Screening for drug related dyspepsia: An analysis of prescription symmetry

Objective. Most patients with severe upper dyspepsia are treated empirically with ulcer drugs. Drug-induced dyspepsia might therefore be reflected in the sequencing of ulcer drugs relative to other medications. Our aim was to screen a large population-base prescription database for evidence of drug-induced dyspepsia. Methods. Prescription data on 31,232 incident users of ulcer drugs were drawn from a research database, covering the county of Funen, Denmark. We identified all individuals who had started their first recorded therapies with an ulcer drug and another non-ulcer drug within a 100 day span. In this selected group, there would normally be an equal number starting either drug first, while a dyspepsia-provoking effect of the non-ulcer drug would manifest as an excess of individuals with the ulcer drug prescribed last. This screening method is robust to confounders that are stable over time. Results. Only non-steroidal antiinflammatory drugs (adjusted rate ratio (RR) 1.8, 95% confidence interval (CI), 1.6-2.0), calcium blockers (RR 1.4, CI 1.2-1.7), corticosteroids (RR 1.1, CI 1.0-1.3), angiotensin converting enzyme inhibitors (RR 1.4, CI 1.1-1.7) and methylxanthines (RR 1.5, CI 1.1-2.2) showed a significant asymmetry suggesting a dyspepsia-provoking effect. An analysis of effect modifiers suggested that the signals for corticosteroids and for angiotensin converting enzyme inhibitors were explained by concurrent use of nonsteroidal anti-inflammatory drugs and by underlying congestive heart failure. The signal for non-steroidal antiinflammatory drugs may be explained by the known reputation of non-steroidal antiinflammatory drugs for causing ulcers. Conclusion. There are hardly any important unknown drug effects that mimic acid related dyspepsia. Drug-induced dyspepsia contributes little to the overall use of ulcer drugs.