Emergence of Macrolide-Resistant Mycoplasma pneumoniae during an Outbreak in a Primary School: Clinical Characterization of Hospitalized Children

Mycoplasma pneumoniae (M. pneumoniae) is a common causative pathogen of community-acquired pneumonia. Here, we report the development of macrolide resistance during a school outbreak of severe M. pneumoniae infections in southwest Germany. We conducted a case series to assess the clinical and laboratory characteristics of hospitalized children with M. pneumonia infection and the prevalence of macrolide-resistant M. pneumoniae (MRMP) in this patient group. We retrospectively analyzed 23 children with serologically (19 patients) and/or PCR (eight patients) confirmed M. pneumoniae infection between October 2019 and December 2019. Most of the 15 hospitalized patients had lower respiratory tract infection (n = 10) and required oxygen therapy (83%). The median length of hospitalization was 7 days (range 3–10 days). In 8/15 patients (53.3%) azithromycin and in 4/15 (26.6%) clarithromycin treatment was applied. However, among the five patients for which extended molecular characterization was performed, sequencing of 23S rRNA revealed no mutation only in the first case, but development of macrolide resistance A2058G in four subsequent cases. Hence, we identified a cluster of hospitalized patients with emerging MRMP. Further studies are warranted to confirm a potential link between macrolide resistance and disease severity.


Introduction
Mycoplasma pneumoniae (M. pneumoniae) is one of the most common pathogens causing community-acquired pneumonia (CAP) worldwide, responsible for up to 40% of cases in the general population during epidemic periods. Severe infections occur in children above 5 years of age, whereas younger children often present with a milder clinical course [1,2].
For treatment, tetracyclines and quinolones are not primarily recommended in children due to potential side effects, leaving macrolides as first therapeutic option [1,2]. The prevalence of macrolide-resistant M. pneumoniae (MRMP) varies from low levels in Europe (Germany: 3%) to high rates in Asia (China: up to 100%) [3,4].

Results
During the observational period, overall, 23 children presented either serologically (19 patients) and/or PCR (eight patients) confirmed M. pneumoniae infections. Of these 23 children, 15 were hospitalized. Patients (12/15) suffering from pulmonary symptoms were subsequently further analyzed (Table 1). Specific IgM was detected in 9/9 serologically tested patients, PCR was positive in 6/8 patients, and positive serology and PCR were found in 2/5 simultaneously tested patients (Table 2).
Of the 12 patients suffering from pulmonary symptoms, 10 were female and two were male. The median age was 8 years (range 3-10 years). Ten cases (83%) needed oxygen, one additionally required a high-flow nasal cannula. In these 10 cases with clinical signs of pneumonia (83%), chest X-rays were performed and radiological signs of pneumonia were detected in eight cases (80%). Two patients (20%) with oxygen demand had only clinical signs of pneumonia (tachypnea, crackles). The median period of oxygen supplementation was 6 days (range 3-9 days) and the median period of hospitalization was 7 days (range 4-10 days).
Patients with pulmonary symptoms were either treated with azithromycin in eight cases (66%) or with clarithromycin in four cases (33%). A positive test result for M. pneumoniae led to a change of antibiotic therapy in four patients, who had been initially treated with β-lactam antibiotics. The first patients with the confirmed wild-type M. pneumoniae strain were treated initially with azithromycin.
Overall, seven of eight PCR-positive samples were retrospectively sent to the national reference center for further analysis. In two specimens, PCR amplifications were negative; the remaining five underwent extended molecular characterization. The strains were identical concerning all described typing approaches (P1-, VNTR-, MLST-and SNP-type), suggesting that the outbreak was monoclonal ( Table 3). Infection of the first inpatient from our observed cluster proved to be due to a macrolide-sensitive wild-type M. pneumoniae strain according to the 23S rRNA sequencing analysis. This changed in the following patients, where an identical nucleotide exchange associated with resistance (A2058G; Escherichia coli (E. coli) numbering) could be identified.
No specific outbreak control measures were applied, since case incidence decreased spontaneously. Local health authorities and local outpatient pediatricians were asked to be alert to identify further MRMP infections.

