Helicobacter pylori infection is mainly acquired in childhood, and studies on the epidemiology of this infection depend on the availability of a noninvasive diagnostic test for use in children. The aim of this study was to determine whether the carbon 13-labeled urea breath test (UBT) can be used in children by evaluating: (1) its sensitivity and specificity compared with either culture or both rapid urease test and histologic examination, (2) whether a test meal or a prolonged fast is required, (3) the usefulness after treatment for H. pylori. Eighty-eight children (mean age, 10.6 +/- 4.19 years) who were undergoing upper endoscopy were studied while fasting, not fasting, and after treatment. Children were given 50 mg of 13C-urea if they weighed less than 50 kg or 75 mg of 13C-urea if they weighed more than 50 kg with 50 mg of a glucose polymer solution in 7.5 ml of water. Breath samples were collected at baseline and at 15, 30, 45, and 60 minutes. In 63 fasting children the UBT was 100% sensitive and 97.6% specific at 30 minutes with a cutoff value of 3.5 delta 13CO2 per mil. Nonfasting tests in 23 children, performed between 1 and 2 hours after their usual meal, were 100% sensitive and 91.6% specific. In 13 children fed directly before the UBT, the sensitivity of the test was reduced to 50%. Thirty minutes was the optimal sampling time. There was a significant decrease in specificity when samples were obtained at 15 minutes, possibly caused by the interference of oral urease-producing organisms. The test was 100% sensitive and specific in 20 children after treatment for H. pylori infection. The UBT is a highly sensitive and specific test for the diagnosis of H. pylori infection in children. Neither a prolonged fast nor a test meal is required.