Abstract: Objective To establish a new model for predicting short-term prognosis of patients with acute liver failure. Methods Relevant laboratory indicators of the patients with acute liver failure were collected to screen out meaningful clinical indicators. Independent risk factors related to prognosis were screened by univariate and multivariate Cox regression analysis, and a model for predicting short-term prognosis was established. The risk score was calculated according to the new model, and the median risk score was taken as the critical point to divide the patients into high-risk group and low-risk group, and the difference in survival rate between the two groups was compared. ROC analysis was used to compare the predictive efficacy of the model and MELD scoring system in different time periods. Results A total of 103 patients were included in the study. Four independent risk factors including prealbumin (PA) (HR=0.99,95%CI:0.98~1.00) , lactic acid (Lac) (HR=1.22,95%CI:1.13~1.31) , platelet count (PLT) (HR=0.99,95%CI:0.99~1.00) and International normalized ratio (INR) (HR=1.14,95%CI:1.02~1.27) were screened out by univariate and multifactorial Cox regression analysis. A prognosis prediction model equation for acute liver failure was constructed. The expression formula of the model is as follows: Risk score = 0.195*Lac + 0.178*INR-0.012*PA-0.008*PLT. The median survival rate of the low-risk group was significantly higher than that of the high-risk group (P<0.01). The new model was superior to MELD score in predicting the 30 d outcome[AUROC 0.88 (95%CI 82.25~94.45) vs 0.75 (95%CI 65.56~84.74), P<0.05].The new model was better than MELD score in predicting 90 d outcome [AUROC 0.84 (95%CI 76.67~91.59) vs 0.75 (95%CI 64.87~84.25), P<0.05]. Conclusion The new model including PA, Lac, PLT and INR could predict the short-term prognosis of acute liver failure successfully.

Key words: Acute liver failure, Prognosis, Nomogram, MELD score