Establishment And Validation Of Clinical Prediction Models In Patients With Early Hepatic Encephalopathy Undergoing Transjugular Intrahepatic Portosystemic Shunt

Background/AimsTransjugular intrahepatic portosystemic shunt(TIPS)could rapidly reduce portosystemic pressure through portosystemic shunt.It had become the highlight in clinical treatment,especially in the treatment of the patients with decompensated portal hypertension due to cirrhosis,since it had the advantages of minimal trauma,rapid recovery and remarkable curative effect.But at the same,the incidence rate of hepatic encephalopathy(HE)was actually increasing,which seriously affected the treatment effects of patients undergoing TIPS.In order to reduce the incidence of HE after TIPS,many scholars had made great efforts to actively explore and seek for the relevant risk factors of HE after TIPS.Unfortunately,there were no quantitative tools for forecasting risk and assessing benefit of the patients with HE after TIPS.In this regard,the paper took patients with HE in the early stage after TIPS as the research objects and assessed comprehensively the relevant risk factors of HE after TIPS in clinical treatment at present.At the same time,a clinical prediction model of early HE after TIPS based on logistic regression was constructed to provide a scientific quantitative tool for forecasting the individual risks and assessing the benefits of HE patients.MethodsIt collected the clinical data of 262 patients in Nanfang hospital and 75 patents in Guizhou provincial people's hospital for further retrospective analysis,then 70%of the data collected from 262 patients in Nanfang hospital(183 in total)was randomly chosen as the training set for establishing the clinical prediction model and the rest 30%(79 in total)was used for internal verification.All the data of 75 patients collected from Guizhou provincial people's hospital were used for external verification.Then,the data in the training set was used for single and multiple factors logsitic regression analysis and the screened independent affecting factors were used for establishing the clinical prediction model.Afterwards,the assessment on clinical efficacy in three aspects as discrimination ability,accuracy and clinical practicability was carried out by adopting the area under the ROC curve,calibration plot,decision curve analysis respectively.Furthermore,the prediction model was shown by three forms as formula,risk scoring system and nomogram.Finally,the internal and external verification was conducted respectively for the prediction model to evaluate the accuracy and stability of the model.Results1.Multivariate logistic regression analysis showed that age(OR=7.207,P<0.05),Child-Pugh grade(OR=3.034,P<0.05)and ln(fibrinogen)(ORF=0.178,P<0.05)were the independent affecting factors of the patients with HE in the early stage after TIPS.Logistic regression model was logitP=-4.829+1.975�age+1,110�Child-Pugh grade-1.724�ln(fibrino gen).2.The 3-year cumulative survival rates of patients with and without hepatic encephalopathy within 3 months after TIPS were 59%and 85%,respectively.Multivariate COX regression survival analysis and survival curve showed that patients without early HE after TIPS had better long-term survival than patients with HE.3.The predictive model constructed by training set was used to evaluate the clinical efficacy:the area under the ROC curve was 0.805;the calibration Hosmer-Lemeshow test was ?2=5.98,P=0.817.The internal and external validation of the prediction model showed that the area under the ROC curve was 0.816 and 0.765,respectively.The calibration Hosmer-Lemeshow test was ?2=5.20,P=0.877 and?2=12.26,P=0.268,respectively.Decision curve analysis in both internal and external validation groups showed that the predictive model had good clinical practicability.Conclusions1.Age,Child-Pugh grade and fibrinogen were the independent affecting factors of the patients with hepatic encephalopathy in the early stage after TIPS.2.The long-term survival of patients without HE within 3 months after TIPS was better than those with HE.3.The clinical prediction model based on age,Child-Pugh grade and fibrinogen is accurate,stable and of clinical application value.4.This project constructs a nomogram for predicting early hepatic encephalopathy after TIPS,which provides an effective reference tool for clinical quantitative assessment of risk and benefit.It is helpful for early warning and treatment of HE after TIPS,and for doctors and patients to make scientific clinical decisions.