The long-term effect (7 wk) of the addition in the diet of 20 g of wheat bran was studied in five healthy subjects. Blood glucose and plasma insulin were assayed during test-meals taking place before (I), then 10 days (II), 24 days (III), and 45 days (IV) after daily ingestion of wheat bran. Bran addition leads to a time progressive reduction of meal-induced glucose response. This effect was the strongest during meal III, at 30 min (from 7.38 +/- 0.11 to 5.0 +/- 1.11 mmol/l; p less than 0.05) and at 60 min (from 5.94 +/- 0.61 to 3.88 +/- 0.78 mmol/l; p less than 0.05). Insulin response remained identical during the first three test meals. It increased then in a significant manner during meal IV, at 60 min (From 45.5 +/- 12.3 to 61.8 +/- 8.9 microU/ml; p less than 0.02), and at 120 min (from 41.2 +/- 16.3 to 61.0 +/- 16.0 microU/ml; p less than 0.05). The causes of the reduction of glucose response could be potential gastric emptying and intestinal transit increases. The exact mechanism of the improved long-term glucose tolerance seen in brain is not obvious nor is the reason for the enhanced insulin response seen later.