Expert Review of Cardiovascular Therapy Downloaded from informahealthcare.com by National Taiwan University on 06/27/15 For personal use only.
Review
Narrow QRS systolic heart failure: is there a target for cardiac resynchronization? Expert Rev. Cardiovasc. Ther. 13(7), 783–797 (2015)
Tom Jackson*1, Simon Claridge1, Jonathan Behar1, Eva Sammut1, Jessica Webb1, Gerald Carr-White2, Reza Razavi1 and Christopher Aldo Rinaldi1,2 1 Department of Cardiovascular Imaging, 4th Floor Lambeth Wing, St Thomas’ Hospital, London, SE1 7EH, UK 2 Cardiology Department, Guy’s and St. Thomas’ Hospitals, London, UK *Author for correspondence: Tel.: +44 20 7188 7188; extn. 50871 [email protected]
informahealthcare.com
Cardiac resynchronization therapy has revolutionized the management of systolic heart failure in patients with prolonged QRS during the past 20 years. Initially, the use of this treatment in patients with shorter QRS durations showed promising results, which have since been opposed by larger randomized controlled trials. Despite this, some questions remain, such as, whether correction of mechanical dyssynchrony is the therapeutic target by which biventricular pacing may confer benefit in this group, or are there other mechanisms that need consideration? In addition, novel techniques of cardiac resynchronization therapy delivery such as endocardial and multisite pacing may reduce potential detrimental effects of biventricular pacing, thereby improving the benefit/harm balance of this therapy in some patients. KEYWORDS: biventricular pacing . cardiac resynchronization therapy . dyssynchrony . heart failure . narrow QRS
Heart failure is a modern day epidemic, with its prevalence expected to increase by 25% by 2030 [1]. Advances in neurohormonal modification have improved survival and reduced hospitalization rates but still over one-third of patients will die within a year of diagnosis [2]. Cardiac resynchronization therapy (CRT) is an established therapy for patients with systolic heart failure and a prolonged QRS duration, which usually manifests as left bundle branch block (LBBB) on the surface electrocardiogram. Early CRT trials, MUSTIC [3] and MIRACLE [4], showed improvements in symptoms, exercise tolerance and quality of life, in the broad QRS heart failure population, as well as a reduction in adverse ventricular remodeling. Subsequent larger randomized clinical trials, COMPANION [5] and CARE-HF [6], showed reduction in mortality and hospitalization with heart failure in this group. Recently, benefits have been demonstrated in patients with milder degrees of heart failure [7,8]. Since 2003, there have been a number of studies investigating the impact of CRT on patients with a narrow QRS duration. These studies have produced varying results; however, they do have diverse inclusion criteria and endpoints. These differences are particularly apparent in whether and how mechanical
10.1586/14779072.2015.1049945
dyssynchrony is assessed and also what criteria are used to define response to CRT. The definition of response is a particular shortcoming of CRT studies, Fornwalt et al. [9] assessed 15 different primary response criteria from the 26 most cited publications on predicting response to CRT and applied them to 426 patients from the PROSPECT study [10]. They showed that the overall response rate ranged from 32 to 91% with poor agreement between two criteria occurring 75% of the time and strong agreement only 4%. Although this study involves patients with a prolonged QRS duration, its result does nevertheless have relevance for the studies in this review. The key question that needs to be considered is, by which biologically plausible mechanism or mechanisms could CRT improve patients with a narrow QRS duration. Biventricular pacing (BiVP) was initially proposed as a technique to treat intraventricular conduction delay and its effect on left ventricular (LV) contraction [3]. In the absence of conduction delay, it is unclear whether there is a target by which CRT may confer benefit. If the aim of BiVP is to improve coordination of a dyssynchronous ventricle, then we need to consider whether dyssynchrony exists in narrow QRS patients, what mechanisms could be
2015 Informa UK Ltd
ISSN 1477-9072
783
Review
Jackson, Claridge, Behar et al.
Expert Review of Cardiovascular Therapy Downloaded from informahealthcare.com by National Taiwan University on 06/27/15 For personal use only.
