BACKGROUND: In contrast with kidneys, transplantation of livers originating from non-heart-beating donors remains rare, mainly because warm ischemia causes a higher rate of potentially lethal primary graft nonfunction. Little is known on the tolerance of liver grafts to warm ischemia. No techniques are available to assess the viability of ischemic livers before implantation. Therefore, experimental models are needed to address these questions before non-heart-beating liver transplantation can be more widely applied. This study aims to develop a reproducible large animal model of liver transplantation using non-heart-beating donors and, in this model, to define the tolerance of the liver to warm ischemia. METHODS: Pigs weighing 25to 30 kg are used. In donors, cardiac arrest is caused by ventricular fibrillation. After increasing lengths of warm ischemia (0, 15, 30, 45, and 60 min), the liver is flushed in situ with 4 degrees C histidine tryptophan ketoglutarate preservation solution and procured. The liver is transplanted after a 4-hour cold storage period. RESULTS: Control livers (no warm ischemia) and livers exposed to 15 minutes of warm ischemia function normally after transplantation, whereas all livers submitted to 60 minutes of warm ischemia display primary nonfunction and cause recipient death. Graft function and survival are occasionally observed after 30 and 45 minutes of warm ischemia. CONCLUSIONS: A reproducible model of non-heart-beating liver transplantation is described. We found that the liver tolerates 15 minutes of warm ischemia. This preclinical model is a valid tool to develop techniques to predict the quality of ischemic livers before implantation and to design interventional strategies to improve the tolerance of the liver to warm ischemia.