BACKGROUND: Gastroduodenal ulcers are common in patients taking low-dose aspirin. However, the factors predisposing to mucosal erosions, the precursor lesions, are not well known. AIMS: To examine the potential risk factors for the development of erosions in patients chronically taking low-dose aspirin. METHODS: Patients included were taking aspirin 75-325 mg daily for >28 days. Exclusion criteria included use of nonsteroidal anti-inflammatory and ulcer-healing drugs. Demographic data were collected at baseline, prior to endoscopy to determine the frequency and number of erosions and Helicobacter pylori status. In those without ulcer or other exclusions, endoscopy was repeated at 3 months. RESULTS: Fewer patients had gastric erosions if they were H. pylori +ve (48.5% vs. 66.4% in H. pylori-ve patients at baseline, P = 0.17; 40.0% vs. 64.1% at 3 months, P = 0.029). If gastric erosions were present, they were also less numerous in H. pylori +ve patients (3.61 +/- 0.83 vs. 4.90 +/- 0.53 at baseline, P = 0.026; 2.17 +/- 0.68 vs. 5.68 +/- 0.86 at 3 months, P = 0.029). There was a trend (0.1 > P > 0.05) for more gastric erosions in those taking >100 mg/day aspirin. Males had more duodenal erosions at baseline (25.2% vs. 7.5%, P = 0.016). Patient age did not affect the presence or number of erosions. H. Pylori was not significantly associated with duodenal erosion numbers. CONCLUSIONS: Helicobacter pylori infection may partially protect against low-dose aspirin-induced gastric erosions; damage to the stomach appears weakly dose-related; and older age does not increase the risk of erosions.