Cardiac resynchronization therapy: results, challenges and perspectives for the future

Abstract

Heart failure (HF) is considered as an epidemic and affects 2% of the population in the Western world. About 15–30% of patients with HF and reduced ejection fraction (HFrEF) also have prolonged QRS duration on the surface ECG, most commonly as a result of left-bundle branch block (LBBB). Increased QRS duration is a marker of a dyssynchronous activation, and subsequent contraction, pattern in the left ventricle (LV). When dyssynchrony is superimposed on the failing heart it further reduced systolic function and ultimately worsens outcome. During the past 15 years several randomized controlled clinical trials have documented that resynchronization of the dyssynchronous failing heart with a biventricular pacemaker – cardiac resynchronization therapy (CRT) – which can restore a more synchronous activation and contraction pattern. This translates in halted or reversed disease progression and improved clinical outcome, including reduced mortality.

In this review, we will discuss several aspects of CRT including mechanisms of dyssynchrony and resynchronization in the failing heart, evidence of CRT efficacy derived from clinical trials and current challenges in CRT including patient selection and optimization of therapy delivery. Last, we will discuss future perspectives including the role of CRT to prevent adverse events in patients with an indication for antibradycardia pacing, the role of leadless pacing in the CRT setting as well as a new clinical arena where dyssynchrony and resynchronization may be important.

Keywords:

• Heart failure
• cardiac resynchronization therapy
• pacing

Introduction

Heart failure (HF) affects approximately 2% of the population in the western world and is associated with severe morbidity and high mortality rates, which approaches 60–75% during the 5 years following initial diagnosis [1].

Left-bundle branch block (LBBB) and/or prolongation of the QRS complex on the 12-lead ECG marks the presence of a dyssynchronous activation and contraction pattern and affects around 30% of patients with HF and reduced ejection fraction (HFrEF) [2]. Dyssynchrony superimposed to a failing heart has adverse effects on cardiovascular health on different levels ranging from deleterious effects on the cellular and molecular level to worse clinical outcome.

In the late 1990s, it became evident that ventricular dyssychrony in part could be ameliorated by pacemaker stimulation of the right and LV simultaneously – cardiac resynchronization therapy (CRT). Subsequently, during the past 15 years several randomized controlled trials have documented that CRT substantially reduces morbidity and mortality in patients with HFrEF and dyssynchrony.

However, despite a convincingly demonstrated clinical benefit with CRT some challenges remain. In particular, these include patient selection, optimization of therapy delivery and CRT implementation and utilization.

In this comprehensive review we will discuss (i) mechanisms of dyssynchrony and resynchronization in HF from a cellular and molecular level to clinical outcome, (ii) the documented efficacy of CRT derived from randomized controlled clinical trials, (iii) current challenges in CRT delivery including patient selection, therapy optimization and implementation and (iv) future perspectives of CRT. In terms of future perspective, we will discuss the development of a system for leadless pacing, the feasibility of CRT to prevent reverse remodeling and HF development in patients with a bradycardia indication for pacing and a new clinical arena where dyssynchrony and resynchronization may be important.

Mechanisms of dyssynchrony and resynchronization in HF: from bedside to bench

A prolongation of the QRS duration on the 12-lead ECG is a readily available marker of a dyssynchronous activation and contraction pattern. The most common causes of QRS-prolongation in HFrEF are LBBB, nonspecific QRS prolongation and right ventricular pacing (RVP). The prevalence of QRS prolongation and/or LBBB is around 30% in different studies and hence constitutes a substantial clinical entity in HFrEF [2].

During LBBB activation (Figure 1(a)), electrical activation of the heart spreads rapidly through the intact right bundle branch, which results in early activation of the septum while activation of the LV lateral wall is delayed. The lateral wall then constitutes a late activated region of the LV (Figure 1(b)). By implanting a CRT device with an additional ventricular pacing lead inserted through the coronary sinus late activated LV regions can be stimulated simultaneously (or near simultaneously) as the right ventricle (Figure 1(c)). This homogenizes LV activation and contraction and thereby ameliorates dyssynchrony (Figure 1(b)).

Figure 1. Combined explanatory figure of dyssynchrony and resynchronization in HF. Electrical dyssyncrhony, evident as LBBB on ECG (A), results in an abnormal left ventricular (LV) electrical activation with early- and late-activated regions. Figure 1(b) displays a homogenous electrical activation pattern in the heart with normal intrinsic conduction (left panel). During LBBB activation (middle panel), septum is activated early while the LV lateral wall is activated late. However, these regional differences in electrical activation are ameliorated when CRT is activated (right panel) [3]. Figure 1(c) depicts a CRT-device with one atrial and two ventricular leads. Figure 1(b) is reprinted from ref. [3]. Copyright © (2012), with permission from Elsevier. Figure 1(c) is reproduced with permission of Medtronic, Inc.

