Survival bias revisited

摘要

We feel compelled to respond to the recent study by Brown et al.' entitled "Debunking the survival bias myth: Characterization of mortality during the initial 24 hours for patients requiring massive transfusion." In this study, the authors claim to provide evidence that the observed mortality benefit associated with high component transfusion ratios is not related to survival bias. The authors report a survival advantage associated with higher ratios for the first 24 hours of hospitalization but fail to present any longer-term survival data, such as 30-day mortality. While we appreciate the interest in early survival as an outcome measure because it may be directly attributable to transfusion strategy during hemorrhage control, survival data beyond 24 hours are necessary for a complete assessment of treatment efficacy. This omission of these data essentially precludes any meaningful comparison of survival outcomes with those of Snyder et al.2 and Magnotti et al.3 because the outcome reported in both of those studies was, in fact, in-hospital mortality and not survival at 24 hours.