Afterload effects: Leading to heterogeneous ventricular contractility
Regional variability: Cytokines Stress kinases Cacium handling Structural proteins Action potential duration Myocardial blood flow
Effect of scar
Figure 1. Possible mechanisms for mechanical dyssynchrony in narrow QRS patients.
leading to this dyssynchrony and whether these mechanisms are amenable to improved timing with BiVP? A further consideration is whether reversal of mechanical dyssynchrony is the overarching mechanism to target for these patients, as there may be other possible benefits to BiVP such as manipulation of atrioventricular delay or the ability to further titrate medications post pacing, which could play an important role. Measuring mechanical dyssynchrony
Mechanical dyssynchrony is a term used to describe systolic intraventricular mechanical variability. This can be measured by a selection of imaging modalities. These modalities include echocardiographic techniques such as M-mode, tissue Doppler, speckle tracking for strain and strain rate, 3D echocardiography and 3D speckle tracking. Other imaging methods used consist of cardiac magnetic resonance (CMR) protocols, radionuclide imaging and cardiac computed tomography [11]. Each of these approaches has performed well at predicting responders to CRT in single-center studies. Furthermore, each has strengths and weaknesses pertaining to the spatial and temporal performance of the modality used, and whether or not it is a global timing dispersion measure or a measure of segmental delay. However, no single approach is yet to be adopted as the method of choice by the CRT community as none have demonstrated a consistent superiority for response prediction [12]. One possible explanation of the failure of a single measure to outperform the others is the test re–test variability for all dyssynchrony measures and the variability in the measurement of all remodeling indices [13]. Evidence for mechanical dyssynchrony in narrow QRS patients
Several studies have suggested that patients with narrow QRS may have ‘mechanical dyssynchrony’. Yu et al. in 2003 used echocardiography tissue Doppler imaging (TDI) to measure the 784
time to peak contraction and relaxation of the 12 LV segments in 67 patients with heart failure and narrow QRS, 45 with heart failure and wide QRS and 88 controls [14]. They defined significant systolic asynchrony as a maximum difference in time to peak systolic contraction of any two of 12 LV segments of over 100 ms; this was present in 51% of the narrow QRS group, 73% of the wide QRS and not present in the control group. Bleeker et al. (2005) also assessed dyssynchrony with TDI in 64 patients with severe LV impairment and a narrow QRS. They showed that 33% of these patients had significant dyssynchrony with a lateral to septal systolic delay of > 60 ms [15]. Chalil et al. (2007) developed a CMR-derived tissue synchronization index (CMR-TSI) from maximal segmental radial wall motion and defined significant dyssynchrony as a > 110 ms [16]. They assessed 66 patients with heart failure with varying QRS durations (147.8 ± 25 ms), and approximately 30% of the narrow QRS patients had significant dyssynchrony. This CMR protocol was further tested in 225 with heart failure patients [17]. The CMR-TSI in this study was calculated as the standard deviation (SD) of time to peak inward endocardial motion in up to 60 myocardial segments, and significant dyssynchrony was defined as 2 SDs above the mean of normal controls. Dyssynchrony was present in 91% of patients with a QRS 6 times as many genes differentially expressed between nonfailing and NICM canine hearts in anterior wall compared with lateral wall myocardium [28]. All of these mechanisms have been seen to homogenize following CRT [29,30]. In addition to regional molecular changes, LBBB NICM is characterized by regional heterogeneities in cellular and tissue electrophysiological properties, in particular, prolongation of the action potential duration in the lateral wall [31]. Action potential duration change is associated with changes in contractility [32], which may lead to dyssynergy and dyssynchrony. Regional variations in afterload have also been described and may play a role in heterogeneous contractility [19,22], or potentially regional variations in response to afterload may lead to dyssynchrony [33]. Other potential mechanisms for regional heterogeneity in NICM include heterogeneous metabolism [20] and imbalance of myocardial blood flow and metabolism [34,35]. There are, therefore, multiple systems that express regional heterogeneity in NICM demonstrated both in humans and in animals (in vitro and in vivo), and although these studies are not primarily in the context of a normal activation pattern (narrow QRS), it is possible to consider these as potential mechanisms