This study examines the effect of moderate intake of red wine, tannic acid, or ethanol during a meal in type 2 diabetic patients and the influence of tannic acid on the digestibility of starch by alpha-amylase. Thirty non-insulin-dependent diabetes mellitus (NIDDM) patients aged 53 +/- 6 years were studied (in vivo study) 10 of whom received red wine (200 mL), 10 tannic acid (150 mg), and 10 ethanol (16 g) with their midday meal (600 calories, 65 g carbohydrate, 20 g lipid, and 34 g protein). All patients were tested on two occasions (water or placebo v wine, alcohol, or tannic acid). The influence of tannic acid (0.25, 0.5, and 1 mg) on the digestibility of starch (100 mg) by alpha-amylase (100 U) was tested in vitro by sequential incubation at 37 degrees C (in vitro study). The maximum glucose excursion after lunch was 2.6 +/- 0.8 mmol/L at 90 minutes (T90) for water and 1.8 +/- 0.9 mmol/L at T90 for red wine taken with the meal. The values at T60 and T90 were significant (P < .01). Comparable results were obtained with tannic acid alone (nonalcoholic component of wine): the maximum glucose excursion after lunch was 2.76 +/- 0.9 mmol/L at T120 for placebo and 1.97 +/- 0.9 mmol/L at T90 for tannic acid (P < .01); no difference in glucose and insulin excursion was observed between water and ethanol. No interaction between tannic acid and starch was observed in the in vitro experiments, although after preincubation of alpha-amylase with tannic acid, digestion was slowed in a dose-dependent manner (6.1 +/- 1.1 minutes for 0.25 mg tannic acid and 13.1 +/- 1.59 minutes for 1 mg tannic acid). Drinking red wine with a meal did not increase blood glucose in NIDDM patients, and led to a slight decrease in some instances. The effect appeared to be mediated by the nonalcoholic compounds in wine such as tannic acid. Ethanol itself had no effect on plasma glucose or insulin levels.