Growing Threat of Macrolide-Resistant Mycoplasma pneumoniae Among Children: What We Know and What We Need

Abstract

Mycoplasma pneumoniae is a unique, wall-less bacterium and a leading cause of community-acquired pneumonia in school-aged children. Epidemiologically, M. pneumoniae causes endemic and epidemic infections every 3–5 years, with recent surges noted globally since mid-2023. Diagnostic challenges persist; although polymerase chain reaction and serology are widely used, both carry risks of false positives or negatives owing to prolonged shedding or delayed antibody responses, respectively. Novel approaches such as ELISpot assays for antibody-secreting cells may improve diagnostic accuracy. Macrolides remain the first-line treatment in children because of the toxicity of tetracyclines and fluoroquinolones. However, macrolide-resistant M. pneumoniae (MRMP) has emerged as a major concern, especially in the Western Pacific Region (South Korea, China, and Japan), where resistance rates have exceeded 80% in recent years. Resistance is developed by 23S rRNA mutations (notably A2063G) and driven primarily by macrolide over usage and clonal expansion—particularly in sequence type 3 strains. Despite high resistance rates, many M. pneumoniae pneumonia cases remain self-limiting, and clinical outcomes are not substantially different between macrolide-sensitive and macrolide-resistant infections. Clinical studies show that while MRMP is associated with longer fever and hospital stays, it rarely increases severity. Tetracyclines and quinolones have been shown to lead to faster fever resolution in MRMP patients, and treatment guidelines now recommend switching to second-line agents when macrolide-refractory disease is suspected. Nevertheless, a considerable proportion of patients recover spontaneously, regardless of antibiotic choice, likely owing to the self-limiting nature of the pathogen, misdiagnosis of viral infections, or the immunomodulatory effects of macrolides. Amid rising MRMP, this review highlights the complex interplay between bacterial resistance, diagnostic limitations, and host factors in shaping M. pneumoniae disease outcomes.

Graphical Abstract

HISTORY AND MICROBIOLOGY

Mycoplasma pneumoniae was first identified in the 1940s by Monroe D. Eaton and colleagues and was initially referred to as the Eaton agent. At the time, the organism was believed to be a type of virus because it could pass through sterile filters and failed to grow on conventional bacterial culture media.1 In 1963, Dr. Robert Chanock2 demonstrated that M. pneumoniae was a major cause of atypical pneumonia in children and adolescents, and it was subsequently named Mycoplasma pneumoniae. M. pneumoniae is very small, measuring approximately 1–2 µm in length and 0.1–0.2 µm in width, and contains a genome of approximately 816,400 base pairs comprising 687 genes. Thus, M. pneumoniae is one of the smallest self-replicating bacteria. It is a prokaryotic pathogen that lacks a cell wall and relies on the host for nutrient acquisition and survival. Its triple-layered cell membrane is structurally supported by sterols. The P1 protein, along with accessory proteins such as P30, P90, and P40, facilitates attachment and binding to host cells via receptors.3

The pathogenic effects of Mycoplasma infections result from localized tissue disruption, cytotoxicity, and host immune responses. M. pneumoniae initially attaches to the surface of airway mucosal epithelial cells and releases toxic molecules that damage host cells.4 Reactive oxygen species, including H₂O₂ and HO₂ produced by M. pneumoniae, have been shown to damage epithelial cells and cilia, and community-acquired respiratory distress syndrome (CARDS) toxin is known to promote local tissue damage and cytotoxicity. Furthermore, interactions between respiratory epithelial cells and the surface lipoproteins of M. pneumoniae may stimulate the host immune system via Toll-like receptor (TLR)-2 or TLR-4 pathways, potentially inducing intercellular adhesion molecule receptor synthesis.5,6

EPIDEMIOLOGY

M. pneumoniae is one of the most common causes of community-acquired pneumonia (CAP) in school-aged children. With improvements in hygiene and the widespread use of effective antibiotics and vaccines, the incidence of pneumonia caused by classic pyogenic bacteria, such as Streptococcus pneumoniae, has markedly decreased.7 As a result, M. pneumoniae has become the leading bacterial pathogen causing pneumonia in children, particularly in developed countries. In prospective multicenter studies on pediatric CAP conducted in the United States from 2010–2012 and 2015–2018, M. pneumoniae was identified as the etiology in approximately 8–9% of cases, with a higher detection rate of 12–20% among children aged 5 years and older.8,9 Similarly, in a retrospective multicenter study conducted in Korea from 2015 to 2020, M. pneumoniae accounted for 16.8% of the cases, making it the most common single pathogen identified; the detection rate was even higher among children aged 5 years and older, ranging from 23.8% to 28.4%.10

In general, respiratory syncytial virus (RSV), parainfluenza virus (PIV), and human metapneumovirus (hMPV) are common causes of CAP in children. However, owing to the characteristics of M. pneumoniae, which tends to cause clinically significant pneumonia in older children and leads to large-scale epidemics every 3–5 years,11 the proportion of CAP attributable to M. pneumoniae can vary widely depending on the study population and period included. Although endemic infections may occur annually in the fall and winter, large epidemics can significantly increase the disease burden, making M. pneumoniae a major cause of pneumonia during such periods.8,10

