摘要
目的:基于临床特征与生化检验指标构建慢性乙型肝炎(CHB)发生肝硬化(LC)的风险预测模型。方法:回顾性分析2020年1月至2023年2月青海红十字医院367例CHB患者的临床资料,均随访至2024年4月。按照随访期间LC发生情况分为LC组(76例)和无LC组(291例)。比较两组临床特征及生化检验指标,经Logistic回归法分析CHB发生LC的影响因素。构建CHB患者发生LC的风险预测列线图模型,以受试者工作特征(ROC)曲线评估模型预测发生LC的效能,以Hosmer-Lemeshow拟合优度检验评估模型预测发生LC的校准度,以决策曲线分析(DCA)评估模型预测发生LC的获益情况。结果:随访8~49个月,中位数33(15,42)个月;LC发生率为20.71%(76/367);HBV标志物乙肝表面抗原(HBsAg)+乙型肝炎e抗原(HBeAg)+乙肝核心抗体(HBcAb)阳性(OR=2.522,95%CI:1.572~4.045)、红细胞体积分布宽度(RDW)(OR=2.208,95%CI:1.396~3.493)、总胆红素(TBIL)(OR=3.171,95%CI:1.554~6.472)、γ-谷氨酰转移酶(GGT)/白蛋白(ALB)(OR=2.252,95%CI:1.316~3.854)、凝血酶原时间(PT)(OR=2.085,95%CI:1.477~2.945)及肝纤维化4因子指数(FIB-4)(OR=3.428,95%CI:1.834~6.406)是发生LC的危险因素(P<0.05),胆碱酯酶(CHE)(OR=0.443,95%CI:0.307~0.639)是发生LC的保护因素(P<0.05);基于Logistic回归分析筛选出的有统计学意义的变量构建CHB患者发生LC的风险预测列线图模型,模型预测CHB患者发生LC的ROC曲线下面积(AUC)为0.946(95%CI:0.918~0.967),灵敏度为86.84%,特异度为91.41%;模型预测发生LC概率及实际发生LC概率比较差异无统计学意义(P>0.05);模型预测发生LC的阈值概率在0.23~0.98区间可实现临床获益。结论:CHB患者发生LC的危险因素包括HBV标志物HBsAg+HBeAg+HBcAb阳性、RDW、TBIL、GGT/ALB、PT及FIB-4,保护因素为CHE,以这些因素为基础构建的风险预测列线图模型对发生LC的预测价值高。
Objective:To establish a risk prediction model of liver cirrhosis(LC)in chronic hepatitis B(CHB)based on clinical features and biochemical indices.Methods:The clinical data of 367 patients with CHB in Qinghai Red Cross Hospital from January 2020 to February 2023 were retrospectively analyzed,and followed up until April 2024.According to the incidence of LC during the following-up period,the patients were divided into LC group(76 cases)and non-LC group(291 cases).The clinical characteristics and biochemical indices of the two groups were compared,and the influencing factors of LC in CHB were analyzed by Logistic regression method.A risk prediction nomogram model for LC in CHB patients was constructed.Receiver operating characteristic(ROC)curve was used to evaluate the model s efficacy in predicting LC,Hosmer-Lemeshow goodness of fit test was used to evaluate the model s calibration degree in predicting LC,and decision curve analysis(DCA)was used to evaluate the model s benefit in predicting LC.Results:The median following-up was 33(15,42)months for 8 to 49 months.The incidence of LC was 20.71%(76/367).HBV marker hepatitis B virus surface antigen(HBsAg)+hepatitis B virus e antigen(HBeAg)+hepatitis B virus core antibody(HBcAb)positive(OR=2.522,95%CI:1.572-4.045),red blood cell distribution width(RDW)(OR=2.208,95%CI:1.396-3.493),total bilirubin(TBIL)(OR=3.171,95%CI:1.554-6.472),γ-glutamyl-transferase(GGT)/albumin(ALB)(OR=2.252,95%CI:1.316-3.854),prothrombin time(PT)(OR=2.085,95%CI:1.477-2.945)and fibrosis index based on the 4 factor(FIB-4)(OR=3.428,95%CI:1.834-6.406)were risk factors for LC(P<0.05),and cholinesterase(CHE)(OR=0.443,95%CI:0.307-0.639)was protective factors for LC(P<0.05).Based on the statistically significant variables selected by Logistic regression analysis,the risk prediction nomogram model of LC in CHB patients was constructed.The area under ROC curve(AUC)was 0.946(95%CI:0.918-0.967),the sensitivity was 86.84%,and the specificity was 91.41%.There was no significant difference between the predicted probability of LC and the actual probability of LC(P>0.05).The model predicted that the threshold probability of LC in the range of 0.23 to 0.98 could achieve clinical benefit.Conclusion:The risk factors for LC in CHB patients include HBV marker HBsAg+HBeAg+HBcAb positive,RDW,TBIL,GGT/ALB,PT and FIB-4,and the protective factors is CHE,based on which the risk prediction column graph model constructed has a high value on predicting LC.
作者
赵仕荣
铁顺录
马国巍
ZHAO Shirong;TIE Shunlu;MA Guowei(Department of Clinical Laboratory,Qinghai Red Cross Hospital,Xining 810000,China;Department of Clinical Laboratory,the Fifth People's Hospital of Qinghai Province,Xining 810000,China;Department of Clinical Laboratory,the Fourth People's Hospital of Qinghai Province,Xining 810000,China)
出处
《东南大学学报(医学版)》
2025年第4期627-634,共8页
Journal of Southeast University(Medical Science Edition)
关键词
慢性乙型肝炎
肝硬化
临床特征
生化检验
chronic hepatitis B
liver cirrhosis
clinical features
biochemical test
