Clinical Prediction of Necrotizing Pneumonia in Children With Mycoplasma pneumoniae: A Model Based on Early Response to Therapy
TL;DR
Among children with MUMPP, higher WBC count, prolonged fever, pleural effusion, and higher CRP identify increased risk of NP, and decision curve analysis suggested net clinical benefit across relevant threshold probabilities.
Abstract
Objective This study aimed to identify factors that predict progression to necrotizing pneumonia (NP) in children with Mycoplasma pneumoniae pneumonia (MPP) who fail to respond to initial therapy and to develop a predictive model to support early clinical decision‐making. Methods We retrospectively analyzed clinical and laboratory data of children with macrolide‐unresponsive Mycoplasma pneumoniae pneumonia (MUMPP) admitted to Tianjin Children’s Hospital from January 2021 to January 2025. MUMPP was defined as persistent fever (≥ 38.5°C, lasting ≥ 4 h/day) after ≥ 72 h of standardized intravenous macrolide therapy. Patients were classified into an observation group (n = 106, those who developed NP) and a control group (n = 130, those who did not). Independent predictors of NP were identified by multivariable logistic regression and incorporated into a nomogram. Model performance was evaluated using ROC analysis, calibration plots, and decision curve analysis. Results In multivariable analysis, increased white blood cell (WBC) count (OR = 2.484, 95% CI: 1.406–4.613), longer fever duration (OR = 1.618, 95% CI: 1.217–2.193), pleural effusion (OR = 3.955, 95% CI: 1.664–10.00), and elevated C‐reactive protein (CRP) (OR = 2.152, 95% CI: 1.547–3.074) were independently associated with NP among children with MPP after initial treatment failure. Calibration curves indicated close agreement between predicted and observed risks in both the derivation cohort (p = 0.588) and the validation cohort (p = 0.424). The model showed good discrimination, with an AUC of 0.848 (95% CI: 0.789–0.908) in the derivation cohort and 0.889 (95% CI: 0.815–0.963) in the validation cohort. Decision curve analysis suggested net clinical benefit across relevant threshold probabilities. Conclusion Among children with MUMPP, higher WBC count, prolonged fever, pleural effusion, and higher CRP identify increased risk of NP.