Peptic ulcer disease (PUD) in RA patients is associated with NSAID use. This study aimed to validate the predictive value of presumed risk factors for NSAID-associated PUD in a prospective gastroscopic study in RA patients. Eighty-one NSAID using RA patients were prospectively divided into four presumed risk groups according to Helicobacter pylori status and history of PUD. As additional risk factors the following were analysed: upper gastrointestinal GI complaints; disability; daily dose of NSAID and antral gastritis. The presence of PUD in the four risk groups did not differ. Additionally it was found that a history of PUD was predictive for current PUD [odds ratio (OR) 3.9; 95% CI 1.1-14]. H. pylori status was not predictive. Transformation from one ulcer type to another was rare. NSAID dose was not a risk factor, while disability was of borderline importance (OR 2.1; 95% CI 1-4.8). Current upper GI complaints were bad predictors. PUD only occurred with a concomitant antral gastritis. A history of PUD, disability and antral gastritis were the most important predictors for current PUD. When an ulcer relapsed it was of the same ulcer type as had been present earlier. This may have practical implications for prophylaxis enabling stratification by previous ulcer type.