Prospective comparison of three risk scoring systems in non-variceal and variceal upper gastrointestinal bleeding.
J Gastroenterol Hepatol. 2015 Oct 30;
Authors: Thanapirom K, Ridtitid W, Rerknimitr R, Thungsuk R, Noophun P, Wongjitrat C, Luangjaru S, Vedkijkul P, Lertkupinit C, Poonsab S, Ratanachu-Ek T, Hansomburana P, Pornthisarn B, Thongbai T, Mahachai V, Treeprasertsuk S
BACKGROUND AND AIM: Data regarding the efficacy of the Glasgow Blatchford score (GBS), full Rockall score (FRS) and pre-endoscopic Rockall scores (PRS) in comparing non-variceal and variceal UGIB are limited. Our aim was to determine the performance of these 3 risk scores in predicting the need for treatment, mortality and re-bleeding among patients with non-variceal and variceal UGIB.
METHODS: During January, 2010 and September, 2011, patients with UGIB from 11 hospitals were prospectively enrolled. The GBS, FRS and PRS were calculated. Discriminative ability for each score was assessed using the receiver operated characteristics curve (ROC) analysis.
RESULTS: A total of 981 patients presented with acute UGIB, 225 patients (22.9%) had variceal UGIB. The areas under the ROC (AUC) of the GBS, FRS and PRS for predicting the need for treatment were 0.77, 0.69 and 0.61 in non-variceal vs.0.66, 0.66 and 0.59 in variceal UGIB. The AUC for predicting mortality and re-bleeding during admission were 0.66, 0.80 and 0.76 in non-variceal vs. 0.63, 0.57 and 0.63 in variceal UGIB. AUC score were not statistically significant for predicting need for therapy and clinical outcome in variceal UGIB. The GBS≤2 and FRS≤1 identified low-risk non-variceal UGIB patients for death and rebleeding during hospitalization.
CONCLUSION: In contrast to non-variceal UGIB, the GBS, FRS and PRS were not precise scores for assessing the need for therapy, mortality and re-bleeding during admission in variceal UGIB.
PMID: 26514879 [PubMed - as supplied by publisher]