Résumé : Objectives: In paediatric cardiac surgery, red blood cell (RBC) transfusions are given to treat haemorrhage or to increase oxygen delivery (therapeutic transfusion). Sometimes, RBCs are added to the prime solution to avoid a too low haematocrit on bypass (CPB-driven transfusion). Our study investigated whether the reason for RBC transfusion might affect severe postoperative morbidity or mortality. Methods: This retrospective cohort study was conducted in a tertiary care level, children's hospital. The charts of all patients admitted between 2006 and 2009 were reviewed. Among transfused patients, children receiving a therapeutic transfusion were compared with those receiving a CPB-driven transfusion. The primary outcome was severe postoperative morbidity or mortality. Statistically significant variables founded with univariate analysis were incorporated in the multivariable logistic regression analysis to build a model that predicts severe postoperative morbidity or mortality. This model was then adjusted for clinically relevant variables that may interfere with 'indication for transfusion'. Results: One hundred and one (48%) patients in the therapeutic and 92 (26%) patients in the CPB-driven transfusion group developed severe postoperative morbidity or died (P < 0.001). A multivariable logistic regression showed that weight, ASA, calculated blood loss, intraoperative fluid balance, duration of CPB time and the indication group for transfusion [1.64 (1.03-2.62); P = 0.039] were independent risk factors for severe postoperative morbidity or mortality. The indication group for transfusion remained significant [2.0 (1.16-3.45); P = 0.013] after adjustment for significant predictors of indication for transfusion. Number of patients with infections, or neurological deficit, and length of mechanical ventilation, PICU and hospital length of stay, were significantly higher in the therapeutic compared with the CPB-driven transfusion group one. Conclusions: The indication group for transfusion has an impact on the occurrence of severe postoperative morbidity or mortality in children undergoing cardiac surgery. © The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.