Living-related segmental liver transplantation is now accepted as a method of liver transplantatior in pediatric group. But if graft failure occurs, mortality is high. Hence the early prediction of graft failure may contribute to the improvement of the chance of survival. The ketone bodies (acetoacetate, B-hydroxybutyrate) are known to be produced in the hepatic mitochondria. KBR (ketone body ratio) is a ratio of acetoacetate/B-hydroxybutyrate, which ref1ects the NAD+/NADH, i.e. oxidoreduction state of NAD of hepatic mitochondria. Because the viability of grafted liver is ultimately depeadant on the oxidoreduction potential of hepatic mitochondria, KBR might reflect the viability of grafted liver. The purpose of this experiment is to study whether KBR is a useful indicator of initial grat failure after liver transplantation. The experiment consisted of ten autologous left lobe liver transplants in dogs. The arterial blood was taken at preoperative period, anhepatic stage, reperfusion-15 minutes, 1 hour, 3 hour, 6 hour anl daily during their survival, and then KBR was calculated by ketone body concentration. The result of KBR measurement was as follows; it was 1.16+-0.33 at preoperative period, and abruptly decreased to 0.46+- 0.21 at anhepatic stage, and recovered to various levels according to initial graft function after reperfusion. To evaluate the initial graft function, the dogs were divided into two groups. A group (n=3) had KBR below 0.7 at reperfusion 1 hour, and B group (n=7) above 1.0. In group A, all of them showed early graft failure and died within 12 hours postoperatively. KBR changes in group A were as follows; it was 1.16 at preoperative period, and decreased to 0.25 at anhepatic stage, and recorded 0.69, 0.54, 0.61, 0.29 at reperfusion-15 min, 1 hour, 3 hour, 6 hour respectively. It never recovered above 0.7 at any period after reperfusion. In group B. they showed good early-graft function with survival of 16 hours to 8 days and died of other causes, ln group B, preoperative KBR was 1.16, and decreased to 0.54 at anhepatic stage, but promptly recovered to 0.95 at reperfusion-15 minute, 1.32 at 1 hour. It slightly decreased to 0.83 at 1 hour but maintained above 1.0 after 6 hour. In conclusion, the decreased KBR at anhepatic stage should promptly recover after reperfusion for the graft viability, and the value below 0.7 of KBR after reperfusion-1 hour represents the early graft failure, and above 1.0 of KBR represents good functioning graft. The KBR is a useful indicator of early graft function in canine autologous segmental liver transplantation, and it is assumed that KBR is further applicable on canine and human allogenic segmental liver transplantation.