Table 1 outlines the adverse effects of dyssynchrony, and their reversal or improvement by CRT, on different levels ranging from cellular and molecular cardiac effects to clinical outcome. These adverse effects include changes in gene and protein expression, cellular signaling pathways, structural, hemodynamic, metabolic and ultimately clinical aspects.

Table 1. Effects of dyssynchrony and resynchronization in HF: from cellular to clinical effects.

	Effect of dyssynchrony	Effect of resynchronization
Cellular and molecular – regional LV (late vs. early activated regions)
Ca^2+ handling proteins	Heterogenic expression [4]	Normalized [4]
Connexin-43	Downregulated [4]	Normalized [4]
Stress kinases	Upregulated [4]	Normalized [4]
Action potential duration	Delayed [5]	Partially restored [5]
Cellular and molecular – global LV
Fetal gene program	Activated [6]	Deactivated [6]
β-adrenaline signaling	Depressed [5]	Improved [5]
Ca^2+ sensitivity	Reduced [5]	Improved [5]
Metabolic
Regional differences in  glucose uptake	Present [5]	Normalized [5]
Global MVO2 consumption	Unchanged [7]	Unchanged [7]
Cardiac efficiency	Reduced [7]	Improved [7]
Structural
Later wall hypertrophy	Increased [8]	Decreased [4]
LV volume	Increased [9]	Decreased [10]
Hemodynamic
Mitral valve regurgitation	Increased [11]	Decreased [11]
Systolic function	Depressed [11]	Improved [12]
Diastolic time interval	Shortened [11]	Prolonged [11]
Force frequency relationship	Blunted [13]	Restored [13]
Clinical
Hospitalization rates	Increased [11]	Reduced [14,15]
Mortality	Increased [2]	Reduced [14–16]

The late activation of the LV later wall put excessive stress on this wall segments resulting in a series of regional cellular and molecular changes within the LV. In particular, stress-activated kinases and Connexin-43 are upregulated in the late-activated region [4], which contribute to increased regional hypertrophy [8] and altered metabolic profile mirrored by an increase in glucose metabolism [5]. Furthermore, the expression of genes and proteins that are important for Ca²⁺ homeostasis, and thereby contractility and arrhythmias, are heterogenic in the LV [4]. Similarly, ion channels are affected regionally resulting in prolonged action potential duration in the later wall [5]. The latter may contribute to increased susceptibility to ventricular arrhythmias, which are common in HFrEF. Importantly, CRT diminishes most of these regional differences on the cellular and molecular level [4,5].

Although reversal of the regional heterogenic cellular and molecular footprint of dyssynchrony is likely an important mechanism responsible for CRT efficacy, some important effects of resynchronization are actually global (evident in both early- and late-activated or remote regions, unaffected by dyssynchrony). In particular, CRT has been shown to reduce global apoptosis, reverse dyssynchrony-induced fetal gene reprogramming [6], improve β-adrenergic signaling and sarcomere Ca²⁺sensitivity [5]. Downregulation of β-adrenergic receptors and a blunted response to catecholamine stimulation occur as a response to chronic overactivation of the sympathetic nervous system and are important pathophysiological features of HFrEF. In experimental studies it has been shown that CRT results in upregulation of β-adrenergic receptors and improves myocyte response to isoproterenol stimulation [5]. The improved catecholamine response has been linked to upregulation of regulator of G-protein signaling 3 (RGS3), which is seemingly a unique mechanism of CRT, not observed with other HFrEF therapies [5]. Importantly, these beneficial changes are also observed in tissue samples from patients with successful CRT treatment and are associated with improved autonomic balance in CRT responders [17].

CRT improves contractility in patients with dyssynchronoy [12], which likely is an important therapy objective in HFrEF. However, increased contractility may actually have adverse long-term effects if it is associated with increased intracellular Ca²⁺levels and myocardial O₂ demand. Therefore, it is notable that CRT exerts its positive inotropic effects without increasing cellular Ca²⁺transients but instead sensitizes sarcomeres to Ca²⁺[5]. This effect may be contributed to the activation of the kinase Glycogen Synthase kinase 3β, which phosphorylates residues on the sarcomeres and thereby improves their contractile response to any given Ca²⁺level. Notably, this may serve as an explanation as to why the improvement in systolic function with CRT is not associated with increased myocardial O₂ demand, but improved energy efficacy [7].

In addition, this and other CRT-induced changes in Ca²⁺homeostasis and timing of cardiac cycle intervals may explain the improved force frequency relationship observed in synchronous pacing over dyssynchronous pacing modes [13]. A blunted force frequency relationship is another hallmark in HFrEF and restoration of a positive inotropic state with increased heart rate is likely an important mechanism of CRT efficacy.

Taken together, these cellular and molecular, metabolic and hemodynamic CRT effects translate in to reverse remodeling in terms of decreased LV volumes [10] and reduced hospitalization rates and mortality [14,15].

Clinical trials of CRT in heart failure with reduced ejection fraction and prolonged QRS-duration

During the last 15 years, close to 9000 patients have been included in the key randomized controlled clinical trials that have documented the efficacy of CRT in patients with HFrEF and QRS-prolongation. A summary of these key trials is shown in Table 2.

