Meta-analysis of clinical studies of the purported deleterious effects of "old" (versus "fresh") red blood cells: are we at equipoise?

Abstract

Background: A meta-analysis examined whether the available data support an adequate suspicion that transfusion of old red blood cells (RBCs) is associated with increased mortality, organ failure, infection, prolonged mechanical ventilation, and prolonged stay in the hospital or the intensive care unit. Such suspicion is required for intentionally exposing patients enrolled in randomized controlled trials (RCTs) to the known or probable--but rare--risks of old RBCs, to document (and prevent) purported common adverse effects of old RBCs.

Study design and methods: Observational studies presenting adjusted results were eligible for analysis if the adequacy of the adjustment for confounding factors could be assessed. Three RCTs and 24 observational studies were retrieved. Medically and statistically homogeneous studies were integrated by fixed-effects methods. Otherwise homogeneous studies conducted in different clinical settings were integrated by random-effects methods.

Results: Based on "as-treated" analysis, transfusion of old RBCs was associated with a significant reduction in mortality (summary odds ratio, 0.38; 95% confidence interval, 0.14-0.99; p < 0.05) across two small RCTs. Integration of adjusted findings on the same outcome, from observational studies conducted in the same setting, produced summary results that were either negative (in six analyses) or impossible to evaluate owing to uncontrolled confounding by the number of transfused RBCs (in two analyses).

Conclusion: The available data do not support an adequate suspicion that old RBCs may be associated with common adverse morbidity and/or mortality outcomes, so as to justify exposing experimental subjects to the other known or probable--but rare--risks of old RBCs.