Restrictive Versus Liberal Strategy for Red Blood-Cell Transfusion: A Systematic Review and Meta-Analysis in Orthopaedic Patients.

BACKGROUND

Current guidelines recommend restrictive criteria for red blood-cell transfusion in most clinical settings. However, patients undergoing orthopaedic surgery may require distinct transfusion criteria since benefits and potential harm often vary considerably based on patient characteristics and surgical procedures. We aimed to assess the efficacy and safety of restrictive transfusion in patients undergoing orthopaedic surgery, especially in important subgroups.

METHODS

Electronic databases were searched to identify randomized controlled trials investigating restrictive (mostly a hemoglobin level of 8.0 g/dL or symptomatic anemia) versus liberal (mostly a hemoglobin level of 10.0 g/dL) transfusion in patients undergoing orthopaedic surgery. For the primary outcome of cardiovascular events, we performed random-effects meta-analyses to synthesize the evidence and to assess the effects in different subgroups according to patient characteristics (with versus without preexisting cardiovascular disease) and surgical procedures (hip fracture surgery versus elective arthroplasty).

RESULTS

Ten trials involving 3,968 participants who underwent hip or knee surgery were included. Mean participant age ranged from 68.7 to 86.9 years. Compared with liberal transfusion, restrictive transfusion increased the risk of cardiovascular events (8 trials; 3,618 participants; relative risk [RR], 1.51; 95% confidence interval [CI], 1.16 to 1.98; p = 0.003; with no heterogeneity across all trials), irrespective of preexisting cardiovascular disease (pinteraction = 0.63). In a subgroup analysis, the increase was observed in patients undergoing hip fracture surgery (RR, 1.51; 95% CI, 1.08 to 2.10; p = 0.02), but did not reach significance in those undergoing elective arthroplasty (RR, 1.53; 95% CI, 0.96 to 2.44; p = 0.07). To minimize the bias caused by variations in transfusion threshold, we conducted an analysis that only included trials using 8.0 g/dL hemoglobin or symptomatic anemia as the threshold for restrictive transfusion and obtained identical results (6 trials; 2,872 participants; RR, 1.51; 95% CI, 1.09 to 2.08; p = 0.01; I = 0%). The 2 arms did not differ with respect to the rates of all infections, 30-day mortality, thromboembolic events, wound infection, pulmonary infection (mainly pneumonia), and cerebrovascular accidents (mainly stroke).

CONCLUSIONS

In patients undergoing orthopaedic surgery, when compared with liberal transfusion, restrictive transfusion increases the risk of cardiovascular events irrespective of preexisting cardiovascular disease. Importantly, the increased risk was observed in patients undergoing hip fracture surgery but did not reach significance in those undergoing elective arthroplasty.

LEVEL OF EVIDENCE

Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.