Table 2. Clinical trials of CRT in patients with HFrEF.

Study (Reference)	Main inclusion criteria	Study design	Sample size	Follow-up (M)	CRT-effect
CRT in NYHA III–IV
MUSTIC SR [18]	EF ≤35% LVEDD >60mm QRS ≥150 ms	1:1 CRT ON/OFF crossover	67	3/12	Improved 6MWT, QOL, pVO2, Reduced HF-hospitalization
MUSTIC AF [19]	EF ≤35% LVEDD >60mm QRS ≥150 ms	1:1 CRT ON/OFF crossover	37	3/12	Improved 6MWT, QOL, pVO2, reduced HF-hospitalization
PATH-CHF [20]	EF ≤35% QRS ≥120 ms PR ≥150ms	CRT vs. RV pacing vs. epicardial LV pacing crossover	41	12	Improved pVO2 and 6MWT
CONTAK CD [21]	EF ≤35% QRS ≥120 ms	1:1 CRT ON/OFF	290	6	Improved NYHA, VE/CO2, LV volumes, LVEF
MIRACLE [22]	EF ≤35% LVEDD ≥55 mm QRS ≥130 ms	1:1 CRT ON/OFF	453	6	Improved NYHA, pVO2, 6MWT
MIRACLE-ICD [23]	NYHA III and IV EF < =35% QRS ≥130	1:1 CRT ON/OFF	555	6	Improved NYHA, VE/CO2, LV volumes, EF
COMPANION [24]	EF ≤35% QRS ≥120ms	1:2:2 OMT/CRTP/CRTD	1520	15	Reduced primary endpoint (death or hospitalization)
CARE-HF [14]	QRS ≥120 ms EF ≤35%	1:1 OMT vs. CRTP	813	30	Reduced death and CV hospitalization
CRT in NYHA (I)-II
REVERSE [10]	EF ≤40% QRS ≥120ms LVEDD ≥55mm	2:1 CRT ON/OFF	610	12-24	Reversed remodeling and longer time to first hospitalization
MADIT-CRT [25]	EF <30% QRS ≥130ms	3:2 CRT-D vs. ICD	1820	28.8	Reduced combined hospitalization and mortality
MADIT-CRT [16] (extended)	EF <30% QRS ≥130ms	3:2 CRT-D vs. ICD	854	84	Reduced mortality
RAFT [15]	EF ≤30% QRS ≥120ms (>200ms if paced)	1:1 CRTD vs. ICD	1798	40	Reduced mortality and hospitalization

CV: cardiovascular; EF: ejection fraction; LVEDD: left ventricular end diastolic diameter; M: months; ms: milliseconds; NYHA: New York Heart Association Functional Class; OMT: Optimal Medical Therapy; pVO₂: peak oxygen consumption; QOL: quality of life; 6MWT: Distance covered in the 6-minute Walk Test.

CRT in moderate-to-severe (NYHA III-IV) heart failure and electrical dyssynchrony

Early-randomized controlled clinical trials focused on investigating CRT efficacy in patients with moderate-to-severe HFrEF (EF≤ 35%), evident as NYHA class III–IV symptoms, and with electrical dyssychrony (QRS duration ≥120 ms). The first of such studies, the MUSTIC trial, included 67 patients in a randomized, controlled, double-blind and cross-over study designs [18]. In this study, CRT was associated with significant improvements in the 6-minute walk test (the primary endpoint) and quality of life. Following this early trial, several similar studies established the beneficial effect of CRT in similar patients in terms of reduced morbidity [19–23].

Subsequently, two pivotal randomized controlled clinical trials were able to document beneficial CRT effects on morbidity and mortality in similar patients. In the COMPANION trial, 1520 patients with EF ≤35%, NYHA III-IV HF with QRS ≥120 ms were randomized to receive CRT with a defibrillator (CRT-D), CRT without a defibrillator (CRT-P) or optimal medical therapy. At 12 months follow-up CRT, either alone or in combination with an ICD, significantly reduced the primary outcome: combined death or hospitalization for any cause [24]. In the CARE-HF study [14], 813 patients were randomized to CRT-P or optimal medical therapy and followed for an average 30 months. For the first time, CRT was associated with a significant and substantial reduction in all-cause mortality (Figure 2(a)). Importantly, long-term follow up of the MUSTIC study and CARE-HF showed that CRT resulted in progressive reverse remodeling which opened the question whether CRT might also be beneficial in less severe HF by halting or stopping disease progression [19].

Figure 2. Effect of CRT on mortality in patients with (A) NYHA III-IV and mainly (B) NYHA II HFrEF. (A) Effect of CRT on all-cause mortality compared to optimal medical therapy from the CARE-HF trial.(B) Effect of CRT-D vs. ICD alone on all-cause mortality from the RAFT-trial. Figure 2(a) is from Ref. [14]. Copyright ? (2005). Reprinted with permission from Massachusetts Medical Society. Figure 2(b) is from Ref. [15]. Copyright ? (2010). Reprinted with permission from Massachusetts Medical Society.

