Background : Aspirin is valuable for preventing vascular events, but information about ulcer frequency is necessary to inform risk-benefit decisions in individual patients.

Aim : To determine ulcer prevalence and incidence in a population representative of those given aspirin therapy and evaluate risk predictors.

Methods : Patients taking aspirin 75–325 mg daily were recruited from four countries. Exclusions included use of gastroprotectant drugs or other non-steroidal anti-inflammatory drugs. We measured point prevalence of endoscopic ulcers, after quantitating dyspeptic symptoms. Incidence was assessed 3 months later in those eligible to continue (no baseline ulcer or reason for gastroprotectants).

Results : In 187 patients, ulcer prevalence was 11% [95% confidence interval (CI) 6.3–15.1%]. Only 20% had dyspeptic symptoms, not significantly different from patients without ulcer. Ulcer incidence in 113 patients followed for 3 months was 7% (95% CI 2.4–11.8%). Helicobacter pylori infection increased the risk of a duodenal ulcer [odds ratio (OR) 18.5, 95% CI 2.3–149.4], as did age >70 for ulcers in stomach and duodenum combined (OR 3.3, 95% CI 1.3–8.7).

Conclusions : Gastroduodenal ulcers are found in one in 10 patients taking low-dose aspirin, and most are asymptomatic; this needs considering when discussing risks/benefits with patients. Risk factors include older age and H. pylori (for duodenal ulcer).