BACKGROUND: This review considers management strategies (combinations of initial investigation and empirical treatments) for dyspeptic patients. Dyspepsia was defined to include both epigastric pain and heartburn. OBJECTIVES: To determine the effectiveness, acceptability, and cost effectiveness of the following initial management strategies for patients presenting with dyspepsia (a) initial pharmacological therapy (including endoscopy for treatment failures) (b) early endoscopy (c) testing for Helicobacter pylori and endoscope only those positive (d) H.pylori eradication therapy with or without prior testing. SEARCH STRATEGY: Trials were located through electronic searches and extensive contact with trialists. SELECTION CRITERIA: All randomised controlled trials of dyspeptic patients presenting in primary care. DATA COLLECTION AND ANALYSIS: Data was collected on dyspeptic symptoms, quality of life and use of resources. MAIN RESULTS: Eighteen papers reporting 20 comparisons were found. Trials comparing proton pump inhibitors (PPI) with antacids (two trials) and H2 receptor antagonists (three trials), early endoscopy with initial acid suppression (five trials), H.pylori 'test and scope' v usual management (three trials) and H.pylori test and treat v. endoscopy (three trials) were pooled. PPIs were significantly more effective than both H2RA s and antacids. Relative risks (RR) and 95% CI were, for PPI: antacid 0.72 (0.64-0.80), PPI: H2RA 0.63 (0.47-0.85). Results for other drug comparisons were either absent or inconclusive. Initial endoscopy was associated with a small reduction in the risk of recurrent dyspepstic symptoms compared with initial prescribing (RR 0.89 (0.77-1.02). H.pylori test and endoscopy increases costs in primary care, but does not improve symptoms. H.pylori test and eradicate may be as effective as endoscopy- based management and reduces costs, by decreasing the proportion of patients that are endoscoped. Further primary care-based trials are needed to compare 'test and treat' with empirical acid suppression. REVIEWER'S CONCLUSIONS: PPIs are effective in the treatment of dyspepsia in these trials which may not adequately exclude patients with gastro-oesophageal reflux disease. The relative efficacy of H2RA and PPI is uncertain. Early investigation by endoscopy or H.pylori testing may benefit some patients with dyspepsia. The review will be updated in 2002 with an individual patient data meta-analysis of the economic data, and a subgroup analysis by age and predominant dyspeptic symptom.