Background: WBC-replete blood transfusion has been suggested as an independent cause of increased postoperative infection. Study design and methods: A total of 597 patients undergoing elective coronary artery or heart valve surgery were randomly assigned to receive plasma- reduced (PR), buffy coat-depleted (BCD), or WBC- filtered (WCF) RBCs in the event of requiring blood transfusion. Details of postoperative course were recorded. Further information was collected from the patient's general practitioner 3 months after discharge. Results: No significant difference in inpatient infection rates was observed among patients randomly assigned to receive PR, BCD, or WCF RBCs. When only those receiving transfusion were analyzed (n = 509), use of PR RBCs was associated with more events coded as infections (p ≤ 0.05) compared with BCD or WCF RBCs. However, when events coded as urinary tract infections were excluded, there was no significant difference among the three groups. Follow-up performed after discharge showed no difference in readmission rates, but a higher reported rate of infection in those randomly assigned to receive WCF RBCs (p <0.02). Conclusion: No evidence has been found, analyzed by intention to treat, that use of WBC-reduced, BCD, or WCF RBCs reduces postoperative inpatient infection in patients undergoing cardiac bypass surgery.