Blood transfusion associated with shoulder arthroplasty

Abstract

Background: Studies have reported high rates of transfusion in shoulder arthroplasty. This study was conducted to evaluate the rate of transfusion at our institution, to confirm reported risk factors for transfusion, and to look for changes over time.We hypothesized that transfusion rates associated with shoulder arthroplasty at our institution are lower than those recently reported and that the incidence of transfusion is higher in individuals with low preoperative hemoglobin, with revision arthroplasty, and in older individuals.

Materials and methods: A retrospective review of 366 shoulder arthroplasties (323 patients) was performed. This included total shoulder arthroplasties, hemiarthroplasties, revision arthroplasties, and reverse total shoulder arthroplasties. Logistic regression analysis evaluated the association of clinical variables with transfusion. Early (1996-2005) and late (2006-2009) groups were compared to evaluate changes in demographics and transfusion rates over time.

Results: The overall transfusion rate was 7.4% (27 of 339). Predictors of transfusion were higher intraoperative blood loss, low preoperative hemoglobin level, and humeral cement fixation. Procedure type was not predictive of transfusion. There was no difference in transfusion rates between the early and late groups, but the late group had an increased use of general anesthesia combined with a regional block, increased intraoperative blood loss, and increased use of sequential compression devices for venous thromboembolism prophylaxis.

Conclusions: Lower preoperative hemoglobin, higher intraoperative blood loss, and humeral cement fixation were predictors of transfusion, but not female sex, increasing age, type of procedure, or comorbidities.