Elsevier

Journal of Electrocardiology

Volume 49, Issue 4, July–August 2016, Pages 603-609
Journal of Electrocardiology

Scar burden assessed by Selvester QRS score predicts prognosis, not CRT clinical benefit in preventing heart failure event and death: A MADIT-CRT sub-study,☆☆

https://doi.org/10.1016/j.jelectrocard.2016.04.008Get rights and content

Highlights

  • In this sub-study of the MADIT-CRT population, we investigate the predictive value of electrocardiographic scar burden estimation by Selvester QRS scoring.

  • Scar burden by QRS scoring predicts clinical prognosis in ICD and CRT receipients, but not CRT efficacy in preventing heart failure events or death, in the LBBB subgroup (no interaction between higher QRS score and CRT assignment).

  • Higher QRS score is associated with less LV remodeling and less recovery of LVEF by echocardiography at 6-month follow-up after CRT.

Abstract

Background

There is a need for improved selection criteria for cardiac resynchronization therapy (CRT). High myocardial scar burden has been associated with worse outcome in CRT patients. However, it is unclear whether high scar burden prevents CRT clinical benefit or is merely predictive of prognosis in heart failure (HF) patients regardless of CRT implantation. We aimed to study the predictive value of scar burden estimated by electrocardiographic Selvester QRS scoring in determining CRT benefit in the multicenter automatic defibrillator implantation trial-cardiac resynchronization therapy (MADIT-CRT) population.

Methods

Selvester QRS scoring was performed on all 1820 ECGs of the MADIT-CRT population by a single observer. In both arms and in their respective LBBB subgroups, QRS score was analyzed in comparison to echocardiographic volumes and in relation to time to HF event or death using Cox proportional hazard ratios. To determine effect on CRT clinical benefit, we tested for interaction between the effects of CRT assignment and QRS score on time to HF event or death.

Results

In the CRT-D arm, a significant correlation was found between higher continuous QRS score and less increase of left ventricular ejection fraction (LVEF) as well as less decrease of left ventricular end-systolic volume (LVESV) (multivariate -p-values: < 0.001). QRS score was significantly correlated with HF event/death in the left bundle branch block (LBBB) subgroup (n = 1037, multivariate HR 1.07 per point, p = 0.046). Scar extent estimated by QRS scoring was neither predictive of CRT clinical benefit in the total study population (interaction -p-value = 0.25) nor in the LBBB subgroup (interaction p-value = 0.86).

Conclusion

High scar burden estimated by Selvester QRS score is predictive of adverse overall prognosis in LBBB patients regardless of CRT implantation. However, QRS score does not identify patients who benefit clinically from CRT-D compared to implantation of ICD only.

Introduction

Cardiac resynchronization therapy (CRT) improves prognosis and ameliorates symptoms in selected heart failure (HF) patients. However, up to one third of patients may not benefit from CRT but are still subjected to adverse effects and costs [1]. Several factors that may predict favorable CRT response determined by echocardiography parameters and/or event-free survival have been identified, including QRS duration ≥ 150 ms [2], non-ischemic heart failure etiology [3], female sex [2], left ventricular (LV) lead placement (non-apical [4], at site of latest activation [5] and away from scar [6]) and left bundle branch block (LBBB) QRS morphology. In addition, newer strict LBBB criteria indicative of complete left bundle branch block [7] have been shown to predict CRT success better than traditional LBBB criteria [8]. Finally, low myocardial scar burden either measured by cardiac magnetic resonance imaging with late gadolinium enhancement (CMR-LGE) or estimated by ECG Selvester QRS score has been linked to favorable CRT outcome [9], [10]. Importantly, previous studies on the value of both ECG estimation of scar using the Selvester QRS score and scar burden by CMR-LGE in predicting CRT response have not featured any non-CRT-D arm and thus could not evaluate the effect of scar burden on CRT clinical benefit.

We aimed to test whether QRS score was a significant predictor of echocardiographic and clinical benefit after CRT implantation in the MADIT-CRT population.

We specifically hypothesized that among CRT recipients, high myocardial scar burden estimated by ECG Selvester QRS score would be predictive of: a) less increase of left ventricular ejection fraction (LVEF), b) less reduction of left ventricular end-systolic volume (LVESV), and c) a smaller improvement in event-free survival compared to implantable cardioverter-defibrillator (ICD).

Section snippets

Study patients

MADIT-CRT enrolled subjects of either sex who were more than 21 years of age with ischemic cardiomyopathy (New York Heart Association [NYHA] class I or II) or non-ischemic cardiomyopathy (NYHA class II only), sinus rhythm, an LVEF of 30% or less and QRS duration of 130 ms or more as described previously [2], [11]. Further exclusion criteria included existing indication for CRT at time of enrollment, previously implanted pacemaker, implantable cardioverter defibrillator (ICD) or CRT, previous

Study patients

A total of 1733 patients were included in this substudy, 1029 in the CRT-D arm and 704 in the ICD arm. The CRT-D arm and the overall population had similar baseline characteristics and similar use of cardiac drugs at time of enrollment as shown in Table 1. Mean follow-up was 2.4 years. There were 441 that reached the clinical endpoint of HF event or death. The average QRS score was 5.2 in the overall study population (4.2 in the LBBB subgroup). The distribution of conduction types in the

Discussion

The major finding of the present study was that in the LBBB subgroup, scar burden estimated by ECG using the Selvester QRS score did not predict CRT clinical benefit, but rather prognosis regardless of CRT implantation.

Because CRT has been shown to be most effective in LBBB, coupled with the notion that extensive scar could prevent CRT effect, one may hypothesize that scar burden would be most important for predicting prognosis in the CRT-D LBBB subgroup. However, we found that CRT was equally

Conclusion

In the LBBB subgroup, myocardial scar estimated by electrocardiographic Selvester QRS scoring predicts prognosis in heart failure patients and may identify echocardiographic responders to CRT. However, QRS score cannot be used to determine who will benefit clinically from CRT-D over implantation of ICD only. Future research should determine if these findings are also present with other markers of myocardial scar, such as cardiac magnetic resonance imaging.

References (26)

Cited by (0)

Financial disclosures: No extramural funding was used to support this work.

☆☆

Author contributions: WZ and JPD acquired data as part of the MADIT-CRT study. DGS, ZL and GSW conceived of this substudy. ZL and SM designed the statistical analysis plan, which was critically reviewed by DGS, WZ and JPD. BW later added complementary analyses to the statistical analysis plan. ZL applied Selvester QRS scoring to all ECGs. DGS scored ECGs that ZL flagged as difficult to analyze. SM conducted all statistical analyses. All co-authors participated in interpretation of results. BW and ZL produced the initial draft of the manuscript, which was then revised and approved by all remaining co-authors.

1

Co-lead authors.

View full text