Predictive parameters of left ventricular reverse remodeling in response to cardiac resynchronization therapy in...

Cardiac resynchronization therapy (CRT) is useful for the treatment of severe congestive heart failure. Unfortunately up to 30% of patients could be non-responders. The aim of our study was to find parameters to predict responsiveness to CRT.

Fifteen patients (9 males, 6 females, mean age 67.3 +/- 7.8 years, range 52-83 years) with dilated cardiomyopathy, NYHA functional class III-IV, left ventricular (LV) ejection fraction < 35% and QRS > or = 110 ms, underwent CRT. All the patients had echocardiographic evidence of systolic dys-synchrony.

One patient died of electromechanical dissociation. The remaining 14 patients maintained biventricular stimulation at 6 months; mean QRS width decreased from 156 to 132 ms (p < 0.001). Ten patients (71%) were considered responders because of a reduction in LV end-systolic volume > 15%. In non-responders (4 patients, 29%) LV end-systolic volume was stable in 3 patients and increased in 1. LV ejection fraction significantly increased only in responders (p < 0.001). Responders had more severe pre-pacing dyssynchrony than non-responders (p < 0.001). Inter- (p = 0.002) and intraventricular dyssynchrony (p = 0.003) did significantly reduce after CRT only in responders. On multiple regression analysis there were two independent predictors of reverse remodeling after pacing: the baseline mitral QS-tricuspid QS (QSm-QSt) time (B = -1.7, p = 0.005) and the intraventricular dyssynchrony index (B = -1.55, p = 0.007). Pre-implant QSm-QSt of 38 ms correctly identified the two groups: responders had a value > 38 ms and non-responders < 38 ms. The pre-implant intraventricular dyssynchrony index of 28 ms was the cut-off value: responders had an index > 28 ms, non-responders < 28 ms.

In the literature a tissue Doppler imaging index of intraventricular dyssynchrony evaluated before implantation is used to select responders to CRT. In our work we studied interventricular and intraventricular dyssynchrony, and both the QSm-QSt time and the standard deviation of the 12 LV segment QS time were correctly able to identify responders.