488. Association between the Respiratory Microbiota and Response to Antibiotic Therapy in Children with Macrolide-resistant Mycoplasma pneumoniae Pneumonia

Abstract Background Along with the overuse of macrolides as the empirical treatment for M. pneumoniae infection, macrolide-resistant M. pneumoniae (MRMP) emerged up to 90 prevalence rate in Asian countries. We investigated whether respiratory microbiota at the beginning of hospitalization was associated with disease severity and response to antibiotic treatment in children with MRMP pneumonia. Methods Prospectively enrolled children aged < 18 years with community-acquired pneumonia caused by MRMP from 2017/4-2020/3 at Linkou Chang Gung Memorial Hospital. Oropharyngeal samplings were collected within 48 hours after admission. Total 66 children were included: 23 had not received macrolide or gave response to macrolide (defined as without effective treatment group); 43 needed doxycycline treatment (defined as doxycycline treatment group). Compared the oropharyngeal microbiota between groups using the 16S-rRNA-based sequencing. Patients who were administered antibiotics other than augmentin or Azithromax were excluded. The difference in microbial composition will be examined using microbial community analysis. Statistically significant species for each group were evaluated by the linear discriminant analysis of effect size analysis to identify differentially abundant biomarkers, comparing disease severity scores, CRP and febrile days lasting. Scheme of sample collection and analyses Augmentin with or without Azithromax was the most common antimicrobial agent used. If fever persisted under the impression of suspected MRMP infection, doxycycline was used following three days of Azithromax usage. Scheme of sample collection and analyses Correlation between microbiome and clinical feature. Results Doxycycline treatment group had longer duration of total febrile days (p< 0.001) and total hospitalization days (p=0.002). Without effective treatment group had higher microbial richness, different overall community composition (p< 0.01). The lowest respiratory alpha diversity had prolonged febrile days (p=0.032). Without effective treatment group had more abundant spices or genera. Classification of samples to detect patients with effective treatment or not based on these selected microbiota genera features achieved the ROC curve of 0.86. From the list of most relatively abundant species in MRMP patient without doxycycline, Fusobacterium periodonticum most negatively correlated with total febrile days and severity score. Clinical characteristics of patients enrolled in this study Data were shown as n (%) or mean ± standard deviation. N/A, not applicable. Taxonomic assignment of the all throat swab samples. Fraction of predominant microbiota with a relative abundance of at least 2% at the order level (A) and genus level (B) in all sample. Higher respiratory microbial diversity is associated with MRMP patients without effective treatment than MRMP patients who needed doxycycline treatment Microbial richness base on the different alpha-diversity index in MRMP patients without effective treatment (blue); MRMP patients with doxycycline treatment (red). (A)(B)The box represented the interquartile range (IQR, and the midline represented the median. Tukey’s honest significance test of ANOVA, p<0.05.) (C)Kaplan-Meier (KM) plot of febrile time by respiratory diversity. (D)Principal coordinates analysis (PCoA) for MRMP patients without effective treatment (blue), MRMP patients with doxycycline treatment (red) by Unweighted UniFrac distance. Significant differences were observed between MRMP patients without effective treatment and MRMP patients with doxycycline treatment. A PERMANOVA (α=0.05) with 999 random permutations, p = 0.001. Conclusion Respiratory dysbiosis was associated with persistent fever and disease severity in MRMP pneumonia. In our study, F. periodoticum negatively related to fever duration and disease severity. Compositional differences in the respiratory microbiome are associated with MRMP patients without effective treatment than MRMP patients who needed doxycycline treatment by LEfSE analysis LDA scores revealed differentially abundant bacteria in the respiratory microbiome of without effective treatment (blue) and doxycycline treatment (red) groups at genus level (A) and species level (B).The mean relative abundance of genera at right side. The mean relative abundance of genera at right side. (C) Receiver operating characteristic (ROC) curves of ASVs to predict MRMP patients without effective treatment or who needed doxycycline treatment. Pairwise Spearman rank correlation heat map of significantly different taxa in throat swabs at initial sampling and clinical severity score in MRMP patients without effective treatment/with doxycycline treatment The genus (A) and species (B) level. Genus and species were in rows; the correlation was indicated by color gradient. Disclosures All Authors: No reported disclosures.