慢性阻塞性肺疾病合并多重耐药铜绿假单胞菌感染的危险因素分析及预测模型构建  

Risk factor analysis and prediction model construction for chronic obstructive pulmonary disease combined with multi-drug resistant Pseudomonas aeruginosa infection

作　　者：阿呷略布莫 谭笑 王荣丽 AGA Luebumo;TAN Xiao;WANG Rongli(Department of Respiratory and Critical Care Medicine,the Affiliated Hospital of Southwest Medical University,Luzhou,Sichuan 646000,China)

机构地区：[1]西南医科大学附属医院呼吸与危重症医学科,四川泸州646000

出　　处：《临床肺科杂志》2025年第5期748-755,共8页Journal of Clinical Pulmonary Medicine

摘　　要：目的探讨慢性阻塞性肺疾病(慢阻肺)患者合并多重耐药铜绿假单胞菌(MDR-PA)感染的危险因素,并构建慢阻肺患者感染MDR-PA的风险列线图模型。方法回顾性分析本院收治住院的慢阻肺合并铜绿假单胞菌(PA)感染的144例患者,根据药物敏感试验结果分成多重耐药组与非多重耐药组。采用单因素分析(t检验、秩和检验、χ^(2)检验)对临床资料初筛危险因素,再通过多因素Logistic回归分析出独立危险因素。最后根据筛选出的独立因素采用R语言软件(R3.5.3)构建风险列线图模型,并通过受试者工作特征(ROC)曲线下面积(AUC)、校准曲线和决策曲线分析(DCA)对列线图进行多方面评估。结果目前临床上慢阻肺合并PA感染的发病率逐渐升高,对多种临床上常用的抗菌药物如环丙沙星、左氧氟沙星的耐药率较高。多因素Logistic逐步回归分析结果显示,中性粒细胞比率、过去一年内住院次数≥2次、合并充血性心力衰竭、抗菌药物联用是慢阻肺患者合并MDR-PA感染的独立危险因素(P<0.05)。风险列线图模型预测值和实测值之间的一致性较高,即说明建立的预测模型对预测慢阻肺患者合并MDR-PA感染具有较高的准确性;预测模型的C-index指数值为0.861(95%CI:0.789~0.934),即说明该预测模型的精准度和区分度良好;DCA曲线结果则表明该列线图预测模型具有实际价值。结论慢阻肺患者合并MDR-PA的独立危险因素包括中性粒细胞比率、过去一年内住院次数≥2次、合并充血性心力衰竭以及抗菌药物联用,基于独立危险因素构建的预测模型具有较高的预测效能和临床应用价值,可为临床医生制定相关的预防和治疗措施提供依据。Objective To investigate the risk factors of combined multidrug-resistant Pseudomonas aeruginosa(MDR-PA)infection in patients with chronic obstructive pulmonary disease(COPD),and to construct a columnar graph model of the risk of MDR-PA infection in patients with COPD.Methods It retrospectively analyzed 144 patients admitted to our hospital with COPD combined with Pseudomonas aeruginosa(PA),and then they were divided into the multi-drug-resistant group and the non-multi-drug-resistant group based on the results of drug susceptibility tests.The clinical data were initially screened for risk factors using one-way analysis(t-test,rank-sum test,χ^(2) test),and then independent risk factors were analyzed by multi-factorial logistic regression.Finally,a risk column chart model was constructed based on the screened independent factors using R language software(R3.5.3),and the column charts were assessed by the area under the curve(AUC)of the subject′s work characteristics(ROC),calibration curves,and decision curve analysis(DCA).Results The incidence of PA infection in COPD patients was gradually increasing,and the resistance to various commonly used antimicrobial drugs,such as ciprofloxacin and levofloxacin,was high.The results of multi-factorial logistic stepwise regression analysis showed that neutrophil ratio,≥2 hospitalisations within the past year,comorbid congestive heart failure,and antimicrobial drug coadministration were independent risk factors for MDR-PA infection in COPD patients with comorbid MDR-PA infection(P<0.05).The consistency between the predicted and measured values of the risk column-line diagram model was high,which meant that the established prediction model had high accuracy in predicting the combined MDR-PA infection in COPD patients.The C-index value of the prediction model was 0.861(95%CI:0.789~0.934),which meant that the accuracy and differentiation of the prediction model were good,and the results of the DCA curve indicated that the column-line graph prediction model had practical valu

关 键 词：慢性阻塞性肺疾病 铜绿假单胞菌 多重耐药 危险因素 列线图 预测模型 

分 类 号：R563.1[医药卫生—呼吸系统]