Abstract: Objective To investigate etiology, clinical characteristics, artificial liver treatment and prognosis-related influencing factors of patients with liver failure in intensive care unit (ICU). Methods A retrospective study was used to analyze the etiology, laboratory tests, major complications/comorbidities, mode and timing of artificial liver therapy and prognosis-related influencing factors in 140 patients with liver failure who received artificial liver therapy in our hospital. Results A total of 140 patients with liver failure who received artificial liver support system were followed up for 60 days, 73 patients (52.1%) in the survival group and 67 patients (47.9%) in the death group. There were statistically significant differences in age (t=-2.325, P=0.022), length of stay (t=-3.439, P=0.001), etiology (χ²=50.239, P<0.001) and classification of liver failure (χ²=12.779, P=0.002) between the two groups (P<0.05). APACHE Ⅱ score (t=-6.792, P<0.001), end-stage liver disease model (MELD) (t=-4.079, P<0.001), total bilirubin (TBil) (t=-3.115, P=0.002), international standardized ratio (INR) (t=-3.839, P<0.001), serum creatinine (SCr) (t=-2.408, P=0.018), urea nitrogen (UREA) (t=-2.042, P=0.043), white blood cell (WBC) (t=-4.853, P<0.001) and procalcitonin (PCT) (Z=-3.11, P=0.002)in survival group were significantly lower than those in death group. Prothrombin activity (PTA) (t=3.443, P=0.001) was significantly higher than that in death group (P<0.05). The incidence of acute kidney injury (χ²=7.312, P=0.007) in survival group was significantly lower than that in death group (P<0.05). There was no significant difference in the incidence of hepatic encephalopathy and electrolyte disorder between the 2 groups (P>0.05). Patients in the 2 groups underwent bilirubin adsorption (PBA), plasma exchange (PE), dual plasma molecular adsorption (DPMAS) and dual plasma molecular adsorption combined with plasma exchange (DPMAS+PE) in four different modes of artificial liver treatment, and the difference in their artificial liver treatment modes was not statistically significant (P>0.05), but there was statistical significance in the period of artificial liver therapy (χ²=10.419, P=0.005) between 2 groups (P<0.05). Multivariate COX regression model showed that age (HR=1.04, 95%CI 1.017-1.065, P=0.001), advanced liver failure (HR=4.889, 95%CI 1.103-21.676, P=0.037), APACHE Ⅱ score (HR=1.085, 95%CI 1.032-1.141, P=0.001), INR (HR=3.089, 95%CI 1.178-8.097, P=0.022), TBil (HR=1.006, 95%CI 1.002-1.01, P=0.006), SCr (HR=1.011, 95%CI 1.001-1.021, P=0.032) and PCT (HR=1.023, 95%CI 1.006-1.04, P=0.009) were independent risk factors for prognosis. Conclusion Different modes of artificial liver support system have no significant effect on the survival of patients with liver failure in ICU, but are associated with the timing of patients receiving artificial liver treatment. Age, advanced liver failure, APACHE Ⅱ score, INR, TBIL, SCr and PCT are independent risk factors for prognosis.

Key words: Liver failure, Artificial liver support system, Survival analysis