for the MADIT-CRT InvestigatorsBackground-Cardiac resynchronization therapy (CRT) plus implantation of an implantable cardioverter defibrillator (ICD) reduced the risk of death or heart failure event in patients with mildly symptomatic heart failure, left ventricular dysfunction, and wide QRS complex compared with an ICD only. We assessed echocardiographic changes in patients enrolled in the MADIT-CRT trial (Multicenter Automatic Defibrillator Implantation Trial: Cardiac Resynchronization Therapy) to evaluate whether the improvement in outcomes with CRT plus an ICD was associated with favorable alterations in cardiac size and function. Methods and Results-A total of 1820 patients were randomly assigned to CRT plus an ICD or to an ICD only in a 3:2 ratio. Echocardiographic studies were obtained at baseline and 12 months later in 1372 patients. We compared changes in cardiac size and performance between treatment groups and assessed the relationship between these changes over the first year, as well as subsequent outcomes. Compared with the ICD-only group, the CRT-plus-ICD group had greater improvement in left ventricular end-diastolic volume index (Ϫ26.2 versus Ϫ7.4 mL/m 2 ), left ventricular end-systolic volume index (Ϫ28.7 versus Ϫ9.1 mL/m

We thank Drs Hamilton and Filardo for their interest in our work. 1 In a prospective randomized trial in which the same patient is subjected to repeated measurements over time, missing data and repeated measures are known statistical problems for which it is difficult to account. As suggested by Drs Hamilton and Filardo, a longitudinal analysis is the most appropriate model for accounting for missing data. 2 A longitudinal model was used initially to analyze the data, essentially yielded similar clinical results, and did not result in any substantive change to our published conclusions. However, at the suggestion of our reviewers and other readers, we presented the data in a way that is more conventional and easily understood and that facilitates comparison with other trials. To account for repeated measures in the same patient, paired data were used in the published analysis. This should have been more clearly stated in our Methods. We would be happy to provide the data as analyzed by the longitudinal model. With regard to multiplicity, it is correct that additional tests increase the likelihood of type I error. Although it is common practice to provide a comprehensive analysis using the available data, our analysis should be considered exploratory in nature. The limitations of multiple retrospective analyses should have been clearly emphasized; however, the results are valid and are derived from a relatively large trial that directly compared all 3 pacing modalities in a prospective randomized fashion. Disclosures

The previous findings that SPWMD predicts reverse remodeling or clinical improvement with CRT were not reproducible in patients randomized in the CONTAK-CD trial.