leading to mechanical dyssynchrony in this group. In the context of ischemic cardiomyopathy, it is more obvious how areas of scar can lead to regional variety in contractility and therefore a dyssynchronous ventricular contraction. However, when an infarct does not lead to a full thickness scar, there remains contractile reserve; therefore, if a partial lateral infarct were to lead to reduced contractile strength in comparison with the septal wall, the lateral wall would be overcome and pushed out during mid- to end-systole while the intraventricular pressure is elevated [18,36]; this mechanism of contractility variability leading to dyssynchrony may also occur in the context of NICM by any of the processes outlined previously. It is also possible to conceive how mechanical dyssynchrony may also occur in the setting of small heterogeneous areas of full-thickness myocardial scar and fibrosis, which produce dyssynchronous contraction while having a negligible electrical impact on the QRS morphology [37]. Recent in vitro work has also demonstrated an interdependence in electromechanical properties between adjacent heterogeneous myocardium with opposing changes in contractility. This could, and is likely to, lead to an exaggeration of contractility differences between areas of differing myocardium, thereby encouraging further dyssynchrony [38]. Studies of CRT in narrow QRS patients Acute hemodynamic studies
The ESTEEM-CRT study by Donahue et al. included 68 patients with a narrow QRS, EF of 28.7 ms) and an indication for an ICD (TABLE 1) [39]. An acute hemodynamic study was performed in 47 patients at the time of CRT-D implant. There was no significant acute hemodynamic response (average increase in LV dP/dtmax of 2 ± 2%); however, New York Heart Association (NYHA) and quality of life scores were substantially improved at 6 and 12 months (p < 0.001), but exercise capacity and LV volumes were unchanged. Ploux et al. performed acute hemodynamic studies in 82 consecutive heart failure patients who underwent CRT implantation irrespective of their QRS duration [40]. These patients were not assessed for mechanical dyssynchrony: 34 had a QRS < 120 ms, 11 had a QRS ‡ 120 to < 150 ms and 37 had a QRS ‡ 150 ms. There was a high correlation between changes in LV dP/dtmax and baseline QRS duration and for the narrow QRS patients no significant change was seen from baseline (+0.4 ± 6.1%; p = ns). In the moderately prolonged QRS group, a nonsignificant hemodynamic increase was seen (+4.4 ± 6.9%; p = 0.06) and in the patients with the broadest QRS there was a significant increase in LV dP/dtmax (+17.1 ± 13.4%; p < 0.001). Work by Frenneaux’s group has also offered a mechanism by which LV pacing may benefit the acute hemodynamic state of patients with a narrow QRS duration and heart failure. Their work has comprehensively investigated the concept of external constraint and the effects of biventricular and LV pacing upon it. External constraint is the negative effect on LV filling from the right ventricle (RV) via the interventricular septum with direct diastolic ventricular interaction and pericardial stretch leading to pericardial constraint [41]. This is calculated during an acute hemodynamic study, where the inferior vena cava is obstructed while LV pressure and volume are measured using a conductance catheter. In patients without external constraint, there is a reduction in both LV end-diastolic pressure (LVEDP) and LV end-diastolic volume (LVEDV) with IVC obstruction; however, there is an initial increase and then deterioration in LVEDV despite a reduction in LVEDP if external constraint is present. The amount of external constraint is the difference in LVEDP between initial LVEDP and its partner of equivalent LVEDV (FIGURE 2). Bringing the timing of LV activation forward with respect to the RV by LV pacing also prioritizes LV filling ahead of RV filling, thereby reducing the volume and pressure the RV applies within the pericardium. In 2004, Turner et al. demonstrated in 20 patients with a QRS duration £ 120 ms and an EF £ 40% that a pulmonary capillary wedge pressure (PCWP) > 15 mmHg was predictive of an acute hemodynamic response to acute LV pacing with a cardiac output increase from 3.9 to 4.5 l/min (p < 0.01) and a fall in PCWP from 24.7 to 21.0 mmHg (p < 0.001); these changes were not seen in patients with a resting PCWP < 15 mmHg [42]. They postulated that a reduction in external constraint was the most likely explanation for this improvement. This was investigated further by Williams et al. who undertook acute pressure–volume studies in 30 patients with an EF £ 35%, a QRS < 120 ms and no evidence of interventricular or 785
786
Single arm 6-month FU Single site
Two arms 3-month FU Single site
Two arms 6-month FU Single site
