The clinical use of the H2 breath test is limited by the finding that a variable fraction of the population fails to excrete appreciable H2 during colonic carbohydrate fermentation. Therefore, we assessed the ability to increase breath H2 excretion in 371 patients (224 female, 147 male) by administering the nonabsorbable sugar lactulose. Following 12g of lactulose, 27% of 94 patients did not increase their breath H2 concentration over 20 ppm and were considered low H2 excretors. Ingestion of 20 g of lactulose in 277 patients yielded a frequency of low H2 excretors of 14%. Six of 10 patients that were low H2 excretors after 12g of lactulose increased their breath H2 levels over 20 ppm when tested with 20g. In 35 patients tested with the same amount of lactulose on two separate occasions, the subject frequently altered his or her H2 producing status over a period of a few weeks. Low H2 excretors had a significantly higher breath CH4 concentration, both fasting (22 +/- 34 ppm) and after lactulose (51 +/- 58 ppm) compared to the remaining patients (5 +/- 13 ppm and 16 +/- 40 ppm, respectively). While the mean age of low excretors (54 +/- 17 years) was significantly higher than the others (44 +/- 17 years), no difference was found for sex prevalence and stool pH. This study demonstrates that respiratory H2 excretion following lactulose ingestion is not consistent and suggests that the application of too restrictive criteria could lead to improper interpretation of the H2 breath test.