoxin, or amiodarone, can adequately reduce heart rate in the majority of patients with AF; however, strict rate control targets are difficult to reach in about one-third of patients, mainly because of side effects of drugs.13 Ablation of the atrioventricular node is most effective since the atria are electrically separated from the ventricles, leading to complete pacemaker dependency, and thus complete heart rate control and rhythm regularity. Pacemaker dependency is also the most important disadvantage of atrioventricular node ablation, since lead failure is not uncommon and is potentially life threatening in pacemaker-dependent patients. In addition, complete atrioventricular node ablation may lead to severe chronotropic incompetence (inability to increase heart rate during exercise) that has been suggested to be an independent predictor of mortality. At this point, rate-adaptive pacing algorithms and sensors of CRT seem not to be capable of full recovery of the physiological chronotropic response. The use of atrioventricular node ablation in AF patients requiring CRT has been highly variable across the published observational studies, and ranges from 22% to 100%, with variable effects on outcome.6 In light of the above, Tolosana et al.14 have now reported the results of a subanalysis of the Spanish Atrial Resynchronization Study II. They followed 202 patients with CRT of which 47 (23%) were having AF. After institution of rate control drugs, only 28% of the AF patients had ≤85% biventricular pacing, and required atrioventricular node ablation. After 1-year follow-up, the authors observed similar CRT response rates (both echocardiographically and clinically) between AF and sinus rhythm patients, and within the AF patients no differences were found between the atrioventricular node-ablated patients and the pharmacologically rate-controlled patients. These results can be interpreted as evidence that pharmacological rate control of AF may be successful in selected patients, and patients with pharmacological rate control may have a similar response to CRT as those with atrioventricular node ablation. The observational design is an important limitation of this and previous studies on this topic, which inherently harbour selection bias in which clinicians choose certain rate control drugs, and use different rate control targets, different device settings, use different tests to evaluate rate control (percentage biventricular pacing by device interrogation, 12-lead Holter monitoring,12 or exercise tests15). Furthermore, the stepwise approach may select certain patient categories that are non-responsive to pharmacological rate control treatment. Taking the above into consideration, it seems most reasonable to start with pharmalogical therapy to optimize rate control in AF patients requiring CRT. When after careful evaluation by device interrogation, Holter monitoring, and exercise testing, the amount of ‘true’ biventricular pacing is suboptimal, atrioventricular node ablation should be considered. However, well-designed randomized trials are urgently needed to evaluate the optimal approach to control heart rate and ensure optimal biventricular pacing in patients with AF. Whether there is a role for device monitoring and alarming when high ventricular rates are detected needs to be determined. Conflict of interest: none declared.