Fifty-three patients with previously uninvestigated chronic dyspepsia symptoms in the absence of gastrointestinal or extra-gastrointestinal disease (functional dyspepsia) underwent antral and duodenal mucosal biopsies to detect the role of such samplings in the presence of normal endoscopic findings. Patients were enrolled in a randomized, placebo-controlled, double-blind trial, receiving either eradicating treatment (colloidal bismuth subcitrate plus metronidazole) or placebo if they had Helicobacter pylori-associated gastritis (20 patients), or cisapride or placebo if they had normal antral mucosa (28 cases). Unsuspected celiac sprue was found in one patient. Eradicating treatment ameliorated histological gastritis (p = 0.01). However, owing to great placebo efficacy, symptom remission rates following a 1-month wash-out period in both treatment groups were no higher than that in controls. Independent of the initial randomization, an extremely low symptom recurrence rate was observed during a drug-free follow-up study equivalent to the mean duration of symptoms before enrollment. We conclude that in functional dyspepsia, bulbar and antral biopsies are not useful in clinical management, equivalent symptom relief can be achieved in patients randomly assigned to both drugs and placebos, and such improvement can be long lasting in the absence of any maintenance treatment. We believe the prevalence of unsuspected villous atrophy and the therapeutic role of investigation-based reassurance deserve further assessment.