The uncertain accuracy of methods for detecting Helicobacter pylori infection in young children complicates research on this infection in early life. The aim of the present report was to describe the correspondence between positive serology and positive urea breath test (UBT) in children followed from age 0 to 24 months in the Pasitos Cohort Study, conducted along the US-Mexico border at El Paso and Juarez. Children were recruited before birth during 1998-2000 and examined at target ages of 6, 12, 18 and 24 months. H. pylori infection was detected using an enzyme immunoassay for serum immunoglobulin G antibodies and the (13)C-urea breath test corrected for age-dependent variation in CO(2) production. Of 472 children, 125 had one or more positive UBT results and 46 had one or more positive serology results. The prevalence of H. pylori infection at target ages of 6, 12, 18 and 24 months was 7%, 14%, 16% and 19%, respectively, by UBT and 8%, 2%, 3% and 3%, respectively, by serology. Few (<1%) of those tested on both tests were positive on both at any age. Among UBT-positive children, 6% were concurrently seropositive and 6% became seropositive later. Because UBT positivity cut points were selected to minimise false positives, these results suggest that H. pylori infection occurred frequently in this cohort, but rarely produced detectable antibodies. For clinical or epidemiological investigations, serology should not be used as the sole method for detecting H. pylori infection in children aged 2 years or less.