Statistical ranking of electromechanical dyssynchrony parameters for CRT

Abstract

ObjectiveMechanical evaluation of dyssynchrony by echocardiography has not replaced ECG in routine cardiac resynchronisation therapy (CRT) evaluation because of its complexity and lack of reproducibility. The objective of this study was to evaluate the potential correlations between electromechanical parameters (atrioventricular, interventricular and intraventricular from the dyssynchrony model presented in 2000), their ability to describe dyssynchrony and their potential use in resynchrony.Methods455 sets of the 18 parameters of the model obtained in 91 patients submitted to various pacing configurations were evaluated two by two using a Pearson correlation test and then by groups according to their ability to describe dyssynchrony, using the Column selection method of machine learning.ResultsThe best parameter is duration of septal contraction, which alone describes 25% of dyssynchrony. The best groups of 3, 4 and ≥8 variables describe 59%, 73% and almost 100% of dyssynchrony, respectively. Left pre-ejection interval is highly and significantly correlated to a maximum of other variables, and its decrease is associated with the favourable evolution of all other correlated parameters. Increase in filling duration and decrease in duration of septum to lateral wall contraction difference are not associated with the favourable evolution of other parameters.ConclusionsNo single electromechanical parameter alone can fully describe dyssynchrony. The 18-parameter model can be simplified, but still requires at least 4–8 parameters. Decrease in left pre-ejection interval favourably drives resynchrony in a maximum of other parameters. Increase in filling duration and decrease in septum–lateral wall difference do not appear to be good CRT targets.