Résumé : Objective: To determine the impact of a restrictive vs. a liberal transfusion strategy on new or progressive multiple organ dysfunction syndrome in children post cardiac surgery. The optimal transfusion threshold after cardiac surgery in children is unknown. Design: Randomized, controlled trial. Setting: Tertiary pediatric intensive care units. Patients: Participants are a subgroup of pediatric patients post cardiac surgery from the TRIPICU (Transfusion Requirements in Pediatric Intensive Care Units) study. Exclusion criteria specific to the cardiac surgery subgroup included: age <28 days and patients remaining cyanotic. Intervention: Critically ill children with a hemoglobin ≤95 g/L within 7 days of pediatric intensive care unit admission were randomized to receive prestorage leukocyte-reduced red-cell transfusion if their hemoglobin dropped either <70 g/L (restrictive) or 95 g/L (liberal). Measurements and Main Results: Postoperative cardiac patients (n = 125) from seven centers were enrolled. The restrictive (n = 63) and liberal (n = 62) groups were similar at baseline in age (mean ± standard deviation = 31.4 ± 38.1 mos vs. 26.4 ± 39.1 mos), surgical procedure, severity of illness (Pediatric Risk of Mortality score = 3.4 ± 3.2 vs. 3.2 ± 3.2), multiple organ dysfunction syndrome (46% vs. 44%), mechanical ventilation (62% vs. 60%), and hemoglobin (83 vs. 80 g/L). Mean hemoglobin remained 21 g/L lower in the restrictive group after randomization. No significant difference was found in new or progressive multiple organ dysfunction syndrome (primary outcome) in the restrictive group vs. liberal group (12.7% vs. 6.5%; p =.36), pediatric intensive care unit length of stay (7.0 ± 5.0 days vs. 7.4 ± 6.4 days) or 28-day mortality (3.2% vs. 3.2%). Conclusion: In this subgroup analysis of cardiac surgery patients, a restrictive red-cell transfusion strategy, as compared with a liberal one, was not associated with any significant difference in new or progressive multiple organ dysfunction syndrome, but this evidence is not definitive. Copyright © 2010 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins.