Breath hydrogen (H2) exhalation after xylose administration reflects the malabsorbed portion of the pentose and thus might facilitate the application of the D-xylose test. Therefore, as a complementary parameter, breath H2-exhalation in response to 25 g D-xylose was assessed in control subjects, in patients with coeliac disease, with chronic pancreatitis and with the irritable bowel syndrome. Patients with coeliac disease showed significantly higher breath H2 concentrations than the controls. Specificity and the positive predictive value of peak H2-increments greater than 56 ppm (i.e. greater than mean + 2 SD of controls) were 100%, but sensitivity was only 40%. In all patients with a positive H2 breath test, urinary D-xylose excretion and serum D-xylose increments were also abnormal. Apart from great overlap between controls and patients with coeliac disease, the failure to produce H2 in response to D-xylose in 12% of the 57 investigated subjects was the major factor limiting diagnostic efficiency of the test. Non H2 production could be shown to reflect a specific metabolic disability of the colonic flora and did not prove complete absorption of the substrate. It is concluded, that the 25 g D-xylose H2 breath test is of no clinical relevance for the diagnosis of celiac sprue but exaggerated breath H2 increases (greater than 56 ppm) with normal urinary and D-xylose tests were indicative for the irritable bowel syndrome in 5 out of 10 patients. The diagnostic impact of this constellation thus merits further investigation.