Analysis of risk factors of liver cirrhosis complicated with mild hepatic enc-ephalopathy and construction of nomogram prediction model
Objective:To carry out targeted screening for the subgroup of patients with high MHE risk in liver cirrhosis patients and provide a clinical basis for early intervention of MHE patients.Methods:Selected patients with liver cirrhosis who were between 18 and 65 years old in the Inpatient Department of Infectious Diseases,Yan'an University Affiliated Hospital from August 2020 to December 2021.According to the test results(NCT-A,NCT-BC,SANT-1),the research subjects were divided into the MHE group and the non-MHE group,the risk factors related to MHE in patients with liver cirrhosis were analyzed,and ROC curve analysis was performed for each independent risk factor to evaluate the predictive effect of each independent risk factor,and further construct the liver cirrhosis combined with the risk factors.The nomogram of MHE predicts the model,and the model is internally validated and the ROC curve is drawn to evaluate the prediction effect of the model.Results:Univariate analysis showed that age,years of education,course of liver cirrhosis,ascites,history of splenectomy,nutritional risk,total protein,aspartate aminotransferase,albumin,prothrombin time.There were statistically significant differences in plasma prothrombin activity,hyaluronic acid,aspartate/glutathione,Child-Pugh grade,and MELD score.The multivariate analysis showed that age,hyaluronic acid greater than 120 ng/ml,Child-Pugh B,Grade C,and less than seven years of education are independent risk factors for MHE in patients with liver cirrhosis.And the predictive ability of the nomogram prediction model for liver cirrhosis with MHE is better than that of independent risk factors.Conclusion:Age,Child-Pugh grade,less than seven years of education,and hyaluronic acid greater than 120 ng/ml were found to be independent risk factors for MHE in patients with liver cirrhosis.The predictive ability of the nomogram prediction model has better calibration and discrimination.
NETL
NSTL
万方数据
维普
