Cardiac Resynchronization Revisited: What is the Next Step?

This article refers to ‘Early intervention and long-term outcome with cardiac resynchronization therapy in patients without a history of advanced heart failure symptoms’, by R. Herscovici et al., and ‘Reduced risk of life-threatening ventricular tachyarrhythmias with cardiac resynchronization therapy: relationship to left ventricular ejection fraction’, by V. Kutyifa et al. Cardiac resynchronization therapy has become established therapy on top of optimal medical therapy for symptomatic heart failure patients with electrical dyssynchrony. In recent years subgroup analysis of randomized controlled studies (RCT) in mild heart failure (HF) patients indicated that the benefit may be greater in patients with LBBB morphology, which is why such patients have a Class I indication in the most recent European Heart Failure Association (HFA)1 and European Heart Rhythm Association (EHRA) guidelines.2 Moreover, recently it has been suggested that CRT benefits may extend over a wider range of reduced LVEFs than currently recommended,3, 4 but one randomized trial with this particular focus was terminated early due to poor patient recruitment (MIRACLE-EF study).5 Two papers in this journal, by Herscovici et al.6 and Kutyifa et al.,7 illustrate some current questions in CRT therapy such as how early in the heart failure evolution CRT may exert a benefit and whether an implantable cardioverter defibrillator (ICD) is beneficial in patients with baseline LVEFs higher than presently recommended in the guidelines for primary preventive ICDs. In a post-hoc analysis of the MADIT-CRT trial, Herscovici et al.6 compared the value of CRT in those with and without prior advanced heart failure (HF). Non-advanced HF was defined as no HF hospitalization ≥3 months before study enrolment and no history of NYHA class III–IV. Not surprisingly, such patients had better renal function, less history of arrhythmia, and received fewer antiarrhythmic drugs, mineralocorticoid receptor antagonists (MRAs), diuretics, and digitalis than those with prior advanced HF. The geographical distribution was also different, with nearly twice as many of those with prior advanced HF recruited from European centres. No information about baseline echo parameters was given except the percentage of patients with LVEF <25%, which was similar in both groups. Echocardiographic benefit after 1 year was similar in the two HF groups assigned to CRT with a defibrillator (CRT-D). By multivariate analysis, CRT-D treatment was associated with a significant reduction in HF hospitalizations and mortality regardless of the presence or absence of prior advanced HF symptoms. In contrast, mortality per se was not significantly improved by CRT-D in those with no prior advanced HF over the study period of 5.6 years despite a trend to improvement after 3 years. Of the previous major randomized controlled studies of CRT in NYHA class I–II, only the RAFT study8 had a long enough follow-up period of 40 months to demonstrate that CRT per se reduced mortality compared with control. The follow-up time in the present substudy by Herscovici et al.6 was even longer. The results of this post-hoc substudy of course need to be interpreted with caution. Moreover, this substudy included only LBBB patients constituting 70% of the patients in MADIT-CRT. The remodelling response consistently has been reported to be much smaller in non-LBBB subjects. Nonetheless, significant reverse remodelling by CRT associated with improved survival has also been reported in non-LBBB patients.9 In MADIT-CRT, repeat echocardiograms were not performed, but long-term survival benefits were limited to patients with LBBB10 which is why the substudy by Herscovici et al. was limited to such patients. Whether LBBB or wide QRS (the inclusion criteria in all RCTs) is the best predictor of CRT response has been disputed. In a case-based meta-analysis comprising 3831 patients, Cleland et al.11 reported a response to CRT with QRS >140 ms for both a combination of HF hospitalization and mortality as well as mortality per se. Only a wide QRS width independently predicted response to CRT in this meta-analysis. The results of Herscovici et al. favour early CRT intervention. CRT therapy overall is underimplemented both in Europe and in the USA. Since early intervention has more impact on a population level, the present findings further underline the importance of improving CRT implementation in particular in mild HF patients. Although most of the randomized clinical trials of prophylactic