The possibility has been investigated that, after admission to hospital with acute upper gastrointestinal bleeding, patients who have been users of aspirin and non-aspirin non-steroidal anti-inflammatory drugs have poorer baseline renal. function, a greater improvement in renal function during their hospital stay, and a larger transfusion requirement than non-users. Patients over 50 years of age admitted to public hospitals with acute upper gastrointestinal bleeding were studied. Creatinine clearance was estimated from serum creatinine and the transfusion requirement was recorded as the number of units of blood transfused on Day 1 and throughout the entire hospital stay. Data were obtained prospectively from case notes and by structured interview. Users of non-steroidal anti-inflammatory drugs were significantly older than non-users. The estimated creatinine clearance on admission to hospital declined with age. Creatinine clearance was 13.2 (95 % CI 6.0 to 20.4) ml · min' lower in users than non-users of non-aspirin non-steroidal anti-inflammatory drugs. However, the difference was attributable to the older age of the drug users rather than to the drugs themselves. On average, the increase in creatinine clearance during hospital stay was the same in users and non-users of non-aspirin non-steroidal anti-inflammatory drugs. Prior use of aspirin had no effect on any measure of renal function. The incidence of blood transfusion was higher in older than in younger patients but neither the incidence of transfusion, nor the transfusion requirement, was different between users and non-users of non-aspirin non-steroidal anti-inflammatory drugs and aspirin. Although the study has not excluded certain adverse effects of prior use of these drugs on renal and haemostatic function after an episode of upper gastrointestinal bleeding, it does indicate that such change are unlikely to be of major clinical significance.