作者：董鲜桃，张永杰，朱姝，孙倩，席从林，李娟，顾润环

单位：

223001江苏省淮安市第二人民医院肿瘤科

Units：

Department of Oncology, Huaian Second People's Hospital, Huaian 223001, China

关键词：

肺癌；化疗；经外周静脉穿刺的中心静脉导管；上肢深静脉血栓；影响因素分析；列线图模型

Keywords：

Lung cancer; Chemotherapy; Peripherally inserted central venous catheter; Upper extremity deep venous thrombosis; Root cause analysis; Nomogram model

CLC：

DOI：

10.12114/j.issn.1008-5971.2022.00.216

Funds：

江苏省自然科学基金资助项目（13KJB350006）

摘要：

目的　探讨肺癌化疗患者经外周静脉穿刺的中心静脉导管（PICC）置管后发生上肢深静脉血栓（UEDVT）的危险因素，并构建其风险预测列线图模型。方法　选取2019年9月至2021年10月在淮安市第二人民医院肿瘤科行PICC置管的296例肺癌化疗患者为研究对象，根据PICC置管后是否发生UEDVT将患者分为UEDVT组和非UEDVT组。收集两组患者临床资料、穿刺置管情况及实验室检查指标，肺癌化疗患者PICC置管后发生UEDVT的影响因素分析采用多因素Logistic回归分析。应用R软件构建肺癌化疗患者PICC置管后发生UEDVT的风险预测列线图模型，绘制ROC曲线以评价该列线图模型对肺癌化疗患者PICC置管后UEDVT发生风险的区分能力，采用拟合优度检验评价该列线图模型对肺癌化疗患PICC置管后UEDVT发生风险的校准能力。结果　296例患者中，51例发生UEDVT，UEDVT发生率为17.2%。UEDVT组中有糖尿病、血栓史、肿瘤分期＞Ⅱ期、导管末端位置在上腔静脉上2/3者占比及血浆D-二聚体（D-D）高于非UEDVT组，置管时间长于非UEDVT组（P ＜0.05）。多因素Logistic回归分析结果显示，糖尿病、肿瘤分期＞Ⅱ期、导管末端位置在上腔静脉上2/3、置管时间延长、血浆D-D升高是肺癌化疗患者PICC置管后发生UEDVT的危险因素（P ＜0.05）。将上述危险因素引入R软件，构建肺癌化疗患者PICC置管后发生UEDVT的风险预测列线图模型。ROC曲线分析结果显示，该列线图模型预测肺癌化疗患者PICC置管后发生UEDVT的曲线下面积为0.787〔95%CI （0.718，0.856）〕，最佳截断值为0.186，灵敏度为0.863，特异度为0.894。拟合优度检验结果显示，该列线图模型预测肺癌化疗患者PICC置管后发生UEDVT的预测概率与肺癌化疗患者PICC置管后发生UEDVT的实际概率比较，差异无统计学意义（P ＞0.05）。结论　糖尿病、肿瘤分期＞Ⅱ期、导管末端位置在上腔静脉上2/3、置管时间延长、血浆D-D升高是肺癌化疗患者PICC置管后发生UEDVT的危险因素，而基于上述危险因素构建的列线图模型具有较好的区分能力和校准能力。

Abstract：

Objective　To investigate the risk factors of upper extremity deep venous thrombosis (UEDVT) afterperipherally inserted central venous catheter (PICC) catheterization in lung cancer patients receiving chemotherapy, and toconstruct a risk prediction nomograph model. Methods　A total of 296 lung cancer patients receiving chemotherapy who receivedPICC catheterization in the Department of Oncology, Huaian Second People's Hospital from September 2019 to October 2021 wereselected as the research objects. According to whether UEDVT occurred after PICC catheterization, the patients were divided intoUEDVT group and non UEDVT group. The clinical data, puncture and catheterization conditions and laboratory examination indexesof the two groups were collected. The influencing factors of UEDVT after PICC catheterization in lung cancer patients receivingchemotherapy were analyzed by multivariate Logistic regression analysis. The risk prediction nomograph model of UEDVT afterPICC catheterization in lung cancer patients receiving chemotherapy was constructed by R software, the ROC curve was drawnto evaluate the distinguishing ability of the nomogram model for predicting the risk of UEDVT after PICC catheterization in lungcancer patients receiving chemotherapy, the goodness of fit test was used to evaluate the calibration ability of the nomogram modelfor predicting the risk of UEDVT after PICC catheterization in lung cancer patients receiving chemotherapy. Results　Among 296patients, 51 cases developed UEDVT, and the incidence of UEDVT was 17.2%. The proportion of patients with diabetes mellitus,history of thrombosis, tumor stage > Ⅱ, the position of the catheter tip above the superior vena cava in two-thirds and plasmaD-dimer (D-D) in UEDVT group were higher than those in non UEDVT group, and the catheterization time was longer than thatin non UEDVT group (P < 0.05) . Multivariate Logistic regression analysis results showed that diabetes mellitus, tumor stage > Ⅱ,the position of the catheter tip above the superior vena cava in two-thirds, plolonged catheterization time, plasma D-D elevationwere the risk factors for UEDVT after PICC catheterization in lung cancer patients receiving chemotherapy (P < 0.05) . The aboverisk factors were introduced into R software to construct a risk prediction nomograph model of UEDVT after PICC catheterizationin lung cancer patients receiving chemotherapy. ROC curve analysis results showed that the area under the curve of the nomogrammodel for predicting UEDVT after PICC catheterization in lung cancer patients receiving chemotherapy was 0.787 [95%CI (0.718,0.856) ] , the optimum cut-off value was 0.186, the sensitivity was 0.863 and specificity was 0.894. The goodness of fit test showedthat there was no significant difference between probability of UEDVT predicted by the nomogram model and the actual probabilityof UEDVT in lung cancer patients receiving chemotherapy after PICC catheterization (P > 0.05) . Conclusion　 Diabetes mellitus,tumor stage > Ⅱ, the position of the catheter tip above the superior vena cava in two-thirds, prolonged catheterization time,plasma D-D elevation are the risk factors for UEDVT after PICC catheterization in lung cancer patients receiving chemotherapy.The nomograph model constructed based on the above risk factors has good discrimination ability and calibration ability.

ReferenceList：