GLP-1 lowers blood glucose in fasting type 2 diabetic patients. To clarify the relation of the effect of GLP-1 to obesity, blood glucose, beta-cell function, and insulin sensitivity, GLP-1 (1.2 pmol/kg.min) was infused iv for 4-6 h into 50 fasting type 2 diabetic patients with a wide range of age, body mass index, HbA1c, and fasting plasma glucose. The effectiveness of GLP-1 was evaluated by calculation of a glucose disappearance constant for each individual (Kg, linear slope of log-transformed plasma glucose), and by the lowest stable glucose level (Nadir plasma glucose) obtained during the infusion. Grouped according to fasting plasma glucose (<10, 10-15, >15 mmol/liter), Kg values were 0.45 +/- 0.03, 0.38 +/- 0.04, and 0.28 +/- 0.04%/min (P = 0.005), and Nadir plasma glucose values were 4.7 +/- 0.1 (3.9-5.9), 5.8 +/- 0.4 (4.3-8.4), and 8.7 +/- 1.4 (6.2-18.7) mmol/liter (P = 0.0003). Nonresponders were not identified. Multiple regression analysis with Kg or Nadir plasma glucose as the dependent parameter and body mass index, age, gender, diabetes duration, and significantly correlated parameters (in multiple regression for Kg: fasting plasma glucose, fasting nonesterified fatty acid, dipeptidyl peptidase activity, peak insulin, and the logarithm of beta-cell function; and for Nadir plasma glucose: fasting plasma glucose, fasting nonesterified fatty acid, dipeptidyl peptidase activity, delta glucagon decrement, F-GLP-1 total, logarithm of beta-cell function, and Kg) as independent parameters resulted in fasting plasma glucose as the only significant predictor of Kg, and fasting plasma glucose and Kg as predictors of Nadir plasma glucose. Kg and Nadir plasma glucose were neither influenced by treatment nor by neuropathy per se. In conclusion, GLP-1 lowers plasma glucose in type 2 diabetes regardless of severity, but glucose elimination is faster and obtained glycemic level lower in patients with the lower fasting plasma glucose. Not all patients can be expected to reach normoglycemia.