; Red blood cell transfusions are an extremely common medical intervention in both the United States and worldwide; over 11 million units are transfused in the United States. Between 60% and 70% of all blood is transfused in the surgical setting. Despite the common use of red blood cell transfusions, the threshold for transfusion has not been adequately evaluated and is very controversial. A decade ago the standard of care was to administer a peri-operative transfusion whenever the hemoglobin (Hgb) level fell below 10 g/dl (the """"""""10/30 rule""""""""). Concerns about the safety of blood, especially with respect to HIV and hepatitis, and the absence of data to support a 10 g/dl threshold led to current standard of care today to administer blood transfusions based on the presence of symptoms and not a specific Hgb/hematocrit level. However, there are no randomized clinical trials in surgical patients that have tested the efficacy and safety of withholding blood until the patient develops symptoms or the """"""""10/30"""""""" approach to transfusion. Patients with underlying cardiovascular disease are at greatest risk of adverse effects from reduced Hgb levels. We propose to conduct a multi-center randomized trial to test if a more aggressive transfusion strategy that maintains postoperative Hgb levels above 10 g/dl improves patient outcome as compared to a more conservative strategy that withholds blood transfusion until the patient develops symptoms of anemia. Eligible patients for the trial will have undergone surgical repair for a hip fracture and have a postoperative Hgb level below 10 g/dl within three days of surgery. Only patients with cardiovascular disease will be entered into the study. Patients will be randomized to one of the two transfusion strategies. The 10 g/dl threshold strategy will use enough red blood cell units to maintain Hgb levels at or above 10 g/dl through hospital discharge. Symptomatic transfusion strategy patients will receive red blood cell transfusions for symptoms of anemia, although transfusion is also permitted but not required if the Hgb level falls below 8 g/dl. Outcomes will include functional recovery (primary outcome: ability to walk ten feet across a room without human assistance at 60-days post-randomization), long-term survival, nursing home placement, and postoperative complications (death in hospital or within 30 days, pneumonia, myocardial infarction, thromboembolism, stroke, delirium). We will randomize 2,600 patients from 25 centers over a 3.5-year period. This will allow us to detect a 16% relative risk reduction in the loss of ability to walk independently with power about 0.90. A pilot study in 84 patients demonstrated the feasibility of the study. Ambulation at 60 days is known to be highly predictive of ultimate functional outcome as well as of mortality at one year. Because inability to walk again has such important implications for quality office, and because, unfortunately, it is a common problem, it far outweighs the remote chance of viral infection or other complications from transfusion in these elderly patients. Also, this study will measure the frequency and 95% confidence intervals of the medical errors that are important in this patient population and are poorly documented in the literature. The medical errors that will be measured are: transfusion errors (blood transfusion to the wrong patient, mislabeling of samples for type and cross match, use of whole blood instead of packed red cells), failure to use thromboembolism prophylaxis, incorrect antibiotic prophylaxis, wrong site surgery and femoral shaft fracture.
