Predicting major hemorrhage in patients with pelvic fracture.

BACKGROUND Pelvic fractures can be an important source of major hemorrhage in victims of blunt trauma. However, no rapid and reliable noninvasive method exists for predicting which subjects will have major hemorrhage. The objective of this study is to use information available upon presentation to the trauma center to develop a clinical prediction rule to identify subjects with pelvic fracture who are at high risk of major hemorrhage. METHODS A retrospective cohort study was performed on all subjects with pelvic fracture from blunt force mechanism at a single level one trauma center during a 4.3 year period. Chart review identified findings from initial pelvic radiographs and from emergency department care including mechanism of injury, and hemodynamic status. Major hemorrhage was defined by angiographic findings, transfusion requirement and pelvic hemorrhage volume. Logistic regression was used to formulate a clinical prediction rule to stratify subjects based on probability of major hemorrhage. RESULTS Complete data were available on 627 of 783 eligible subjects. Predictors of major hemorrhage included emergency department hematocrit 30 or less, pulse rate of 130 or greater, displaced obturator ring fracture and pubic symphyseal wide diastasis. Combinations of predictors defined groups with probability of major hemorrhage from 1.6% to 66%. CONCLUSIONS Probability of major pelvic fracture related hemorrhage can be estimated from initial pelvic radiograph, pulse, and hematocrit.