Severe hypoglycemia, the most serious side effect of sulfonylurea therapy, has been reported to occur more frequently with glyburide than glimepiride. The present studies were undertaken to test the hypothesis that a differential effect on glucagon secretion may be involved. We performed hyperinsulinemic hypoglycemic (approximately 2.5 mmol/L) clamps in 16 healthy volunteers who received in randomized order placebo, glyburide (10 mg), and glimepiride (4 mg) just before beginning the insulin infusion and measured plasma glucagon, insulin, C-peptide, glucagon, epinephrine, cortisol, and growth hormone levels during the clamp and during a 3-hour recovery period after discontinuation of the insulin infusion. Neither sulfonylurea altered glucagon responses or those of other counterregulatory hormones (except cortisol) during the clamp. However, glyburide delayed plasma glucose recovery from hypoglycemia (plasma glucose at end of recovery period: control, 4.9 +/- 0.2 mmol/L; glyburide, 3.7 +/- 0.2 mmol/L; P = .0001; glimepiride, 4.5 +/- 0.2 mmol/L; P = .08). Despite lower plasma glucose levels, glyburide stimulated insulin secretion during this period (0.89 +/- 0.13 vs 1.47 +/- 0.15 pmol x kg(-1) x min(-1), control vs glyburide; P = .001), whereas glimepiride did not (P = .08). Short-term administration of glyburide or glimepiride did not alter glucagon responses during hypoglycemia. In contrast, during recovery from hypoglycemia, glyburide but not glimepiride inappropriately stimulates insulin secretion at low plasma glucose levels. This differential effect on insulin secretion may be an important factor in explaining why glyburide causes severe hypoglycemia more frequently than glimepiride.