The Importance of Electrocardiographic Follow-Up in Heart Failure

This article refers to ‘Prevalence and incidence of intra-ventricular conduction delays and outcomes in patients with heart failure and reduced ejection fraction: insights from PARADIGM-HF and ATMOSPHERE’ by S.L. Kristensen et al., published in this issue on pages 2370–2379. Heart failure with reduced ejection fraction (HFrEF) has long been seen as a mainly progressive disease with remodelling as a key pathophysiological concept. The opposite – i.e. reverse remodelling – was considered to be present only if the cause of heart failure could be adequately addressed (e.g. after revascularisation in case of hibernation, treatment of myocarditis). The paper by Kristensen et al.1 in this issue of the Journal follows the line of progression as an important pathophysiological mechanism by investigating the new incidence of ventricular conduction delay in HFrEF, based on results of two large treatment trials. The paper nicely illustrates that progressive conduction delay can be seen as a marker of progression of HFrEF, but also as a contributing factor to a worsened outcome. Nowadays however, we know that improvement of ejection fraction and reduction of left ventricular diameters with intensified treatment is relatively common, even in elderly patients with long-standing disease.2 Since one of the pharmacological interventions studied (i.e. sacubitril/valsartan) has improved clinical outcome as compared to standard therapy, it would have been interesting to know in how many patients QRS duration decreases over time, i.e. the incidence of QRS narrowing during follow-up. This would help to better understand the pathophysiology of (reverse) remodelling and the relationship with conduction delay. This aspect has not yet been widely investigated beyond the effect of QRS narrowing by biventricular pacing on clinical outcome.3 However, the effect of reverse remodelling after cardiac resynchronisation therapy (CRT) on the intrinsic QRS duration has not yet been studied. There is one study that investigated QRS narrowing in patients treated with assist devices to unload the left ventricle. In these patients, there was a significant shortening of QRS duration from a mean of 150 to 118 ms after 6 months.4 Interestingly, part of the reduction in QRS duration was already seen after 1 week (mean 134 ms), suggesting that both ventricular (un)loading and structural alterations may influence conduction delay. Therefore, it would be interesting to know if medical treatment that results in reverse remodelling of the left ventricle such as sacubitril/valsartan is also accompanied by reduction of conduction delay. This has only been reported in small case series (e.g.5) but not investigated systematically. Still, the treatment with sacubitril/valsartan was reported to result in electromechanical remodelling in animals6 and is also accompanied by a reduced risk of clinically significant arrhythmias.7 These findings suggest that a reduction in QRS duration might be possible with better treatment (apart from CRT as direct effect), but this needs to be properly investigated. If this is the case, it would be also interesting to know if reduction in QRS duration is associated with better outcome and if such a reduction could be a marker or even target for better treatment. An important finding in the paper of Kristensen et al.1 is that an intermediate prolonged QRS duration was one of the best predictors of incident wide QRS. The event rate of any new-onset wide QRS was 14.1 per 100 patient-years in patients with minimal conduction delay, i.e. those with a QRS duration between 110 and 129 ms, and only 3.4 in patients with a narrow QRS. Patients with minimal conduction delay had almost equally poor outcome as those with broader QRS.1 However, these patients do not qualify for CRT, which may even result in worse outcome.8 Further progression results in worse outcome,1 but the question arises as this may be prevented with better medical therapy. Kristensen et al. report risk factors for the progression, among other and not surprisingly minimal conduction delay. However, they do not report if medical treatment that results in better outcome and reverse remodelling is also able to reduce the risk of progression of conduction delay or even a reduction as discussed above. If true, it might indicate QRS duration as a potential marker for effectiveness of treatment, which might help to individualise the treatment in HFrEF. Obviously, such an approach needs prospective testing before it is applicable, but the first step would be a better insight in the interaction between treatment and electrical (re)modelling. The authors have these data available and it may be interesting to investigate these aspects in an additional