Background: A large variety of drugs have been implicated in causing dyspepsia. Due to the high background incidence of dyspepsia it is impossible to distinguish between spontaneous and truly drug-related symptoms. Most patients with dyspeptic symptoms are treated empirically. Drug-induced dyspepsia might therefore be reflected in the sequencing of prokinetics relative to other medications. Aim: To screen a large prescription database for signs of drug-induced functional dyspepsia, applying a symmetry principle. Methods: Prescription data on all incident users of cisapride and metoclopramide were used to identify individuals who had started their first therapies with a prokinetic drug and an index drug within a 100-day span. A dyspepsia-provoking effect of the index drug would manifest as an excess of persons with the prokinetic drug prescribed last in this selected population. Relative to conventional analyses based on case-control or cohort design, this principle is robust to confounders that are stable over time. Results: In the cisapride analysis (1825 persons) no single drug had adjusted rate ratios significantly above unity. An inverse signal for antidepressants (rate ratio 0.57; 95% CI: 0.39-0.84) suggests that these drugs may have a therapeutic effect against functional dyspepsia. In the metoclopramide analysis (6126 persons) positive signals were found for 14 drugs, all well-known for causing nausea as a side-effect, with the exception of insulin (rate ratio 2.91, 95% CI: 1.40-8.11). Conclusions: Drug-induced symptoms of functional dyspepsia are rare and do not contribute to the use of cisapride. The start of insulin treatment may induce nausea.