中心静脉导管相关静脉血栓风险预测模型构建和评价  

Construction and evaluation of risk prediction model of central venous catheter related venous thrombosis

作　　者：郑美琼[1] 程春燕 闫凡 蔡林霖 曹政芳 付莹[1] 张帆[1] 陈春晓 Zheng Meiqiong;Cheng Chunyan;Yan Fan;Cai Linlin;Cao Zhengfang;Fu Ying;Zhang Fan;Chen Chunxiao(Henan Provincial People's Hospital,Zhengzhou 453600,Henan,China)

机构地区：[1]河南省人民医院,河南郑州453600

出　　处：《临床心身疾病杂志》2024年第5期119-124,共6页Journal of Clinical Psychosomatic Diseases

基　　金：河南省医学科技攻关计划联合共建项目(编号LHGJ20190585)。

摘　　要：目的构建血液内科化疗患者超声平面技术下贵要静脉置入中心静脉导管(PICC)相关静脉血栓(RVT)风险的预测模型,并评价其效果。方法将175例血液内科化疗患者作为研究对象,均行超声平面技术下贵要静脉PICC置管,收集临床数据,分析PICC-RVT发生的影响因素,构建列线图预测模型,评价其预测价值。结果175例化疗患者PICC-RVT发生率为19.43%。Logistic回归分析显示,合并糖尿病、PICC置管史、血栓史、化疗史、导管尖端位置、导管留置时间、院外维护场所、置管前D-二聚体(D-D)是PICC-RVT发生的独立影响因素(P<0.01)。构建血液内科化疗患者超声平面技术下贵要静脉PICC-RVT发生的个体化列线图预测模型,经检验,该预测模型拟合良好,校准度、预测能力均较高。受试者工作特征(ROC)曲线显示,该模型预测血液内科化疗患者超声平面技术下贵要静脉PICC-RVT的曲线下面积为0.926(95%CI:0.882~0.971),灵敏度为91.18%,特异度为79.43%。结论合并糖尿病、PICC置管史、血栓史、化疗史、导管尖端位置、导管留置时间、院外维护场所、置管前D-D是血液内科化疗患者超声平面技术下贵要静脉PICC-RVT发生的影响因素,根据上述因素构建列线图模型具有可行性,且预测价值可靠。Objective To build a risk prediction model for venous thrombosis caused by peripherally inserted central catheter(PICC)guided by ultrasound planar technique(UPT)in patients undergoing chemotherapy in the department of hematopathology and evaluate its effect.Methods A total of 175 patients undergoing chemotherapy in the hematology department were selected as research objects,all patients underwent ultrasound-guided PICC placement in basilic vein,clinical data were collected,influencing factors of PICC-related venous thrombosis(RVT)analyzed,nomogram prediction models constructed,and their predictive values evaluated.Results The incidence of PICC-RVT was 19.43%in 175 patients undergoing chemot herapy.Logistic regression analysis showed that comorbid diabetes,the history of PICC catheterization,thrombosis and chemotherapy,catheter tip position,catheter retention time,off-hospital maintenance place,and pre-catheterization(D-d imer,D-D)were independent factors influencing PICC-RVT(P<0.01),construction of individualized nomogram prediction model guided by the UPT for the occurrence of PICC-RVT of basilic vein in patients undergoing chemotherapy in the department of hematology had good fit and higher calibration and predictive ability via testing,the receiver operating characteristic(ROC)curve showed that the area under ROC curve of the model predicting PICC-RVT of basilic vein was 0.926(95%CI:0.882~0.971),sensitivity 91.18%,and specificity 79.43%.Conclusion Comorbid diabetes,the history of PICC catheterization,thrombosis and chemotherapy,catheter tip position,catheter retention time,off-hospital maintenance place,and pre-catheterization D-D were factors influencing the occurrence of PICC-RVT in basilic vein,the nomogram model based on the above factors is feasible and has reliable prediction value.

关 键 词：化疗 超声平面技术 贵要静脉 中心静脉导管 静脉血栓 影响因素 列线图模型 

分 类 号：R453[医药卫生—治疗学]