The ability of portal vein insulin to control hepatic glucose production (HGP) is debated. The aim of the present study was to determine, therefore, if the liver can respond to a selective decrease in portal vein insulin. Isotopic ([3H]glucose) and arteriovenous difference methods were used to measure HGP in conscious overnight fasted dogs. A pancreatic clamp (somatostatin plus basal portal insulin and glucagon) was used to control the endocrine pancreas. A 40-min control period was followed by a 180-min test period. During the latter, the portal vein insulin level was selectively decreased while the arterial insulin level was not changed. This was accomplished by stopping the portal insulin infusion and giving insulin peripherally at half the basal portal rate (PID, n=5). In a control group (n=5), the portal insulin infusion was not changed and glucose was infused to match the hyperglycemia that occurred in the PID group. A selective decrease of 120 pmol/l in portal vein insulin was achieved (basal, 150+/-36 to last 30 min, 30+/-12 pmol/l) in the absence of a change in the arterial insulin level (basal, 30+/-3 to last 30 min, 36+/-4 pmol/l). Neither arterial nor portal insulin levels changed in the control group (30+/-6 and 126+/-30 pmol/l, respectively). In response to the selective decrease in portal vein insulin, net hepatic glucose output (NHGO) increased significantly, from 8+/-1 (basal) to 30+/-6 and 14+/-2 micromol x kg(-1) x min(-1) by 15 min and the last 30 min (P < 0.05) of the experimental period, respectively. Arterial plasma glucose increased from 5.9+/-0.2 (basal) to 10.5+/-0.4 micromol/l (last 30 min). Three-carbon gluconeogenic precursor uptake fell from 11.2+/-2.9 (basal) to 5.9+/-0.7 micromol x kg(-1) x min(-1) (last 30 min), and thus a change in gluconeogenesis could not account for any of the increase in NHGO. With matched hyperglycemia (basal, 5.5+/-0.3 to last 30 min, 10.5+/-0.8 micromol/l) but no change in insulin, NHGO decreased from 12+/-1 (basal) to 0 (-1+/-6 micromol x kg(-1) x min(-1), last 30 min, P < 0.05) and hepatic gluconeogenic precursor uptake did not change (basal, 8.0+/-1.7 to last 30 min, 8.9+/-2.2 micromol x kg[-1] x min[-1]). Thus, the liver responds rapidly to a selective decrease in portal vein insulin by markedly increasing HGP as a result of increased glycogenolysis. These studies indicate that after an overnight fast, basal HGP (glycogenolysis) is highly sensitive to the hepatic sinusoidal insulin level.