摘要
目的筛选无创性预测原发性胆汁性胆管炎(PBC)患者熊去氧胆酸应答不佳的指标,并建立logistic回归模型。方法选取106例确诊PBC的患者为建模组,回顾性分析患者熊去氧胆酸治疗前及治疗后1年的临床资料。根据巴黎Ⅱ标准,将患者分为应答不佳组(n=38)和完全应答组(n=68)。采用非条件logistic回归分析的方法,分析患者的一般状况、化验指标及超声检查结果,筛选无创性预测应答不佳的指标,建立回归模型。另选取初次就诊于地坛医院的47例PBC患者为验证组,给予熊去氧胆酸治疗并随访1年,验证该模型。结果单因素分析表明,应答不佳组和完全应答组治疗前的胆碱酯酶、碱性磷酸酶、直接胆红素、γ-谷氨酰转肽酶、球蛋白、甘油三酯、总胆固醇、LDL-C、ESR、CD4^(+)T淋巴细胞计数、CD8^(+)T淋巴细胞计数比较差异均有统计学意义(P均<0.05)。而多因素非条件logistic回归分析表明,治疗前的胆碱酯酶、碱性磷酸酶、直接胆红素可有效预测应答不佳。回归模型为:Y=0.000339×胆碱酯酶(U/L)^(+)2.227292×(是否低于2倍正常上限的ALP;1=是,2=否)^(+)0.024151×直接胆红素(μmol/L)。该模型对应答不佳预测的灵敏度为88.6%,特异度为84.8%。经验证组检验,该模型的受试者操作特征曲线下面积为0.81(95%CI 0.69~0.94)。结论治疗前的胆碱酯酶、碱性磷酸酶、直接胆红素可预测PBC患者对熊去氧胆酸的应答不佳。
Objective To screen non-invasive indicators for predicting poor response to ursodeoxycholic acid in patients with primary biliary cholangitis(PBC)and establish a logistic regression model.Methods A total of 106 patients with PBC were selected in the modeling group.Clinical data of PBC patients before and after receiving 1 year ursodeoxycholic acid therapy were retrospectively analyzed.According to ParisⅡcriteria,all patients were divided into two groups:poor response group(n=38)and complete response group(n=68).The general condition,laboratory indicators and ultrasound examination results of the patients were analyzed by nonconditional logistic regression model before and after treatment.Non-invasive indicators for predicting poor response to ursodeoxycholic acid were screened,and the regression model was established.To validate this model,47 patients with PBC who visited Ditan Hospital for the first time were selected in the validation group,and were treated with ursodeoxycholic acid for 1 year.Results Univariate analysis showed that in the modeling group,there were significant differences in cholinesterase,alkaline phosphatase,direct bilirubin,γ-glutamyltranspeptidase,globulin,total triglyceride,total cholesterol,low-density lipoprotein cholesterol,erythrocyte sedimentation rate,CD4^(+)T cel count and CD8^(+)T cel count between the poor response and complete response groups before treatment(al P<0.05).Multivariate unconditional logistic analysis showed that cholinesterase,alkaline phosphatase and direct bilirubin before treatment were effective predictors of poor response.The regression model was:Y=0.000339×cholinesterase(U/L)^(+)2.227292×(whether the ALP is lower than 2 times of normalupper limit;1=yes,2=no)^(+)0.024151×DBIL(μmol/L).The sensitivity and specificity of the model for predicting poor response were 88.6%and 84.8%,respectively.In this validation cohort,the area under the receiver operating characteristic curve was 0.81(95%CI 0.69-0.94).Conclusion Cholinesterase,alkaline phosphatase and direct bilirubin before treatment can predict poor response to ursodeoxycholic acid in PBC patients.
作者
高丽丽
潘仕达
张亦瑾
高学松
苏楠
徐琍
段雪飞
Gao Lili;Pan Shida;Zhang Yijin;Gao Xuesong;Su Nan;Xu Li;Duan Xuefei(Department of General Medicine,Beijing Ditan Hospital,Capital Medical University,Beijing 100015,China;不详)
出处
《新医学》
CAS
2023年第2期127-131,共5页
Journal of New Medicine
基金
北京市医管局培育计划(PZ2022030)。
关键词
原发性胆汁性胆管炎
应答不佳
无创性预测
熊去氧胆酸
Primary biliary cholangitis
Poor response
Non-invasive prediction
Ursodeoxycholic acid
