Background: Studies in younger patients with diabetes have shown that insulin profiles are more physiologic and postprandial glucose levels are lower with repaglinide than with glyburide. We conducted this study to determine if the differences in insulin/glucose profiles between repaglinide and glyburide were similar or different in the elderly. Methods: Thirteen elderly patients with diabetes were given glyburide or repaglinide in randomized crossover design, followed by a meal tolerance test. Samples were taken at regular intervals to measure glucose and insulin values. Results: The 0-30 min area under the curve (AUC) for insulin was higher (glyburide, 294 +/- 37 pM; repaglinide, 382 +/- 39 pM) (P < 0.01), and the 180-240 min AUC for insulin was lower (glyburide, 325 +/- 50 pM; repaglinide, 196 +/- 20 pM) (P < 0.01) with repaglinide. The 0-240 min AUC for glucose was not different between treatments (glyburide, 7.4 +/- 0.5 mM; repaglinide, 7.1 +/- 0.3 mM) (P not significant). Five subjects treated with glyburide and no subjects treated with repaglinide required glucose from 180 to 240 min (P < 0.05, Fisher's exact test). Conclusions: We conclude that repaglinide results in a more physiologic insulin profile and less frequent hypoglycemia than glyburide in the elderly.