When cardiac resynchronization therapy may be harmful: time to wake up

This editorial refers to ‘Cardiac resynchronization therapy in patients with systolic heart failure and QRS interval <130 ms: insights from a meta-analysis’ by R.M. Shah et al., on page 267–273. Cardiac resynchronization therapy (CRT) is one of the major breakthroughs in heart failure (HF) management which has proven to reduce mortality and HF hospitalizations in patients with symptomatic systolic HF and electrical dyssynchrony.1–3 In later years, it has become apparent that the response to CRT is related to the magnitude of QRS prolongation and the presence of left bundle branch block (LBBB).4,5 Importantly, it is also becoming increasingly clear that CRT may be harmful in patients with HF and narrow QRS. However, the clinical implications of these observations have not been extensively discussed or adopted. Shah et al.6 in a meta-analysis of four randomized controlled trials (RCTs) published in this issue of EP-Europace corroborate that CRT in patients with systolic HF and narrow QRS is associated with increased all-cause mortality. Since many patients in Europe have received CRT outside of the guideline recommendations7 including those with narrow QRS, the consequences of potential harm by CRT in such patients must be further discussed and acted upon. The most recent European Herat Rhythm Association (EHRA) and European Society of Cardiology guidelines on cardiac pacing and CRT2 already state that CRT is not indicated in HF patients with narrow QRS (Class III, level of evidence B). Instead, CRT is indicated for HF patients with LBBB with QRS ≥150 ms (Class I, level of Evidence A) and for those with QRS 120–150 ms (Class I, level of Evidence B). For patients without LBBB, the class of recommendations are lower. In these patients, CRT should be considered (Class IIa, level of Evidence B) if the QRS width is ≥150 ms and may be considered (Class IIb, level of Evidence B) in patients with the QRS duration of 120–150 ms. It is worthwhile reviewing the inclusion criteria in the RCTs on CRT over the last 15 years. In MUSTIC (MUltisite STimulation In Cardiomyopathy)1 and CARE-HF (CArdiac REsynchronization-Heart Failure)2 trials, the inclusion criterion was QRS ≥150 ms based on the assumption that electrical dyssynchrony would be amenable to electrical therapy that is CRT. In the subsequent studies,2 the inclusion criteria were more lenient with QRS as low as ≥120 ms reflecting the common border for abnormal QRS in the electrocardiogram literature. However, QRS in patients included in these trials were significantly wider, with an average width of 168 ms in New York Heart Association (NYHA) III–IV HF trials and 158 ms in trials in NYHA I–II.2 This important difference between the QRS widths in patients included in RCTs and the inclusion criteria may explain why incremental benefits of CRT with increasing baseline QRS duration have been reported in RCTs in mild HF patients4,5 in which more patients had somewhat smaller QRS widths than in the RCTs in patients with NYHA II-III. A recent case-based meta-analysis of five major RCTs in 3872 patients with NYHA II–IV has demonstrated a clear benefit of CRT on mortality and the composite outcome of mortality and HF-related hospitalizations in individuals with a baseline QRS of at least 140 ms, but no advantage in patients with narrow QRS (<120 ms) highlighting the importance of QRS prolongation as a prerequisite for CRT benefit.8 The evidence of benefit in the smaller range of QRS prolongation (120–1404–150 ms9) thus remains weak with the possible exception of females who innately have smaller QRS widths than men.10 Much research in the field of CRT research over the last 10 years has been focused on improving the response rate to CRT above the commonly reported one of 60–70%.2 Adding mechanical dyssynchrony criteria on top of wide QRS was shown to be futile when tested in a multicentre setting despite encouraging single-centre results from highly skilled centres. Furthermore, the value of CRT in HF patients with mechanical dyssynchrony but without electrical dyssynchrony, i.e. narrow QRS, has been tested in several RCTs.11–13 The rationale for this assumption has been that mechanical dyssynchrony may exist in HF patients with normal QRS and may exert harm. In the Resynchronization therapy in normal QRS (RethinQ) study,11 no clinical benefit from CRT on peak VO2 and reverse left ventricular remodelling has been shown in NYHA III–IV HF patients with QRS <130 ms. The much larger Echocardiography Guided Cardiac Resynchronization Therapy (EchoCRT) study12 also compared CRT-D (CRT with an implantable cardioverter-defibrillator (ICD) with ICD in NYHA III–IV HF patients with QRS <130 ms. The primary endpoint was the composite of death from any cause or first hospitalization for HF. This study was prematurely interrupted by the Data Safety and Monitoring Board due to futility. But the study results clearly go beyond futility and demonstrate a significant 81% relative risk increment for total mortality