ICDs have required an LVEF of ≤30–35% for enrolment, the median LVEF of enrolled patients was well below 30%. When ICD is prescribed as part of CRT therapy, the reverse remodelling induced by CRT may soon put a responder in an LVEF zone in which no ICD is indicated for primary prevention. One-third of the MADIT-CRT3 study patients had a higher LVEF (range 31–45%) than the inclusion criterion of LVEF ≤30%. In the paper by Kutyifa et al.7 in this issue of the journal, a baseline LVEF >30% in MADIT-CRT predicted less ventricular tachyarrhythmias and less benefit from CRT-D in reducing fast ventricular tachcardia (VT) or ventricular fibrillation (VF) compared with those with LVEF <30%. When LVEF was assessed as a continuous variable, each 5% higher baseline LVEF was associated with a significant 30% (P < 0.001) reduction in the risk for fast VT/VF and with a corresponding linear reduction in benefit from CRT-D in reducing fast VT risk (P-value for interaction treatment by LVEF = 0.003). A proportion of patients in the MADIT-CRT trial had secondary preventive ICDs. It is tempting to speculate whether a concomitant ICD should be avoided in similar ‘high’ LVEF patients. The extent of reverse remodelling has been shown to decrease ventricular arrhythmia in both the REVERSE and MADIT-CRT12, 13 studies, and also total mortality,9 again suggesting that a decrease in ventricular arrhythmia is to be expected in NYHA I–II patients who respond to CRT. In super-responders to CRT, need for ICD therapies is small,14 but super-response needs time to evolve and therefore does not permit use of CRT only at the time of implantation. Still, previous meta-analyses have indicated that the survival contribution of concomitant ICD to CRT therapy is not extensive, but less in NYHA III–IV than in NYHA I–II patients.15 By these results, patients with mild HF then should receive an ICD. In patients with no prior myocardial infarction or ischaemic heart disease, the value of a primary preventive ICD is less well documented but still recommended in the guidelines.1 Though the decision to implant a CRT-P rather than a CRT-D in a mild HF patient is thus a very difficult one, the immediate conclusion from the study by Kutyifa et al.7 is to question concomitant ICD use in a recipient who at the time of implant has a borderline LVEF, does not have fully up-titrated HF medication, has not had a previous myocardial infarction, and does not have a secondary preventive indication for an ICD. Recently Ruwald14 reported very low risk for ventricular arrhythmias over 22 months in those in whom LVEF improved to >50% during CRT and also in those with an intermediate improvement to LVEF of 36–50%. Female sex, higher baseline LVEF, and smaller ventricular and atrial volumes predicted normalization of LVEF. However, even though these factors and other prediction scores such as the Validate-CRT16 may be useful, they are only of limited help when the choice of device is being made in an individual patient. There are three important implications of the recent CRT studies3, 4 and the present studies by Herscovici et al.6 and Kutyifa et al.7 The first is to extend the indication to CRT also to patients with cardiac dysfunction and no history of symptoms or of episodes of HF decompensation. The second is to stress further the importance of a stringent up-titration of HF medication to target dose before device implantation. The extent of reverse remodelling is clearly related to dose, in particular for beta-blockers.5 However, just like for CRT, reverse remodelling takes time to develop, often exceeding 1 month. In mild HF patients it should be easier to up-titrate to target or near target doses than in moderate to severe HF patients, stressing that this longer waiting period to achieve maximal drug-induced reverse remodelling may be worthwhile in order to avoid unnecessary ICDs. Thirdly, we must make efforts to ensure a wider implementation of CRT therapy overall and in particular in mild HF patients. Gasparini and Galimberti suggested a sweet spot for CRT,17 and pushing for more implantations in the healthier patients. The results of the present papers suggest that the next step may be to study CRT benefit in patients with asymptomatic HF in stage B (structural heart disease—no symptoms). Conflicts of interest: C.L. was principal investigator of REVERSE and MIRACLE EF, CRT studies sponsored by Medtronic; reports research grants, speaker honoraria, and consulting fees from Medtronic; and speaker honoraria and consulting fees from St. Jude Medical, Vifor, Cardio3, and Novartis.