[Clinical study of survival time for chronic liver failure.]

To identify independent risk factors influencing the survival time of patients with chronic liver failure and construct a predictive model.

Retrospective analysis was applied to clinical data of 362 patients with chronic liver failure treated with artificial liver in Tianjin third centre hospital between May 2002 and May 2007. Data were analyzed with SPSS 13.0 statistic software, t test and rank test were used on quantitative data, chi-square test was used on qualitative data, Cox regression analysis was used to select the independent risk factors influencing the survival time. According to independent risk factors from Cox regression model, a prognostic model was established.

1. Independent risk factors (P less than 0.05) influencing the survival time were: Child-Pugh score, bilirubin separation ALT, ascites, arginine, age, tyrosine and serum sodium. 2. By receiver operating characteristic curves (ROC) analysis, the area under ROC (AUR) to predict the outcome of chronic liver failure patients was 0.782, and the cutoff score was 27.69.

1. Child-Pugh score, bilirubin separation ALT, ascites, arginine, age, tyrosine and serum sodium are independent risk factors affecting survival time of patients with chronic liver failure. 2. Cox model we constructed can reliably predict the survival time of patients with chronic liver failure.