消化道肿瘤患者输液港导管相关血流感染列线图模型构建

Construction of a nomogram prediction model for catheter-related bloodstream infections in infusion ports implanted in patients with gastrointestinal tumors

目的分析与验证影响消化道肿瘤化疗患者输液港导管相关血流感染(CRBSI)的危险因素,为针对性干预提供参考。方法回顾性收集573例植入静脉输液港的消化道肿瘤化疗患者为建模组,统计CRBSI发生率,通过单因素分析与多因素Logistic回归分析筛选影响CRBSI发生的危险因素,将危险因素纳入R软件建立CRBSI的列线图预测模型。前瞻性选取201例植入静脉输液港的消化道肿瘤化疗患者为验证组,对列线图预测模型进行外部验证。结果建模组CRBSI发生率为10.82%,验证组发生率为12.94%。Logistic回归分析显示,卡氏评分(KPS)≤60分、胃肠外营养支持、糖尿病史、蝶形针每次使用时间是输液港CRBSI发生的危险因素(均P<0.01);基于筛选出的危险因素构建CRBSI的列线图预测模型,内部验证的ROC曲线下面积为0.865,外部验证的ROC曲线下面积为0.775。结论影响消化道肿瘤化疗患者输液港CRBSI发生的独立危险因素包括患者身体健康状态、肠外营养、糖尿病史及蝶形针单次使用时间,基于这些危险因素构建的列线图模型具有良好的预测效能。医护人员应针对危险因素加强干预,以预防CRBSI的发生。

Objective To analyze and validate the risk factors of catheter-related bloodstream infections(CRBSI)in infusion ports implanted in patients with gastrointestinal tumors,and to provide references for intervention strategies.Methods A total of 573 gastrointestinal cancer patients with an infusion port implantation for chemotherapy were retrospectively enrolled and served as the derivative group,whose CRBSI rate was calculated.Then through univariate and multivariate Logistic regression analysis,the risk factors of CRBSI were identified,and used to build up a nomogram prediction model in R software.Another 201 patients were prospectively enrolled and served as the validation group for external verification,in which,the validity of the nomogram prediction model was validated.Results The incidence rate of CRBSI was 10.82%in the derivative group,and 12.94%in the validation group.Multivariate Logistic regression analysis showed that KPS score≤60 points,use of parenteral nutrition,diabetes history,and single use time of butterfly needle were independent risk factors of CRBSI in the infusion port(P<0.01 for all).The CRBSI nomogram prediction model was constructed based on the selected risk factors,and subjected to internal and external validation respectively,with the former yielding a result of 0.865,and the latter 0.775,for the area under the ROC curve.Conclusion The risk factors of CRBSI in the infusion portare associated with patients′health status,use of parenteral nutrition,diabetes history,and single use time of butterfly needle.The CRBSI nomogram prediction model constructed based on the risk factors enjoys good predictive efficacy.Medical staff should beef up intervention targeting the risk factors,in an effort to prevent the occurrence of CRBSI.