Background In patients with insulinoma, biochemical proof of inappropriately elevated insulin secretion during hypoglycemia is required prior to surgery. Because circulating insulin levels usually vary widely, we have used the combined OGTT-fasting test to define new normative criteria for a retrospective systematic analysis. Methods We retrospectively analyzed insulin concentrations from OGTT-fasting tests of 64 patients with surgically removed insulinomas. In addition, the response to intravenous somatostatin infusions was estimated. Normative criteria were defined to obtain comparable estimates of insulin concentrations: basal, glucose-stimulated maximum, postglucose plateau, and secretory bursts. Results Three types of insulin secretion patterns were identified: (1) the autonomous secretion pattern (type 1, N = 17) with basal and post-OGTT plateau insulin concentrations of approximately 50 mU/L, suppression after OGTT by 41%, virtual absence of distinctive secretory bursts, and resistance to somatostatin-mediated suppression (25 %); (2) the inadequate suppression pattern (type 2, N = 28) with moderately elevated basal and post-OGTT insulin concentrations of approximately 20 mU/L, suppression after OGTT by 73%, absence of secretory bursts, and incomplete somatostatin-induced suppression (56 %); (3) the late-burst secretion pattern (type 3, N = 19) with similar basal and post-OGTT insulin concentrations of 17 mU/L, suppression after OGTT by 76%, true insulin bursts of D 13 ± 11 mU/L (184%), and nearly complete somatostatininduced suppression by 64%. Conclusions By means of a new normative analysis of the combined OGTT-fasting test, three different patterns of insulin secretion can be described in patients with insulinoma: the autonomous secretion type, the inadequate suppression type, and the late-burst secretion type.