Isolated hepatocyte studies demonstrated that leucine can be a precursor of ketone bodies. In this study we examine the relative contribution of leucine to hepatic ketogenesis in vivo. Three groups of conscious dogs with long-term indwelling catheters in the femoral artery, hepatic vein, and portal vein were studied. Group I (n = 3) animals were fasted overnight for 24 hours, and those in groups II and III (n = 4, each) were fasted for 62 to 68 hours (designated 3-day fast). Groups I and III received intravenous saline solution (0.9%) and served as controls. In group II selective acute insulin deficiency (SAID) was induced by a peripheral intravenous somatostatin (SRIF) infusion and intraportal glucagon (0.55 ng/body weight/min). Net hepatic production (NHP) of ketone bodies (kb) and leucine (leu) was measured by the arteriovenous difference technique. Hepatic conversion of leucine to ketone bodies was measured by continuous infusion of L-U-[14C]-leucine and by determination of the appearance of [14C]-ketone bodies across the liver. In the group fasted overnight NHPleu was 0.02 +/- 0.01 mumol/kg/min, a value not different from zero. NHPkb was 3.1 +/- 0.1 mumol/kg/min and hepatic conversion of leucine to ketone bodies accounted for 3.5% of NHPkb. Insulin deficiency after 3 day's fasting resulted in a near 70% increase in NHPleu (from basal values of 0.31 +/- 0.1 mumol/kg/min to 0.52 +/- 0.06 mumol/kg/min during SAID, p less than 0.01). NHPkb increased from 11.0 +/- 1.0 to 15.5 mumol/kg/min (p less than 0.05). The rate of leucine conversion to ketone bodies (L-C) increased from 1.1 +/- 0.25 to 2.4 +/- 0.3 mumol/kg/min (p less than 0.01) with SAID. We conclude that as the dog progresses to fasting, the contribution of leucine carbon to hepatic production of ketone bodies increases from 3.5% to 10% (p less than 0.01), and this value increases to 15% (p less than 0.01 versus groups I and II) after SAID. Furthermore, the amount of leucine carbon taken up by the liver was not sufficient to account for all [14C]-labeled leucine to ketone bodies. The data suggest that the leucine carbon converted to ketone bodies must have been derived from intrahepatic protein sources of possibly from the keto acids of leucine, which are derived by the breakdown of leucine at distant sites, such as skeletal muscle or adipose tissue.