Is it all about cutoffs? Can DIC scores predict bleeding in APL?

Abstract

Predicting outcome with a scoring system should be interpreted with caution. In the application of the disseminated intravascular coagulation (DIC) scoring system, a key point is to use well-established cutoff values for the variables. In the article by Mitrovic et al., activated partial thromboplastin time was presented by seconds, while prothrombin time (PT) was expressed as percentage. Such expressions appear confusing and contradictory. In the DIC scoring system, scoring of PT is determined by prolongation in seconds. I cannot see the reason to use the percentage of PT in this article. Furthermore, the PT cutoff was defined as 50%, which may be entirely arbitrary. We had conducted an investigation in our hospital cohort. We had analyzed our cohort by the chi-square method to determine the correlation between DIC scores and bleeding events. We found no relationship between scores and events of bleeding or fatal bleeding. However, reanalysis using a cutoff DIC ≧6 revealed a marginally significant difference in bleeding risk between high-score and low-score patients (P = 0.046). The difference was insignificant for life-threatening bleeding. While our experience appears to support authors' conclusion that DIC score ≧6 is associated with bleeding in acute promyelocytic leukemia patients, we remain skeptical in regard to such manipulations. We believe subsequent studies on a large-scale basis or more accumulated data critically reviewed by experts are needed to shed lights on this important issue.