in the CRT compared with the ICD arm and a non-significant 20% relative risk increment for the composite primary endpoint. The LESSER-EARTH (Evaluation of Resynchronization Therapy for Heart Failure) trial included NYHA III–IV HF patients with narrow QRS ≤120 ms,14 but no mechanical dyssynchrony criteria required for inclusion. In this double-blind trial, HF patients with severe left ventricular dysfunction were randomized to CRT-D or ICD with exercise duration as the primary endpoint. Importantly, this study was also stopped prematurely by the Data Safety and Monitoring Board due to futility and safety concerns because of worse exercise tolerance and a non-significant trend to more HF hospitalizations in the CRT arm. Importantly, CRT resulted in a prolongation of the QRS duration of 40 ms compared with unpaced intrinsic rhythm. By imaging technique a significant leftward shift in lateral septal delay after 6 months of CRT was demonstrated, suggesting that CRT had induced intraventicular dyssynchrony. Even though it has not been shown in this study, it is fair to speculate that induced dyssynchrony over time may deteriorate left ventricular function and affect prognosis. To investigate the effects of CRT in patients with narrow QRS on systolic HF mortality, all-cause mortality, and HF-related hospitalizations, Shah et al.6 performed a meta-analysis of four RCTs.11–14 Out of 1186 patients, 595 were treated with CRT-D and 591 with ICD only. The EchoCRT study12 dominated the contribution with 809 patients. All but one14 RCT used mechanical dyssynchrony criteria in addition to narrow QRS as the inclusion criteria. Included patients were somewhat younger than those in previous RCTs on CRT2 and the age ranged between 58 (EchoCRT12) and 68 (Narrow CRT13) years. Otherwise the patients were very similar to those in previous RCTs. Most patients were men, with NYHA III HF and with a mean left ventricular ejection fraction of around 27%. In contrast, the average QRS width was clearly normal and ranged between 104 and 107 ms. The results of this meta-analysis confirm that CRT is harmful in HF patients with narrow QRS. Patients treated with CRT-D had a 66% increased mortality risk compared with ICD [relative risk, (RR), 1.66; 95% confidence interval (CI), 1.096–2.515; P = 0.017]. There was a non-significant trend towards increased HF mortality (RR, 1.29; CI, 0.68–2.45; P = 0.429), whereas the incidence of HF hospitalizations did not differ between the groups. When limiting the analysis to RCTs that required mechanical dyssynchrony, almost identical results were obtained. The results from Shah et al.6 suggest that CRT-D does not reduce HF mortality or HF hospitalizations in symptomatic patients with systolic HF and QRS ≤130 ms compared with ICD alone. Instead, CRT-D is associated with higher all-cause mortality in comparison with ICD therapy. The presence of mechanical dyssynchrony does not influence outcome. Taken together, these results confirm that the patient eligibility for CRT should be based on electrical dyssynchrony. However, the findings raise the question what to do with HF patients with narrow QRS who have already been implanted with a CRT device. The results of the EHRA and Heart Failure Association (HFA) CRT Survey7 conducted in 2009 encompassing 2438 recipients in 13 countries indicated that CRT was implanted in patients with QRS width <130 ms in 19% and was based solely on mechanical dyssynchrony in 9%, suggesting that many patients in Europe may have ongoing CRT that may be either potentially harmful or already causing harm. Much has been achieved in the identification of responders to CRT. The effect of long QRS duration on the CRT benefit has already been demonstrated in RCTs4,5 and in meta-analysis.9 Subgroup analysis has indicated that CRT benefit is confined to patients with LBBB.4,5 Recently, the long-term Multicenter Automatic Defibrillator Implantation Trial-Cardiac Resynchronization Therapy (MADIT-CRT) results suggested harm by CRT in patients who did not have LBBB at baseline when followed over 7 years.15 It is now high time to emphasize in which patients CRT should not be given and to define when therapy should be suspended. The ongoing non-responder CRT registry may help further identify the characteristics of such patients. It may be difficult to stop ongoing CRT therapy for various patient- and physician-related reasons.16 Therefore, a randomized study whether to continue or turn off CRT in patients with narrow QRS and existing CRT devices is needed. Conflict of interest: C.L. was the principal investigator of REVERSE, a CRT study sponsored by Medtronic reports research grants, speaker honoraria, and consulting fees from Medtronic, and received speaker honoraria and consulting fees from St. Jude Medical. C.L. was supported by the Swedish Heart Lung Foundation [grants 20080498 and 20110406] and the Stockholm County Council [grants 20090376 and 20110610. No funding organization had any role in the collection, analysis or interpretation of data or approval/disapproval for publication.