Purpose: Acute kidney injury (AKI) occurs in up to 20% of adult patients after cardiac surgery, and renal medullary hypoxia is an important cause. We have shown in previous human studies that persistently low intra-operative urinary pO2 is correlated with a high risk of AKI. However, before attempting to intervene in humans, by altering the CPB variables (flow, pressure, haematocrit and temperature) we determined the most effective methods to improve oxygenation in an animal model. Methods: Merino ewes (n=12, weight 35–45 kg) underwent aseptic surgery to implant monitoring probes. 5 days later the sheep were anaesthetized again, the chest was opened, and the sheep placed on CPB, commencing with a flow rate at 80 ml/kg/min (approx 2.6 L/min/m2) and MAP at 70 mm Hg. Different protocols were performed, altering CPB flow while maintaining a constant MAP, and then altering MAP at a steady pump flow. Results: Anaesthesia was associated with significant reduction in renal blood flow and cortical perfusion. CPB caused further reduction in renal blood flow (61% less than the conscious sheep), reduced cortical perfusion (44% fall), and reduced medullary perfusion (40% fall) and medullary pO2. Renal oxygen delivery (DO2) was reduced by more than 50% on CPB compared to the conscious sheep, while renal O2 extraction increased. Increased CPB flow modestly increased the renal blood flow and the medullary pO2. Infusion of metaraminol at 0.2 mg/min to achieve MAP at 80 mm Hg resulted in a significant improvement in renal blood flow, and both cortical and medullary pO2. Simultaneous increase in flow and MAP optimized renal blood flow and oxygenation. Increased haematocrit and moderate hypothermia had no effect. Conclusion: CPB may result in significant fall in renal blood flow and in renal oxygenation, which is the likely cause of clinical AKI. This can be ameliorated by increased CPB flow (at 3 L/min/m2) and in particular with increased MAP, the optimal perfusion being at a MAP of 80 mmHg