Discussion
The M. pneumoniae strain found in the first investigated patient was sensitive to macrolides, whereas samples taken later during the outbreak from subsequent patients revealed macrolide resistance. To our knowledge, this is the first report about a macrolideresistant strain emerging during an outbreak in children. Development of resistance during antimicrobial treatment with macrolides has been reported for single patients [5]. Recently, a child with acute lymphoblastic leukemia and chronic M. pneumoniae carriage was reported to have a de novo macrolide resistance mutation (mix of A2058T and A2058G) and subsequent breakthrough pneumonia [6].
It has been previously discussed that MRMP may cause a more severe clinical course compared to macrolide-susceptible strains [7]. The pneumonia cases in our analysis were hospitalized longer than other cases we usually see in the respective age groups. All sequenced patients were previously healthy and had no macrolide treatment before. Therefore, we suspect either an increased virulence of the strains or that the inadequate treatment with macrolides led to delayed clinical improvement. Moreover, no extraordinary extended infiltrates nor pleural effusion could be detected in the chest X-rays. In a recent large school outbreak, an attack rate of 4% and an extended length of stay of patients in hospital for up to 4 weeks (median 11 days) were observed [8]. Pathogen characterization indicated that M. pneumoniae P1 type 1 was the causative agent during this outbreak and the strain harbored an exchange of A to G at position 2058, the most common point mutation in domain V of the 23S rRNA. A recent meta-study showed that the duration of fever in patients infected with macrolide-resistant strains lasts 1.71 days longer and the length of stay is 1.61 days longer [9]. To date, no specific factor has been identified to predict the clinical course of disease, and no evidence supports the causal relationship between macrolide resistance and disease severity.
Due to the retrospective nature of the study, we neither could identify other potential "first" MRMP cases in the outpatient community nor track how many patients in the nearby outpatient community were treated with macrolides. Since the first patient with MRMP in our study cohort was found in October, other potential M. pneumoniae infections, either with wild-type or MRMP strains, may have already emerged in the community in September. However, unfortunately, due to a lack of active surveillance and patient material, potential macrolide resistance in these patients could not be further analyzed.
Finally, the necessity and efficacy of antibiotic therapy in M. pneumoniae infections is still under debate [9]. All our patients showed a clinical improvement despite proven macrolide resistance, indicating the self-limiting course of the disease. The potential anti-inflammatory effect of macrolides as well as of steroids may contribute to clinical improvement in infections with MRMP [10].

Study Design and Participants
Between October 2019 and December 2019, we observed a cluster of hospitalized pediatric patients with M. pneumoniae infections, who went to same primary school in Mannheim. Children with M. pneumoniae infection were identified either by serology (IgM/IgG; Institute Virion/Serion, Würzburg, Germany) and/or real-time PCR (FilmArray ® , Biofire ® , Nürtingen, Germany) for M. pneumoniae (nasopharyngeal swabs). In children for whom patient material was initially available, we analyzed for the presence of macrolide resistance via sequencing of 23S rRNA (German reference center for mycoplasma, Dresden, Germany).

Data Analysis
Clinical and laboratory characteristics of hospitalized patients with M. pneumonia infections were retrospectively analyzed. Due to the small case numbers, a statistical analysis could not be conducted. Instead, a detailed case series description has been performed.

Conclusions
Our study highlights the probable development of antimicrobial resistance of M. pneumoniae during an outbreak of severe cases of CAP among school children. Even in countries with a low rate of resistance, clinicians should be aware of the emergence of resistance of primarily susceptible strains and its potential contribution to the course of subsequent infections. Timely molecular characterization of MRMP strains can help to improve the efficiency of treatment.