Mycoplasma pneumoniae Carriage With De Novo Macrolide-Resistance and Breakthrough Pneumonia

Our patient is a 3-year-old girl with high-risk pre-B acute lymphoblastic leukemia that was diagnosed in September 2017. In May 2018, the patient was admitted to the Children’s of Alabama hospital because of a fever of 101°F and worsening cough that had been lingering for few weeks in the setting of continued treatment of her acute lymphoblastic leukemia with doxorubicin and vincristine. Respiratory physical examination was normal. Her complete blood count showed neutropenia (absolute neutrophil count of 150/µL). Chest radiograph revealed left lower lobe subsegmental atelectasis. The treating team started her on cefepime and continued her home fluconazole for a previously diagnosed Candida tropicalis osteomyelitis. Nasopharyngeal aspirate tested positive for M pneumoniae by using a multiplexed polymerase chain reaction (PCR) test for multiple respiratory pathogens (GenMark Diagnostics). The panel result was negative for 14 viruses and Chlamydia pneumoniae. Nasopharyngeal culture later grew M pneumoniae in the University of Alabama at Birmingham Diagnostic Mycoplasma Laboratory (isolate No. 72749).⁶  M pneumoniae macrolide resistance real-time PCR assay demonstrated lack of macrolide-resistance mutations in the 23S rRNA gene.⁷  The minimum inhibitory concentration (MIC) for erythromycin was 0.004 µg/mL, confirming macrolide susceptibility.⁸  Accordingly, the child received high-dose azithromycin (10 mg/kg per day) for 3 days with subsequent resolution of the fever, improvement of the cough, and discharge from the hospital.

In October 2018, the patient was readmitted for a febrile illness associated with a worsening cough that was present for 2 weeks before admission. The family denied any sick contacts or family members with similar symptoms. The patient was neutropenic (absolute neutrophil count 470/µL), and her nasopharyngeal aspirate test was positive for rhinovirus and enterovirus but negative for M pneumoniae by multiplexed PCR. She had a normal examination of her lungs. Her chest radiograph showed scattered interstitial infiltrates in both lungs. Over the next 7 days, she received cefepime and her home fluconazole. She continued to suffer high-spiking fevers and a subsequent decline in respiratory status, which included tachypnea and hypoxemia. Chest computed tomography on day 8 of admission revealed right upper lobe and right lower lobe pneumonia with right hilar and paratracheal lymphadenopathy (Fig 1).

A repeat nasopharyngeal multiplexed PCR assay was negative for the previously recovered viruses and for M pneumoniae. To uncover the causative infection, the pulmonary team obtained bronchoalveolar lavage (BAL), which tested positive for M pneumoniae by real-time PCR and by culture.⁹  Azithromycin and vancomycin were then added because of continued clinical worsening. The following day, a macrolide-resistance mutation in the 23S rRNA gene was identified in the isolate by M pneumoniae macrolide resistance real-time PCR assay. Sequencing of the amplicon confirmed 2 different point mutations conferring macrolide resistance in a mixed population (A2063G or A2063T; Escherichia coli numbering 2058) (Fig 2). BAL culture grew M pneumoniae with an erythromycin MIC of 32 µg/mL, verifying macrolide resistance. In view of the clinical pneumonia with a positive M pneumoniae PCR assay and culture test, serology was not performed. The patient was hypogammaglobulinemic (immunoglobulin G: 145 mg/dL, normal: 316–1148) and received intravenous immunoglobulin, which masked any further serological testing. Accordingly, azithromycin was changed to levofloxacin, with eventual improvement of the fever, hypoxemia, and respiratory symptoms. Four days later, the patient was discharged to complete the antimicrobial course at home.

Molecular genotyping was performed on the isolates recovered from the 2 admissions in 2 ways: P1 subtyping (determining sequence variations in the polymorphic P1 adhesin gene)¹⁰  and multiple-locus variable number tandem repeat analysis (MLVA) (which counts the copy number of naturally varying tandem repeated DNA sequences found in different loci throughout the M pneumoniae genome).¹¹  The P1 gene was amplified by using primer pairs ADH1/ADH2¹²  and Mp5f/Mp16r¹³  to cover the 2 repetitive element loci, RepMP4 and RepMP2/3. MLVA typing was performed by using the 5-locus scheme (Mpn1, Mpn13, Mpn14, Mpn15, and Mpn16).^11,14  The highly variable Mpn1 locus was included to assure the discrimination power for strain tracking.^15,16  DNA sequences were analyzed and compared with reference sequences from the National Center for Biotechnology Information. Results showed that both strains were P1 subtype 2 with a novel variant (that has not been reported) denoted 2bv that has a 12-bp deletion (CCAATTCCCAAA) at the RepMP2/3 locus (position 3738, AP017318.1_ [179335-184254]) compared with variant 2b. The MLVA type of both strains was 53662. Thus, except for the macrolide-resistance mutation, the genetic fingerprints of the 2 M pneumoniae organisms obtained from the 2 clinical encounters were novel and identical, suggesting the persistence of the same strain.

