Defects in hepatic insulin action in type 2 diabetes and its possible underlying mechanisms were assessed in euglycemic-hyperinsulinemic clamp studies, using improved tracer methods (constant specific activity technique). Ten obese diabetic patients (age 54 years, BMI 29 ± 0.5 kg/m²) and ten matched control subjects were studied at baseline (after an overnight fast) and during insulin infusions of 20- and 40-mU · m^–2 · min^–1. In the diabetic patients, plasma glucose levels were normalized overnight before the studies by low-dose insulin infusion. Hepatic sinusoidal insulin levels were estimated, and plasma levels of free fatty acids (FFAs) and glucagon were determined to assess the direct and indirect effects of insulin on hepatic glucose production (HGP) in type 2 diabetes. Baseline rates of HGP (86 ± 3 vs. 76 ± 3 mg · m^–2 · min^–1, P < 0.05) were slightly elevated in the diabetic patients compared with control subjects, despite much higher hepatic sinusoidal insulin levels (26 ± 3 vs. 12 ± 2 mU/l, P < 0.001). Consequently, a marked defect in the direct (hepatic) effect of insulin on HGP appeared to be present at low insulin levels. However, in response to a small increase in baseline hepatic sinusoidal insulin levels of 11 mU/l (26 ± 3 to 37 ± 3 mU/l, P < 0.05) in the 20-mU clamp, a marked suppression of HGP was observed in the diabetic patients (86 ± 3 to 32 ± 5 mg · m^–2 · min^–1, P < 0.001), despite only minimal changes in FFAs (0.33 ± 0.05 to 0.25 ± 0.05 mmol/l, NS) and glucagon (14 ± 1 to 11 ± 2 pmol/l, P < 0.05) levels, suggesting that the impairment in the direct effect of insulin can be overcome by a small increase in insulin levels. Compared with control subjects, suppression of HGP in the diabetic patients was slightly impaired in the 20-mU clamp (32 ± 5 vs. 22 ± 4 mg · m^–2 · min^–1, P < 0.05) but not in the 40-mU clamp (25 ± 2 vs. 21 ± 3 mg · m^–2 · min^–1, NS). In the 20-mU clamp, hepatic sinusoidal insulin levels in the diabetic patients were comparable with control subjects (37 ± 3 vs. 36 ± 3 mU/l, NS), whereas both FFA and glucagon levels were higher (i.e., less suppressed) and correlated with the rates of HGP (R = 0.71, P < 0.02; and R = 0.69, P < 0.05, respectively). Thus, at this insulin level impaired indirect (extrahepatic) effects of insulin seemed to prevail. In conclusion, hepatic insulin resistance is present in obese type 2 diabetic patients but is of quantitative significance only at low physiological insulin levels. Defects in both the direct and the indirect effects of insulin on HGP appear to contribute to this resistance.