Two arms 28-month mean FU FU review at 6 months/ 12 months and then every 12 months Single site
Two Arms Randomized blinded 6-month FU Multisite
Achilli et al. (2003)
Yu et al. (2006)
Bleeker et al. (2006)
Gaspirini et al. (2007)
RethinQ (2007)
Improvement in all endpoints except MLHFQ in narrow group No difference between groups except MLHFQ
NYHA class 6MWT MLHFQ Maximal metabolic equivalent on treadmill LVESV LVEF NYHA class 6MWT MLHFQ LVESV/LVEDV LVEF NYHA class 6MWT LVESV LVEF Mortality
Primary endpoint: Increase of peak VO2 by >1 ml/kg Secondary Endpoints: MLHFQ NYHA class
LVEF
Review
Narrow QRS systolic heart failure: is there a target for cardiac resynchronization? Expert Rev. Cardiovasc. Ther. 13(7), 783–797 (2015)
Tom Jackson*1, Simon Claridge1, Jonathan Behar1, Eva Sammut1, Jessica Webb1, Gerald Carr-White2, Reza Razavi1 and Christopher Aldo Rinaldi1,2 1 Department of Cardiovascular Imaging, 4th Floor Lambeth Wing, St Thomas’ Hospital, London, SE1 7EH, UK 2 Cardiology Department, Guy’s and St. Thomas’ Hospitals, London, UK *Author for correspondence: Tel.: +44 20 7188 7188; extn. 50871 [email protected]
informahealthcare.com
Cardiac resynchronization therapy has revolutionized the management of systolic heart failure in patients with prolonged QRS during the past 20 years. Initially, the use of this treatment in patients with shorter QRS durations showed promising results, which have since been opposed by larger randomized controlled trials. Despite this, some questions remain, such as, whether correction of mechanical dyssynchrony is the therapeutic target by which biventricular pacing may confer benefit in this group, or are there other mechanisms that need consideration? In addition, novel techniques of cardiac resynchronization therapy delivery such as endocardial and multisite pacing may reduce potential detrimental effects of biventricular pacing, thereby improving the benefit/harm balance of this therapy in some patients. KEYWORDS: biventricular pacing . cardiac resynchronization therapy . dyssynchrony . heart failure . narrow QRS
Heart failure is a modern day epidemic, with its prevalence expected to increase by 25% by 2030 [1]. Advances in neurohormonal modification have improved survival and reduced hospitalization rates but still over one-third of patients will die within a year of diagnosis [2]. Cardiac resynchronization therapy (CRT) is an established therapy for patients with systolic heart failure and a prolonged QRS duration, which usually manifests as left bundle branch block (LBBB) on the surface electrocardiogram. Early CRT trials, MUSTIC [3] and MIRACLE [4], showed improvements in symptoms, exercise tolerance and quality of life, in the broad QRS heart failure population, as well as a reduction in adverse ventricular remodeling. Subsequent larger randomized clinical trials, COMPANION [5] and CARE-HF [6], showed reduction in mortality and hospitalization with heart failure in this group. Recently, benefits have been demonstrated in patients with milder degrees of heart failure [7,8]. Since 2003, there have been a number of studies investigating the impact of CRT on patients with a narrow QRS duration. These studies have produced varying results; however, they do have diverse inclusion criteria and endpoints. These differences are particularly apparent in whether and how mechanical
10.1586/14779072.2015.1049945
dyssynchrony is assessed and also what criteria are used to define response to CRT. The definition of response is a particular shortcoming of CRT studies, Fornwalt et al. [9] assessed 15 different primary response criteria from the 26 most cited publications on predicting response to CRT and applied them to 426 patients from the PROSPECT study [10]. They showed that the overall response rate ranged from 32 to 91% with poor agreement between two criteria occurring 75% of the time and strong agreement only 4%. Although this study involves patients with a prolonged QRS duration, its result does nevertheless have relevance for the studies in this review. The key question that needs to be considered is, by which biologically plausible mechanism or mechanisms could CRT improve patients with a narrow QRS duration. Biventricular pacing (BiVP) was initially proposed as a technique to treat intraventricular conduction delay and its effect on left ventricular (LV) contraction [3]. In the absence of conduction delay, it is unclear whether there is a target by which CRT may confer benefit. If the aim of BiVP is to improve coordination of a dyssynchronous ventricle, then we need to consider whether dyssynchrony exists in narrow QRS patients, what mechanisms could be
2015 Informa UK Ltd
ISSN 1477-9072
783
Review
Jackson, Claridge, Behar et al.
Expert Review of Cardiovascular Therapy Downloaded from informahealthcare.com by National Taiwan University on 06/27/15 For personal use only.
