A Comparison of Child-Turcotte-Pugh Classification with Four Models of End-Stage Liver Disease-Based Prognostic System for Spontaneous Bacterial Peritionitis in Cirrhosis

Background and Aims: For the prediction of the prognosis of spontaneous bacterial peritonitis (SBP) in patients of cirrhosis, no direct comparisons have been made among the five models, Child-Turcotte-Pugh classification (CTP), the model for end-stage liver disease (MELD), the model for end-stage liver disease with the incorporation of serum sodium (MELD-Na), the integrated model for end-stage liver disease (iMELD) score, and the model for end-stage liver disease to sodium (MESO) index. Materials and Methods: Between January 2005 and December 2006, 190 patients who met the criteria for liver cirrhosis with SBP for the first time were enrolled in this retrospective study. Patients' clinical and laboratory data were obtained at diagnosis, and the Child-Turcotte-Pugh (CTP) and 4 MELD-based scores were calculated accordingly. Patients were followed up until November 2009 or until death. Results: The in-hospital mortality rate was 38.9%. The cumulative 3-month and 6-month mortality rates were 52.6% and 63.1%, respectively. The CTP score and 4 MELD-based model scores could all significantly predict the in-hospital, 3-month, and 6-month mortality. For in-hospital mortality, only iMELD had a significantly higher area under curve (AUC) in comparison with the CTP score (p=0.039), and the other comparisons showed no significant difference. For 3-month and 6-month mortality, all 4 MELD-based models had significantly better prediction abilities than the CTP classification. iMELD had the best AUC, followed by MELD-Na, MESO, and MELD, but there were no statistical differences between them. Conclusions: All of the 4 MELD-based models, iMELD, MELD-Na, MESO, and MELD, have significantly better prediction abilities than the CTP classification on 3-month and 6-month mortality of SBP patients. For the prediction of in-hospital mortality, only iMELD had significantly superior prediction ability over the CTP score.