Figure 3. CRT induces reverse remodeling in both patients with ischemic and nonischemic cardiomyopathy.

Reprinted from Ref. [28], Copyright © (2010), with permission from Elsevier.

CRT in mild (NYHA I-II) heart failure and electrical dyssynchrony

The abovementioned clinical trials paved the way to establish CRT as an important therapy option in selected patients with moderate-to-severe HFrEF. Subsequent trials focused on whether CRT was effective also in slowing disease progression and reducing clinical events in patients with less severe HFrEF (i.e. NYHA class I–II symptoms). The clinical efficacy of CRT in these patients has been documented in three large clinical trial programs [10,15,16,25].

These studies included patients with HFrEF (EF ≤30, 35 or 40%), QRS-prolongation (cut-off: 120–130 ms) and predominantly NYHA-class II symptoms. In the REVERSE study, CRT did not improve HF clinical composite score, which was the primary endpoint [10] but was associated in significant reverse remodeling (powered secondary endpoint) and significantly increased the time to first hospitalization. In the MADIT-CRT trial, 1820 patients with EF ≤ 30%, QRS duration ≥130ms and NYHA class I–II symptoms were randomized to either CRT-D or ICD alone. During an average follow up of 2.4 years the addition of resynchronization on top of ICD-treatment resulted in a significant reduction in the combined primary endpoint of all-cause death or HF event (Hazard ratio [95%confidence interval]: 0.66 [0.52–0.84]) [25]. This was mainly driven by a reduction in HF events and all-cause mortality was not reduced with CRT in the first trial analysis. However, in a study extension with 7 years follow-up, CRT was associated with a reduction in all-cause mortality (Hazard ratio [95%confidence interval]: 0.56 [0.43–0.80] in patients with LBBB [16].

In the RAFT trial, 1798 patients were randomly assigned to receive CRT-D or ICD alone. Inclusion criteria were similar to the MADIT-CRT trial but with three major exceptions. First, a small number of patients with NHYA III symptoms were included (20%). Second, patients with HFrEF and RVP for bradycardia indication were enrolled and upgraded to CRT. Last, patients with atrial fibrillation (AF) were included for the first time in a major CRT trial. With a mean follow-up of 40 months, longer than in MADIT-CRT, the primary endpoint (all-cause mortality and HF hospitalization) was significantly reduced in the CRT treated group (hazard ratio [95% confidence interval]: 0.75 [0.64–0.87] [15]. Importantly, all-cause mortality alone was reduced with CRT, which may reflect the longer follow-up period (Figure 2(B)). Subgroup analyses of the CRT effect in patients with AF and those upgraded from conventional RVP were limited by the small number of included patients (13% with AF and 7% with upgrade from RVP). However, there were no significant or substantial differences in CRT effect between these subgroups.

Current challenges in CRT delivery

Although CRT is clearly a major contributor to improved management of selected patients with HFrEF, some challenges remain. One is that a substantial proportion of CRT patients (30–40%) do not respond to the therapy in terms of improved symptoms and/or echocardiographic evidence of reverse remodeling. Below, we will discuss if improved patients selection and/or optimization of therapy delivery can increase the response rate. Other important current challenges of CRT are the lack of trial-based evidence of CRT efficacy in certain clinical settings and low implementation of CRT in appropriate patients. These two latter challenges will also be discussed below.

Patient selection and prediction of therapy response

QRS width and bundle branch morphology

CRT became an established therapy in HFrEF, based on which electrical dyssynchrony was measured as QRS widening on the 12 lead ECG. Wide QRS was thus the inclusion criteria in all conducted randomized controlled studies. Nonetheless, 2/3 of patients in these trials also had LBBB meaning that the bulk of evidence is within this subgroup. In contrast, right-bundle branch block (RBBB) was the least common conduction disturbance (5–10%) [10]. Subgroup analysis from clinical trials have highlighted that CRT is more effective in patients with LBBB compared to RBBB [15,16]. These results were subsequently adopted in the clinical guidelines from the European Society of Cardiology [1]. Since the precise mechanisms by which CRT reduces morbidity and mortality are incompletely understood, targeting the right patients is difficult. Thus, whether it is only the QRS width itself or if, bundle-branch morphology better identifies patient response is unclear. Although observational studies suggest that a non-LBBB QRS morphology is associated with a worse outcome [16] this may reflect the very different clinical profile of patients with different QRS morphologies. Furthermore, QRS duration and QRS morphology are strongly related. Patients with intraventricular conduction delay (IVCD) usually have shorter QRS duration and therefore would not be expected to benefit from CRT. Patients with RBBB on the other hand, generally have QRS duration similar to patients with LBBB; since such patients usually have left anterior or posterior hemiblock. Meta-analyses using aggregate data are confounded by their inability to investigate the relationship between QRS duration and morphology. Case-based meta-analysis overcomes this because QRS duration and morphology are known for each individual patient and their interaction can be studied. Importantly, a large case-based meta-analysis [26], found prolonged QRS-duration to be strongly predictive of the response to CRT. Moreover, neither etiology, ejection fraction (EF) nor NYHA class were independent predictors of response to CRT.

