Breath hydrogen (H2) analysis was used to study lactose malabsorption in Southern Chinese children and infants. End-expired air was collected in 85 children using a modified anaesthesia bag system; while in infants, a novel automated end-expired sampling device was constructed and tested on 45 term and 27 preterm infants. Hydrogen and other respiratory gases were measured in the expired air using standard gas chromatograph equipped with a thermal conductivity detector. The system was found to have a detection limit of 0.5 ppm for H2. Both sampling methods were found to be reproducible, with intra-individual coefficient variations of less than 10%. Using 5% carbon dioxide as the expected alveolar concentration, the samples obtained by the bag system represented 85% of the end-expired air, while those obtained by the automated machine corresponded to 75%-100% end-tidal air. Taking 20 ppm rise in breath H2 as a cutoff criterion, the incidence of lactose malabsorption in the children was 78%; while in term and preterm infants this was 17.8% and 63% respectively.