Carvedilol in Patients with Chronic Heart Failure

Publisher Website
Google Scholar
Abstract
In their collection of studies evaluating the efficacy of carvedilol in congestive heart failure (May 23 issue),1 Packer et al. replaced the analyses of primary end points with an analysis of mortality. The data in the article come from four different studies, none of which used overall mortality as a prospective primary or secondary end point. The primary end points of the studies were measures of morbidity (exercise tolerance in three and hospitalization in the fourth). The analyses of these primary end points are not provided, analyses that were shown to have nonsignificant results (P>0.05) in three of the four studies. The failure to report nonsignificant end points in clinical trials has been criticized,2 and prespecifying the anticipated analysis in a study protocol is now an accepted standard.3 The investigators in the Carvedilol Heart Failure Study Group have opened themselves to the criticism that they selected the analysis of overall mortality because of its favorable results, thereby biasing the conclusions. The first and most important analysis in a clinical trial must be that of the prospectively stated primary end point.