The Clinical Significance Of Rockall And Glasgow-Blatchford Scoring Systems In Upper Gastrointestinal Hemorrhage Patients

Aims:The performance of the Rockall and Glasgow-Blatchford score in predicting outcomes and the need for clinical intervention in upper gastrointestinal hemorrhage patients was compared, in order to investigate their clinical value.Methods:A retrospective observational study was performed in all patients with upper gastrointestinal hemorrhage (UGIH) who attended the Second Affiliated Hospital of Dalian Medical University from January2010to December2012. Calculated the Pre-endoscopic Rockall score (PRS) as well as the Complete Rockall score (CRS) and Glasgow-Blatchford score (GBS) for each patient, and compared the difference of the scores, outcomes and the need for clinical intervention between two groups (variceal and nonvariceal). Comparisons of these three risk scores in predicting recurrent bleeding, mortality and the need for clinical intervention(blood transfusion and/or endoscopic/surgical intervention) were made by calculation of the areas under the receiver-operator characteristic curves(AUC) with95%CI’s.Results:In316patients, there were79(25%) variceal bleeds. The variceal group had higher scores, more recurrent bleeding and mortality, and higher endoscopic intervention rate (P<0.05). For prediction of mortality, AUC was obtained for CRS(0.840), PRS(0.805), GBS(0.779), respectively (P<0.05); for prediction of rebleeding, AUC was obtained for CRS(0.799), PRS(0.713), GBS(0.651), respectively (P<0.05). CRS was superior to both the PRS and GBS in predicting rebleeding and mortality. In predicting the need for overall clinical intervention, the GBS performed better than both the PRS and CRS (AUC0.760vs0.744vs0.690, respectively). In the separate prediction endoscopic or surgical intervention, the GBS was not effective (AUC=0.574, P>0.05). When a cut-off value of2was used, sensitivity and specificity of GBS for predicting the need for clinical intervention was99%and17%.Conclusions:Pre-endoscopic Rockall score as well as Complete Rockall score and Glasgow-Blatchford score system are all suitable for UGIH patients including variceal bleeds. Among these, CRS maybe the best predictor for adverse outcomes, the GBS performed best in predicting the overall clinical intervention but not in the separate prediction endoscopic or surgical intervention, the risk stratification of optimal threshold was2.