Model for end-stage liver disease score and serum natrium level in predicting short-term prognosis of liver failure

Objective To evaluate the model for end-stage liver disease (MELD) and MELD combined with serum natrium level (MELD-Na) in predicting short-term prognosis of liver failure. Methods Clinical data of 322 patients with liver failure admitted in the First Affiliated Hospital of Xinjiang Medical University from April 2003 to April 2012 were retrospectively analyzed. MELD and MELD-Na scores were calculated at diagnosis and one week after the diagnosis, and then ΔMELD and ΔMELD-Na were determined. Receiver operating characteristics (ROC) curve and Kaplan-Meier survival curve were used to evaluate the value of the above scores in predicting 3-month prognosis. Results The 3-month mortality rates of acute/sub-acute, acute-on-chronic and chronic liver failure were 77.4%(24/31), 41.7%(50/120) and 56.1%(96/171), respectively, and the difference was of statistical significance (χ2=14.273, P 0.05). For acute-on-chronic liver failure, AUCs were 0.889 and 0.897 for ΔMELD and ΔMELD-Na in predicting short-term prognosis, which were higher than those of MELD and MELD-Na scores at the baseline (Z=3.110 and 3.500, P 0.05); Kaplan-Meier survival curve showed that the 3-month mortality rate for patients with ΔMELD>3.5 was 87.8%, and the average survival time was 34.05 d. For chronic liver failure, AUC of ΔMELD was 0.871 in predicting short-term prognosis, which was higher than that of ΔMELD-Na (Z=4.229, P 4.5 was 89.9%, and the average survival time was 29.08 d. Conclusion For acute/sub-acute liver failure, MELD, MELD-Na, ΔMELD and ΔMELD-Na are all satisfactory in predicting short-term prognosis; for acute-on-chronic liver failure, ΔMELD and ΔMELD-Na are better than MELD, MELD-Na scores at the baseline; and for chronic liver failure, ΔMELD is the best indicator. Key words: Liver failure; Prognosis; Model for end-stage liver disease; Model for end-stage liver disease-Na