Objective To explore the predictive value of lactate⁃to⁃albumin ratio (LAR) on short⁃term prognosis in patients with liver failure. Methods A total of 842 patients with liver failure who were treated in ICU were screened from the MIMIC⁃Ⅳ database. Basic population information, relevant scores at admission, comorbidity, vital sign at first monitoring and laboratory indices at first detection after admission, treatment conditions, and prognostic status were collected. Patients were assigned to death group (346 cases) or survival group (496 cases) according to death status 28 days after admission. The univariate and multivariate COX regression model were used to analyze the influencing factors for all⁃cause mortality of patients with liver failure 28 days after admission. Through drawing the receiver operating characteristic (ROC) curve, the prediction efficiency between LAR and other scores on all⁃cause mortality of patients with liver failure 28 days after admission was compared. Patients with liver failure were assigned to high LAR group (n=278) or low LAR group (n=564), and their survival curve was compared between the two groups. The influencing factors for the correlation between LAR and all⁃cause mortality risk in patients with liver failure 28 days after admission were evaluated by the subgroup analysis. Results The results of COX regression analysis revealed that LAR was the independent predictive factor for all⁃cause mortality of patients with liver failure 28 days after admission (P<0.05). Area under the curve (AUC) of LAR for predicting all⁃cause mortality of patients with liver failure 28 days after admission was 0.669, and there was no statistically significant difference in it compared with Acute Physiology Score Ⅲ, Oxford Acute Severity of Illness Score, Sequential Organ Failure Assessment score, and Simplified Acute Physiology Score Ⅱ (P>0.05), but it was superior to systemic inflammatory response syndrome score (P<0.05); in addition, AUC of various scores combined with LAR was superior to AUC of non⁃combination (P<0.05). The all⁃cause mortality rate of patients 28 days after admission in the high LAR group was higher than that in the low LAR group (P<0.05). The results of subgroup analysis indicated that comorbidity condition of cardiac infarction and prothrombin time could affect relational degree between LAR and all⁃cause mortality risk of patients with liver failure 28 days after admission (interaction effect P value <0.05). Conclusion LAR can be regarded as the independent predictive indicator for all⁃cause mortality of patients with liver failure 28 days after admission, and its prediction efficiency is relatively ideal, which is helpful for the elevation of prediction efficiency of current scores.

无