Establishment Of A Survival Prediction Model For Patients With Cirrhosis Combined With Hepatic Encephalopathy In The Intensive Care Unit

Objective:The purpose of this study was to establish a comprehensive survival prediction model for patients with cirrhosis combined with hepatic encephalopathy(HE)in the intensive care unit(ICU)and to evaluate the predictive performance of the model to improve the prediction of 90-day survival of patients.Methods:This study analyzed 620 patients with liver cirrhosis complicated with HE from the America Medical Information Mart for Intensive Care III(MIMIC-III)database.The patients were randomly divided into two groups in a 7-to-3 ratio to form a training cohort(n = 434)and a test cohort(n = 176).Cox regression analyses were used to identify associated risk variables.Based on the results of the multivariate Cox regression model,a nomogram was established using associated risk predictor variables to predict the 90-day survival rate of patients with cirrhosis complicated with HE.The new model was compared with the Sequential Organ Failure Assessment(SOFA)scoring model in terms of the concordance index(C-index),the area under the curve(AUC)of receiver operating characteristic(ROC)analysis,the net reclassification index(NRI),the integrated discrimination improvement(IDI),calibration curve,and decision curve analysis(DCA).Results:1.Eight independent potential predictors for 90-day mortality were screened by univariate and multifactorial Cox regression analysis: just as age,mean heart rate,mean temperature,mean arterial pressure,SOFA,red blood cell distribution width(RDW),the use of proton pump inhibitor and albumin infusion.Among them:advanced age,higher SOFA score,higher RDW,lower mean body temperature,faster mean heart rate,and the use of albumin were risk factors for patient prognosis.Higher mean arterial pressure and proton pump inhibitor use were protective factors.2.An eight-factor nomogram about the 90-day survival prediction model for patients with liver cirrhosis and HE was constructed based on the results of multivariate Cox regression analysis.The nomogram could match the exact value of each predictor to the corresponding score on the score scale,and sum all indicator scores to obtain a total score,which matches the corresponding predicted 90-day survival rate.3.Evaluation of prediction model performance:(1)AUC of the 90-day survival ROC curve(0.743 for training cohort and 0.733 for test cohort)and C-index(0.704 for training cohort and 0.695 for test cohort)showed that the new model had a better discriminatory ability.(2)In comparison with the SOFA alone model,the NRI value of the 90-day survival prediction probability of the new model was 0.560 in the training cohort and 0.364 in the test cohort;in addition,the IDI value of the 90-day survival prediction probability of the new model was 0.119 in the training cohort and 0.083 in the test cohort,respectively.The NRI and IDI values indicated that the predictive ability and overall performance of the new model were better than that of the SOFA model alone.(3)Calibration curves were shown: the calibration curves of the new model on the training and test cohort were very close to the standard 45-degree diagonal,with the associated four tangent points evenly distributed.The finding showed that the new model had excellent calibration capabilities.(4)The results of the decision curves demonstrated that the new model decision curves produced net clinical values for both the training and test cohort.Conclusion:This study is the first study of the 90-day mortality prediction of cirrhosis patients with HE in ICU through the data of the MIMIC-III database.Independent factors influencing the short-term survival of patients with cirrhosis combined with hepatic encephalopathy are identified,and a 90-day survival prediction model nomogram for patients is successfully constructed.It is confirmed that the eight-factor nomogram has a good efficiency in predicting the 90-day survival rate of patients.