To compare the performances of the AIMS65 and Glasgow Blatchford risk score (GBRS) as risk assessment tools in elderly patients with upper gastrointestinal bleeding (UGIB).

A retrospective study was performed in 293 elderly patients with UGIB in the gastroenterology department. The primary outcome was inpatient mortality. Secondary outcomes were rebleeding, blood transfusion and a composite clinical endpoint of inpatient mortality, rebleeding, and endoscopic, radiologic or surgical intervention. The GBRS and AIMS65 scores were respectively calculated for all elderly patients. And the area under the receiver operating characteristic curve (AUROC) was calculated to evaluate the predictive value of the two scoring systems.

Of the 293 elderly patients, 27 (9.2%) died, 31 (10.6%) rebleeding, 164 (55.9%) received blood transfusion, and 100 (34.1%) experienced the composite clinical endpoint. The AUROCs of the AIMS65 score for inpatient mortality, rebleeding, blood transfusion and the composite clinical endpoint were 0.833 (95%CI: 0.785–0.874), 0.646 (95%CI: 0.588–0.700), 0.666 (95%CI: 0.609–0.720), 0.702 (95%CI: 0.645–0.754), respectively. The AUROCs of the GBRS were 0.681 (95%CI: 0.624–0.734), 0.746 (95%CI: 0.692–0.795), 0.753 (95%CI: 0.700–0.802), 0.744 (95%CI: 0.690–0.793), respectively.

For the elderly patients with UGIB, the AIMS65 score is superior to GBRS in predicting inpatient mortality, and the GBRS is superior in predicting rebleeding and blood transfusion. Both scores are similar in predicting the composite clinical endpoint.