The effects of a chronic load of nonabsorbable sugars on intracolonic bacterial metabolism of carbohydrates and on H2 breath excretion are disputed. However, most of the discussion relies on indirect evidence or on results of in vitro studies. Thus, we attempted to assess directly and in vivo the effects on intracolonic metabolism of lactulose of a chronic oral load of this nonabsorbable disaccharide. 20 g of lactulose was given orally twice daily during 8 d to eight normal volunteers. In all, breath H2 concentration was measured on days 1 and 8 after ingestion of the morning lactulose dose. In four subjects, stools were collected during 2 d at the beginning and at the end of the lactulose maintenance period to measure fecal pH and daily outputs of carbohydrates and P-galactosidase. The four other subjects were intubated on days 1 and 8 to measure the pH and the concentrations of carbohydrates, lactic acid, and volatile fatty acids (VFA) in the distal ileum and cecal contents.Moreover, '4C-lactulose was added to cold lactulose and "CO2 breath outputs determined. Pulmonary H2 excretion fell from day 1 to day 8 (P < 0.05), whereas '4CO2 excretion increased (P < 0.01). Fecal water pH, lactic acid, and VFA concentrations did not vary between the two stool collection periods. 24-h fecal weight, fecal water, and carbohydrate outputs showed a trend to decrease between days 1 and 2 and days 7-8, whereas jlgalactosidase activity rose markedly (P < 0.01). No significant variations were observed for all parameters measured in ileal fluid. In the cecum, areas under the concentration curves decreased from day 1 to day 8 for lactulose, galactose, and fructose (P < 0.01), while an increase was found for lactic acid (P < 0.001), acetic acid (P < 0.0001), and total VFA (P < 0.001). Cecal fluid pH dropped faster (P < 0.05) and to a lower level (P < 0.05) on day 8 than on day 1. These data clearly show that a chronic load of a nonabsorbable sugar induces changes in colonic bacterial metabolic pathways resulting in a better efficiency of the flora to digest the carbohydrate. Introduction Measurement of pulmonary hydrogen (H2) excretion is widely used to diagnose and quantify carbohydrate malabsorption (1-6). However the accuracy of the H2 breath test in the study of chronic carbohydrate malabsorption is debated (1). Perman et al. (7) showed that an 8-d oral administration of lactulose resulted in acidic stools and in a dramatic fall of H2 breath excretion after an oral load of this sugar or of lactose in lactase-deficient patients. From stool homogenate incubations with various sugars, these authors concluded that the fall of H2 production was due to an inhibition of bacterial carbohydrate metabolism by the low colonic pH (7). Unfortunately, they did not measure stool sugar outputs. If their statement was always true, lactose-H2 breath test could be erroneously normal in lactase-deficient patients ingesting every day large amounts of incompletely absorbed carbohydrates, including lactose itself. This is contradictory to the...