Prognostic factors and timing of endoscopy in acute non-variceal upper gastrointestinal bleeding

Abstract

Purpose and objectives: Identification of causes, endoscopic findings and clinical course of acute non-variceal upper-GI bleeding, determining the rate of re-bleeding, surgical treatment and mortality within 40 days, followed by the identification of independent prognostic factors and the importance of the timing of endoscopy.

Methods: Consecutive patients with acute non-variceal bleeding over a period of 2 years were studied, patients with variceal bleeding and lower-GI bleeding were excluded. We recorded medical history, hemodynamic variables, endoscopic findings and degree of bleeding lesions, number of transfusions, interval from admission to endoscopy, the rate of re-bleeding, surgical treatment and mortality.

Results: 163 patients with a mean age of 65.14 years, with women accounting for 42.9%, were studied. Peptic lesions of the stomach or duodenum were the source in 78.5% of cases, active bleeding represented 19%, endoscopic treatment was applied in 49.1% of patients. Re-bleeding was observed in 5.5% and 4.3% of patients underwent surgical treatment, mortality was 12.9%. In multivariate analysis, independent predictors of death were the clinical Rockall score > 3 (OR, odds ratio = 12.54; p = 0.0014) with 90.5% sensitivity and 97.8% negative predictive value, treatment with steroids or selective serotonin reuptake inhibitors (SSRI) (OR = 5.14; p = 0.045) and the lowest recorded haemoglobin concentration (OR = 2.96; p = 0.031). Endoscopy within 6 hours of admission identified more higher-risk lesions and more frequently led to endoscopic treatment. However, no difference was found in transfusion requirements, recurrence of bleeding, surgery or death.

Conclusion: Rockall score > 3 was the strongest independent prognostic factor of death and these patients should merit urgent management. In a subgroup of patients with endoscopy within 6 hours of admission, endoscopic treatment was applied more often with no advantages in the clinical outcome.