Three end-expiratory breath hydrogen (H2) sampling methods were compared in a patient group (n = 12) and a laboratory staff group (n = 12) on two separate occasions. H2 samples obtained with each method showed significantly different concentrations (p less than 0.001) but no significant differences in coefficient of variation when individual triplicate samples were evaluated. There was a high correlation between the breath H2 concentrations obtained by the three methods (r = 0.93-0.96). Fasting breath H2 values after an overnight fast and an unrestricted diet the day before the investigation were compared with values obtained after an overnight fast and a low-fibre diet the day before the test in two patient groups (n = 39 and 39) with a comparable distribution of diagnoses and in one group of healthy volunteers (n = 17). Fasting breath H2 concentrations were significantly lower after a low-fibre diet in the patient groups (p less than 0.005) and in healthy volunteers (p less than 0.02). We conclude that each of the three end-expiratory sampling methods can be chosen for use in H2 breath tests depending on suitability and convenience and that a low-fibre diet the day before the H2 breath test lowers fasting breath H2 concentration.