ABSTRACT
Objective To compare Glasgow-Blatchford score (GBS) with full Rockall score system (full RS) in predicting clinical outcomes of upper gastrointestinal bleeding(UGIB).Methods Four hundred hospitalized UGIB patients were enrolled in this study in Baoji Central Hospital from March 2014 to June 2016. Patients were scored by full RS and GBS,respectively.The receiver operating characteristic(ROC)curve analysis was used to compare the predictive values of the two score systems for mortality,re-bleeding,transfusion,ICU admission and endoscopic intervention. Results Thirty-eight cases were lost in follow-up. The one-month mortality rate, re-bleeding rate, blood transfusion rate, ICU admission rate and endoscopic intervention rate of 362 patients were 11.6%, 9.4%, 17.7%, 13.3% and 16.0%, respectively. GBS showed better predictive values in ICU admission (AUC=0.731, P=0.009), re-bleeding (AUC=0.707, P=0.019), blood transfusion (AUC=0.704, P=0.021) and endoscopic intervention (AUC=0.742, P=0.006). Full RS showed better predictive values in one-month mortality (AUC=0.681, P=0.040). There was significant difference in predicting value of clinical outcomes of UGIB by ROC curve between GBS and full-RS systems (P<0.05). Conclusion The GBS system is more suitable for predicting clinical outcomes of ICU admission, re-bleeding, blood transfusion and endoscopic intervention in patients of UGIB.The full-RS system is more suitable for predicting one-month mortality of UGIB patients.