A predictive model of super response to cardiac resynchronization therapy in short-term period.

Abstract

The left bundle branch block, nonischemic heart failure (HF) and female gender are the most powerful predictors of a super response to cardiac resynchronization therapy (CRT). It is important to identify super responders who can derive most benefits from CRT. We aimed to establish a predicting model that could be used for prognosis of a super response to CRT in short-term period. Patients with QRS ≥ 130ms, New York Heart Association (NYHA) II-III class of HF, left ventricle ejection fraction (LVEF) ≤ 35% and indications for CRT were included in the study. Before and 6month after CRT the electrocardiography, echocardiography and cardiac scintigraphy were performed. The study's primary endpoint was the NYHA class improvement ≥ 1 and left ventricle end systolic volume decrease > 30% or LVEF improvement > 15% after 6month CRT. Based on collected data, we developed a predictive model regarding a super response to CRT. Of 49 (100.0%) patients, 32 (65.3%) had a super response to CRT. Patients with a super response were likelier to have a lower cardiac index (p = 0.007), higher rates of interventricular delay (IVD) (p = 0.003), phase standard deviation of left ventricle anterior wall (PSD LVAW) (p = 0.009) and ∆QRS (p = 0.02). Only IVD and PSD LVAW were independently associated with a super response to CRT in univariate and multivariate logistic regression. We created a logistic equation and calculated a cut-off value. The resulting ROC curve revealed a discriminative ability with AUC of 0.812 (sensitivity 90.62%; specificity 70.59%). Our predictive model is able to distinguish patients with a super response to CRT.