We have recently found that in nondiabetic dogs and humans, suppression of glucose production (GP) is mediated by both peripheral and hepatic effects of insulin. We have also found that both nonesterified fatty acids (NEFA) and glucagon are important determinants of the peripheral effect of insulin on GP. However, in moderately hyperglycemic depancreatized dogs, suppression of GP appeared to be mediated by peripheral but not hepatic insulin. In this latter study, insulin concentrations were in the high postprandial range (approximately 300 pmol/L) and suppression of GP may have been close to maximum. The aim of the present study was to determine whether GP can be regulated by hepatic insulin in depancreatized dogs at low insulin concentrations in the postabsorptive range. Depancreatized dogs were maintained at moderately hyperglycemic levels (approximately 10 mmol/L) by subbasal insulin infusions. In paired experiments, additional low-dose equimolar insulin infusions (0.75 pmol/kg x min) were administered peripherally (PER, n = 6) or portally (POR, n = 6) during glucose clamps. This resulted in a minimal increase in peripheral insulin levels, which was greater in PER versus POR, 29.0 +/- 3.7 versus 11.7 +/- 2.2 pmol/L. Also, we infused insulin peripherally at half this rate (1/2 PER, n = 6) to match the increase in peripheral insulin levels in POR (1/2 PER, 14.6 +/- 2.2) and thus obtain a selective POR versus 1/2 PER difference in hepatic sinusoidal insulin levels. PER suppressed GP more than POR (45.4% +/- 4.0% v 35.3% +/- 6.8%, P < .001), whereas POR did not suppress GP more than 1/2 PER (35.6% +/- 6.3%). Therefore, suppression of GP was proportional to peripheral rather than hepatic sinusoidal insulin levels, as in our previous study at higher insulin concentrations. In conclusion, during glucose clamps in moderately hyperglycemic depancreatized dogs, (1) suppression of GP was dominated by insulin's peripheral effects not only at postprandial but also postabsorptive insulin levels, and (2) we found no evidence for a hepatic effect of insulin in suppressing GP. We hypothesize that this effect is reduced in the depancreatized dog model of diabetes due to hepatic insulin resistance and/or hyperglycemia.