Our case is the first to prove that M pneumoniae carriage can result in subsequent invasive pulmonary infection months after the initial M pneumoniae pneumonia in the setting of immunodeficiency. The authors could not find any reports of macrolide-susceptible M pneumoniae pneumonia followed by de novo mutation, chronic quiescent carriage, and breakthrough macrolide-resistant M pneumoniae pneumonia. A relatively recent study of ICU patients with ventilator-associated pneumonia revealed a surprisingly high prevalence of M pneumoniae (41%), and that subgroup had fewer ventilator-free days and lower average oxygenation. In this report, we raise the possibility that individuals colonized with M pneumoniae may have a more severe hospital course during a respiratory illness.¹⁷ 

The rate of nasopharyngeal M pneumoniae carriage in asymptomatic children varies significantly (0%–56%) in different epidemiological studies, and the carriage rate has been reported to be as high as 65% in children with asthma.^18,19  In a prospective study of children with M pneumoniae carriage in the Netherlands, researchers found that M pneumoniae persisted in the upper airways for up to 4 months,²⁰  and similar reports have originated from Sweden and the United States.^1,21  Even with macrolide therapy, prolonged shedding of M pneumoniae can last for weeks.²²  However, M pneumoniae persistence observations were based on M pneumoniae positivity, and whether the recovered M pneumoniae is the same strain or a new infection was not investigated. In our case, molecular genotyping was performed on the 2 isolates and revealed that the same M pneumoniae strain persisted in the patient and caused a second pneumonia more than 5 months after the initial infection.

In this case, 2 subpopulations of M pneumoniae harboring different macrolide-resistance point mutations (A2063T and A2063G) apparently emerged after the short course of azithromycin treatment of the initial pneumonia and poor response to macrolides during the subsequent M pneumoniae pneumonia. This likely reflects rapid evolution of the M pneumoniae strain in adaptation to the antimicrobial pressure. The short duration of treatment, use of bacteriostatic antibiotics, and immunocompromised status may have increased the chance of developing macrolide resistance and chronic M pneumoniae carriage.²³ 

Failure to detect M pneumoniae in the nasopharyngeal swab by PCR during the second pneumonia episode is consistent with the lower sensitivity of nasopharyngeal M pneumoniae PCR relative to sputum or BAL M pneumoniae PCR, as previously suggested.^24,25  It could also be due to a less-sensitive multiplexed PCR used to test the sample compared with real-time PCR that was not performed. In addition to the clinical and radiologic findings consistent with M pneumoniae pneumonia,²⁶  pulmonary infection by macrolide-resistant M pneumoniae was confirmed by isolation of the organism from BAL with broth microdilution and MIC determination and 23S rRNA sequencing. Clinical deterioration while on broad-spectrum antibiotics, including azithromycin, and subsequent improvement with fluoroquinolone support the causative role of macrolide-resistant M pneumoniae in the clinical course. This case documents that breakthrough M pneumoniae pneumonia from chronic bacterial carriage can occur in the setting of immunodeficiency.

The macrolide resistance rate of M pneumoniae in patients treated at Children’s of Alabama between 2015 and 2018 was 20.8% based on laboratory surveillance data, which is higher than most other areas of the United States.²⁷  Macrolide-resistant M pneumoniae can result in prolonged respiratory symptoms that necessitate changing antimicrobial agents to fluoroquinolones because of failure in response to macrolides, even in immunocompetent children.²³  PCR to detect macrolide-resistance gene mutations should be obtained, if possible, to confirm resistance. This is especially true in immunocompromised children who lack the phagocytic cells where the macrolides tend to concentrate.²⁸ 

The initial diagnosis of M pneumoniae pneumonia was based on clinical suspicion, radiologic abnormalities, positive M pneumoniae PCR assay in nasopharyngeal secretions, and response to macrolides. BAL is not the standard of care to diagnose M pneumoniae pneumonia because of the risk of sedation and invasiveness of bronchoscopy.²⁹  However, after failure of the clinically appropriate therapy of pneumonia in immunocompromised patient, BAL is often performed. We did not obtain respiratory samples between the 2 episodes to confirm M pneumoniae carriage because of the lack of respiratory symptoms. Nevertheless, the exact match in the genetic makeup (including a novel mutation) of the 2 M pneumoniae isolates (acquired 5 months apart from the same patient) highly suggests chronic carriage of the organism. However, reinfection of the child from a family member cannot be excluded, although we view this as unlikely. Prolonged M pneumoniae carriage and reactivation in our immunocompromised child may not be generalizable to the immunocompetent population. Immune dysfunction may increase carriage rate or reactivation of latent M pneumoniae, leading to invasive infection.²³ 

Given the high prevalence of M pneumoniae carriage and widespread use of macrolides, breakthrough macrolide-resistant M pneumoniae pneumonia may not be an uncommon event in immunocompromised hosts. Accordingly, effective treatment of M pneumoniae pneumonia with bactericidal antibiotics may reduce the risk of chronic carriage and subsequent life-threatening infections.

Technical support for Mycoplasma pneumoniae culture, PCR, and MIC measurements was provided by Donna Crabb, Amy Ratliff, and Melanie Fecanin at the University of Alabama at Birmingham Diagnostic Mycoplasma Laboratory.

BAL

bronchoalveolar lavage

MIC

minimum inhibitory concentration

MLVA

multiple-locus variable number tandem repeat analysis

PCR

polymerase chain reaction

rRNA

ribosomal RNA