Ventricular arrhythmias after cardiac resynchronization therapy: does reverse remodelling reverse risk?

This editorial refers to ‘Reduced appropriate implantable cardioverter defibrillator therapy after cardiac synchronization therapy-induced left ventricular function recovery: a meta-analysis and systematic review’†, by N.A. Chatterjee et al., on page 2780. The term ‘remodelling’ was first used >30 years ago to describe pathological left ventricular (LV) structural and functional changes occurring after myocardial infarction.1 LV remodelling was subsequently adopted to describe the complex interplay of molecular, myocyte, fibroblast, and interstitial mechanisms influenced by haemodynamic load, neurohormonal activation, and inflammatory signalling. Clinically, LV remodelling is assessed by measuring increases in LV volumes and reductions in ejection fraction (EF).2 Reverse remodelling may then be defined as the process by which an injured left ventricle may return towards normal. Reverse remodelling may occur spontaneously, such as in acute non-ischaemic cardiomyopathy,3 or in response to a therapeutic intervention. Therapy associated with reverse remodelling may be pharmacological, such as renin–angiotensin–aldosterone system inhibition and beta-adrenergic receptor blockade, or device-based, such as cardiac resynchronization therapy (CRT) and mechanical circulatory support. The effects of CRT on LV reverse remodelling have been of particular interest because the precise mechanisms remain unclear, CRT response has been variable, and associated risks of ventricular arrhythmias have been uncertain. In this issue, Chatterjee et al. report an interesting meta-analysis focused on patients with LV reverse remodelling after CRT-defibrillator (CRT-D) therapy.4 Their objective was to determine the rate of implantable cardioverter defibrillator (ICD) detected ventricular tachyarrhythmias after CRT-D in patients with LV recovery (defined by an improvement in EF). They pooled six retrospective cohort studies including 1740 heart failure patients with QRS widening and LVEF ≤35% before CRT who also had follow-up of LVEF recorded. These combined patient groups had an average baseline LVEF of 20–29%; 63% of patients demonstrated an LVEF ≥35%, and 10% had an LVEF ≥45% after CRT-D implant. Their major findings were that patients with LVEF recovery (defined as ≥35% or ≥45%) had significantly lower rates of ICD therapy for ventricular arrhythmias compared with patients without LVEF recovery (P < 0.001). In particular, patients with recovery of LVEF ≥45% after CRT and those with CRT-D for primary prevention had very low rates of ventricular arrhythmia burden (0.4–0.8/100 person-years). Some of the limitations worthy of consideration with this meta-analysis included: (i) the lack of a standardized interval for LVEF determination after CRT (ranging from 4 to 20 months); (ii) different definitions of LV recovery (LVEF ≥35% or ≥45%); and (iii) assessment of LV recovery limited to LVEF measurements. We agree with the authors that prospective randomized studies of long-term duration will be needed to determine confidently arrhythmia risk and whether ICD therapy must be included indefinitely in patients with LV reverse remodelling. These noteworthy findings of Chatterjee et al. invite us to reflect on the pathological features of LV reverse remodelling and ventricular arrhythmogenesis after CRT-D, where the myocardial substrate appears to play an important role. In a broad sense, factors that may influence LV remodelling include ischaemia, infarction, fibrosis, inflammation, and others still to be defined.3,5 These elements which influence the composition of myocardial substrate may be considered irreversible, such as fibrosis,5 or reversible, such as resolution of acute inflammation in non-ischaemic cardiomyopathy or reperfusion of ischaemic tissue. Because CRT is thought to treat the electromechanical derangement most commonly associated with left bundle branch block, the interplay of irreversible and reversible factors of myocardial substrate may be complex.6 Furthermore, the response to CRT is widely variable, ranging from no LV reverse remodelling observed in 30–50% of patients, called non-responders, to near normalization of LV function in 10–20% of patients, called super-responders.7 Chronic right ventricular (RV) pacing resulting in a cardiomyopathy may occur from an electromechanical derangement alone.8 This may occur in patients who begin with a normal LVEF before RV pacing, have a decline in LVEF after RV pacing, and subsequently have return of normal LVEF after upgrade to CRT.9 Although excluded from several large randomized clinical trials, patients with RV pacing-induced cardiomyopathy and recovery of LV function with CRT have a very favourable long-term outcome and minimal risk of ventricular arrhythmias. Intense investigation continues to advance understanding of CRT response and arrhythmia risk, including means to differentiate myocardial substrate with reversible from irreversible LV dysfunction, the interaction of LV lead position, and its relationship to regional scar, global scar burden,10 and abnormal regional mechanical forces.11,12 Cardiac imaging will probably play an increasing role in assessing LV reverse remodelling and risk of ventricular arrhythmias. LVEF is an important selection criterion for both ICD and CRT implantations by current guidelines. However, it is a common frustrating clinical experience to observe differences in LVEF determinations in the same patient by echocardiographic, radionuclide, angiographic, and even cardiac magnetic resonance (CMR) imaging techniques. Echocardiographic endocardial tracing of biplane images using Simson's rule is the current standard for LVEF,13 but this approach is not uniformly implemented in many clinical centres. The most highly reproducible measures of echocardiographic LVEF have been with randomized clinical trials utilizing a core laboratory. However, important trials such as SCD-HeFT did not use a core lab or a uniform method for LVEF determination.14 Reproducibility of LVEF may be improved by 3D imaging, which is not widely utilized at present.13 In addition to methodological considerations, LVEF is imperfect for determining risk of ventricular arrhythmias. Most victims of sudden cardiac death have EFs which exceed current primary prevention guidelines for ICD implantation. Furthermore, Zecchin et al. reported that appropriate ICD interventions for ventricular arrhythmias were documented in 11% of CRT-D recipients who were super-responders (LVEF >50%) after implant.7 Accordingly, it was not advised to downgrade a CRT-D to CRT-pacing device at the time of generator change in a patient with recovery of LVEF because the risk of ventricular arrhythmias appears to remain. Future advances to refine risk for arrhythmias include quantification of myocardial fibrosis and extracellular matrix by late gadolinium enhancement CMR imaging5 and characterization of the myocardial substrate by echocardiographic speckle tracking measures of global longitudinal strain and global circumferential strain.13 Heterogeneity in timing of regional mechanical forces in the radial axis or by longitudinal dispersion using strain imaging have also been shown to be markers for ventricular arrhythmias after CRT.11,12 Finally, myocardial meta-iodobenzylguanidine imaging has been used to localize abnormalities in the adrenergic norepinephrine transporter terminals associated with increased risk for ventricular arrhythmia and sudden death.15 This meta-analysis supports that benefit from CRT in terms of LV reverse remodelling by LVEF is associated with reduced arrhythmia burden.4 This report also highlights that several important questions remain about how to define the complex interaction of factors which may influence arrhythmic risk after CRT. There is clearly room for improvement for LVEF as the sole clinical standard for risk stratification. Several advances in cardiac imaging have promise to refine determination of the risk of ventricular arrhythmias after CRT and aid in clinical decisions for device therapy beyond LVEF. Conflict of interest: J.G. reports grants from Biotronik and GE during the conduct of the study, and from Medtronic and St. Jude outside the submitted work. D.S. reports being a consultant to Atricure, Avery-Dennison, Biosense, and Medtronic. He receives grants funded by Biosense, Boston Scientific, and Medtronic outside the submitted work.