OBJECTIVE To determine the efficacy of the alpha-glucosidase inhibitor miglitol (BAYm 1099) regarding the starch content of food. METHODS Thirty-six non-insulin-dependent diabetes mellitus (NIDDM) subjects were studied in a double-blind randomized study comparing treatment with a single dosage of 100 mg miglitol or placebo and a single-blind crossover comparison of three test meals in which the carbohydrate contained either 30, 50, or 70% starch, and quantities of fat and protein were kept constant. RESULTS Postprandial blood glucose excursions were reduced by approximately 50% with miglitol after all test meals. In contrast, after miglitol treatment, maximum postprandial serum C-peptide and insulin values reached the same levels as after placebo treatment, although the time to reach these maximum levels was delayed. Free fatty acid values decreased after both miglitol and placebo similarly. Twenty-eight untoward events in 15 patients were reported in the miglitol treatment group and 11 events in 7 patients in the placebo treatment group. CONCLUSIONS Miglitol reduces postprandial blood glucose excursions independent of the starch content of the meal. Because no effects were found on incremental postprandial maximal levels of serum insulin and C-peptide, it may be that miglitol exerts, in addition to a delay of intestinal carbohydrate absorption, extraintestinal effects as well, particularly effects on disposition of glucose or anti-insulin counterregulatory factors.