Short-term assays in animals and humans were carried out to determine
the ''in vivo'' starch availability from different Brazilian foods. Th
e digestion of these starch samples was also followed by an ''in vitro
'' hydrolysis technique. In animals the glycemic response of polenta,
sandwich loaf and white rice was significantly different from beans (C
arioca and Fradinho) and wheat bran (p<0.05). For extruded corn, white
bread, curau, potato, corn flour, beiju macaroni, canjica, raw manioc
flour and manioc the glycemic response was similar. In humans, beans
with and without hull produced a significantly lower glycemic response
compared to polenta, rice and rice with beans. Plasma insulin levels
were significantly lower in individuals fed all diets containing beans
compared to others test diets. The secretion of glucagon like peptide
-1 7-36 (GLP-1 7-36) was the same for polenta or bean diets. The hull
of the bean had no effect on the glucose and insulin responses The pos
itive correlation between ''in vitro'' digestibility and glycemic resp
onse in animals (y = 721 + 3.7x; r = 0.824, p<0.01) and humans (y = 78
7 + 3.9x; r = 0.935, p<0.01) allowed the use of the ''in vitro'' techn
ique for screening starchy foods.