Abstract

BACKGROUND: Gastrointestinal bleeding is related to non-steroidal anti-inflammatory drug (NSAID) use, especially aspirin, but only a small subset of users bleed. AIM: To look for risk factors or mechanisms whereby aspirin may promote gastrointestinal bleeding. PATIENTS: Sixty one patients with previous aspirin related upper gastrointestinal bleeding and 61 matched controls. METHODS: Patients and controls were given 375 mg of aspirin and sequential skin bleeding time and blood aspirin levels were measured. Additional studies included platelet lumiaggregation, von Willebrand factor, Factor VIII, and coagulation studies. RESULTS: Baseline skin bleeding time was similar in bleeders and controls, but bleeders had a more prolonged skin bleeding time after aspirin use. Hyper-response was more frequent in bleeders (30% v 9.3%; p < 0.01) and was associated with more than one previous separate bleeding event and a lower packed cell volume during the preceding bleeding episode. No differences were found in other factors studied. Logistic regression analysis identified prolonged skin bleeding time after aspirin use as an independent factor contributing to aspirin related gastrointestinal bleeding (RR = 5.4; 95% CI: 1.8 to 17.1). CONCLUSIONS: 30% of patients with a history of aspirin related gastrointestinal bleeding have an exaggerated prolongation of skin bleeding time in response to aspirin, which may be a risk factor for bleeding. This intrinsic defect or to subclinical von Willebrand disease or different aspirin metabolism.