Predictive Value of Traditional and Novel Composite Inflammatory Indicators for Severe and Refractory Mycoplasma pneumoniae Pneumonia in Children.

OBJECTIVE

This study aimed to comprehensively evaluate the predictive efficacy of traditional single inflammatory indicators and novel composite inflammatory indicators (CLR, LMR, NLR, NPR, PIV, PLR, SII, SIRI) for severe Mycoplasma pneumoniae pneumonia (SMPP) and refractory MPP (RMPP) in children.

METHODS

This study retrospectively enrolled 1791 children with MPP and collected their case data. A phased modeling strategy (univariate analysis, LASSO regression, multivariate logistic regression) was employed to construct prediction models. Model performance was evaluated using area under the curve (AUC) from receiver operating characteristic (ROC) curves, calibration curves with the Hosmer-Lemeshow test, bootstrap resampling with 1000 repetitions, and decision curve analysis (DCA).

RESULTS

The cohort included 512 SMPP, 269 RMPP, and 1180 general MPP cases; mentiontly, 170 children met both SMPP and RMPP criteria. The SMPP prediction model identified nine independent risk factors (Hb, PLT, D-D, FIB, LMR, NPR, SII, duration of cough and fever), achieving an AUC of 0.803. The RMPP model identified seven factors (Hb, CRP, FIB, LMR, NPR, duration of cough and fever) with an AUC of 0.889. The calibration curves, Hosmer-Lemeshow test, bootstrap internal validation, and DCA curve together confirmed the robustness and clinical applicability of the models.

CONCLUSION

This multi-parameter integration strategy enables precise MPP risk stratification, holding significant implications for clinical treatment planning and antibiotic selection.