analysis. The poor outcome in patients with minimal conduction delay may have practical implications as QRS widening should be more carefully evaluated in these patients. These patients might even benefit from CRT implantation. It remains, however, a challenge to predict which patients will develop new-onset wide QRS. Careful analysis other than measuring QRS duration and describing the type of conduction disturbance – i.e. right bundle branch block (RBBB), left bundle branch block (LBBB), or intra-ventricular conduction delay (IVCD) – could be helpful. Often, there are subtle electrocardiogram (ECG) characteristics such as the absence of the initial Q-wave in I, or aVL that could be hints that the patient might develop new-onset LBBB. These kinds of analysis are, however, cumbersome and hard to systematically study as subtle ECG characteristics are often rather subjective, such the analysis of the various ECG criteria in the different definitions that exist for describing LBBB.9 Importantly, preventive implantation of CRT in patients likely to develop LBBB cannot be advocated based on current evidence, but it may be interesting to prospectively investigate such an approach. Cardiac resynchronisation therapy has been shown to be most effective in patients with LBBB or QRS duration of >150 ms in patients with IVCD.10 Nevertheless, individual evaluation might result in implantation of CRT in patients with RBBB or moderately prolonged QRS in patients with IVCD. An important consideration in this respect is the presence of atrio-ventricular conduction delay, which is evidenced by a prolonged PR interval on the ECG. Patients with HFrEF and wide QRS often have a prolonged PR interval as well. A prolonged PR interval may result in elevated left ventricular end-diastolic pressure, end-diastolic mitral regurgitation, and reduced left ventricular pump function. This seems especially the case in patients with HFrEF in whom it is associated with an increased risk for atrial fibrillation, advanced atrio-ventricular heart block, heart failure, and death.11 As it is remarkable that the outcome described in the paper by Kristensen et al. is comparable between the subgroups of patients with wide QRS,1 it would have been interesting to know the incidence of prolonged PR interval, and especially in the patients with new-onset wide QRS; particularly as the negative effects of prolonged PR interval, referred to atrio-ventricular dromotropathy could potentially be treated by improving the atrio-ventricular coupling by CRT. In a subanalysis of the MADIT-CRT trial,12 it was demonstrated that non-LBBB patients with a long PR interval experienced a significant clinical benefit from CRT implantation in contrast to patients with non-LBBB but without a prolonged PR interval. The presence of a wide QRS should, therefore, be evaluated in relation to the presence of a prolonged PR interval to optimise treatment for the individual patients with HFrEF. Additionally, the ECG hides more information besides QRS widening that serves as prognostic marker13 and may even guide HFrEF treatment, such as atrial fibrillation and atrial flutter/tachycardia, atrio-ventricular conduction delay or block, the morphology of the (ventricular paced) QRS complex, and the presence of frequent ventricular premature beats (Figure 1). Thus, additional analyses regarding incident ECG findings from the study by Kristensen et al.1 could be of interest. The importance of repeated ECGs in patients with HFrEF is additionally emphasized by the findings that delayed implantation of CRT might result in less response to CRT as summarized in a recent review,14 although the underlying evidence for this notion is relatively poor. Thus, the most important clinical consequences of the paper by Kristensen et al.1 in this issue of the Journal is not the fact that non-LBBB or even small increases in conduction delay, either prevalent or incident, have (almost) equally poor prognosis as complete LBBB. Much more importantly, incident LBBB in HFrEF is one of the limited examples where a marker of poor outcome has direct therapeutic consequences, i.e. implantation of CRT. Besides ECG parameters, there are countless prognostic factors in HFrEF (and heart failure in general), but the clinical consequences of knowing that an individual patient is at high risk, are largely unknown and not prospectively investigated.15 The real value is therefore the fact that 2.6% of HFrEF patients with QRS duration <130 ms developed complete LBBB each year. In these patients, the evidence for treatment to improve outcome is very clear. The authors do not report in how many of these patients, a CRT device has been implanted. It may be speculated that it has been less than those that should have received such a device. Conflict of interest: none declared.