Consideration in treatment decisions for refractory Mycoplasma pneumoniae pneumonia

1. Difficulties in laboratory diagnosis of macrolide-resistant MP infection

The diagnosis of macrolide-resistant MP infection can be made by measuring the MIC of strains isolated by culture or by detecting the point mutation of 23S rRNA by polymerase chain reaction (PCR). The MICs of each antimicrobial can be determined by a microdilution method [3]. The criteria for drug-resistant MP, proposed by the Japanese Society for Mycoplasmology [18], is MICs of ≥16 µg/mL for erythromycin, clarithromycin, and azithromycin [18]. Among the 36 isolates of macrolide-resistant MP from Korean children, 2010–2015, the MIC₅₀/MIC₉₀ values for erythromycin, clarithromycin, and azithromycin were 16/128 µg/mL, 64/128 µg/mL, and 8/16 µg/mL, respectively, which were higher than those for 23 macrolide-susceptible MP isolates, with MIC₅₀/MIC₉₀ values of 0.001/0.002 µg/mL for erythromycin, 0.001/0.002 µg/mL for clarithromycin, and 0.001/0.001 µg/mL for azithromycin [3]. In a recent study on MP isolates retrieved from Japanese children, 2011–2016, MIC₅₀/MIC₉₀ of erythromycin, clarithromycin, and azithromycin for macrolide-resistant MP was >128/>128, >128/>128, and 32/64 µg/mL, respectively, while those for macrolide-sensitive MP was 0.004/0.008, 0.002/0.004, and 0.00025/0.0005, respectively [19]. However, this test is not useful to clinical practice because it requires long time (>2 weeks) for bacterial growth.

The PCR method can be performed to identify macrolide resistance using cultured MP isolates or DNA extracted from nasopharyngeal samples [3,20]. The PCR products can be sequenced to identify the transitions in domain V of the 23S rRNA gene that are associated with macrolide resistance. Multiplexed PCR test has been widely used to detect MP from respiratory samples in many hospitals in Korea, but the sequencing test to identify macrolide resistance is not available in clinical practice. In Japan, it is recommended that clinicians should assess the efficacy of macrolides and consider to collect samples available to clarify the susceptibility of the causative organism in the absence of defervescence 48–72 hours after the start of macrolide therapy [21]. It is needed to introduce clinically available tests to detect macrolide resistance of MP.

2. Uncertain impact of macrolide resistance on clinical course

Since the first macrolide resistance of MP was isolated from clinical samples obtained from children with pneumonia in Japan in 2000, it has been suggested that macrolide resistance can be associated with severe respiratory symptoms in children. However, the clinical relevance of resistant strains is largely debated, because it is not clarified whether resistant strains are related with severe or prolonged disease [22-27]. A recent study reported that some of macrolide-resistant MP pneumonia patients who received macrolide treatment showed clinical failure with progress of disease [27]. In contrast, it has been reported that macrolide resistance alone does not affect clinical course [3]. A multicenter study showed that the time to defervescence after initiation of macrolide therapy in MP pneumonia children was longer in the macrolide-resistant group than in the macrolidesusceptible group (5.7 days vs. 4.1 days). However, multivariate logistic regression analysis revealed that the presence of extrapulmonary signs, homogeneous lobar consolidation, and parapneumonic effusion was responsible for prolonged fever ≥7 days following macrolide treatment regardless of macrolide resistance [3]. In a Taiwan study, while intensive care unit (ICU) cases (n=34) were associated with more pleural effusion, longer fever duration, and longer hospital stay, compared with non-ICU cases (n=291), macrolide resistance was similar in ICU and non-ICU groups [23].

1. Serologic diagnosis

Diagnosis of MP infection in routine clinical practice has been mainly based on serology, since bacterial culture of this organism takes long time and lacks sensitivity. Several specific antibody tests can be used to measure serum antibodies to MP. The complement fixation test, which was the standard serologic method for the diagnosis of MP infection, is currently not widely used due to nonspecific reactions and lack of sensitivity [28]. Alternative formats adapted for commercial serologic assays include indirect immunofluorescence assay, particle agglutination (PA) assay, and enzyme-linked immunoassay (EIA). The PA assay was the most widely used method in Korea until recently, because it is easy to perform and provide quantitative results with acceptable sensitivity [29]. However, the PA assay has some disadvantages: ambiguity in the determination of agglutination, nonspecific reactions, and inability to discriminate between IgG and IgM [30]. The EIA can detect IgG and IgM separately, which can be somewhat helpful to distinguish current from past infections [29]. Recently, many diagnostic laboratories have replaced the PA tests with EIA for demonstration of MP antibodies [20]. The current widely used IgM EIA test has the advantage of being able to diagnose using a single sample, but it is likely to have a false-positive result and can be detected for a long time after infection. Therefore, in Japanes Guiding Principle, even if IgM is positive, it is recommended to confirm the diagnosis using IgG tests from paired samples [21]. Testing paired samples is practically impossible and is not always necessary for routine MP pneumonia care. In the case of refractory pneumonia, testing IgG from paired samples can be used to confirm the diagnosis of MP infection since it is often more than 1–2 weeks after the onset of illness when determining whether to change the treatment.

