基于肺部超声评分的儿童难治性肺炎支原体肺炎预测模型的构建  

Construction of a prediction model for refractory mycoplasma pneumoniae pneumonia in children based on lung ultrasound score

作　　者：李登峰 黄家虎 李廷俊 吕勇[3] 金珍珍 连少峰 Li Dengfeng;Huang Jiahu;Li Tingjun;Lv Yong;Jin Zhenzhen;Lian Shaofeng(Department of Pediatrics,Jieshou People's Hospital,Jieshou 236500,China;Department of Emergency,Shanghai Children's Hospital,School of Medicine,Shanghai Jiao Tong University,Shanghai 200062,China;Department of Pediatrics,the First Affiliated Hospital of University of Science and Technology of China(Anhui Provincial Hospital),Hefei 230001,China)

机构地区：[1]界首市人民医院儿科,安徽界首236500 [2]上海交通大学医学院附属儿童医院急诊科,上海200062 [3]中国科学技术大学附属第一医院(安徽省立医院)儿科,安徽合肥230001

出　　处：《临床荟萃》2025年第2期153-157,共5页Clinical Focus

基　　金：阜阳市卫生健康委员会科研课题——多中心研究肺部超声在儿童重症肺炎患儿中的临床应用(FY2021-119)。

摘　　要：目的基于部超声评分(Lung ultrasound score,LUS)联合炎症指标构建儿童难治性肺炎支原体肺炎(refractory mycoplasma pneumoniae pneumonia,RMPP)预测模型。方法回顾性分析2023年6月-2024年1月在上海市儿童医院呼吸科、中国科学技术大学附属第一医院儿科及界首市人民医院儿科住院的肺炎支原体肺炎病历数据,根据诊断标准将患者分为RMPP组和非难治性支原体肺炎组(non-refractory mycoplasma pneumoniae pneumonia,n-RMPP)组。通过单因素及多因素Logistic回归分析筛选出RMPP的独立危险因素,构建预测模型,并采用受受试者工作特征曲线(receiver operating characteristic,ROC)对模型预测效能进行验证。结果共纳入876例肺炎支原体肺炎患儿,其中RMPP组498例,n-RMPP组378例。二元Logistic回归分析显示LUS评分(OR=1.276,95%CI 1.196~1.362)、C-反应蛋白(CRP)(OR=1.098,95%CI 1.062~1.119)、D-二聚体(D-D)(OR=2.683,95%CI 1.923~3.742)、乳酸脱氢酶(LDH)(OR=1.015,95%CI 1.013~1.018)是儿童RMPP独立危险因素(P<0.05);ROC显示肺部超声评分或各炎症指标单独预测RMPP均有较好预测价值,联合预测效能最佳,ROC曲线下面积为0.969(95%CI 0.958~0.979),敏感度97.4%,特异度84.7%。结论基于LUS联合炎症指标的回归模型具有较好的预测效能,为临床诊断儿童RMPP提供很好的依据。便于儿科医生早期识别及干预,改善RMPP的危险结局。Objective To construct a prediction model of pediatric refractory mycoplasma pneumoniae pneumonia(RMPP)based on lung ultrasound score(LUS)and inflammation index.Methods The clinical data of children with mycoplasma pneumoniae pneumonia(MPP)hospitalized in the Respiratory Department of Shanghai Children's Hospital,Pediatrics Departmen of the First Affiliated Hospital of University of Science and Technology of China or Jieshou People's Hospital from June 2023 to January 2024 were retrospectively analyzed,and the children were assigned into RMPP group and non-RMPP group according to the diagnostic criteria.The independent risk factors of RMPP were screened by univariate and multivariate logistic regression analysis,and the predictive model was constructed,and the predictive efficiency of the model was verified by receiver operating characteristic curve(ROC).Results A total of 876 children with MPP were included,including 498 in RMPP group and 378 in non-RMPP group.Binary Logistic regression analysis showed that LUS(OR=1.276,95%CI 1.196-1.362),C-reactive protein(CRP,OR=1.098,95%CI 1.062-1.119),D-dimer(D-D,OR=2.683,95%CI 1.923-3.742)and lactate dehydrogenase(LDH,OR=1.015,95%CI 1.013-1.018)were independent risk factors for RMPP in children(P<0.05).The ROC showed good predictive value of LUS or each inflammatory index alone in predicting RMPP,while,the combined predictive efficacy was the best,with an area under the ROC curve(AUC)of 0.969(95%CI 0.958-0.979),a sensitivity of 97.4%and a specificity of 84.7%.Conclusion The regression model based on LUS combined with inflammatory indicators has a good predictive efficacy and provides a good basis for clinical diagnosis of children with RMPP.It is convenient for pediatricians to identify and intervene early and improve the dangerous outcome of RMPP.

关 键 词：肺炎 支原体 难治性 儿童 肺部超声评分 炎症指标 预测模型 

分 类 号：R563.1[医药卫生—呼吸系统]