儿童大环内酯类药物无反应性肺炎支原体肺炎早期预测列线图模型的构建与验证OA

Objective:Mycoplasma pneumoniae is one of the most common pathogens causing community-acquired pneumonia in children.This study aims to explore the independent risk factors for macrolide-unresponsive Mycoplasma pneumoniae pneumonia(MUMPP)in children and to construct a nomogram prediction model. Methods:A total of 800 children with Mycoplasma pneumoniae pneumonia(MPP)hospitalized in the Department of Pediatrics of The Second People's Hospital of Futian District,Shenzhen from January 2023 to May 2024 were retrospectively selected as the model development cohort.Univariate and multivariate Logistic regression analyses were used to identify independent risk factors for the occurrence of MUMPP,and a nomogram model was constructed based on these factors.Another 200 children with MPP admitted from October 2024 to June 2025 were used as the external validation cohort.Receiver operating characteristic(ROC)curves,decision curve analysis(DCA),and calibration curves were used to test and evaluate the discrimination,accuracy,and clinical applicability of the nomogram diagnostic model. Results:A total of 1 000 children were included in the study.Among the 800 cases in the model development cohort,700 were general Mycoplasma pneumoniae pneumonia(GMPP)and 100 progressed to MUMPP;among the 200 cases in the validation cohort,147 were GMPP and 53 were MUMPP.Multivariate logistic regression analysis showed that in the model development cohort,high-sensitivity C-reactive protein[CRP,odds ratio(OR)=1.107,95%confidence interval(CI)1.084 to 1.135,P<0.001],interleukin-6(IL-6)(OR=1.105,95%CI 1.044 to 1.174,P<0.001),D-dimer(D-D)(OR=82.070,95%CI 27.858 to 288.699,P<0.001),alanine aminotransferase(ALT)(OR=1.069,95%CI 1.023 to 1.119,P=0.003),atelectasis on chest imaging(OR=3.062,95%CI 1.144 to 8.378,P=0.027),pleural effusion(OR=5.949,95%CI 2.011 to 17.866,P=0.001),and positive detection of macrolide resistant genes(OR=7.374,95%CI 2.644 to 21.261,P<0.001)were 7 independent risk factors for MUMPP.A nomogram model was constructed based on these variables.The area under the ROC curve(AUC)of the nomogram model was 0.973 in the model development cohort and 0.904 in the validation cohort,indicating good ability to distinguish between GMPP and MUMPP in children at an early stage.The calibration curve and decision curve also demonstrated good performance of the nomogram prediction model. Conclusion:Based on CRP,IL-6,D-D,ALT,atelectasis,pleural effusion,and positive detection of macrolide resistant genes,the nomogram model established in this study can predict the risk of MUMPP in children with MPP early,rapidly,and effectively.It may serve as a quantitative tool for early assessment of the risk of MUMPP in children and help clinicians identify pediatric MUMPP at an early stage.