CineECG repolarization gradients predict acute hemodynamical response in CRT patients

Abstract

Abstract Background Cardiac resynchronization therapy (CRT) is widely used for patients with conduction disorders and heart failure. The proportion of the CRT-non-responders can exceed 30%, which warrants search for new approaches to optimize the CRT device implantation and follow-up. CineECG is a novel cardiac anatomy-based ECG modality proposed for better visualization of the myocardial depolarization and repolarization processes. This method yielded promising results in patients with sinus rhythm but has not been tested in ventricular pacing. Purpose The aim of the present study was to evaluate CineECG-derived mean temporo-spatial isochrone (TSI) trajectory in different pacing protocols and test its association with acute hemodynamical responses in CRT patients. Methods CineECG uses a generic heart/torso model to construct the average electrical pathway within the cardiac anatomy from the 12 lead ECG. CineECG and invasive hemodynamical parameters were tested in 15 patients undergoing multisite CRT system implantation. ECG recording and hemodynamical measurements were performed simultaneously, while the patients were undergoing 4-6 biventricular pacing protocols with the same interventricular pacing delay (30 ms) and different left ventricular (LV) lead localizations (apical, basal, multisite). LV dP/dt max was tested for association with the parameters of standard ECG (QRS duration, QT duration, Tpeak-Tend duration) and CineECG pathway parameters. The latter pathway directions were computed for QRS and ST-T intervals for anteroposterior (X), left-to-right (Y), and apicobasal (Z) directions. Results Biventricular protocols did not differ between each other in LV dP/dt. In univariate linear regression analysis, LV dP/dt max was associated with QRS and ST-T pathway directions (Figure 1 - table). The left-to-right (Y) and apicobasal (Z) ST-T pathway directions remained independent predictors of LV dP/dt max in multivariate linear regression analysis with the adjustment for a pacing protocol [regression coefficient 571.6 (319.3 – 823.8), p<0.001; regression coefficient -406.9 (-636.7 – -177.2), p=0.001, respectively]. Optimal repolarization sequences associated with better acute hemodynamic responses assessed by LV dP/dt max were from right to left and from base to apex (Figure 2). Conclusion CineECG repolarization pathway gradients in the left-to-right and apicobasal directions outperform temporal ECG parameters and predict the acute hemodynamical response in biventricular pacing protocols in CRT patients.