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

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. (C) 2016 Elsevier Inc. All rights reserved.