To investigate the pathophysiology of warm ischemia (WI) of the liver, the changes in hemodynamics and energy metabolism were studied during and after 60-min complete WI induced by total hepatic vascular exclusion (HVE) in the canine model. Hepatic arterial blood flow after WI was maintained at 76% of the pre-ischemic level, while portal blood flow was only 27% of the pre-ischemic level associated with increased portal vein pressure, which was twice the pre-ischemic level, resulting in a decrease of total hepatic blood flow to 46% of the pre-ischemic level. Concentration of tissue lipid peroxide increased after WI. Arterial blood ketone body ratio (AKBR), which reflects the hepatic mitochondrial redox state, could not recover to the pre-ischemic level after termination of WI. However, when 100 mg/kg of allopurinol (xanthine oxidase inhibitor) was administered intravenously 10 min prior to initiating WI, AKBR was restored to the pre-ischemic level at 30 min after WI in spite of the fact that allopurinol administration to one group produced no remarkable changes in the hepatic hemodynamics compared with the group without allopurinol treatment. Concentration of adenine nucleotides was significantly higher for the treated group at the end of and after WI than for the group without allopurinol treatment and was maintained at a higher level even after WI. Lipid peroxide production was suppressed. Electron microscopic examination revealed that allopurinol treatment could not prevent mitochondrial swelling. It is suggested that WI causes injury primarily to the portal sinusoidal circulation, resulting in portal congestion concomitant with high portal pressure after the release of WI. Allopurinol could prevent the deterioration of mitochondrial ATP metabolism, and was able to inhibit lipid peroxide production, resulting in the rapid recovery of mitochondrial redox state in spite of the fact that it produced no amelioration of hepatic hemodynamics and morphological alterations.