Risk factors for mortality in severe upper gastrointestinal bleeding.

Abstract

Factors related to mortality after upper gastrointestinal (GI) bleeding may differ in importance in high- and low-risk populations. This retrospective study was undertaken to define risk factors of mortality in patients needing endoscopy for upper gastrointestinal bleeding at a tertiary care centre. Three hundred and sixty-two patients with upper gastrointestinal bleeding were identified from endoscopy charts. Patients' characteristics, bleeding parameters, clinical presentation, pre-existing medication and laboratory data were retrieved from hospital charts and patients who survived and those who died in hospital were compared. The mean Rockall score was 6.6+/-1.3 with 92.5% of patients belonging to a Rockall high-risk group. In hospital, mortality was 26.5%, with 6.4% of patients dying as a direct consequence of bleeding. Variceal bleeding was associated with a high risk of recurrent bleeding and death. Renal disease, liver disease, coagulopathy and immunosuppression were more frequently found in non-survivors than in survivors. Accordingly, serological tests of renal or liver function and coagulation were more disturbed in non-survivors. On average, heart rate was higher, and blood pressure and haemoglobin levels lower in non-survivors. Heparin, glucocorticoids, and anti-ulcer drugs were more frequently used in patients dying in hospital. By logistic regression analysis, in-patient status at the time of bleeding, renal disease and coagulopathy, as well as glucocorticoid use, were risk factors for hospital mortality. In tertiary care, a high mortality rate is observed in upper GI bleeding. Teams involved in treating such patients should be aware of the setting in which treatment is performed and its related risk factors.