Afterload effects: Leading to heterogeneous ventricular contractility
Regional variability: Cytokines Stress kinases Cacium handling Structural proteins Action potential duration Myocardial blood flow
Effect of scar
Figure 1. Possible mechanisms for mechanical dyssynchrony in narrow QRS patients.
leading to this dyssynchrony and whether these mechanisms are amenable to improved timing with BiVP? A further consideration is whether reversal of mechanical dyssynchrony is the overarching mechanism to target for these patients, as there may be other possible benefits to BiVP such as manipulation of atrioventricular delay or the ability to further titrate medications post pacing, which could play an important role. Measuring mechanical dyssynchrony
Mechanical dyssynchrony is a term used to describe systolic intraventricular mechanical variability. This can be measured by a selection of imaging modalities. These modalities include echocardiographic techniques such as M-mode, tissue Doppler, speckle tracking for strain and strain rate, 3D echocardiography and 3D speckle tracking. Other imaging methods used consist of cardiac magnetic resonance (CMR) protocols, radionuclide imaging and cardiac computed tomography [11]. Each of these approaches has performed well at predicting responders to CRT in single-center studies. Furthermore, each has strengths and weaknesses pertaining to the spatial and temporal performance of the modality used, and whether or not it is a global timing dispersion measure or a measure of segmental delay. However, no single approach is yet to be adopted as the method of choice by the CRT community as none have demonstrated a consistent superiority for response prediction [12]. One possible explanation of the failure of a single measure to outperform the others is the test re–test variability for all dyssynchrony measures and the variability in the measurement of all remodeling indices [13]. Evidence for mechanical dyssynchrony in narrow QRS patients
Several studies have suggested that patients with narrow QRS may have ‘mechanical dyssynchrony’. Yu et al. in 2003 used echocardiography tissue Doppler imaging (TDI) to measure the 784
time to peak contraction and relaxation of the 12 LV segments in 67 patients with heart failure and narrow QRS, 45 with heart failure and wide QRS and 88 controls [14]. They defined significant systolic asynchrony as a maximum difference in time to peak systolic contraction of any two of 12 LV segments of over 100 ms; this was present in 51% of the narrow QRS group, 73% of the wide QRS and not present in the control group. Bleeker et al. (2005) also assessed dyssynchrony with TDI in 64 patients with severe LV impairment and a narrow QRS. They showed that 33% of these patients had significant dyssynchrony with a lateral to septal systolic delay of > 60 ms [15]. Chalil et al. (2007) developed a CMR-derived tissue synchronization index (CMR-TSI) from maximal segmental radial wall motion and defined significant dyssynchrony as a > 110 ms [16]. They assessed 66 patients with heart failure with varying QRS durations (147.8 ± 25 ms), and approximately 30% of the narrow QRS patients had significant dyssynchrony. This CMR protocol was further tested in 225 with heart failure patients [17]. The CMR-TSI in this study was calculated as the standard deviation (SD) of time to peak inward endocardial motion in up to 60 myocardial segments, and significant dyssynchrony was defined as 2 SDs above the mean of normal controls. Dyssynchrony was present in 91% of patients with a QRS 6 times as many genes differentially expressed between nonfailing and NICM canine hearts in anterior wall compared with lateral wall myocardium [28]. All of these mechanisms have been seen to homogenize following CRT [29,30]. In addition to regional molecular changes, LBBB NICM is characterized by regional heterogeneities in cellular and tissue electrophysiological properties, in particular, prolongation of the action potential duration in the lateral wall [31]. Action potential duration change is associated with changes in contractility [32], which may lead to dyssynergy and dyssynchrony. Regional variations in afterload have also been described and may play a role in heterogeneous contractility [19,22], or potentially regional variations in response to afterload may lead to dyssynchrony [33]. Other potential mechanisms for regional heterogeneity in NICM include heterogeneous metabolism [20] and imbalance of myocardial blood flow and metabolism [34,35]. There are, therefore, multiple systems that express regional heterogeneity in NICM demonstrated both in humans and in animals (in vitro and in vivo), and although these studies are not primarily in the context of a normal activation pattern (narrow QRS), it is possible to consider these as potential mechanisms leading to mechanical dyssynchrony in this group. In the context of ischemic cardiomyopathy, it is more obvious how areas of scar can lead to regional variety in contractility and therefore a dyssynchronous ventricular contraction. However, when an infarct does not lead to