Many animal and human studies have shown aspirin to cause gastric mucosal erosions and enhanced spontaneous microbleeding. Chronic use is associated with uncomplicated gastric ulcer (but not clearly with uncomplicated duodenal ulcer) and with presentation with haematemesis and melaena. Sub-group analysis suggests that the risks of bleeding duodenal ulcer as well as bleeding gastric ulcer (and possibly non-ulcer bleeding) are increased. These data imply that an anti-haemostatic as well as an ulcerogenic effect contributes to such presentation and acute data support this proposition. A broken mucosa and inhibition of thromboxane synthesis both appear to be necessary for bleeding to occur. Aspirin enhances the bleeding associated with mucosal biopsy analogous to its prolongation of the skin bleeding time. Anti-haemostatic properties could account for the apparent increase in presentation with haematemesis and melaena in patients taking low doses of aspirin for cardiovascular prophylaxis. However, acute studies also show aspirin 300 mg to cause a level of mucosal injury which is substantial, though significantly less than with aspirin 600 mg q.d.s. The risk attached to aspirin consumption needs to be kept in perspective. It can be calculated that aspirin use is associated with approximately 700 bleeds per annum in the UK. The relative risk for upper gastrointestinal bleeding is roughly similar to that for non-aspirin non-steroidal anti-inflammatory drugs but those affected are generally younger, in an age group where death from upper gastrointestinal bleeding is very uncommon. In keeping with this and in contrast to non-aspirin non-steroidal anti-inflammatory drugs, no study has shown increased mortality from aspirin-related upper gastrointestinal complications.