Clinical characteristics and logistic regression analysis of macrolide-resistant Mycoplasma pneumoniae pneumonia in children.

Abstract

To investigate macrolide-resistant Mycobacterium pneumoniae (MRMP) pneumonia in children and construct a logistic regression model for mutations in the Mycoplasma pneumoniae drug-resistant gene. Clinical data of 281 children were analyzed. Sequencing confirmed a mutation at the A2063G locus of the 23S rRNA gene in 227 children (A2063G group); 54 children showed no mutations (non-MRMP [NMRMP] group). We compared clinical features, laboratory tests, imaging, and bronchoscopy results and constructed a multifactorial logistic regression model to analyze risk and protective factors. The A2063G group had longer durations of fever and hospitalization before admission, a higher proportion of treatment with sodium methylprednisolone succinate (MPS)/dexamethasone, longer time to discontinue hormones, and higher probability of combined infections. Monocyte percentage was significantly higher in the A2063G group. Imaging suggested a higher incidence of infections in the right lung compared to both lungs. Univariate analysis revealed fever duration before admission, hormone dose and duration, monocyte percentage, and mixed infections as risk factors for Mycoplasma pneumoniae infection with the A2063G mutation. The logistic regression model showed that mixed infections were an independent risk factor for the A2063G locus mutation, whereas hormone dose was a protective factor. A prevalence of macrolide resistance of 80.8% among children was observed in the region. Logistic regression analysis revealed that co-infection with other respiratory pathogens is an independent risk factor for the development of resistance genes, while the use of hormone dosage acts as a protective factor.