Introduction Upper gastrointestinal bleeding (UGIB) developing while in hospital is traditionally thought to be due to stress ulceration affecting critically ill patients. The role of potentially damaging drugs is not clear.
Aims Given the increasing use of low-dose aspirin and other antithrombotic drugs—clopidogrel, dipyridamole, and warfarin—for vascular protection, we aimed to clarify their impact on UGIB affecting hospitalised patients.
Methods Between 2008 and 2009, we prospectively assessed the clinical details and outcomes of all inpatient bleeders at our hospital, and compared these with outpatients admitted because of bleeding. For each inpatient we included three outpatients as controls. Patients were excluded if they had a variceal cause. The Charlson comorbidity score and the complete Rockall score for non-variceal upper gastrointestinal bleeding were also calculated. The χ2 test, Fisher's exact test, Mann–Whitney U test, and logistic regression analyses with ORs and 95% CIs were used as appropriate.
Results Abstract PWE-157 table 1 summarises the main characteristics of patients who bled while in hospital (inpatient bleeders) vs those who were admitted after bleeding (outpatient bleeders). Inpatient bleeders were older and more likely to be males and to have cardiovascular disease. After adjustment for age and sex, the logistic regression analysis still showed that patients who bled as inpatients were significantly more likely to be taking antithrombotic medication, especially non-aspirin agents, than those who were admitted after bleeding, [OR (95% CIs), 2.15 (1.25 to 3.68); p=0.006]. The adjusted odds of death within 30 days of bleeding was almost twice as high in the inpatient group, [OR 1.88 (0.74 to 4.77)]. The endoscopic abnormalities in more than 80% of patients in both groups included erosive oesophagitis, gastric, or duodenal ulcers or erosions.
Conclusion The use of antithrombotic drugs is a significant risk factor for UGIB developing in hospitalised patients. The endoscopic lesions, whether caused by these agents or not, are potentially preventable.
Competing interests None declared.