Studies in animals have shown that NSAID/ASA-induced erosions have an ischaemic pathogenesis. We therefore studied the question of whether such erosions in human gastric biopsy material can be identified on the basis of the ischaemic necrosis. Histological sections prepared from forceps biopsy material obtained from 122 patients with erosions (at least three biopsy specimens from the erosion and two from antrum and corpus each) were classified by a pathologist blinded to the endoscopic findings and the medication used by the patients. NSAID/ASA erosions were diagnosed when a homogeneous eosinophilic ischaemic necrosis blending into the adjoining lamina propria presented. Helicobacter pylori (Hp)-induced erosions were diagnosed when, in the presence of Hp gastritis, erosive defects were covered with a non-homogeneous fibrinoid necrosis containing granulocytes and cell debris. Finally, the histological classification was compared with data on medication usage. The histological diagnosis was Hp-induced erosions in 59 patients, NSAID/ASA-induced erosions with no Hp gastritis in 23, and NSAID/ASA-induced erosions with concomitant Hp gastritis in 40. A comparison of this histological classification with the data provided by the referring physicians on patient medication revealed that 70% of the patients with histological diagnosis of NSAID/ASA-induced erosions in the absence of Hp gastritis, and 65% of those diagnosed to have NSAID/ASA-induced erosions and concomitant Hp gastritis, had been taking such drugs. Among the erosions diagnosed as H. pylori-induced, 81% of the patients were reported not to take such medication. The sensitivity of the diagnosis of NSAID/ASA-induced erosions was 72.9%, and specificity 79.6%. The results of the present study show that a high percentage of the NSAID/ASA-induced erosions of the gastric mucosa can indeed be correctly diagnosed at histology.