The insulinotropic gut hormone glucagon-like peptide (GLP)-1 increases secretory burst mass and the amplitude of pulsatile insulin secretion in healthy volunteers without affecting burst frequency. Effects of GLP-1 on secretory mechanisms in type 2 diabetic patients and subjects with impaired glucose tolerance (IGT) known to have impaired pulsatile release of insulin have not yet been studied. Eight type 2 diabetic patients (64+/-9 years, BMI 28.9+/-7.2 kg/m2, HbA1c 7.7+/-1.3%) and eight subjects with IGT (63+/-10 years, BMI 31.7+/-6.4 kg/m2, HbA1c 5.7+/-0.4) were studied on separate occasions in the fasting state during the continued administration of exogenous GLP-1 (1.2 pmol x kg(-1) x min(-1), started at 10:00 P.M. the evening before) or placebo. For comparison, eight healthy volunteers (62+/-7 years, BMI 27.7+/-4.8 kg/m2, HbA1c 5.4+/-0.5) were studied only with placebo. Blood was sampled continuously over 60 min (roller-pump) in 1-min fractions for the measurement of plasma glucose and insulin. Pulsatile insulin secretion was characterized by deconvolution, autocorrelation, and spectral analysis and by estimating the degree of randomness (approximate entropy). In type 2 diabetic patients, exogenous GLP-1 at approximately 90 pmol/l improved plasma glucose concentrations (6.4+/-2.1 mmol/l vs. placebo 9.8+/-4.1 mmol/l, P = 0.0005) and significantly increased mean insulin burst mass (by 68%, P = 0.007) and amplitude (by 59%, P = 0.006; deconvolution analysis). In IGT subjects, burst mass was increased by 45% (P = 0.019) and amplitude by 38% (P = 0.02). By deconvolution analysis, insulin secretory burst frequency was not affected by GLP-1 in either type 2 diabetic patients (P = 0.15) or IGT subjects (P = 0.76). However, by both autocorrelation and spectral analysis, GLP-1 prolonged the period (lag time) between subsequent maxima of insulin concentrations significantly from approximately 9 to approximately 13 min in both type 2 diabetic patients and IGT subjects. Under placebo conditions, parameters of pulsatile insulin secretion were similar in normal subjects, type 2 diabetic patients, and IGT subjects based on all methodological approaches (P > 0.05). In conclusion, intravenous GLP-1 reduces plasma glucose in type 2 diabetic patients and improves the oscillatory secretion pattern by amplifying insulin secretory burst mass, whereas the oscillatory period determined by autocorrelation and spectral analysis is significantly prolonged. This was not the case for the interpulse interval determined by deconvolution. Together, these results suggest a normalization of the pulsatile pattern of insulin secretion by GLP-1, which supports the future therapeutic use of GLP-1-derived agents.