Evaluating Electrocardiography-Based Identification of Cardiac Resynchronization Therapy Responders Beyond Current Left Bundle Branch Block Definitions

Abstract

Objectives: This study aimed to evaluate the association of 4 left bundle branch block (LBBB) definitions and their individual ECG characteristics with clinical outcome. Furthermore, it aimed to combine relevant outcome-associated electrocardiographic (ECG) characteristics into a novel outcome-based definition. Background: LBBB morphology is associated with positive response to cardiac resynchronization therapy. However, there are multiple LBBB definitions. Associations with outcomes may differ between definitions and depend on varying contributions of the individual ECG characteristics that these LBBB definitions are composed of. Methods: A retrospective multicenter study was conducted in 1,492 cardiac resynchronization therapy patients. Patients were classified as LBBB or non-LBBB according to definitions provided by the European Society of Cardiology, American Heart Association, MADIT-CRT (Multicenter Automatic Defibrillator Implantation with Cardiac Resynchronization Therapy) trial, and according to Strauss et al., the primary endpoint was left ventricular assist device implantation, cardiac transplantation, and all-cause mortality. Results: LBBB classification differed significantly between the 4 definitions (kappa coefficients ranging from 0.09 to 0.92). The American Heart Association definition correlated the least (0.09 to 0.12) with the other definitions. Only 13.8% of patients were classified as LBBB by all definitions. During a follow-up period of 3.4 ± 2.4 years, 472 (32%) patients experienced the primary endpoint. For each LBBB definition survival analysis showed a significant association of LBBB with outcome, with relative risk reduction ranging from 39% to 43%. Each LBBB definition included characteristics that were not associated with outcome. Combining outcome-associated ECG characteristics into a novel prediction model did not significantly improve diagnostic performance (relative risk reduction 43%). Conclusions: The classification of LBBB is highly dependent on the LBBB definition used. However, each LBBB definition provides a comparable difference in risk of adverse clinical events between LBBB and non-LBBB patients. Combining individual outcome-associated ECG-characteristics into a novel prediction model does not improve association with outcome.