BACKGROUND: There is interest in noninvasive H pylori testing as a means of predicting diagnosis and determining management in dyspeptic patients. AIMS: To assess the value of the 14C urea breath test as a predictor of peptic ulcer disease in patients presenting with dyspepsia. PATIENTS AND METHODS: 327 consecutive patients referred for investigation of dyspepsia had a 14C urea breath test performed before endoscopy. Patients were not included if they had previously confirmed ulcer disease, previous gastric surgery, or were taking non-steroidal anti-inflammatory drugs. RESULTS: Of the 182 patients with a positive 14C urea breath test, duodenal and/or gastric ulcers were present in 45% and erosive duodenitis in a further 2%. Oesophagitis was present in 12% of the breath test positive patients with two thirds of the oesophagitis patients having co-existent ulcer disease. The prevalence of ulcer disease in the H pylori positive dyspeptic patients was independently related to smoking and previous investigation status. If previously uninvestigated, the prevalence of ulcers was 67% in smokers and 46% in non-smokers. If previous upper gastrointestinal investigations were negative, the prevalence of ulcers was 53% in smokers and 28% in non-smokers. Of the 136 patients with a negative breath test, only 5% had peptic ulcers. The most frequent endoscopic finding in these H pylori negative subjects was oesophagitis, being present in 17%. CONCLUSIONS: This study demonstrates that a positive H pylori test is a powerful predictor of the presence of underlying ulcer disease in dyspeptic patients, especially if smokers, and that a negative H pylori test is a powerful predictor of the absence of ulcer disease. It also indicates that a negative upper gastrointestinal investigation does not preclude subsequent presentation with ulcer disease.