摘要
目的探讨乙肝肝硬化患者食管胃底静脉曲张(EGV)的影响因素,构建列线图并进行验证。方法将2018年1月—2022年12月江苏大学附属镇江三院诊断为乙肝肝硬化的患者303例作为研究对象,按照7∶3的比例随机分为训练集和验证集,根据胃镜结果将患者分为无EGV组与EGV组,收集患者的一般资料、实验室及影像学指标。通过LASSO回归筛选变量及多因素Logistic回归分析影响因素,以此构建列线图并进行验证,通过绘制受试者操作特征曲线(ROC)、校准曲线(Calibration)、决策曲线(DCA)、临床影响曲线(CIC)评估模型的临床实用价值。结果将纳入研究的乙肝肝硬化患者分为训练集211例与验证集92例。LASSO回归筛选出7个最可能的非零系数特征变量:性别、肝硬度值(LSM)、血小板计数(PLT)、总胆红素(TBIL)、血红蛋白(HGB)、红细胞计数(RBC)、凝血酶原时间(PT)。多因素Logistic回归分析显示,性别(男性)(OR=4.830,95%CI:1.479~15.775,P=0.009)、高肝硬度值(LSM)(OR=1.132,95%CI:1.009~1.270,P=0.034)、高胆红素(TBIL)(OR=1.133,95%CI:1.057~1.214,P<0.001)是乙型肝炎肝硬化患者EGV的独立危险因素,高血小板(PLT)(OR=0.978,95%CI:0.967~0.989,P<0.001)是乙型肝炎肝硬化患者EGV的保护因素。通过该4项指标构建列线图,模型在训练集和验证集中ROC曲线下面积分别为0.845(95%CI:0.793~0.896,P<0.001)和0.932(95%CI:0.877~0.987,P<0.001)。训练集和验证集Calibration曲线平均绝对误差分别为0.040和0.036,校准曲线与理想曲线大体一致。训练集与验证集DCA曲线显示列线图有一定的临床实用性。CIC曲线显示该列线图可对高风险人群进行分层预测。结论研究构建的列线图具有较高的临床实用价值,可为临床医师初步判断乙肝肝硬化患者EGV的风险提供依据。
Objective To explore the influencing factors of esophagogastric varices(EGV)in patients with hepatitis B cirrhosis,and to establish and validate the nomogram model.Methods Three hundred and three patients who were diagnosed with hepatitis B cirrhosis in the Third People's Hospital of Zhenjiang from January 2018 to December 2022 were selected as the research objects.The training set and validation set were randomly constructed in a ratio of 7:3.According to the results of endoscopic examination of esophageal and gastric varices,the patients were divided into varices group and non-varices group.The general information,laboratory and imaging indicators were collected.The variables were screened by LASSO regression,and the nomogram were constructed and verified by multivariate Logistic regression analysis of influencing factors.The clinical utility value of the model was assessed by receiver operator characteristic(ROC)curve,calibration curve,decision curve(DCA),and clinical impact curve(CIC).Results The patients with hepatitis B cirrhosis in the study were divided into the training set(n=211)and validation set(n=92).The seven most likely non-zero coefficient characteristic variables were screened by LASSO regression:gender,LSM,PLT,TBIL,HGB,RBC,PT.Multivariate Logistic regression showed that gender(male)Multivariate Logistic regression showed that gender(male)(OR=4.830,95%CI:1.479~15.775,P=0.009),high LSM(OR=1.132,95%CI:1.009~1.270,P=0.034)and high TBIL(OR=1.133,95%CI:1.057~1.214,P<0.001)were independent risk factors for the EGV in patients with hepatitis B cirrhosis,while High PLT(OR=0.978,95%CI:0.967~0.989,P<0.001)was protective factor was protective factor.The nomogram was constructed with the above 4 influencing factors,ROC curves showed that the area under the curve(AUC)was 0.845(95%CI:0.793~0.896,P<0.001)in the training set,and 0.932(95%CI:0.877~0.987,P<0.001)in the validation set.The average absolute error of the training set was 0.040,and the validation set was 0.036,the calibration curve and the ideal curve were basically fitted.DCA curves of the training and validation sets showed that the nomogram had certain clinical practicability.CIC curves showed that the nomogram model can be used to forecast the high-risk population in different levels.Conclusion The nomogram constructed in this study has a high clinical utility and can provide evidence for clinicians to preliminary judge the risk of EGV in patients with hepatitis B cirrhosis.
作者
唐舒
邹圣强
董邓波
唐超
姚雅极
TANG Shu;ZOU Shengqiang;DONG Dengbo;TANG Chao;YAO Yaji(North China University of Science and Technology,Tangshan 063210,China)
出处
《中国煤炭工业医学杂志》
2024年第3期297-307,共11页
Chinese Journal of Coal Industry Medicine
基金
江苏省学生体质健康促进研究中心研究课题(编号:2022B008)
2021年度镇江市重点研发项目(编号:SH2021030)
镇江市社会发展指导性科技计划项目(编号:FZ2021068)。
