Introduction and aim: The role of serum albumin level in patients with non-variceal upper gastrointestinal bleeding (NVUGB) has not been extensively studied. Our aim was to evaluate the role of serum albumin on admission in terms of in-hospital mortality in patients with NVUGB.
Materials and methods: Patients admitted with NVUGB during a 4-year period were prospectively included. Demographic, clinical, and laboratory data were collected. ROC curve analysis was used to determine the cutoff value for serum albumin on admission that made a distinction between deceased patients and survivors with respect to serum albumin on admission, as well as its overall performance compared with the Rockall score.
Results: 185 patients with NVUGB were evaluated. Men predominated (56.7%) and a mean age of 59.1±19.9 years was found. Mean serum albumin on admission was 2.9±0.9g/dl with hypoalbuminemia (< 3.5g/dl) detected on admission in 71.4% of cases. The ROC curve found that the best value for predicting hospital mortality was an albumin level of 3.1g/dl (AUROC 0.738). Mortality in patients with albumin ≥ 3.2g/dl was 1.2% compared with 11.2% in patients with albumin<3.2g/dl (P=.009; OR 9.7, 95%CI 1.2-76.5). There was no difference in overall performance between the albumin level (AUORC 0.738) and the Rockall score (AUROC 0.715) for identifying mortality.
Conclusions: Patients with hypoalbuminemia presenting with NVUGB have a greater in-hospital mortality rate. The serum albumin level and the Rockall score perform equally in regard to identifying the mortality rate.
Keywords: Escala Rockall; Gastrointestinal bleeding; Hipoalbuminemia; Hypoalbuminemia; Mortalidad; Mortality; Rockall score; Sangrado gastrointestinal; Scoring systems; Sistemas de puntuación.
Copyright © 2016 Asociación Mexicana de Gastroenterología. Publicado por Masson Doyma México S.A. All rights reserved.