BACKGROUND Portal vein embolization (PVE) is a technique to increase future remnant liver volume. A standardized animal model, resembling the clinical PVE procedure, is needed to clarify some of the unresolved issues surrounding PVE. For this purpose we developed a new rabbit model for PVE. METHODS Twenty female New Zealand white rabbits were allocated to two protocols, each containing two subgroups. Eighty percent of the liver portal venous system was embolized with polyvinyl-alcohol particles and coils (protocol 1: 300-500 μm particles and one coil; protocol 2: 90-180 μm combined 300-500 μm particles and three coils). In all rabbits CT-volumetry, ICG clearance test, blood sampling, and portography were performed prior to PVE and at d 7 and 14. Additional blood sampling and CT volumetry was done on d 3 and 7. RESULTS PVE was technically feasible in the rabbit. CT-volumetry demonstrated a strong correlation with actual liver weight and volume measured at sacrifice. The hypertrophy response was highest at d 7 in both protocols, which was consistent with the amount of proliferating hepatocytes. Protocol 2 showed less revascularization of the portal venous system and demonstrated the highest hypertrophy response. Comparable to the clinical situation, only a small, transient increase in transaminases was observed. There were no changes in liver function parameters after PVE. Histopathologic findings in the rabbit livers were comparable to those found in human livers. CONCLUSIONS We successfully devised a rabbit model for PVE, which resembles the human clinical situation.