Whether glucagon-like peptide (GLP)-1 requires the hepatic portal vein to elicit its insulin secretion-independent effects on glucose disposal in vivo was assessed in conscious dogs using tracer and arteriovenous difference techniques. In study 1, six conscious overnight-fasted dogs underwent oral glucose tolerance testing (OGTT) to determine target GLP-1 concentrations during clamp studies. Peak arterial and portal values during OGTT ranged from 23 to 65 pM and from 46 to 113 pM, respectively. In study 2, we conducted hyperinsulinemic-hyperglycemic clamp experiments consisting of three periods (P1, P2, and P3) during which somatostatin, glucagon, insulin and glucose were infused. The control group received saline, the PePe group received GLP-1 (1 pmol.kg(-1).min(-1)) peripherally, the PePo group received GLP-1 (1 pmol.kg(-1).min(-1)) peripherally (P2) and then intraportally (P3), and the PeHa group received GLP-1 (1 pmol.kg(-1).min(-1)) peripherally (P2) and then through the hepatic artery (P3) to increase the hepatic GLP-1 load to the same extent as in P3 in the PePo group (n = 8 dogs/group). Arterial GLP-1 levels increased similarly in all groups during P2 ( approximately 50 pM), whereas portal GLP-1 levels were significantly increased (2-fold) in the PePo vs. PePe and PeHa groups during P3. During P2, net hepatic glucose uptake (NHGU) increased slightly but not significantly (vs. P1) in all groups. During P3, GLP-1 increased NHGU in the PePo and PeHa groups more than in the control and PePe groups (change of 10.8 +/- 1.3 and 10.6 +/- 1.0 vs. 5.7 +/- 1.0 and 5.4 +/- 0.8 micromol.kg(-1).min(-1), respectively, P < 0.05). In conclusion, physiological GLP-1 levels increase glucose disposal in the liver, and this effect does not involve GLP-1 receptors located in the portal vein.