RESUMEN
Objective To evaluate the model of end-stage liver disease (MELD) and Child-Tur-cotte-Pugh (CTP) in prognosis for cirrhotic patients with massive hemorrhage from esophageal varieosis, who underwent emergency endoscopic sclerotherapy (EIS). Methods The clinical and follow-up data of a co-hort of 65 liver cirrhotic patients with massive hemorrhage from esophageal varicosis treated by EIS were ana-lyzed retrospectively. Correlation between MELD values and CTP score and classification was analyzed. The area under the receiver operating characteristic (ROC) curve was used to compare MELD with CTP score and classification in predicting mortality risk. The optimal threshold of MELD to determine the prognostic death risk was calculated by ROC curves. Results MELD, CTP score and classification correlated signifi-cantly with each other. Seven patients died within six months and 9 died within the first year of follow-up.There was no significant difference in gender, age and serum bilirubin between the survival and death groups, while the serum creatinine, international normalized ratio (INR), CTP score, and MELD score were significantly different. The areas under the ROC curve of MELD were both greater than 0.8 to predict 6-month and 1 -year survival, and also were greater than those of CTP score and classification. Areas under the ROC curve of MELD is superior to, but not significantly different though, CTP score or classification. The optimal threshold values of MELD and CTP score both showed satisfactory predicting accuracy of mortality risk. Conclusion MELD is an accurate predictive system for 6 months and 1 year in liver cirrhotic patients with massive hemorrhage from esophageal varicosis treated by EIS. The MELD value is very important in the selection of patients for E1S. But it is not more efficient than CTP score and classification.