Experimental diabetes and fasting are both associated with hypoinsulinaemia and share several other metabolic features. We investigated hepatic and peripheral glucose metabolism in young rats after near-total depletion of their fat mass. Conscious rats were fasted for 72 h (n = 13), while 6 h-fasted animals (n = 14) served as controls. Rats were studied either during saline infusion or insulin (18 m-units/kg per min)-clamp studies. In fasting, despite a 2-fold increase in hepatic glucose-6-phosphatase (Glc-6-Pase) Vmax. (from 16 +/- 2 mumol/g of liver per min in control; P < 0.001), the basal hepatic glucose production (HGP) decreased by 47% [from 88 +/- 3 mumol/kg lean body mass (LBM) per min in control; P < 0.01]. The decreased HGP in fasting was associated with a 70% decrease in the hepatic levels of glucose 6-phosphate (Glc-6-P) (from 366 +/- 53 nmol/g wet wt. in control; P < 0.01). Thus Glc-6-Pase activity assayed in the presence of the Glc-6-P levels found in vivo was decreased by 44%. During hyperinsulinaemia, peripheral glucose uptake was decreased by 15% with 3 days of fasting (from 272 +/- 17 mumol/kg LBM per min in control; P < 0.01). This was completely accounted for by a 42% decrease in whole-body glycolysis (P < 0.01), while the rate of glycogen synthesis was unchanged. Thus fasting (after near-total fat depletion) differs from experimental diabetes because: (1) despite markedly increased Glc-6-Pase, HGP is decreased in fasting, due to a marked decrease in the substrate level (Glc-6-P) in vivo; and (2) the impairment in peripheral insulin sensitivity in fasting is due to a decrease in the glycolytic, and not the glycogen-synthetic, pathway.