Preoperative Erythropoietin Alpha Reduces Postoperative Transfusions in THA and TKA but May Not Be Cost-effective

Preoperative erythropoietin alpha (EPO) has been shown to be effective at reducing postoperative blood transfusions in total hip arthroplasty (THA) and total knee arthroplasty (TKA); however, treatment with EPO is associated with additional costs, and it is not known whether these costs can be justified when weighed against the transfusion reductions achieved in patients who receive the drug. The purpose of this study is to investigate (1) efficacy of preoperative EPO in reducing postoperative transfusions in TKA and THA; (2) whether patients treated with EPO have reduced length of stay or a different discharge disposition; and (3) whether EPO use reduces overall blood management costs. Patients undergoing primary THA or TKA over a 10-month period with preoperative hemoglobin < 13 g/dL were recommended to be treated preoperatively with EPO. During that time, 80 of 286 (28%) patients met that inclusion criterion and the treating team recommended EPO to all of them; of that group, 24 (30%) opted to take EPO and 56 (70%) opted not to. Patients receiving at least one dose of EPO and those not receiving EPO were compared in terms of transfusion frequency, length of stay and discharge disposition, and overall blood management costs. Demographics, preoperative hemoglobin, and operative blood loss for both groups were similar (p > 0.05). No transfusion triggers were used; rather, patients with postoperative hemoglobin < 10 mg/dL and who were symptomatic despite fluid boluses were transfused. The clinician responsible for transfusing symptomatic patients was blinded to the patient's EPO treatment status. Costs were defined as direct costs paid or incurred by our institution for EPO, allogeneic blood, and variable costs associated with patient care after THA/TKA. A decision-tree cost analysis was performed using the collected clinical data and cost data collected from our institution; the analysis considered total associated blood management cost for an EPO and a non-EPO strategy with sensitivity analysis of key cost variables. The proportion of patients receiving transfusions was lower in patients who received EPO than in patients who did not (0% [zero of 24] versus 41% [23 of 56]; p < 0.001). The mean length of inpatient hospital stay (EPO: 3.0 +/- A 0.4 versus control: 3.3 +/- A 0.8 days, p = 0.77) and discharge disposition also was not different between the groups. The cost analysis demonstrated that the EPO strategy was more costly compared with no EPO (USD 2632 versus USD 2284) and its cost would need to be less than USD 225/dose for this to change. EPO reduced the need for postoperative transfusions in high-risk patients undergoing THA and TKA; however, it was not found to be cost-effective in our model. Our model could not consider relatively rare complications of blood transfusions, including disease transmission, deep periprosthetic infections, and transfusion reactions, but if surgeons or patients value avoiding these potential but rare factors highly, this could reasonably influence the decision of whether to use EPO despite our findings that it was not cost-effective. Level III, therapeutic study. See Guidelines for Authors for a complete description of levels of evidence.