A New Paradigm in Cardiac Resynchronization Therapy?

We thank Polasek et al for the opportunity to clarify some crucial points of our study.1 First, our study is obser- vational and retrospective, thus needing further confirmation by prospective ad hoc studies as stated in the report. Obvi- ously, given the retrospective and observational nature of the study, a con- trol group cannot be expected unless we performed cardiac resynchronization therapy (CRT) device implantation dis- regarding the guidelines and it is not our case. Accordingly, we just commented on reported results, but we did not discourage CRT implantation in these patients. Second, the hypothesis gener- ated by our data was the existence of an upper cut-off value of QRS duration above which CRT is less effective because of extensive electrical and structural remodeling. Our observations are consistent with findings that showed a worse clinical outcome and prognosis related to either left bundle branch block morphology or prolonged QRS duration in patients with heart failure.2,3 A recent large prospective observational study, totaling 3,319 patients with CRT, showed that cardiac mortality was highest at the upper extremes of QRS duration.4 Moreover, a subanalysis of the Resynchronization Reverses. Remodeling in Systolic Left Ventricular Dysfunction (REVERSE) trial showed a linear reduction of left ventricular vol- umes after CRT as the QRS duration in- creases from 120 to 180 ms, followed by a declining efficacy from 180 ms on- ward.5 Third, it is important to underline that we described and commented the results of meta-analysis as usually is done in the discussion section.6 Further- more, we deliberately chose to comment on the graph showing the combined end point of death and hospitalization for heart failure, rather than death only, because this is more representative of the efficacy of CRT given the well- established deleterious economic and clinical impact of hospitalization for worsening heart failure.7,8 Finally, we would advise Polasek et al that it is hard to discuss unpublished results and compare them with others using different end points: a less validated end point (!10% reduction of the left ventricular end-systolic diameter) in Polasek data and a well-established end point (!15% reduction of the left ventricular end- systolic volume) in our study.