Conversely, later studies have firmly established that CRT is not indicated in patients with narrow QRS. In the ECHO-CRT trial, the value of mechanic dyssynchrony as indication for CRT was tested in patients with narrow or mildly prolonged QRS [27]. This study was a prospective randomized controlled trial, which showed that patients with a QRS duration or <130 ms may be harmed by CRT even if they have evidence of left ventricular dyssynchrony on echocardiography and that this applies not only to those with a QRS duration of <120 ms but also in those with QRS 120–130 ms. Therefore, the most recent guidelines firmly state that CRT should not be given to patients with narrow QRS (<130 ms) [1].

Ischemic vs. non-ischemic heart disease

Ischemic heart disease (IHD) and hypertension are the most common underlying causes of HFrEF. Accordingly 50–60% CRT recipients have underlying IHD with or without a previous myocardial infarction, including with or without scar tissue. Not surprisingly, the magnitude of LV reverse remodeling is less prominent, and the overall outcome worse, than in nonischemic patients [28]. Patients with ischemic etiology are also more likely to have RBBB [10] and to have an intrinsically worse prognosis than those with LBBB. Despite these facts, evidence indicates that patients with underlying IHD derive outcome benefits similar to those with nonischemic etiology.

In the CARE-HF trial, patients with ischemic etiology assigned to CRT experienced a significant reduction in total mortality, which was similar as for those with nonischemic cardiomyopathy [29]. Similar observations were made in the REVERSE trial of patients with milder HFrEF [28]. As in CARE-HF, the CRT-benefit on clinical events was similar comparing ischemic, and nonischemic cardiomyopathy despite less impressive reverse remodeling for those with ischemic cardiomyopathy. Indeed, patients with nonischemic dilated cardiomyopathy patients derive a 2–3 times greater magnitude of reverse remodeling in spite of larger LV volumes at baseline (Figure 3) [28]. It can be speculated that CRT may correct conduction in myocardial tissue dilated by delayed activation but is less effective when contractility is impaired by extensive myocardial scar tissue even in the presence of conduction delay. Importantly and similar to other CRT studies [29], LV function in IHD patients in REVERSE was significantly better at baseline than in non-IHD patients despite the fact that IHD patients less often received at least 50% of target dose of beta-blockers, a drug that in itself induces reverse LV remodeling to an extent related to dose [28]. These observations together with the lesser extent of reverse remodeling by CRT could reflect the less “plasticity” of myocardial scar tissue both to dilate and shrink and confirms what has been previously shown in other randomized CRT-studies or registry studies of patients in NYHA III–IV. It also implies that CRT works by mechanisms beyond reverse LV remodeling.

Atrial fibrillation

AF is common in HFrEF with a prevalence ranging from 10 in NYHA I to 50% in NYHA IV [30]. Notably, with the exception of one early small, randomized clinical trial [19] and a small population of the RAFT trial [15], the large-scale clinical trials have almost exclusively included patients with sinus rhythm (SR) at baseline. Therefore, the evidence for efficacy of CRT is less studied in patients with symptomatic HFrEF, wide QRS and AF. Theoretically, CRT may be less effective in AF compared to SR for two primary reasons. First, at least a smaller part of the CRT effect is likely due to improved atrio-ventricular coupling as a delayed atrio-ventricular timing, compromising ventricular filling, is common in patients with HFrEF and SR. This effect will be absent in AF. Second, the RR interval may be highly variable in AF. At short RR intervals, ventricular pacing may be inhibited, and fusion, or pseudo fusion beats may occur. This may be an important limitation in applying CRT in AF since an even small declines in the percentage of captured (i.e. resynchronized) ventricular beats is associated with worse outcome in CRT recipients. This is observed even if as few as <5–10% of beats are not biventricularly paced [31].

In the clinical setting, patients with HFrEF and AF can be suitable CRT candidates based on two scenarios. First, patients may have QRS prolongation and symptoms warranting a CRT based on the evidence they had in SR. Second, patients may have symptomatic HFrEF and AF where pharmacological rate control is challenging and atrio-ventricular junction ablation (AVJA) is considered. Even though there are no randomized clinical trials that support CRT efficacy in these scenarios, some evidence exists. In two smaller clinical trials CRT implantation after AVJA was associated with improved LV structure and function as well as a reduction in clinical adverse events compared to RVP [32]. In addition, some meta-analyses have compared the effect of CRT in patients with AF and SR. The results from these meta-analyses are conflicting since some show no difference between CRT efficacy in SR vs. AF [33], while other indicate that CRT is less effective in AF [34].

