The contribution of gluconeogenesis (GNG) to endogenous glucose output (EGO) in type 2 diabetes is controversial. Little information is available on the separate influence of obesity on GNG. We measured percent GNG (by the 2H2O technique) and EGO (by 6,6-[2H]glucose) in 37 type 2 diabetic subjects (9 lean and 28 obese, mean fasting plasma glucose [FPG] 8.3 +/- 0.3 mmol/l) and 18 control subjects (6 lean and 12 obese) after a 15-h fast. Percent GNG averaged 47 +/- 5% in lean control subjects and was significantly increased in association with both obesity (P < 0.01) and diabetes (P = 0.004). By multivariate analysis, percent GNG was independently associated with BMI (partial r = 0.27, P < 0.05, with a predicted increase of 0.9% per BMI unit) and FPG (partial r = 0.44, P = 0.0009, with a predicted increase of 2.7% per mmol/l of FPG). In contrast, EGO was increased in both lean and obese diabetic subjects (15.6 +/- 0.5 micromol x min(-1) x kg(-1) of fat-free mass, n = 37, P = 0.002) but not in obese nondiabetic control subjects (13.1 0.7, NS) as compared with lean control subjects (12.4 +/- 1.4). Consequently, gluconeogenic flux (percent GNG x EGO) was increased in obesity (P = 0.01) and markedly elevated in diabetic subjects (P = 0.0004), whereas glycogenolytic flux was reduced only in association with obesity (P = 0.05). Fasting plasma glucagon levels were significantly increased in diabetic subjects (P < 0.05) and positively related to EGO, whereas plasma insulin was higher in obese control subjects than lean control subjects (P = 0.05) and unrelated to measured glucose fluxes. We conclude that the percent contribution of GNG to glucose release after a 15-h fast is independently and quantitatively related to the degree of overweight and the severity of fasting hyperglycemia. In obese individuals, reduced glycogenolysis ensures a normal rate of glucose output. In diabetic individuals, hyperglucagonemia contributes to inappropriately elevated rates of glucose output from both GNG and glycogenolysis.