Clinical Prediction Model For The Treatment Of Ursodeoxycholic Acid In Primary Biliary Cholangitis

Objective: Ursodeoxycholic acid(UDCA)is the first-line drug for the treatment of primary biliary cholangitis(PBC).But there are still some patients with poor response to it.Early identification and intervention can effectively improve the survival rate.Therefore,the purpose of this study was to explore the risk factors of 1-year non-response to UDCA treatment and to establish a predictive model to predict the non-responder to UDCA treatment earlier and provide more evidence for the treatment of PBC patients.Methods: We conducted a retrospective study of 346 patients with PBC who were first diagnosed from January 2017 to December 2021.All included patients were taking UDCA for the first time and continued to take UDCA for at least 1 year.Review regularly.The general information of the patients,relevant laboratory indicators before treatment,and liver function indicators after 1 year of UDCA treatment were collected,and the Paris I criteria were used as the response endpoint,and the patients were divided into a response group and a non-response group.Analyze and compare the baseline differences of indicators,identify independent risk factors,and propose predictive indicators.At the same time,this study proposes a predictive model in the form of scores.Based on the screened risk factors,the prediction is made by scoring against the probability,and finally through the ROC curve and bootstrap internal verification Validate the model.Results: The incidence of PBC was mainly concentrated in the age group of 50-70 years old,with a male-to-female sex ratio of about 1:9.8.Compared with females,male patients tended to develop at a later age.22.83% of patients had no obvious clinical manifestations.Abnormal liver function was found in physical examination or other examinations,and the common clinical symptoms were fatigue(25.14%),yellowing of skin and sclera(17.05%),skin itching(14.45%).28.90% of patients had extrahepatic immune diseases,the most common disease being Sj?gren’s syndrome(54%).According to Paris I criteria,the one-year response rate of UDCA was 78.32%,and the non-responder group was more likely to develop cirrhosis than the responder group,but age and gender differences were not related to the development probability of liver cirrhosis.After univariate and multivariate logistic analysis,age at first diagnosis,gender,gp210 antibody,ALP,ALB,and APRI were 6 independent risk factors associated with non-response to UDCA treatment for 1 year.Based on these factors,a prediction was established.The risk score prediction model with a value of 0.794,that is,the six indicators correspond to different scores at different stages,and the sum of the scores corresponds to the total score-probability table,which is the non-response probability of the patient.The calibration curve established by the Bootstrap method was verified internally.The two curves were in good agreement,and the C statistic was 0.716,indicating that the model had high discriminant and accuracy..Conclusion: We constructed a risk score prediction model,which can predict patients who do not respond to UDCA treatment for 1 year.It is helpful for early identification of patients who do not respond to UDCA treatment biochemically,and provides more evidence for the treatment of PBC patients.