a full thickness scar, there remains contractile reserve; therefore, if a partial lateral infarct were to lead to reduced contractile strength in comparison with the septal wall, the lateral wall would be overcome and pushed out during mid- to end-systole while the intraventricular pressure is elevated [18,36]; this mechanism of contractility variability leading to dyssynchrony may also occur in the context of NICM by any of the processes outlined previously. It is also possible to conceive how mechanical dyssynchrony may also occur in the setting of small heterogeneous areas of full-thickness myocardial scar and fibrosis, which produce dyssynchronous contraction while having a negligible electrical impact on the QRS morphology [37]. Recent in vitro work has also demonstrated an interdependence in electromechanical properties between adjacent heterogeneous myocardium with opposing changes in contractility. This could, and is likely to, lead to an exaggeration of contractility differences between areas of differing myocardium, thereby encouraging further dyssynchrony [38]. Studies of CRT in narrow QRS patients Acute hemodynamic studies
The ESTEEM-CRT study by Donahue et al. included 68 patients with a narrow QRS, EF of 28.7 ms) and an indication for an ICD (TABLE 1) [39]. An acute hemodynamic study was performed in 47 patients at the time of CRT-D implant. There was no significant acute hemodynamic response (average increase in LV dP/dtmax of 2 ± 2%); however, New York Heart Association (NYHA) and quality of life scores were substantially improved at 6 and 12 months (p < 0.001), but exercise capacity and LV volumes were unchanged. Ploux et al. performed acute hemodynamic studies in 82 consecutive heart failure patients who underwent CRT implantation irrespective of their QRS duration [40]. These patients were not assessed for mechanical dyssynchrony: 34 had a QRS < 120 ms, 11 had a QRS ‡ 120 to < 150 ms and 37 had a QRS ‡ 150 ms. There was a high correlation between changes in LV dP/dtmax and baseline QRS duration and for the narrow QRS patients no significant change was seen from baseline (+0.4 ± 6.1%; p = ns). In the moderately prolonged QRS group, a nonsignificant hemodynamic increase was seen (+4.4 ± 6.9%; p = 0.06) and in the patients with the broadest QRS there was a significant increase in LV dP/dtmax (+17.1 ± 13.4%; p < 0.001). Work by Frenneaux’s group has also offered a mechanism by which LV pacing may benefit the acute hemodynamic state of patients with a narrow QRS duration and heart failure. Their work has comprehensively investigated the concept of external constraint and the effects of biventricular and LV pacing upon it. External constraint is the negative effect on LV filling from the right ventricle (RV) via the interventricular septum with direct diastolic ventricular interaction and pericardial stretch leading to pericardial constraint [41]. This is calculated during an acute hemodynamic study, where the inferior vena cava is obstructed while LV pressure and volume are measured using a conductance catheter. In patients without external constraint, there is a reduction in both LV end-diastolic pressure (LVEDP) and LV end-diastolic volume (LVEDV) with IVC obstruction; however, there is an initial increase and then deterioration in LVEDV despite a reduction in LVEDP if external constraint is present. The amount of external constraint is the difference in LVEDP between initial LVEDP and its partner of equivalent LVEDV (FIGURE 2). Bringing the timing of LV activation forward with respect to the RV by LV pacing also prioritizes LV filling ahead of RV filling, thereby reducing the volume and pressure the RV applies within the pericardium. In 2004, Turner et al. demonstrated in 20 patients with a QRS duration £ 120 ms and an EF £ 40% that a pulmonary capillary wedge pressure (PCWP) > 15 mmHg was predictive of an acute hemodynamic response to acute LV pacing with a cardiac output increase from 3.9 to 4.5 l/min (p < 0.01) and a fall in PCWP from 24.7 to 21.0 mmHg (p < 0.001); these changes were not seen in patients with a resting PCWP < 15 mmHg [42]. They postulated that a reduction in external constraint was the most likely explanation for this improvement. This was investigated further by Williams et al. who undertook acute pressure–volume studies in 30 patients with an EF £ 35%, a QRS < 120 ms and no evidence of interventricular or 785
786
Single arm 6-month FU Single site
Two arms 3-month FU Single site
Two arms 6-month FU Single site
Two arms 28-month mean FU FU review at 6 months/ 12 months and then every 12 months Single site
Two Arms Randomized blinded 6-month FU Multisite
Achilli et al. (2003)
Yu et al. (2006)
Bleeker et al. (2006)
Gaspirini et al. (2007)
RethinQ (2007)
Improvement in all endpoints except MLHFQ in narrow group No difference between groups except MLHFQ
NYHA class 6MWT MLHFQ Maximal metabolic equivalent on treadmill LVESV LVEF NYHA class 6MWT MLHFQ LVESV/LVEDV LVEF NYHA class 6MWT LVESV LVEF Mortality
Primary endpoint: Increase of peak VO2 by >1 ml/kg Secondary Endpoints: MLHFQ NYHA class
LVEF
