AIM: Use of nonsteroidal anti-inflammatory drugs (NSAIDs) increases risk and worsens prognosis for patients with complicated peptic ulcer disease. Therefore, patients who are at high risk of peptic ulcer often use tramadol instead of NSAIDs. Tramadol's effect on peptic ulcer prognosis is unknown. The aim was to examine mortality in the 30 days following hospitalization for perforated peptic ulcer among tramadol and NSAID users compared with non-users. METHODS: The study was based on data on reimbursed prescriptions and hospital discharge diagnoses for the 1993-2004 period, extracted from population-based healthcare databases. All patients with a first-time diagnosis of perforated peptic ulcer were identified, excluding those with previous ulcer diagnoses or antiulcer drug use. Cox regression was used to estimate 30-day mortality rate ratios for tramadol and NSAID users compared with non-users, adjusting for use of other drugs and comorbidity. RESULTS: Of 1271 patients with perforated peptic ulcers included in the study, 2.4% used tramadol only, 38.9% used NSAIDs and 7.9% used both. Thirty-day mortality was 28.7% overall and 48.4% among users of tramadol alone. Compared with the 645 patients who used neither tramadol nor NSAIDs, the adjusted mortality rate in the 30 days following hospitalization was 2.02-fold [95% confidence interval (CI) 1.17, 3.48] higher for the 31 'tramadol only' users, 1.41-fold (95% CI 1.12, 1.78) higher for the 495 NSAID users and 1.32-fold (95% CI 0.89, 1.95) higher for the 100 patients who used both drugs. CONCLUSION: Among patients hospitalized for perforated peptic ulcer, tramadol appears to increase mortality at a level comparable to NSAIDs.