In the light of the weak trial evidence of CRT efficacy in patients with AF future large-scale clinical studies are warranted. Nevertheless, the ESC guidelines [1] award a IIa (level of evidence B) recommendation for CRT use in AF patients with NYHA III–IV symptoms, reduced EF and QRS duration ≥120 ms where a high degree of ventricular pacing can be achieved (either pharmacologically or by AVJA), and also in patients with reduced EF where AVJA is planned for rate control of AF.

Gender aspects

Women have usually been underrepresented in clinical trials of CRT. However, it is not clear if this is a result of gender biased inclusion and utilization or if a lower proportion of women with HFrEF meet indication for CRT. However, there is a growing body of evidence from registry studies, randomized controlled clinical trials and meta-analyses that suggest that the CRT effect is larger in women compared to men. Data from consecutive patients from a single center has documented that women have significantly lower all-cause mortality after CRT implantation than men (adjusted hazard ratio: 0.44) [35]. Even more importantly, subgroup analysis from MADIT-CRT and the RAFT trial have revealed that the beneficial effect of CRT on clinical outcome is either significantly (MADIT-CRT) or near-significantly (RAFT) larger in women compared to men, in spite of low female representation [15,25]. The reasons for the apparent superior CRT efficacy in women are not fully understood. However, it is hypothesized that women in general have smaller hearts than men and therefore the same QRS duration may imply greater amount of dyssynchrony in women. This could point to a larger amendable substrate and hence translate into a larger therapy effect.

Renal function

Chronic kidney disease (CKD) is common in patients with HF and leads to worse prognosis and therapeutic challenges. In the randomized controlled clinical trials of CRT, patients with CKD have generally been excluded. In the REVERSE trial, patients with serum creatinine >265 μmol/L were excluded, while patients with milder degree of CKD were included, and the effect of CKD on LV remodeling was studied. Of 561 patients in this substudy, 401 had estimated glomerular filtration rate (eGFR) ≥ 60 mL/min/1.73m² and 160 had CKD defined as eGFR <60 mL/min/1.73m². At baseline, LV function was similar between the two kidney function groups. After the 12-month follow-up period, CKD was associated with adverse cardiac remodeling including worsening function and larger LV size after adjustment for all confounding variables including CRT status. There was, however, no significant interaction between CKD and CRT status on the outcome of clinical adverse events. Analysis of the results from this indicates less extent of reverse remodeling in CKD patients but potentially a comparable effect on clinical outcome [36]. In the MADIT-CRT trial, renal function was assessed in a different way from REVERSE. In this study, patients with an elevated ratio of blood urea nitrogen to serum creatinine experienced a significantly greater reduction in the risk of HF or death with CRT-D as compared with patients with a low ratio. These findings suggest an association between baseline renal function and response to CRT [36]. On the other hand, severe renal dysfunction is associated with an increased risk of device-related complications, which should be taken into noted when CRT is considered. Future studies are needed to evaluate whether CRT improves outcomes and delays HF progression in CKD patients.

Can CRT efficacy be improved by optimizing therapy delivery and/or device programming?

Left ventricular lead position

Theoretically, the region in the LV that is activated latest during LBBB, RVP or other conduction abnormalities, should be the optimal site for lead placement and subsequent stimulation/activation in order to obtain maximal resynchronization. However, this notion is hampered by at least two factors. LV lead placement in areas with scar tissue is undesirable, and it may be difficult to reach desired target LV region depending on the anatomy of the sinus coronaries venous system.

A targeted LV lead placement can be performed on anatomical bases or individualized, based on imaging techniques. Subgroup analyses from the major CRT trials are somewhat conflicting regarding the superiority of aiming for a specific anatomical region during implantation. In the COMPANION trial, the CRT effect was similar in patients with anterior, lateral or posterior lead placement. On the other hand, subgroup analyses from the REVERSE and MADIT-CRT trials have indicated that a lateral lead position is preferential and that an apical position (compared to basal or mid ventricular) should be avoided. However, there is no clear evidence of a supreme anatomical area that routinely should be targeted during CRT implantation although the ESC guidelines recommends against an apical lead position [37].

Studies have investigated if identifying a target for LV lead placement, on an individual patient level, using speckle tracking [38,39], can enhance the CRT effect. In the largest and best-controlled studies it has been shown that speckle tracking guided lead placement to the areas with the latest activation but without scar yields significantly reduced adverse clinical events [39] and higher response rates [38]. However, LV placement guided with speckle tracking may be associated with high costs and may be less effective in centers with less experience. Therefore, such strategy should be further studied before implemented on a large scale.

Bipolar vs. multipolar leads

In recent years, a multipolar pacing lead has been developed to facilitate multipolar pacing (MPP) using a single LV lead. As opposed to conventional pacing (with a bipolar pacing lead) the multipolar lead can be used for tailored programming of active poles in longer segment of the lead. Thereby, it offers more pacing vectors when multiple poles are active simultaneously. Theoretically, this technical advancement offers some potential benefits. The most prominent among these are the possibility to shift pacing configuration in case of phrenic nerve stimulation, avoid anatomical areas with initial high thresholds, and to optimize resynchronization by LV stimulation involving multiple vectors. Regarding the latter, small studies have shown that MPP acutely reduces dyssynchrony and improves hemodynamics compared to conventional CRT with a bipolar electrode [40]. Subsequently, results from a small clinical study suggest that these acute effects translate into improved mid-term echocardiographic and clinical response to CRT [41]. Similarly, in a recent registry study, MMP was associated with higher resolution of phrenic nerve stimulation, less need for lead repositioning and lower all-cause mortality [42].

