Outcome of super-responders to cardiac resynchronization therapy defined by endpoint-derived parameters of left ventricular remodeling: a two-center retrospective study

Abstract

Aims: Various studies have attempted to identify super-responders to cardiac resynchronization therapy (CRT) by echocardiographic parameters of reverse remodeling. However, scientific evidence regarding those parameters is scarce. This study aimed at validating the definition of super-response to CRT based on the following frequently employed echocardiographic parameters: left ventricular ejection fraction (LVEF), end-diastolic volume index (EDVI), and end-systolic volume index (ESVI).

Methods and results: We retrospectively investigated echocardiographic data and outcomes of 542 patients after CRT implantation. The primary endpoint comprised all-cause mortality, heart transplantation, ventricular assist device implantation (VAD), and hospitalization for heart failure. Secondary endpoints were hospitalization for heart failure, and the combination of all-cause mortality, heart transplantation and VAD. Two approaches were employed defining super-response based on improvement of echocardiographic parameters: one derived from the negative predictive value (NPV) for clinical endpoints, and second from best quartiles of improvement. Using the NPV method, an absolute 25 % increase in LVEF, a relative 38 % reduction in EDVI, and 46 % in ESVI were calculated as optimal cut-offs identifying 4.9, 18.5, and 21.3 % as super-responders. The best quartiles method resulted in lower cut-off values, i.e. 14 % increase in LVEF, 26 % reduction in EDVI, and 36 % in ESVI. All cut-offs except LVEF ≥ 25% were significantly associated with improved outcomes after 5 years (median follow-up 35.7 months).

Conclusions: NPV- and best quartile-based cut-offs validate previously applied empirical echocardiographic cut-offs to define super-response to CRT. These data provide evidence for using these empirical cut-offs in daily practice and facilitate inter-study comparability.