We validated a commercial enzyme-linked immunosorbent assay (ELISA), Helico-G, in diagnosing H. pylori in 129 patients (mean age 50 years, range 15-86). We analysed the results of endoscopy against serology to see whether there was a possibility of adopting the strategy of not endoscoping dyspeptic subjects under the age of 45. H. pylori infection was considered present if either histology and/or culture were positive. The ELISA had a sensitivity of 88%, specificity of 72%, positive predictive value of 85%, negative predictive value of 77% and accuracy of 82% in detecting H. pylori. In a subgroup of 52 subjects aged 45 or less (mean age 35 years, range 15-45), 17 out of 25 patients with positive endoscopic findings were H. pylori seropositive while 16 out of 27 patients had normal endoscopic findings. Eighteen out of the 52 patients (35%) were H. pylori seronegative and normal endoscopically except for five patients (10%) who had mild to moderate oesophagitis and two who had non-erosive gastritis (4%). All patients with duodenal ulcer disease (7) were seropositive giving predictive values of positive and negative serology for a diagnosis of duodenal ulcer disease as 28% and 100%, respectively. Therefore adopting a strategy of endoscoping subjects under the age of 45 only if they were H. pylori seropositive would have saved 35% of endoscopies in this age group but missed oesophagitis in 10%. Negative serology would tend to exclude duodenal ulcer disease while positive serology discriminates poorly for it. Serology may be a useful adjunct in screening to reduce endoscopy workload provided that patients with gastro-oesophageal reflux symptoms are excluded.