Disaccharide lactulose is commonly used as a standard to quantitate the colonic fermentation of undigested sugars by means of H2 breath measurements. However, its high hydrogen production rate during fermentation may make it inappropriate for mimicking the fermentation of more complex carbohydrates, such as starch. Indigestible carbohydrates with a higher molecular weight might be more suitable than lactulose as a standard in H2 breath studies of starch digestibility. To test this hypothesis, we measured H2 breath in 8 healthy volunteers after a standard meal supplemented with 5 g or 10 g of lactulose or inulin, an indigestible oligosaccharide with an average degree of polymerization 4.5 times higher than that of lactulose. The results were then compared with those obtained after a standard meal containing a known amount (6.1 g) of resistant starch from high-amylose corn starch. Median H2 breath excretion per gram of reference carbohydrate was lower after the 5 g dose of inulin than after the 5 g dose of lactulose (19.1 vs 26.6 ppm x h x g-1; Wilcoxon's rank test p = 0.021) but similar after the two 10 g doses (inulin 22.4; lactulose 23.6; p = 0.234). Median H2 breath excretion per gram of resistant starch was significantly lower than that for both lactulose and inulin (p < 0.02), being 4.7 ppm x h x g-1. In vitro fermentation for 8 hrs with fecal homogenate showed similar mean hydrogen production rates for inulin and lactulose (30.5 vs 27.7 mL/mg fermented carbohydrate), and a significantly lower rate for starch (9.1 mL/mg) (n = 7; ANOVA p = 0.0007).(ABSTRACT TRUNCATED AT 250 WORDS)