Early Results of Orthopaedic Trauma and Anemia: Conservative Versus Liberal Transfusion Strategy

Abstract

Introduction: The purpose of this study was to determine whether it is safe to use a conservative packed red blood cell transfusion hemoglobin threshold (5.5 g/dL) compared with a liberal transfusion threshold (7.0 g/dL) for asymptomatic patients with musculoskeletal-injured trauma out of the initial resuscitative period. Methods: This was a multicenter, prospective, nonblinded, randomized study done at three level 1 trauma centers. One hundred patients were enrolled. One patient was inappropriately enrolled, withdrawn from the study, and excluded from analysis leaving 99 patients (49 liberal and 50 conservative) with 30-day follow-up. After initial resuscitation, patients were enrolled and randomized to either a liberal or a conservative transfusion strategy. This strategy was followed throughout the index hospitalization. The primary outcome of the study was infection. Superficial infection was defined as clinical diagnosis of cellulitis or other superficial infection treated with oral antibiotics only. Deep infection was defined as clinical diagnosis of fracture-related infection requiring IV antibiotics and/or surgical débridement. Results: Ninety-nine patients were successfully followed for 30 days with 100% follow-up during this time. Seven infections (14%) occurred in the liberal group and none in the conservative group (P < 0.01). Five deep infections (10%) occurred in the liberal group and none in the conservative group (P = 0.03). Three superficial infections (6%) occurred in the liberal and none in the conservative group, which was not a significant difference (P = 0.1). No difference was observed in length of stay between groups. Discussion: Transfusing young healthy asymptomatic patients with orthopaedic trauma for hemoglobin <7.0 g/dL increases the risk of infection. No increased risk of anemia-related complications was identified with a conservative transfusion threshold of 5.5 g/dL. Data availability and trial registration numbers: Data are available on request. IRB protocol number is 1402557771. This study was registered with Clinicaltrials.gov identifier NCT02972593. Level of evidence: Level 2, unblinded prospective randomized multicenter study.