成人患者发生医院内侵袭性念珠菌病的危险因素:一项病例对照研究

Risk factors for hospital-associated invasive candidiasis in adult patients: a case-control study

  • 摘要:
    目的  为了进一步探索成人患者发生医院内侵袭性念珠菌病(HAIC)的危险因素,找到更有效的防控措施,从而降低成人HAIC的发生风险。
    方法  本次研究采用1∶2匹配的病例对照研究,于北京某三甲综合教学医院使用统一设计的调查问卷回顾性收集2015-2019年病例组(103例)及对照组(206例)患者的基础信息及可能的影响因素;通过使用单因素分析、多因素条件logistic回归及XGBoost模型分析成人HAIC发生的危险因素。
    结果  多因素条件logistic回归分析显示念珠菌定植(OR=16.539, 95%CI: 3.133~87.451)、应用全胃肠外营养(TPN)(OR=12.582, 95%CI: 2.275~69.689)、留置尿管(OR=9.887, 95%CI: 1.912~51.025)、留置中心静脉导管(CVC)(OR=4.715, 95%CI: 1.316~16.942)、使用β-内酰胺酶抑制剂复合制剂类药物(OR=3.875, 95%CI: 1.513~9.955)及碳青霉烯类药物(OR=3.788, 95%CI: 1.528~9.421)是成人HAIC的独立危险因素(P<0.05);基于XGBoost模型分析显示,重要性评分前7位依次为:使用碳青霉烯类药物(4.71)、使用有创呼吸机(3.2)、入住重症监护病房(ICU)(2.32)、使用β-内酰胺酶抑制剂复合制剂(1.68)、入住ICU时长(1.54)、念珠菌定植(1.48)、手术(1.42)。对两种模型预测效果进行比较,结果显示XGBoost模型的受试者工作特征曲线下面积(AUC)(0.78)高于多因素logistic回归模型(0.72)。
    结论  本次研究首次对HAIC发生的危险因素进行分析,结果显示今后要重点关注ICU的念珠菌定植感染情况,β-内酰胺酶抑制剂复合制剂类药物、碳青霉烯类等广谱抗菌药物的使用及联合使用等因素,以期采取更有针对性的防控措施,有效防控HAIC。

     

    Abstract:
    OBJECTIVE  To explore the risk factors for hospital-associated invasive candidiasis(HAIC)in adult patients and to find more effective preventive and control measures to reduce the risk of HAIC in adults.
    METHODS  This was a matched case-control study (1∶2), in which basic information and possible influencing factors of patients in the case group (103 cases) and control group (206 cases) were retrospectively collected by a uniformly designed questionnaire from 2015-2019 in a three-A teaching hospital in Beijing; risk factors for the occurrence of HAIC in adults were analyzed by univariate analysis, multifactorial conditions logistic regression, and XGBoost model.
    RESULTS  Multifactorial conditions logistic regression analysis showed that Candida colonization (OR= 16.539, 95%CI: 3.133-87.451), total parenteral nutrition(TPN) support(OR=12.582, 95%CI: 2.275-69.689), indwelling urinary catheter (OR=9.887, 95%CI: 1.912-51.025), indwelling central venous catheter(CVC) (OR=4.715, 95%CI: 1.316-16.942), using β-lactamase inhibitor combination antibiotics (OR=3.875, 95%CI: 1.513-9.955) and carbapenems (OR=3.788, 95%CI: 1.528-9.421) were independent risk factors for HAIC in adults. Analysis based on the XGBoost model showed that the top 7 importance scores in a descending order were the use of carbapenems (4.71), use of invasive ventilator (3.2), admission length to intensive care unit (ICU) (2.32), use of β-lactamase inhibitor combination (1.68), admission to ICU (1.54), Candida colonization (1.48), and surgery (1.42). Comparison of the predictive effects of the two models showed that the Area Under Curve (AUC) of the XGBoost model (0.77) was higher than that of the multifactor logistic regression model (0.72).
    CONCLUSIONS  This study analyzes the risk factors for the occurrence of HAIC for the first time. The results show that Candida colonization in ICUs and the use of broad-spectrum antimicrobial drugs such as β-lactamase inhibitors, carbapenems and the combination should be focused in the future, so as to take more targeted preventive and control measures for HAIC.

     

/

返回文章
返回