SummaryThe effects of an a-glucosidase inhibitor, Acarbose (Bay g 5421) upon the circadian plasma glucose profile were studied in 7 diabetic patients in i hospital. All subjects were totally insulin-dependent as demonstrated by their low preand postprandial plasma C-peptide levels. The patients received Acarbose (5 days) or a placebo (4 days) in a randomized order.The mean plasma glucose levels, the M value and the MAGE index recorded during both periods were compared. Acarbose was given at doses of 200, 100 and 100 mg before breakfast, lunch and dinner, respectively. The doses of regular insulin (4/day) were daily adapted throughout the study in an attempt to reach the best possible glycemic control. Even though the insulin doses did not significantly differ during both periods, a significant dei crease of the mean plasma glucose levels was observed when the patients received Acarbose (8.2±0.4 vs 10.1±mmol/1; P<0.05). Moreover, the drug induced a significant decrease of postprandial plasma glucose measured at 10.30 am and 8.30 pm; the M value was also lower during Acarbose than during placebo (36±6 vs 45±6; P<0.02) while the MAGE index remained unaffected. Except for an increase in the number of hypoglycemic episodes, no side effects were recorded. Thus, Acarbose could represent an useful complementary drugt in the treatment of type 1 diabetes, essentially by lowering the postprandial rise of plasma glucose.