BACKGROUND: Normothermic machine perfusion (NMP) of liver grafts donated after circulatory death (DCD) has shown promise in large animal and clinical trials. Following procurement, initial flush with a cold preservation solution is the standard of care. There is concern that initial cooling followed by warming may exacerbate liver injury, and the optimal initial flush temperature has yet to be identified. We hypothesize that avoidance of the initial cold flush will yield better quality liver grafts. METHODS: Twenty-four anaesthetized pigs were withdrawn from mechanical ventilation and allowed to arrest. After 60-minutes of warm ischemia to simulate a DCD procurement, livers were flushed with histidine-tryptophan-ketoglutarate (HTK) at 4°C, 25°C or 35°C (n = 4 per group). For comparison, an adenosine-lidocaine crystalloid solution (AD), shown to have benefit at warm temperatures in heart perfusions, was also used (n = 4 per group). During 12-hours of NMP, adenosine triphosphate (ATP), lactate, transaminase levels, and histological injury were determined. Bile production and hemodynamics were monitored continuously. RESULTS: ATP levels recovered substantially following 1-hour of NMP reaching pre-ischemic levels by the end of NMP with no difference between groups. There was no difference in peak aspartate aminotransferase (AST) or in lactate dehydrogenase (LDH). Portal vein resistance was lowest in the 4°C group reaching significance after 2 hours (0.13 CI -0.01,0.277, p = 0.025). Lactate levels recovered promptly with no difference between groups. Comparison to AD groups showed no statistical difference in the abovementioned parameters. On electron microscopy the HTK4°C group had the least edema with mean cell thickness of 2.92µm (p = 0.41) while also having the least sinusoidal dilatation with a mean diameter of 5.36µm (p = 0.04). For AD, the 25°C group had the lowest mean cell thickness at 3.14µm (p = 0.09). CONCLUSIONS: Avoidance of the initial cold flush failed to demonstrate added benefit over standard 4°C HTK in this DCD model of liver perfusion.