However, large-scale RCTs documenting efficacy of MMP over conventional CRT in terms of clinical end-points are currently lacking but are actively pursued, e.g. in the MORE CRT-MPP study (Clinical Trials.gov Identifier: NCT02006069).

Individual optimization of device programming

The delay between atrial sensing or stimulation and ventricular stimulation (AV-delay), the timing interval between stimulation in each ventricle (VV-delay) as well as the basic programmed paced heat rate (pHR) are all programmable parameters that may affect cardiac function and thereby the CRT effect. This has led to substantial efforts to prove that optimizing these parameters on an individual patient level improves therapy efficacy.

Several studies have documented that optimization of the AV-delay improves the acute hemodynamic response to CRT [12]. In addition, smaller randomized studies with limited follow-up periods documented that AV-delay optimization could improve symptoms and/or increase the number of responders [43]. However, the only larger randomized controlled clinical study of individual AV-delay optimization failed to demonstrate a significant beneficial effect of this strategy [44].

Similarly, optimization of the VV-delay acutely improves hemodynamic in CRT patients but has not been associated with substantially improved outcome in trials with clinical outcome variables [45].

Considering that individual AV-, and VV-delay optimization is time consuming, and associated with additional costs but not with documented improved outcome routine device optimization is likely not feasible. However, it may still be considered in patients who do not respond to CRT with “out of the box” settings.

Compared to optimization of the AV- and VV-delay, the programming of different pHRs has not gained the same interest in the literature, in spite of being a programmable parameter with possible large hemodynamic and clinical effects. We have previously demonstrated that a higher pHR in CRT patients results in acute hemodynamic improvements in terms of increased cardiac output and reduced filling pressure, which is sustained during 2 weeks of ambulatory living [46]. Moreover, increased pHR is associated with reduced sympathetic activity [47]. It is important to notice that these beneficial effects may be outweighed by increased metabolic demand or other factors when pHR is permanently increased in CRT recipients. Nevertheless, the clinical effect of different pHR should be studied further.

Implementation of CRT in patients with heart failure

In this article we have addressed numerous aspects of evolving CRT technology, CRT optimization and expanding CRT indications. However, health care systems and clinicians caring for HF patients must also consider optimal implementation of existing evidence and optimal utilization of CRT as currently available. In the US, where active quality improvement programs such as the Get With The Guidelines (GWTG) – HF Registry have contributed to the improved utilization of evidence-based HF interventions, a referral center reported an increase in CRT use in HFrEF from 0% in the 1990s to 39% in the 2000s [48]. Similarly, absolute numbers of CRT (and ICD) implantations increased in Europe throughout the 2000s [49].

However, CRT utilization in less selective, more generalizable cohorts is less studied and may be inadequate. In a comprehensive analysis from the nonselective Swedish Heart Failure Registry (SwedeHF), in NYHA II-IV HF of at least 6 months duration (with time to optimize medical management) and EF <30%, CRT use increased nonsignificantly from 2.4% to 8.2% between 2003 and 2012, despite indication in about 30% in HFrEF, and this was also associated with a lack of improvement in risk-adjusted mortality [50].

The reasons for underutilization are not known. Perceptions of high cost may persist. In SwedeHF, the elderly [51] and patients with more recent indications such as those with NYHA II [52] may receive less CRT, but women do not [53]. A low awareness of indication for device therapy in HFrEF among health care providers may also contribute [54]. Certainly efforts to increase awareness are necessary to improve utilization, but as treatment options in HF get more complex while HF patients are cared for extensively in primary care, sufficient awareness may be difficult to achieve and novel approaches such as screening may be necessary.

Future perspectives

Expanding indication for CRT: should CRT be used in patients with an indication for antibradycardia pacing to prevent left ventricular dysfunction induced by right ventricular stimulation?

Conventional ventricular pacing for bradycardia is delivered by RVP. RVP (especially with an apical lead position) results in an activation and contraction pattern that is similar to LBBB, and thereby compromised systolic function [55].

Most patients without structural heart disease tolerate chronic RVP well. However, there is evidence that RVP results in a dose-dependent deterioration in cardiac function and increased clinical events in patients with structural heart disease [56]. In a clinical context, there are two clinical scenarios where this may be important. First, patients may develop LV dysfunction during antibradycardia pacing using RVP. Second, patients with preexisting LV dysfunction may develop an indication for chronic antibradycardia pacing. There are some studies that suggest upgrading a conventional RVP device to a CRT (scenario 1) and de novo implantation of a CRT instead of a conventional RVP device (scenario 2) may be beneficial.

