The effect of glibenclamide treatment on insulin-mediated glucose disposal was studied in eight C-peptide-negative type I diabetic patients. The patients were studied twice by the euglycemic insulin clamp technique. One of the two experiments was preceded by glibenclamide treatment at the dose of 5 mg, three times daily for 15 days; half of the patients had the first test before and the second test after sulfonylurea treatment, and vice versa. Insulin was infused for four periods of 2 h each sequentially at 0.5, 1.0, 2.0, and 5.0 mU kg-1 min-1; for each insulin infusion period, the steady state plasma free insulin levels were comparable with or without glibenclamide. The mean +/- SEM plasma glucose concentration was 88 +/- 2 mg/dl in both experiments. The insulin-mediated glucose disposal rate was greater with glibenclamide during the first insulin infusion period (which generated plasma free insulin levels within the physiological range) 2.68 +/- 0.32 mg kg-1 min-1 with glibenclamide vs. 1.97 +/- 0.20 mg kg-1 min-1 without glibenclamide (P less than 0.005). However, glucose disposal rates did not differ in the diabetic patients with or without glibenclamide treatment during the second, third, and fourth insulin infusion periods, which generated plasma free insulin levels in the supraphysiological range. These results provide evidence for an extrapancreatic effect of glibenclamide at low insulin concentrations during euglycemic clamping in patients with insulin-dependent diabetes mellitus. However, this effect was not reflected clinically in either an increased rate of hypoglycemic reactions or decreased insulin needs during the short term period of treatment.

The effect of previous exposure to glucose on subsequent glucose-stimulated insulin and somatostatin secretion has been investigated using the isolated perfused rat pancreas. As expected, when the pancreases of non-diabetic rats were exposed to 16.7 mM glucose on two occasions, 20 min apart, insulin secretion during the second period of exposure to high glucose was greater than that during the first period. By contrast, there was no potentiation of somatostatin secretion during the second glucose stimulation with respect to that of the first. Indeed, when the basal glucose concentration was low (1.4 or 2.8 m m ) somatostatin secretion during the second glucose stimulation was lower than that during the first. Since exogenous insulin is known to inhibit glucoseinduced somatostatin secretion, it seemed possible that lack of visible potentiation of glucose-induced somatostatin secretion by glucose could have been due to partial D cell inhibition by simultaneously augmented insulin secretion during the second glucose stimulation. In an attempt to exclude such an interaction between B and D cells, somatostatin secretion was also studied in the pancreases of spontaneously diabetic, Wistar (BB) rats (these animals are insulin deficient and are maintained by daily injections of insulin). However, even though insulin secretion was not detectable from these pancreases, glucose potentiation of glucose-induced somatostatin secretion did not occur. Although the pancreatic B and D cells are known to respond in a similar manner to many secretagogues the present results show that glucose potentiation of glucosestimulated somatostatin secretion is not found under circumstances where potentiation of insulin secretion does occur. In addition, the absence of potentiated somatostatin secretion could not be attributed to partial inhibition of the D cell by insulin.