In April, 2005, a 79-year-old Chinese man presented with severe iron defi ciency anaemia (haemoglobin 67 g/L). He had no history of peptic ulcer disease and there were no signs or symptoms of gastrointestinal bleeding or disease. He had had a myocardial infarction in 1999, and had been taking aspirin 80 mg daily since then; he was not using any non-steroidal anti-infl ammatory drugs (NSAIDs) or herbal preparations. Upper and lower gastrointestinal endoscopy failed to identify any bleeding lesions. He was supportively transfused, and misoprostol 400 μg daily was empirically added. In the subsequent 7 months, he had a progressive drop in haemoglobin from 106 g/L (post-transfusion) to 54 g/L. He developed bilateral ankle oedema, with serum albumin concentration dropping from 39 g/L to 23 g/L. There was no proteinuria and the cause of hypoalbuminaemia was suspected to be loss from the gastrointestinal tract. A technetium-99m-labelled human serum albumin scan, which is purported to be a sensitive test for a protein losing enteropathy, was normal. Small bowel capsule endoscopy (Olympus, Tokyo, Japan), however, showed multiple small bowel ulcers, particularly in the ileal region (fi gure A and B). In view of the possible link between aspirin therapy and small bowel enteropathy, aspirin was withheld. A repeat capsule endoscopy 3 months later showed a few small healing erosions only (fi gure C and D). When last seen in July, 2006, our patient remained asymptomatic with no oedema or anaemia (haemoglobin 122 g/dL), and his serum albumin was 41 g/L. Small bowel injury is increasingly recognised to be an important complication of NSAID therapy. NSAIDs cause varying degrees of small bowel injury from increased intestinal permeability, intestinal infl ammation, protein loss, blood loss, and ulcerations, to perforation and diaphragm-like strictures. In our study, comparing the gastrointestinal tolerability of celecoxib with diclofenac and a proton pump inhibitor, we found that 30% of the serious gastrointestinal events were distal to the duodenum with one death due to small bowel perforation. Post-hoc analysis of a large-scale clinical outcome trial showed that lower gastrointestinal events accounted for 40% of all serious gastrointestinal events in patients on NSAIDs. The diffi culty in assessing the severity of NSAIDenteropathy was previously hampered by the lack of reliable tests for examining the small bowel. With the advent of capsule endoscopy and double balloon enteroscopy, small bowel injury can now be directly visualised. Capsule endoscopy corroborates the high prevalence of NSAID-enteropathy demonstrating that about 70% of regular NSAID users have small bowel erosions and ulcers. Here we illustrate a case of severe enteropathy induced by low-dose aspirin that has many striking similarities with NSAID-enteropathy. Although aspirin at anti-infl ammatory drug doses can cause stomach and duodenal ulcers, it is generally believed that aspirin does not cause any small bowel damage based on intestinal permeability and faecal infl ammatory marker studies. Our observation not only challenges the safety of aspirin on small bowel mucosa but also raises questions about the effi cacy of misoprostol as a prophylactic treatment against enteropathy. Low-dose aspirin, used for cardiovascular prophylaxis, is one of the most commonly prescribed medications worldwide. The demonstration of a severe form of enteropathy with this drug suggests the need to defi ne the extent of this potential problem. Increasing clinical awareness and appropriate early investigation are needed to establish the proper diagnosis.