The results of therapeutic trials in functional dyspepsia (FD), a frequently encountered condition, are contradictory. Our aim, then, was to produce a pooled estimate, or meta-analysis, of a series of short-term randomized placebo-controlled clinical trials on the pharmacological treatment of FD with antisecretory and gastrokinetic drugs. We retrieved trials for analysis purposes by consulting computerized data bases and by scanning published reviews, Current Contents, and references cited in the individual studies. We also requested bibliographical updates from the medical departments of the manufacturers of the drugs used in the various trials. Of 74 trials retrieved by these means, 23 proved eligible for meta-analysis on the basis of six selection criteria defined a priori. Results were expressed in terms of "therapeutic success" (TS), which includes "symptom-free patients," patients with "significant improvement in symptoms," "excellent results," and so on. The differences in TS rates between the various drugs and placebo were calculated in each trial as the algebraic difference together with the respective 95% confidence interval (95% C.I.); the pooling of results of all eligible trials was done using Cochran's weighted method. With antisecretory drugs, the mean difference in TS rates versus placebo was +20% (95% C.I.: 14-24%). The therapeutic gain for the respective antisecretory agents was 25% (95% C.I.: 14-36%) in the case of pirenzepine and 18% (95% C.I.: 12-24%) in the case of H2 antagonists. Meta-analysis of trials with gastrokinetic drugs also showed superior efficacy of these agents compared with placebo, with a mean difference in TS rates of +46% (95% C.I.: 40-52%).(ABSTRACT TRUNCATED AT 250 WORDS)