In this issue of JAMA, Rohde and colleagues report results of a meta-analysis evaluating the association between a liberal transfusion strategy vs a restrictive strategy and risk of hospital-acquired infections.1 In the meta-analysis, which included 17 randomized trials that compared these strategies among 7456 patients, the absolute rates of serious hospital-associated infection were 12.7% in the liberal transfusion group and 10.6% in the restrictive transfusion group. The number needed to treat with a restrictive transfusion strategy to prevent serious infection was about 48 patients. Results were consistent when analyses were repeated in trials with concealed randomization and infrequent protocol violations but not for overall infections. A prior meta-analysis of liberal vs restrictive transfusion strategies and infection outcomes reported a nonstatistically significant reduction in infections.2 The current meta-analyses by Rohde et al includes several newly published trials, although these significant differences in serious infection rates may not hold up when the results of other large transfusion trials are reported. Most trials define a restrictive transfusion strategy as the administration of red blood cells (RBCs) once hemoglobin falls below either 7 or 8 g/dL, and most trials define a liberal strategy as the administration of RBCs once hemoglobin level falls below 10 g/dL.
Carson JL. Blood Transfusion and Risk of InfectionNew Convincing Evidence?. JAMA. 2014;311(13):1293-1294. doi:10.1001/jama.2014.2727