Cardiac surgical patients are one of the main users of blood products accounting for approximately half of all transfusions in surgical patients. Annually around 3500 cardiac operations are performed in New Zealand. There is little evidence to guide perioperative transfusion practice and the optimum haemoglobin thresholds for transfusion. Anaemia and transfusions are associated with morbidity and mortality and many unnecessary transfusions are being administered due to lack of evidence of benefit. We propose to conduct a multicentre international randomised controlled trial of 3664 patients to determine what degree of anaemia requires blood transfusion in higher risk patients undergoing cardiac surgery. If this trial demonstrates the safety of a restrictive transfusion strategy, the risks and costs attributable to unnecessary red blood cell transfusion will be reduced. Alternatively, if a restrictive transfusion strategy is found to be inferior, risks of complications related to untreated perioperative anaemia will be reduced.