The fundamental importance of baseline comparisons in a clinical trial

Abstract

THE FUNDAMENTAL IMPORTANCE OF BASELINE COMPARISONS IN A CLINICAL TRIAL To the Editor: We read with interest the study by Hecht-Dolnik and associates in the September 2009 issue of the Journal, ‘‘Hetastarch Increases the Risk of Bleeding Complications in Patients After OffPump Coronary Bypass Surgery: A Randomized Clinical Trial.’’ In this article, they conclude that the intraoperative administration of 1 L of hetastarch was associated with an increase in postoperative chest tube drainage and transfusion of blood products, which resulted in early termination of the study. Although we acknowledge the importance of the clinical question the authors have addressed, we have major concerns regarding the equality of the intervention and the control groups at baseline and the method in which this is reported. Although the authors claim that there were no statistically significant differences in baseline characteristics between groups, Table 1 indicates the baseline international normalized ratio (INR) in the hetastarch group was 1.47 (2.12) whereas in the albumin group baseline INR was 1.07 (0.10), values presented as mean (standard deviation). Although the mean value between these 2 groups may not be statistically different, it is misleading to claim baseline equivalence between the groups. The upper limit of INR is typically quoted at 1.3, indicating the mean INR in the hetastarch group is abnormal and in the albumin group mean INR is normal. Furthermore, the large standard deviation in baseline INR in the hetastarch group can only be explained by the presence of one or more outliers with markedly abnormal baseline value. In contrast, the standard deviation for INR of the