2. Diagnosis using molecular detection

Molecular diagnostic tests using PCR are more advantageous than culture tests and serologic tests since they can detect the presence of bacterial DNA quickly and more sensitively [14,31,32]. In addition, PCR assays can be used to diagnose MP infection earlier than serology tests, particularly in young children and immunocompromised patients who do not produce antibodies [33]. The PCR can also play a complementary role in interpreting the results of serologic tests to exclude the past infections. Because both serology and PCR may occasionally give nonoverlapping false-negative results, a combination of these 2 diagnostic methods would be helpful [20].

However, PCR testing using nasopharyngeal specimen or sputum may be positive also in the presence of upper respiratory carriage or asymptomatic infection. PCR cannot distinguish actual infection from carriage or asymptomatic infection. Unfortunately, there is no current diagnostic test that is able to differentiate between asymptomatic carriage and symptomatic infection [32].

3. Asymptomatic carriage of MP in the upper respiratory tract

Like many other respiratory pathogens, MP can be carried asymptomatically in upper respiratory tract [2]. Recent studies have demonstrated that asymptomatic carriage of MP is highly prevalent. In a Dutch study, MP DNA was detected in 21% of asymptomatic healthy children in 16% of children with respiratory symptoms. The MP detection rates using culture were not significantly different between asymptomatic and symptomatic children (1.0% vs. 1.6%) [32]. Among symptomatic children, the prevalence of MP by real-time PCR did not differ significantly between children with an upper respiratory infection and a lower respiratory infection (15.9% vs. 15.6%) [32]. This suggests that among pneumonia patients with positive MP PCR in nasopharyngeal specimens, some of them are caused by a pathogen other than MP and MP only plays a role as a colonizer. A longitudinal observation of asymptomatic MP carriers revealed that MP bacteria were present for up to 4 months without causing disease [32]. This prolonged shedding of MP seems to be related with that natural immunity against MP does not persist for a long time after MP infection and frequent reinfection is possible in the same person [9].

4. Considering infections caused by other respiratory pathogens

When patients with pneumonia who have been empirically treated with macrolide have fever and have no clinical improvement, the possibility of infection with other pathogens should be considered as well as that of macrolide-resistant MP infection. Children with MP infection were found to simultaneously carry many kinds of pathogens in their upper respiratory tract (Table 2) [32]. These pathogens include the bacteria Staphylococcus pneumoniae, Staphylococcus aureus, Moraxella catarrhalis, and Haemophilus influenzae, and the viruses influenza A/B, human metapneumovirus, respiratory syncytial virus, parainfluenzavirus, rhinovirus, coronavirus, bocavirus, and adenovirus [20]. Recent studies have reported coexistence of MP with other pathogens in children with MP pneumonia. Other pathogens were simultaneously found in 20.5%−56.1% of the patients with MP pneumonia [20,34-36]. However, the impact of coinfections in MP pneumonia on disease severity is not yet determined [20,37,38].

Table 2.

Viral coinfection in children with macrolide-resistant Mycoplasma pneumoniae pneumonia from studies published from 2018–2020

Recently, multiplexed PCR tests for bacterial or viral pathogen of respiratory infections have been widely used in many hospitals, which can be helpful to rule out the possibility of coinfection. However, it can require multiple nasopharyngeal swabs that make young patients and their caregivers discomfort, and additional cost. Since virus panel testing is currently not covered by national health insurance for patients of other ages than neonates, it may not be applied to all patients requiring the tests.

