AIMS/HYPOTHESIS: We compared four methods to assess their accuracy in measuring insulin secretion during an intravenous glucose tolerance test in patients with Type II (non-insulin-dependent) diabetes mellitus and with varying beta-cell function and matched control subjects. METHODS: Eight control subjects and eight Type II diabetic patients underwent an intravenous glucose tolerance test with tolbutamide and an intravenous bolus injection of C-peptide to assess C-peptide kinetics. Insulin secretion rates were determined by the Eaton deconvolution (reference method), the Insulin SECretion method (ISEC) based on population kinetic parameters as well as one-compartment and two-compartment versions of the combined model of insulin and C-peptide kinetics. To allow a comparison of the accuracy of the four methods, fasting rates and amounts of insulin secreted during the first phase (0-10 min) and the second phase (10-180 min) were calculated. RESULTS: All secretion responses from the ISEC method were strongly correlated to those obtained by the Eaton deconvolution method (r = 0.83-0.92). The one-compartment combined model, however, showed a high correlation to the reference method only for the first-phase insulin response (r = 0.78). The two-compartment combined model failed to provide reliable estimates of insulin secretion in three of the control subjects and in two patients with Type II diabetes. The four methods were accurate with respect to mean basal and first-phase secretion response. The one-compartment and two-compartment combined models were less accurate in measuring the second-phase response. CONCLUSION/INTERPRETATION: The ISEC method can be applied to normal, obese or Type II diabetic patients. In patients with deviating kinetics of C-peptide the Eaton deconvolution method is the method of choice while the one-compartment combined model is suitable for measuring only the first-phase insulin secretion.