Whether carbohydrate malabsorption causes diarrhea probably depends on the balance between the osmotic force of the carbohydrate and the compensatory capacity of the colon to dispose of the carbohydrate by bacterial fermentation. The present study evaluated the specific role of the osmolarity by comparing the severity of diarrhea after ingestion of two nonabsorbable carbohydrates, the fructooligosaccharide Idolax and the disaccharide lactulose. Both carbohydrates are readily fermented by the colonic flora but differ in osmolarity, the osmotic force being twice as high for lactulose as for Idolax. Twelve subjects were given increasing doses (0, 20, 40, 80, 160 g/d) of Idolax and lactulose in a crossover design. Every dose level was administered for three days with intervals of one week. Stools were collected on the third day to determine 24-hr volume, concentrations of short-chain fatty acids, L- and D-lactate, residues of Idolax or lactulose, sodium, potassium, pH, osmolarity, and in vitro productions of organic acids. Measured by short-chain fatty acid and lactate formation in a fecal incubation system, the fermentation of Idolax and lactulose was identical and very rapid compared with a range of reference carbohydrates. A laxative effect of both Idolax and lactulose was demonstrated. The increment in fecal volume as a function of the dose administered was twice as high for lactulose (slope of the regression line = 7.3, r = 0.64, P< 10(-5)) as for Idolax (slope = 3.7, r = 0.51, P<10(-3)), i.e., isosmolar doses of lactulose and Idolax had the same effect on fecal volume. The variation in fecal volume was substantial (lactulose 80 g/day: 110-1360 g/day; Idolax 160 g/day: 130-1440 g/day). High responders had earlier and larger fecal excretions of the saccharide compared with low-responders. Fecal volume in carbohydrate-induced diarrhea is proportional to the osmotic force of the malabsorbed saccharide, even though all or the majority of the saccharide is degraded by colonic bacteria. The capacity to modify the diarrhea varies considerably from person to person and is associated with colonic saccharide disposal, whereas the variation in response to isosmolar amounts of different saccharides is small within the same individual.