1. Tetracyclines

Tetracyclines are generally safe in children although concerns about adverse reactions, such as gastrointestinal disturbances, esophagitis, photosensitivity, tooth discoloration, still remain [6,47]. Most concerning side effect is incorporation into calcifying tissues at the time of their administration. The affinity for mineralizing tissue leads to incorporation into calcifying tissues, such as teeth, cartilage, and bone [48]. Studies in the 1960s and 1970s showed that repeated administration of tetracycline to neonates and young children was associated with teeth staining. As a consequence, the use of tetracycline in young children was discouraged. Approved age of this drug use varies from country to country: ≥12 years of age in Korea and the United Kingdom, ≥8 years of age in the United States. However, later studies on doxycycline use reported no or only negligible tooth staining even in young children aged 2–8 years [49,50]. This difference was attributed to a low affinity for calcium of doxycycline [51]. Factors related to tooth discoloration are dosage, duration of treatment, stage of tooth mineralization, and activity of the mineralization process [52]. However, previous studies showed that short cycles with limited courses of treatment (less than 6 courses, 6 days per course) caused negligible tooth discoloration in children treated in the first 5 years of life [53].

Minocycline is more likely to cause other central nervous system effects (e.g., dizziness, lack of concentration, ataxia, vertigo, and tinnitus associated with weakness, nausea, and vomiting) and pigmentation of various body sites. Although overall adverse reactions are reported more frequently for minocycline, minocycline and doxycycline are generally well tolerated in young children [48].

2. Fluoroquinolones

Fluoroquinolones have not been widely accepted as safe drug for use in children because of the observation that young animals consistently develop damages on articular surfaces of cartilage after exposure to these drugs. However, fluoroquinolones are generally well-tolerated as supported by recent reports on musculoskeletal safety concerns. Retrospective and prospective assessments of children exposed to fluoroquinolones have not clearly demonstrated the development of lesions consistent with those observed in young animals [54,55]. A recent comparative, unblinded, observational surveillance study has monitored the occurrence of predefined musculoskeletal disorders (arthralgia, arthritis, tendinopathy, and gait abnormality) in children participating in prospective, randomized efficacy trials. Levofloxacin was well tolerated during and for 1 month after therapy [56]. Even though the incidence of musculoskeletal disorders was greater in levofloxacin group compared with comparator group at 2 months (2.1% vs. 0.9%; P=0.04) and 12 months (3.4% vs. 1.8%; P=0.03) after starting therapy [56], 85% of these disorders consisted of reports of arthralgia, without any objective signs of joint disease documented for the children by the investigator at the time of the follow-up visit examination. Moreover, the results of following-up to 5 years after treatment revealed no clinically detectable difference between levofloxacin- and comparator-treated children in musculoskeletal toxicity [57]. It suggested that risks of cartilage injury with levofloxacin appear to be uncommon, are clinically undetectable during 5 years, or are reversible [57].

Other potential adverse reactions of fluoroquinolones, albeit very uncommon in children, include central nervous system adverse effects (seizures, headaches, dizziness, lightheadedness, sleep disorders, and hallucinations) and peripheral neuropathy. However, recent studies have reported that the rates of neurologic events were statistically similar between fluoroquinolone- and comparator-treated children [56,58].

Although there has been no compelling published evidence that supports the occurrence of sustained injury to developing bones or joints in children treated with currently available fluoroquinolones, the U.S. Food and Drug Administration (FDA) analysis of ciprofloxacin safety data suggests the possibility of increased musculoskeletal adverse events. The rate of arthropathy in the ciprofloxacin group exceeded that of the comparator group by more than 6% [59]. In 2016, FDA restricted the use of fluoroquinolones even in adults only if no other alternative treatment exists [60]. It is because of the potential risks of disabling side effects involving tendons, muscles, joints, nerves, and the central nervous system.

No compelling published evidence to date supports the occurrence of sustained injury to developing bones or joints in children treated with available fluoroquinolone agents; however, FDA analysis of ciprofloxacin safety data suggests the possibility of increased musculoskeletal adverse events.

Tosufloxacin is an oral fluoroquinolone developed by Toyama Chemical Co. Ltd. (Tokyo, Japan) in 1990. The granule form was released for administration to children with otitis media or pneumonia in Japan in January 2010. In Japan, the use of tosufloxacin for children is recommended only for pneumonia when a resistant bacterial infection is suspected and other antimicrobial agents are expected to be ineffective. A Japanese guiding principle has recommended that tosufloxacin may be considered for patients with MP pneumonia who do not respond to macrolide even in children [21]. Although no adverse reaction to the musculoskeletal system has been reported in children yet, it has not been approved for use in children in other countries than Japan due to a potential risk of inducing cartilage and joint toxicity in children. Additionally, because the increased use of fluoroquinolones has a risk of emerging resistant bacteria [61], decisions about antibiotic use should be made carefully.

Therefore, when prescribing a fluoroquinolone to a child or adolescent, the risks and benefits of this drug should be reviewed and discussed with patients and their caregivers or legal guardians. They should be aware of side effects, including rash, diarrhea, joint or tendon pain, confusion, or numbness or tingling of extremities while taking the antibiotic.