OBJECTIVE Previous erosive oesophagitis studies have excluded patients with peptic ulcers or taking aspirin, and conflicting results have been reported concerning the influence of. We aimed to study the possible correlation between erosive oesophageal and gastro-duodenal diseases in patients with or without erosive oesophagitis who are taking low-dose aspirin, or simple analgesics, or those infected with. METHODS Endoscopic oesophageal and gastro-duodenal lesions were graded in 287 patients with reflux oesophagitis, median age 57 years, including 168 with oesophageal erosions, 131 infected with, 45 patients taking aspirin 75 mg daily, and 65 patients taking simple analgesics containing paracetamol/codeine. RESULTS The grades of oesophageal erosions correlated positively with the duodenal scores (r = 0.15; P = 0.01) in the study group as a whole (n = 287), and in patients (n = 131) with (r = 0.169; P = 0.05). Eighty of 168 patients with erosive oesophagitis had (48%), compared with 51 of 119 patients (43%) with non-erosive oesophageal disease (P = 0.47). Oesophageal scores were highest in the aspirin group (P = 0.04; Kruskal-Wallis test), with grades > or =3 being found in 36% of patients on aspirin, 22% on simple analgesics, and in 18% of other patients. CONCLUSIONS The degree of oesophageal damage correlates positively with that in the duodenal mucosa, although the overall prevalence of erosive oesophagitis is not influenced by. Also, patients taking aspirin have a greater degree of oesophageal damage. These indicate the presence of a common process mediating both oesophageal and duodenal diseases in at least some patients with these disorders.