Author Affiliation: Harrington-McLaughlin Heart & Vascular Institute, University Hospitals Case Medical Center, Case Western Reserve University School of Medicine, Cleveland, Ohio.
I disagree with the comments of Magee et al. There was no multiple testing in our meta-analysis. The approach used in this meta-analysis was different from the exploratory subgroup analyses of a single clinical trial. The predefined hypothesis of this meta-analysis was that patients with moderately prolonged QRS interval (ie, 120-150 milliseconds [ms]) do not benefit from cardiac resynchronization therapy (CRT), which was suggested by many investigators even before the major clinical trials of CRT were performed and is biologically plausible.1- 3 Therefore, the only subgroups examined in this analysis were moderately prolonged QRS interval vs severely prolonged QRS interval (ie, >150 ms). Because of this, making corrections for multiple testing does not pertain to this analysis. Moreover, even for a hypothetical situation where multiple testing would be relevant, the significant P value of <.001 for the difference in the response of the 2 QRS subgroups is so robust that the results would not change with any type of correction.
Sipahi I. Limitations of Subgroup Analyses in Meta-analysis of Cardiac Resynchronization Therapy by QRS Duration—Reply. Arch Intern Med. 2012;172(4):375-376. doi:10.1001/archinternmed.2011.1685