This double blind randomised study tested the effectiveness of colloidal bismuth subcitrate (De-Nol) in non-ulcer dyspepsia (NUD) and if any benefit is associated with clearance of Campylobacter pylori (C pylori) from the gastric mucosa. Sixty six patients with dyspepsic symptoms, normal abdominal ultrasound, and upper GI endoscopy, were randomly allocated to placebo or De-Nol for eight weeks. Antral biopsies were taken for bacteriological and histological examination, and endoscopies and clinical questionnaires were administered before and after treatment. Fifty two patients (25 on De-Nol and 27 on placebo) completed the trial. De-Nol cleared C pylori from 10 of the 12 C pylori positive patients (83.3%), whereas placebo did not clear C pylori from any of the eight C pylori positive patients (p less than 0.01). In patients receiving De-Nol gastritis improved (p less than 0.01) and symptomatic response was better (p less than 0.001) compared with placebo. In the placebo group seven of the 19 C pylori negative patients became positive: this was associated with significant deterioration of symptoms, a phenomenon not seen in the De-Nol treated group.