Abstract 18774: Electromechanical and Scar-Mediated Interactions in Cardiac Resynchronization Therapy Based on Cardiac Magnetic Resonance

Abstract

Background: Mechanical dyssynchrony (MD), left ventricular (LV) lead electrical timing relative to QRS onset (QLV, or normalized for QRS width as LVLED), percent LV scar volume (PSV), and scar at the LV lead position (LP-SCAR) are important determinants of the response to cardiac resynchronization therapy (CRT), but their interactions are not well understood. Objective: We sought to determine these electrical, mechanical, and scar-mediated interactions in a clinical CRT cohort. Methods: Patients undergoing CRT first had cardiac magnetic resonance (CMR) with cine DENSE (displacement encoding with stimulated echoes) and late gadolinium enhancement (LGE), echocardiography, and cardiopulmonary exercise testing, followed by repeat testing at 6 months. Factors influencing CRT response (defined as a 15% reduction in LVESV) and clinical events up to 5 years after CRT were analyzed using logistic and Cox proportional hazards regression, respectively. Results: The 70 patients studied (66 ± 13 years old, 26% F) had an LVEF of 22.0 ± 8.2%, and 51% had positive LGE (PSV 10.0% [IQR 6.3%-16.7%]). 36 (51%) were CRT responders, 16 (23%) died during follow-up, 24 (34%) had the combined endpoint of heart failure hospitalization (HFH) or death, and 10 (14%) had sustained ventricular tachycardia (VT). Predictors of CRT nonresponse included CURE ≥ 0.70 (OR 73.2, 95% C.I. 8.8-605.8), QLV ≤ 90 ms (OR 9.0 [2.6-31.0]), and LP-SCAR (OR 2.9 [1.1-7.6]). The logistic model with these 3 covariates was highly predictive of CRT nonresponse (AUC 0.933; p < 0.001). Electromechanical factors also predicted clinical events. The best Cox model for death consisted of CURE ≥ 0.70 (HR 3.88 [1.38-10.88]) and LVLED ≤ 0.50 (HR 5.12 [1.79, 14.69]), while LVLED ≤ 0.50 also predicted the combined endpoint of HFH or death (HR 3.64 [1.57, 8.46]). Of note, LP-SCAR and CURE influenced VT events (p=0.07). Conclusions: The clinical course after CRT-D implantation reflects complex interactions between electrical, mechanical, and scar-mediated factors that can be understood well with CMR. A model combining CMR-derived mechanical and scar-mediated factors with electrical parameters accurately predicts both CRT nonresponse and adverse clinical events, such as heart failure hospitalization and death.