Regarding upgrading to CRT, the RAFT trial was the only major clinical CRT trial that included a small subgroup of these patients [15]. The subgroup was too small to draw firm conclusions on. However, several smaller clinical studies have investigated the effect of CRT vs. RVP in patients with EF ≤40% and symptoms of HF. In general, these studies have shown that CRT results in improvement in EF and symptoms and reduced LV volumes [57]. In addition, data from both observational studies and registries support the upgrade of RVP devices to CRT in patients with HFrEF. Based on this evidence, the most recent pacing guidelines published by the European Society of Cardiology awards upgrading an RVP to CRT, a class I recommendation [37].

So far, a total of seven randomized controlled clinical trials have been published where de novo implantation with CRT was compared to RVP device implantation in patients with a bradycardia indication for ventricular pacing and varying degrees of LV function and HF symptoms.

The HOBIPCAE, COMBAT and BLOCK HF trials included a total of approximately 880 patients with a bradycardia indication for RVP (mostly high degree AV-block) and moderate-to-serve systolic dysfunction [37]. These trials consistently showed improved outcome of de novo CRT compared to RVP device implantation in terms of improved cardiac structure and function, improved quality of life and reduction in clinical events.

Four other trials have randomized patients with preserved EF and a bradycardia indication for ventricular pacing to receive either CRT or RVP [37]. In general, these studies showed that CRT could prevent the deterioration in heart function and progression of remodeling. The impact of CRT on more clinically relevant endpoints in this setting remains largely unknown due to lack of sufficiently powered clinical trials.

More evidence from clinical trials of CRT efficacy in these two clinical settings is therefore warranted.

Leadless pacing and CRT

Recently, a system for leadless RVP has been developed [58]. This technique, which doesn’t require a pacemaker pocket or lead may be an important innovations in cardiac pacing. A lead with a generator is introduced transvenously to the heart and screwed into the endocardium. At present the limitations to widespread use of this technique is the large introduction diameter sheeth (16 Fr), which is linked to bleeding complications, and the lack of atrio-ventricular synchronous pacing. For the future, one may speculate that there are also potentially significant clinical benefits of leadless LV endocardial pacing. Endocardial LV pacing is more physiological and has the potential to be less pro-arrhythmic by reduced dispersion of ventricular repolarization. Moreover, it likely requires lower pacing energy outputs compared with optimally placed coronary sinus leads (epicardial pacing). Furthermore, because it is not limited to those coronary sinus branches able to accommodate a transvenous lead, endocardial pacing offers a larger choice of optimal LV stimulation sites; there is also the added benefit of no phrenic nerve stimulation. Leadless LV endocardial pacing might mitigate these limitations and expand our ability to provide optimal CRT. However, leadless pacing has not yet been tested as an alternative to conventional LV pacing and its’ place in a CRT context is therefore unknown.

Importance of dyssynchrony and resynchronization in a novel clinical arena

Recently, the importance of dyssynchrony and resynchronization has been acknowledged in a novel clinical arena: conduction abnormalities or need for ventricular pacing after transcatheter aortic-valve implantation (TAVI).

Symptomatic severe aortic stenosis is associated with up to 50% yearly mortality and recent trials have demonstrated that TAVI reduces morbidity and mortality in patients with serve aortic stenosis not suitable for surgery. Accidental, procedure-related damage to the conduction system resulting in high degree AV-block or LBBB occurs in approximately 30% of patients during TAVI and seems to be more common with the self-expanding Core valve compared to other devices [59]. Both the need for permanent RVP and de novo LBBB have been linked to increased mortality following TAVI [59,60]. Less is known about the mechanisms linking RVP/LBBB to increased mortality. However, studies have documented an absence of TAVI-induced EF improvement and reverse remodeling in patients who develop LBBB/RVP [59,60]. This suggests that dyssynchrony may be a risk factor linking LBBB/RVP to poor ventricular recovery and subsequent worsened clinical outcome. CRT can be considered for antibradycardia pacing after TAVI-induced AV-block based on current ESC recommendations (see above), but has not been tested in this patient group. Also, whether CRT is beneficial in patients with persistent LBBB following TAVI remains unknown but should be studied.

Disclosure statement

MS: Grant/Speaker: Medtronic Inc., Boston Scientific, St Jude Medical. Consultant: Medtronic, Inc.

FB: Grant/speaker Boston, St-Jude Medical, Medtronic. Consultant: St-Jude Medical, Biotronik, Medtronic.

FG: None. LM: None. LL: Speaker’s or consulting honoraria from St Jude, Novartis, Bayer, ViforPharma, and HeartWare and research grants to author’s institution from Boston Scientific, Medtronic and AstraZeneca.

CL: Grant/Speaker: Medtronic Inc., Boston Scientific, St Jude Medical, Biotronik, LivaNova, Impulse Dynamics. Consultant: Novartis, Vifor Pharma, Cardio-3.