Hydrogen is produced during fermentation in the large intestine and may be excreted in breath and flatus or further metabolized by the flora. However, there is little information about total H2 excretion from different substrates or the extent to which it is metabolized in the colon. We have therefore measured total H2 and methane excretion in 10 healthy subjects using a whole body calorimeter. Breath gases were measured simultaneously with total excretion in response to lactulose, pectin, and banana starch. Metabolic activities of the predominant H2 consuming anaerobes (methanogenic, sulfate reducing, and acetogenic bacteria) were measured in fecal samples. Total H2 excretion on a starch and fiber-free diet was 35 +/- 6.1 mL/24 h +/- SEM. H2 from 7.5 g, 15 g, and 22.5 g lactulose was 88.1 +/- 22.4 mL, 227.0 +/- 60.7 mL, and 321.8 +/- 79.2 mL. Four of the subjects also excreted CH4, which was 51.3 +/- 5.5 mL, 97.3 +/- 18.4 mL, and 157.5 +/- 36.3 mL for the respective lactulose doses. H2 excretion was less in methanogenic subjects (7.9 mL/g lactulose) than in nonmethanogenic (17.3 mL/g), but total H2 excreted as, hydrogen + methane, was 34.9 mL/g. H2 from pectin (20 g) was 14.1% +/- 3.2% and from starch (22.2 g) 38.6% +/- 9.2% of an equivalent lactulose dose. Sixty-five percent of total H2 and CH4 was expired in breath at total excretion rates up to 200 mL/24 h. Over this the proportion decreased to 25% with an overall average of 58%. Only subjects with CH4 excretion in vivo showed methanogenesis in feces, whereas nonmethanogenic subjects showed high sulfate-reducing activity in feces (58.7 +/- 5.6 nmol 35SO4 reduced.h-1.g-1 wet wt vs. 7.9 +/- 2.0 nmol.h-1.g-1 in methanogens). Acetogenesis rates were very low in both groups. It was concluded that H2 excretion varies with different substrates. The proportion of H2 that is exhaled in breath is higher than currently accepted and varies with total excretion rate. Substantial amounts of H2 are consumed by methanogenic and sulfate-reducing bacteria.