In a porcine liver transplant model, a brief period of oxygenated hypothermic machine perfusion (HMP) at the end of simple cold storage (SCS) has been shown to improve the viability of damaged liver grafts. To test the clinical validity of this strategy, we randomized SCS-discarded human liver grafts to either 4 hours of HMP (n = 13) or an additional 4 hours of SCS (n = 14). All livers were then warm reperfused to mimic ischemia-reperfusion injury ex vivo. The settings for HMP were: portal vein: 3 mm Hg, 300 mL/min and hepatic artery: 20 mm Hg, pO(2): 300 mm Hg. Perfusion used Kidney Machine Perfusion Solution at 4°C to 8°C. During warm reperfusion, aspartate aminotransferase (AST) and lactate dehydrogenase (LDH) values were higher (P < .015) among the SCS versus HMP methods at all times. The AST slope was lower in HMP versus SCS (P = .01). The LDH slope tended to be lower for HMP versus SCS (P = .07). Morphological scores were not different between HMP and SCS. At the start of warm reperfusion, MAPK was lower in HMP versus SCS (P = .02). Endothelin-1 (EDN1) and ICAM-1 tended to be lower in HMP versus SCS (P = .1 and .07, respectively). No difference was noted in MAPK, EDN1, and ICAM-1 after 60 or 120 minutes of warm reperfusion. In conclusion, HMP down-regulated MAPK and tended to reduce EDN1 and ICAM-1 mRNA in human liver grafts. During warm reperfusion, HMP versus SCS livers showed reduced AST and LDH release but no morphological difference. Further optimization of liver HMP may require different timing/duration of perfusion and/or an higher perfusion temperature.