Aims/hypothesis. This study sought first to compare the pharmacodynamics and pharmocokinetics of two rapid-onset, rapidly-reversible insulinotropic agents, nateglinide and repaglinide, in pre-diabetic Cynomolgus monkeys and second to use these agents to assess the metabolic effects of early insulin secretion on prandial glucose control. Methods. First, equipotent doses of nateglinide (20 mg/kg) and repaglinide (0.1 mg/kg) or vehicle were given intragastrically to overnight-fasted ketamine-anesthetized pre-diabetic Cynomolgus monkeys and samples were obtained for measurement of plasma glucose, insulin, glucagon, NEFA and drug concentrations. Second, nateglinide, repaglinide or vehicle were administered 10 min before a glucosesupplemented liquid meal and prandial glucose and insulin profiles were compared. Results. Although oral administration of nateglinide and repaglinide elicited similar maximum increments of plasma insulin (+403 and +448 pmol/l, respectively), the effects of nateglinide were more rapidly manifest and less prolonged. With nateglinide, insulin increased within 10 min and returned to baseline within 50 min. After repaglinide, the first increase occurred at 30 min and insulin concentrations remained increased for 3.5 h post-dose. When given 10 min before a meal, nateglinide increased early, but not total insulin release (AUC 0-210 =108 vs 150 nmol/l min for nateglinide and vehicle, respectively) and reduced prandial glucose excursions by 78%. Repaglinide increased total insulin release (AUC 0-210 =298 nmol/l min) and reduced glucose excursions by 53%. Conclusion/interpretation. Nateglinide is more rapidacting and rapidly-reversible than is repaglinide. By restoring a more physiologic insulin profile, nateglinide is more effective than repaglinide in controlling prandial glucose excursions with less hyperinsulinaemia. It has been shown convincingly that in Type 2 (noninsulin-dependent) diabetes mellitus [1,2] and even in precursors thereof [3] the kinetics of insulin secretion are disrupted despite the existence of absolute (if not relative) hyperinsulinaemia. The disrupted insulin secretory kinetics can be confirmed by either the absence of the acute insulin response to i.v. glucose [4] or a sluggish insulin response to oral glucose [5] or a meal [6]. The impaired early insulin response leads to a marked impairment of the suppression of endogenous glucose production normally seen after glucose administration or a meal [5,6] and to a lesser extent, a later decrease of peripheral glucose utilization [6].