Early clinical predictors of macrolide-resistant Mycoplasma pneumoniae pneumonia in hospitalized children during the 2024 outbreak in South Korea.

BACKGROUND

After the COVID-19 emergency phase, Mycoplasma pneumoniae (MP) infections surged globally, heightening concerns over macrolide-resistant M. pneumoniae (MRMP), especially in children. Early detection of MRMP is crucial for reducing morbidity and optimizing antimicrobial use, but developing reliable prediction models remains challenging.

METHODS

This retrospective study evaluated 458 hospitalized children with PCR-confirmed MP pneumonia during the 2024 outbreak at a South Korean university hospital. We compared demographic, clinical, laboratory, radiographic, and treatment data between macrolide-sensitive (MS) and -resistant groups. Multivariable logistic regression and ROC analyses identified independent predictors of resistance and developed a risk model.

RESULTS

The prevalence of MRMP was profoundly high at 83.8%. Compared with MS infections, MRMP was associated with older age, higher inflammatory markers, lower MP PCR cycle threshold (Ct) values, and more pulmonary complications. Multivariable analysis identified elevated erythrocyte sedimentation rate (ESR) and lower PCR Ct as robust, independent predictors of macrolide resistance. Integrating the PCR Ct value into a baseline clinical model significantly improved prediction accuracy, increasing the area under the curve (AUC) from 0.659 to 0.806. Chest radiographic findings provided negligible incremental diagnostic value (AUC 0.808).

CONCLUSIONS

MRMP is highly prevalent among hospitalized children during post-emergency outbreaks in East Asia. Incorporating admission PCR Ct values and ESR into early clinical assessments substantially improves prediction of macrolide resistance, enabling timely, targeted escalation to second-line therapies.