作者：郝其艳，王炜

单位：

224000江苏省盐城市第三人民医院（北院）肿瘤科

Units：

Department of Oncology, Yancheng Third People's Hospital (North Hospital) , Yancheng 224000, China

关键词：

肺肿瘤；中心静脉导管；导管相关性血流感染；影响因素分析；列线图

Keywords：

Lung neoplasms; Central venous catheters; Catheter-related bloodstream infection; Root cause analysis;Nomograms

CLC：

R 734.2

DOI：

10.12114/j.issn.1008-5971.2023.00.174

Funds：

江苏省2017年度省第五期“333工程”科研项目资助（BRA2017222）

摘要：

目的　探讨肺癌化疗患者发生导管相关性血流感染（CRBSI）的影响因素，构建其风险预测列线图模型并进行验证。方法　选取2010年1月至2022年11月盐城市第三人民医院收治的肺癌化疗患者1 169例为研究对象，按照6∶4的比例将其分为建模集（701例）及验证集（468例）。收集所有患者的临床资料，根据是否发生CRBSI将建模集患者分为发生组和未发生组。采用多因素Logistic回归分析探讨肺癌化疗患者发生CRBSI的影响因素；采用R 4.1.2软件包及rms程序包建立肺癌化疗患者发生CRBSI的风险预测列线图模型；采用Hosmer-Lemeshoe拟合优度检验评价该列线图模型的拟合程度；绘制校准曲线以评估该列线图模型预测建模集及验证集肺癌化疗患者发生CRBSI的效能；采用ROC曲线分析该列线图模型对建模集及验证集肺癌化疗患者发生CRBSI的预测价值。结果　建模集701例肺癌化疗患者中，发生CRBSI 71例（10.13%），未发生CRBSI 630例（89.87%）。两组肿瘤分期、营养状况、有糖尿病者占比、化疗次数、有重症监护病房住院史者占比、导管维护时间延长者占比、导管移动者占比、导管留置时间、穿刺次数比较，差异有统计学意义（P＜0.05）。多因素Logistic回归分析结果显示，糖尿病、化疗次数、重症监护病房住院史、导管维护时间延长、导管移动、导管留置时间、穿刺次数是肺癌化疗患者发生CRBSI的影响因素（P＜0.05）。基于多因素Logistic回归分析结果，构建肺癌化疗患者发生CRBSI的风险预测列线图模型。Hosmer-Lemeshoe拟合优度检验结果显示，在建模集中该列线图模型的拟合程度较好（χ2=8.905，P=0.350），在验证集中该列线图模型的拟合程度较好（χ2=8.693，P=0.365）。校准曲线分析结果显示，该列线图模型预测建模集和验证集肺癌化疗患者的CRBSI发生率与实际发生率基本吻合。ROC曲线分析结果显示，该列线图模型预测建模集和验证集肺癌化疗患者发生CRBSI的AUC分别为0.859〔95%CI（0.804，0.914）〕、0.876〔95%CI（0.813，0.940）〕。结论　糖尿病、化疗次数≥5次、有重症监护病房住院史、导管维护时间延长、导管移动、导管留置时间≥30 d、穿刺次数≥2次是肺癌化疗患者发生CRBSI的危险因素，基于上述因素构建的列线图模型对肺癌化疗患者发生CRBSI具有一定预测价值。

Abstract：

Objective To explore the influencing factors of catheter-related bloodstream infection (CRBSI) inpatients with lung cancer undergoing chemotherapy, and construct a nomogram model for predicting its risk and validate it.Methods A total of 1 169 lung cancer patients undergoing chemotherapy admitted to Yancheng Third People's Hospital fromJanuary 2010 to November 2022 were selected as the research subjects. The patients were divided into modeling set (70 cases)and validation set (468 cases) in a ratio of 6 : 4. Clinical data of patients were collected, the patients in modeling set were dividedinto occurrence group and non-occurrence group based on whether CRBSI occurred. Multivariate Logistic regression analysis wasused to analyze the influencing factors of CRBSI in patients with lung cancer undergoing chemotherapy. The nomogram modelfor predicting the risk of CRBSI in patients with lung cancer undergoing chemotherapy was constructed by using the R 4.1.2software package and rms package. Hosmer-Lemeshoe goodness of fit test was used to evaluate the fitting degree of the nomogrammodel. Calibration curve was drawn to evaluate the effectiveness of the nomogram model for predicting CRBSI in patients withlung cancer undergoing chemotherapy in modeling set and validation set, and the ROC curve was used to analyze the predictivevalue of the nomogram model for CRBSI in patients with lung cancer undergoing chemotherapy in modeling set and validation set. Results Among 701 lung cancer patients undergoing chemotherapy in the modeling set, 71 (10.13%) patients had CRBSI, 630(89.87%) patients had no CRBSI. There were significant differences in tumor stage, nutritional status, the proportion of patientswith diabetes, the number of chemotherapy, the proportion of patients with a history of hospitalization in intensive care unit, theproportion of patients with prolonged catheter maintenance time, the proportion of patients with catheter movement, catheterretention time, and the number of punctures between the two groups (P < 0.05) . Multivariate Logistic regression analysis showedthat diabetes, the number of chemotherapy, history of hospitalization in intensive care unit, prolonged catheter maintenancetime, catheter movement, catheter retention time, and the number of punctures were the influencing factors of CRBSI in patientswith lung cancer undergoing chemotherapy (P < 0.05) . The nomogram model for predicting CRBSI in patients with lung cancerundergoing chemotherapy was constructed based on the results of multivariate Logistic regression analysis. The results of HosmerLemeshoe goodness of fit test showed that the nomogram model fit well in modeling set (χ2=8.905, P=0.350) and in validationset (χ2=8.693, P=0.365) . The results of calibration curve analysis showed that the incidence of CRBSI in patients with lungcancer undergoing chemotherapy predicted by the nomogram model was basically consistent with the actual incidence of CRBSI inpatients with lung cancer undergoing chemotherapy in modeling set and validation set. The results of ROC curve analysis showedthat the AUC of the nomogram model for predicting CRBSI in patients with lung cancer undergoing chemotherapy in modelingset and validation set was 0.859 [95%CI (0.804, 0.914) ] , 0.876 [95%CI (0.813, 0.940) ] , respectively. Conclusion Diabetes,the number of chemotherapy ≥ 5, history of hospitalization in intensive care unit, prolonged catheter maintenance time, cathetermovement, catheter retention time ≥ 30 d, and the number of punctures ≥ 2 are the risk factors of CRBSI in patients with lungcancer undergoing chemotherapy. The nomogram model constructed based on the above factors has a certain predictive value forCRBSI in patients with lung cancer undergoing chemotherapy.

ReferenceList：