Etiological characteristics and predictive model construction for poor prognosis in children with severe community-acquired pneumonia

ZENG Yuan; JIA Xiao-dan; SONG Zhe

doi:10.13820/j.cnki.gdyx.20244000

Abstract：Objective To investigate the etiological characteristics of severe community-acquired pneumonia(SCAP)in children and to construct a predictive model for poor outcome(PO).Methods A total of 240 pediatric pa-tients diagnosed with SCAP between January 2021 and December 2023 were enrolled according to inclusion and exclusion criteria.Clinical data and etiological findings were collected.Based on whether the children experienced a poor prognosis(PO),they were divided into a PO group and a non-PO group.Univariate and multivariate Cox proportional hazards re-gression analyses were performed to identify independent risk factors associated with PO.A nomogram model was then de-veloped using R software.Model performance was evaluated using the area under the ROC curve(AUC),concordance in-dex(C-index),calibration curve,and decision curve analysis(DCA),with internal validation conducted via the boot-strap method.Results A total of 350 pathogens were identified in the 240 children,including 212 bacterial strains(60.57％),70 viruses(20.00％),60 Mycoplasma pneumoniae(17.14％),and 8 Chlamydia pneumoniae(2.29％).Mixed infections were observed in 88 cases(25.14％).Variables with significant differences in univariate analysis were included as independent variables,and PO occurrence was the dependent variable in the multivariate logistic regression a-nalysis.The results showed that mixed infection,altered mental status,hemoglobin(Hb)＜90 g/L,albumin(ALB)＜30 g/L,and neutrophil-to-lymphocyte ratio(NLR)≥3.5 were independent risk factors for PO in children with SCAP(P＜0.05).The predictive model had a C-index of 0.754(95％CI:0.677-0.830).The Hosmer-Lemeshow good-ness-of-fit test showed x2=3.525,P=0.897,indicating good model calibration.DC A revealed that when the thresh-old probability ranged from 32％to 94％,the model provided a high net clinical benefit.Conclusion Mixed infections,altered consciousness,Hb＜90 g/L,ALB＜30 g/L,and NLR≥3.5 are associated with poor prognosis in children with SCAP.The nomogram-based predictive model established in this study may help clinicians identify high-risk patients early and implement timely interventions to improve outcomes.