Vanadyl sulfate (VOSO(4)) is an oxidative form of vanadium that in vitro and in animal models of diabetes has been shown to reduce hyperglycemia and insulin resistance. Small clinical studies of 2- to 4-week duration in type 2 diabetes (T2DM) have led to inconsistent results. To define its efficacy and mechanism of action, 11 type 2 diabetic patients were treated with VOSO(4) at a higher dose (150 mg/day) and for a longer period of time (6 weeks) than in previous studies. Before and after treatment we measured insulin secretion during an oral glucose tolerance test, and endogenous glucose production (EGP) and whole body insulin-mediated glucose disposal using the euglycemic insulin clamp technique combined [3-(3)H]glucose infusion. Treatment significantly improved glycemic control: fasting plasma glucose (FPG) decreased from 194 +/- 16 to 155 +/- 15 mg/dL, hemoglobin A(1c) decreased from 8.1 +/- 0.4 to 7.6 +/- 0.4%, and fructosamine decreased from 348 +/- 26 to 293 +/- 12 micromol/L (all P < 0.01) without any change in body weight. Diabetics had an increased rate of EGP compared with nondiabetic controls (4.1 +/- 0.2 vs. 2.7 +/- 0.2 mg/kg lean body mass.min; P< 0.001), which was closely correlated with FPG (r = 0.56; P< 0.006). Vanadyl sulfate reduced EGP by about 20% (P< 0.01), and the decline in EGP was correlated with the reduction in FPG (r = 0.60; P< 0.05). Vanadyl sulfate also caused a modest increase in insulin-mediated glucose disposal (from 4.3 +/- 0.4 to 5.1 +/- 0.6 mg/kg lean body mass x min; P< 0.03), although the improvement in insulin sensitivity did not correlate with the decline in FPG after treatment (r = -0.16; P = NS). Vanadyl sulfate treatment lowered the plasma total cholesterol (223 +/- 14 vs. 202 +/- 16 mg/dL; P < 0.01) and low density lipoprotein cholesterol (141 +/- 14 vs. 129 +/- 14 mg/dL; P < 0.05), whereas 24-h ambulatory blood pressure was unaltered. We conclude that VOSO(4) at maximal tolerated doses for 6 weeks improves hepatic and muscle insulin sensitivity in T2DM. The glucose-lowering effect of VOSO(4) correlated well with the reduction in EGP, but not with insulin-mediated glucose disposal, suggesting that liver, rather than muscle, is the primary target of VOSO(4) action at therapeutic doses in T2DM.