Effects of blood transfusion on clinical and functional outcomes in patients with hip fracture

BACKGROUND: Anemia and transfusion are common among elderly patients requiring surgery. The effects of transfusion on morbidity and mortality are controversial. The influence of transfusion on risk‐adjusted mortality, readmissions, and functional mobility was examined. STUDY DESIGN AND METHODS: A consecutive cohort of 551 patients undergoing surgery for hip fracture at four hospitals was prospectively studied. Outcomes were death, readmission, and functional independence measure‐locomotion scores within 60 days of discharge. The trigger Hb level was defined as the lowest value before the first postoperative transfusion. Multivariate analyses adjusted for a validated, hip‐fracture‐specific risk model and predictors of transfusion. RESULTS: Overall, 54.4 percent of patients received transfusions after surgery. Seventy‐two percent of patients with a lowest postoperative Hb level of less than 10.0 g per dL received transfusions compared to 19.6 percent of those whose lowest measurement was at least 10.0 g per dL (p < 0.0001). In the 60 days after discharge, 3.8 percent of patients died and 16.9 percent were readmitted. Transfusion was associated with lower risk‐adjusted odds of readmission (OR, 0.54; 95% CI, 0.30‐0.97), but it did not influence mortality or mobility functioning. In subgroups analyses, the benefit of transfusion on readmission rates appeared to be concentrated among patients with a trigger Hb level of less than 10.0 g per dL. For patients with a trigger Hb level of at least 10.0 g per dL, transfusion did not affect risk‐adjusted rates of death or readmission, but was associated with better risk‐adjusted functional mobility scores (p < 0.01). CONCLUSIONS: Postoperative transfusion reduced the risk of readmission but did not decrease mortality or improve mobility. Randomized controlled trials of different transfusion strategies will be needed to clarify the true benefits and risks of transfusion in surgical patients.