BACKGROUND Acarbose, a pseudo-tetrasaccharid, inhibits intestinal alpha-glucosidases, effects a reduction of postprandial hyperglycemia and is particularly used in the treatment of patients with type-2 diabetes mellitus. The aim of the study is to analyse by a continuous blood glucose measurement the acute effect of acarbose after a carbohydrate loading and during a 12 hours period. METHODS We examined 10 patients with type-2 diabetes mellitus (mean age 59.2 +/- 3.79, HbA1 9.2 +/- 0.26%) treated with sulfonylureas and/or insulin after a carbohydrate meal and 12 hours during daytime, to test whether the first application of acarbose influences the mean blood glucose or the blood glucose amplitudes. Four measurements were enrolled using a portable continuous blood glucose sensor (Glucosensor, unitec Ulm). A measurement after a carbohydrate loading (Fresubin, 500 kcal, 69 g carbohydrate) with 100 mg acarbose (Glucobay) was followed by a 12-hour measurement during daytime with 3 x 100 mg acarbose and standard diet. These measurements were repeated without acarbose. RESULTS After a carbohydrate loading, the mean blood glucose level (AUC 44,320 +/- 10,660 with acarbose vs. 61,390 +/- 12,590 without acarbose; mean + SD; p = 0.004) decreased by 28%. During daytime blood glucose levels were not significantly decreased (165.7 +/- 50.3 mg/dl vs 183.7 + 67.4 mg/dl; p = 0.1) although the postprandial blood glucose amplitudes after the 3 meals were reduced significantly (85.90 +/- 24.6 mg/dl vs 106.5 +/- 20.5 mg/dl; p = 0.02). CONCLUSIONS Continuous blood glucose monitoring indicated that acarbose diminished mean blood glucose levels after a carbohydrate loading in patients with type-2 diabetes mellitus, but not during 12 hours of standard diet, although blood glucose amplitudes decreased. Long-term improvements of metabolism by acarbose may therefore be related to the reduction of blood glucose amplitudes which is likely to reduce toxic effects of glucose on islet cell function.