Admission hematocrit predicts the need for transfusion secondary to hemorrhage in pediatric blunt trauma patients

BACKGROUND Pediatric trauma uses a substantial amount of resources. Quick and cost-effective measures that can be used to identify children with clinically relevant injuries are essential to resource allocation and optimization of patient care. Admission hematocrit is rapid and inexpensive, causes minimal harm, and can potentially aid in critical decision making. We hypothesize that admission hematocrit predicts the need for transfusion in pediatric blunt trauma patients. METHODS Records of trauma patients age 0 year to 17 years (2005-2013) who presented to a pediatric Level 1 trauma center were retrospectively reviewed. Data collected include demographics, computed tomographic scan findings, need for an intervention secondary to bleeding (blood transfusion, angioembolization, or operation), and admission hematocrit. RESULTS We found a significant decrease in admission hematocrit between patients requiring a transfusion and patients who did not (27% vs. 36%, p < 0.01). We evaluated a subset of patients who had an abdominal computed tomographic scan and found a significant decrease in admission hemocrit between those who required a transfusion for an intra-abdominal injury and those who did not (29% vs 37%, p < 0.01). In this subset, serial hematocrit values remained significantly lower in the patients requiring a transfusion up to 67 hours after admission (p = 0.04). A cutoff admission hematocrit of 35% or less has a sensitivity of 94% and a negative predictive value of 99.9% in identifying children who need a transfusion after blunt trauma. CONCLUSION An admission hematocrit of 35% or less provides a reliable screening test because of its low false negative rate and high specificity for identifying patients at an increased risk of bleeding after injury. Admission hematocrit could be widely implemented to identify patients who may need a transfusion with low expense and minimal harm for our pediatric patients and may be able to alter the entire course of their trauma resuscitation. LEVEL OF EVIDENCE Epidemiologic/prognostic study, level III.