Diphenylhydantoin (DPH) is known to inhibit insulin secretion in vitro and can cause hyperglycemia in man. To determine if there were selective effects of this drug in diabetes, insulin responses to intravenous stimuli were measured before and after ingesting the drug for several days in seven nonobese subjects with mild oral glucose intolerance. These were compared with responses of eight nonobese control subjects with normal oral glucose tolerance. Before DPH, early insulin release after intravenous glucose was similar in both groups, though glucose clearance was subnormal in the glucose-intolerant group. Arginine, given thirty minutes after glucose (postglucose arginine) was followed by prominent insulin release in both groups. After DPH, there was no significant alteration in glucoseinduced earlyinsulin release or clearance of glucose, though downward trends were noted in both groups. However, after DPH, the glucose-intolerant group showed significant reduction of insulin responses to postglucose arginine. These differences in drug effect occurred despite similar serum drug levels and per cent binding of DPH in the normal and in the glucose-intolerant group. Drug levels attained were in the effective range seen in patients taking DPH for epilepsy. In the isolated, perfused rat pancreas, effects of DPH in the presence of a high arginine stimulus were differentiated from those of a glucose stimulus. Whereas DPH eliminated both early and late arginine-induced insulin secretion, only late glucose-induced secretion was abolished by the drug. These data suggest that DPH will unmask early defects of insulin secretion in patients with mild oral glucose intolerance. The major early effect of the drug on insulin secretion in these patients is thus seen after a nonglycemic stimulus, postglucose arginine, which may have an insulinotropic mechanism distinctive from that of glucose.

A special glucose infusion test was used to provide successive steplike increments of glucose stimulation in six normal-weight and seven obese subjects. This "staircase" method of glucose infusion demonstrates that insulin responds in "spike" fashion despite maintenance of a continuous, fixed, submaximal glucose stimulation. Further, the "spike" response recurs as the glucose concentration is stepped up.Obese subjects showed an exaggeration of both the first phase "spike" pattern and the more gradual second phase of insulin release. Mathematical analysis of early phase insulin release indicates that obese persons have greater total amounts of insulin available for release at all glucose concentrations than do normals, yet the proportion of totally available insulin released to a given glucose stimulus is not increased in obesity.These findings imply that the early phase, hyperinsulin response to glucose in obesity is due to a greater quantity of insulin available for release rather than to an increased sensitivity of the beta cell to glucose. DIABETES 23:763-70, September, 1974.In obese patients abnormally high levels of circulating insulin have been found under basal conditions and after pancreatic stimulation with practically all known insulin stimulators, including glucose, 1 tolbutamide, 2 glucagon, 3 arginine, 4 and leucine. 5 Characterization of this hyperinsulinism suggests that it consists predominantly of biologically active insulin 6 of normal molecular size 78 whose elevated levels in obese patients represent hypersecretion rather than impaired clearance. 910 This hypersecretion of insulin in obese patients is a consequence of the obesity, since it becomes normal on reducing to normal weight 11 and can be acquired in normal subjects made obese after forced feeding. 12 The mechanism underlying this hypersecretion of insulin has not yet been determined. It has been attributed to some unknown signal for beta-cell hyperse-

No abstract