The present study examines the effect of the route of insulin delivery on glucose turnover in humans. By using a new noninvasive in vivo method, the acute effect of insulin secreted by the pancreas can be compared with that of insulin delivered by a peripheral vein. Three euglycemic-hyperinsulinemic studies were performed in lean healthy men. In the first study (n = 10), constant portal hyperinsulinemia was produced using a programmed intravenous tolbutamide infusion algorithm, and the insulin secretion rate was mathematically derived by deconvolution from peripheral plasma C-peptide levels. In the second study (n = 10), exogenous insulin was infused by peripheral vein at the same rate as that determined in the first study. In the third study (n = 7), the peripheral insulin levels in the first study were matched by infusing exogenous insulin into a peripheral vein at half that rate. Peripheral insulin levels were higher (P < 0.001) with the full-rate peripheral insulin infusion (266.3 ± 28.1 pmol/l) than with the portal delivery of insulin (171.1 ± 30.4 pmol/l) or the half-rate peripheral insulin infusion (158.6 ± 7.4 pmol/l) (portal versus half-rate peripheral insulin infusion, NS). Calculated hepatic insulin levels were higher (P < 0.001) in the portal insulin study (443.1 ± 52.6 pmol/l) than in the full-rate peripheral insulin study (303.6 ± 30.9 pmol/l) or in the half-rate peripheral insulin study (204.5 ± 9.8 pmol/l). Hepatic glucose production (HGP) was suppressed to a greater extent with the full-rate peripheral insulin infusion (69.3 ± 7.8%, P < 0.001 vs. portal or half-rate peripheral insulin) than portal (50.3 ± 9.8%) or half-rate peripheral insulin infusion (36.8 ± 3.8%). In the portal insulin study, however, suppression was > in the half-rate peripheral insulin study (P < 0.01), in spite of equal peripheral insulin levels. The assumption that tolbutamide, when used in this fashion, has no independent effect on glucose turnover, glucagon, or gluconeogenic precursor and energy substrates for gluconeogenesis was validated in five C-peptide-negative patients with IDDM. We conclude that in nondiabetic humans, 1) peripheral effects of insulin are important in suppressing HGP, as evidenced by the greater suppression of HGP with equivalent rate peripheral versus portal insulin delivery, and 2) because HGP was suppressed to a greater extent with portal verus peripheral insulin delivery at half the rate when peripheral insulin levels were matched, insulin-induced suppression of HGP is also partly mediated by a direct hepatic effect.

This content is only available via PDF.