Risk factors of in-hospital mortality of non-variceal upper gastro-intestinal bleeding

Background: The in-hospital mortality rate of non-variceal upper gastrointestinal bleeding (UGIB) is still considerable, and this figure seems to be remained stable during the past years. The aim of this retrospective study was to evaluate the risk-factors of in-hospital death due to non-variceal UGIB in this setting.

Patients and methods: Between 2005 December and 2006 January, 116 patients with non- variceal UGIB were admitted. The distribution of patients according to the source of bleeding was the followings: gastric ulcer (n=38 [32%]), duodenal ulcer (n=30 [25%]), gastric or duodenal erosions (n=22 [19%]), oesophageal erosions (n=14 [12%]), upper GI cancer (n=8 [6%]), Mallory-Weiss tear (n=7 [6%]). The primary outcome of the study was the in-hospital mortality rate. Independent variables as ulcer localisation, Forrest type of bleeding, presence of Mallory-Weiss lesion, gastric or duodenal erosions, erosive oesophagitis, underlying chronic disease, NSAID/antiplatelet/anticoagulat use, Helicobacterium (Hp) status, previous upper GI bleeding, requirement of surgery, blood transfusion units were entered into multiple logistic regression model to identify variables which had significant influence on in- hospital death.

Results: The overall mortality rate was 19%, surgery was required in 2.6% of patients. Multiple logistic regression analyses showed that presence of underlying chronic disease (p=0.009), requirement of surgery (p<0.001), bleeding from gastric ulcer (p=0.003), Forrest I/a and II/b bleeding (p=0.033 and 0.020) and Hp negative status (p=0.021) were definitive risk factors of mortality by non-variceal UGIB.

Conclusions: Comorbid conditions, active bleeding form gastric ulcer, requirement of surgery for bleeding control and negative Hp status are significant predictors of death of patients with non-variceal UGIB.