Impaired hepatic glucose uptake (HGU) has been implicated in the development of hyperglycemia in type 2 diabetes; the relative impact of plasma glucose and insulin levels on this process remains controversial. We compared the effects of euglycemic hyperinsulinemia on HGU, skeletal muscle glucose uptake, and hepatic influx rate-constant (H-Ki) in 38 diet-treated diabetic patients and 22 nondiabetic controls, using positron emission tomography with (18)F-fluorodeoxyglucose and the insulin clamp technique. Control subjects were divided into two subgroups: one including older, heavier, insulin-resistant controls (whole-body glucose uptake, M = 21.4 +/- 5.4 micromol x min(-1) x kg(-1)) to match characteristics of diabetic patients (M = 20.4 +/- 9.9); the other including younger, leaner, insulin-sensitive controls (M = 48.2 +/- 9.9, P < 0.01). Skeletal muscle glucose uptake showed a similar group distribution as the M value. Insulin clearance rates were lower, whereas glycosylated hemoglobin and clamp plasma insulin levels were higher in diabetic patients than in controls. HGU and H-Ki were similar in the two nondiabetic subgroups and lower in diabetic patients than in controls (1.9 +/- 0.5 vs. 2.3 +/- 0.7 micromol x min(-1) x 100 ml(-1), and 0.37 +/- 0.09 vs. 0.44 +/- 0.14 ml x min(-1) x 100 ml(-1), P < or = 0.01). In the whole dataset, H-Ki was inversely related to fasting plasma glucose (correlation coefficient = -0.40, P = 0.0018). In diabetic subjects, H-Ki was reciprocally related to glycosylated hemoglobin (correlation coefficient = -0.36, P = 0.029). We conclude that insulin-mediated HGU is impaired, in type 2 diabetes, in some proportion to the degree of glycemic control.