There was a global outbreak of M. pneumoniae from mid-2019 to early 2020, just before the start of the coronavirus disease 2019 (COVID-19) pandemic. Following this period, the detection of M. pneumoniae was minimal for approximately three years, but a global resurgence began again in mid-2023.12,13 A similar trend was observed in South Korea; compared with other major respiratory pathogens, such as PIV, RSV, hMPV, and influenza virus, which re-emerged sequentially from autumn 2021 through the second half of 2022, M. pneumoniae reappeared relatively late as a major respiratory pathogen.14,15 This delayed re-emergence is believed to be partly attributable to the microbiological characteristics of M. pneumoniae, including its slow generation time (~6 hours), long incubation period (1–3 weeks), and relatively low transmission rate.13

DIAGNOSTIC DILEMMAS IN M. PNEUMONIAE INFECTION

In the early days, culture and the cold agglutinin test were used to diagnose M. pneumoniae infection. However, culture has a low success rate and requires a long incubation period, while the cold agglutinin test is limited by its high rates of both false positives and false negatives. As a result, specific antibody testing and polymerase chain reaction (PCR) have become the main diagnostic tools in recent years. Specific antibody detection is typically performed using chemiluminescence assays or enzyme-linked immunosorbent assays to measure IgM and IgG separately or by using indirect particle agglutination (PA) testing to detect total specific antibody levels, including IgM, IgG, and IgA.16

According to previous studies, antibody titers measured using the PA method can persist at levels up to 1:320 for as long as 6–7 months after infection; therefore, a titer of at least 1:640 is recommended as the diagnostic threshold for acute infection.11 Since M. pneumoniae typically has a long incubation period, antibodies are often produced during the acute phase, and even a short interval of 5–7 days between paired serum samples can lead to a fourfold increase in titer. However, in clinical practice, it is often impractical to obtain paired serum samples for the rapid diagnosis of M. pneumoniae pneumonia (MPP), so single IgM measurements are frequently used instead. Notably, IgM levels can rise even in mild or asymptomatic infections and may remain elevated at high levels for several weeks to months, which carries a considerable risk of false positives. Conversely, IgM can yield false-negative results or very low titers in the very early stages of symptom onset, complicating clinical interpretation.17,18 Likewise, PCR is not a perfect diagnostic tool; M. pneumoniae DNA can be detected in asymptomatic healthy children, and colonization may persist for weeks, even among pneumonia patients, necessitating caution when interpreting results for acute infection.17,19 Recently, an ELISpot assay that measures antibody-secreting cells (ASCs) rather than serum antibodies has been developed and is expected to serve as a promising marker for acute-phase infection in the future.19

MACROLIDE RESISTANCE OF M. PNEUMONIAE

Mechanisms and trends

M. pneumoniae lacks a cell wall and is thus naturally resistant to beta-lactams but is susceptible to antibiotics that inhibit protein synthesis (macrolides and tetracyclines) or DNA replication (fluoroquinolones).4 Macrolides are considered first-line antibiotics for the treatment of MPP in children because of the toxicity risks of tetracyclines and fluoroquinolones in this age group. Early use of macrolides can help reduce the duration of symptoms. However, treatment does not result in a dramatic response and does not eradicate organisms from the respiratory tract; thus, the beneficial effect of macrolides in the treatment of MPP remains unclear.20 Nevertheless, macrolides are also recommended as the main empirical antibiotics for the treatment of CAP in children.21 As a result, they are frequently used in children, who commonly present to hospitals with respiratory infections.

Macrolides bind to nucleotides in domains II and/or V of the 23S rRNA within the 50S ribosomal subunit of M. pneumoniae. This binding inhibits protein synthesis by causing premature release of peptidyl transferase RNA from the ribosome. Mutations at the target site alter the binding site’s structure.20 Macrolide resistance in M. pneumoniae was first experimentally demonstrated by Norio Okazaki and colleagues in the 1990s, who identified mutations at nucleotide positions 2063 or 2064 in the 23S rRNA gene as the underlying mechanism. In particular, the A2063G mutation was shown to confer high-level resistance to macrolides, whereas susceptibility to tetracyclines remained unchanged.22 Since then, the proportion of macrolide-resistant M. pneumoniae (MRMP) strains carrying these mutations has increased markedly among clinical isolates, especially in Japan from 2003–2006 and, similarly, in South Korea from 2003–2011.23,24 MRMP has been particularly prevalent in countries in the Western Pacific Region, namely, South Korea, China, and Japan, where a systematic review of studies published up to September 2021 reported an average macrolide resistance rate of 53.4%, compared with 1.4–9.8% in other continents; the vast majority (96.8%) of these cases were associated with the A2063G mutation.25,26 Trends in the 2000s have varied among these countries. In China, the resistance rate has consistently maintained a high prevalence at 80–90%; in Japan, the rates increased until 2011–2012, declined between 2013 and 2018, and began rising again after 2019, whereas in South Korea, the rate has shown a continuous upward trend except during the 2019 epidemic (Fig. 1 and Supplementary Table 1).

Fig. 1. Annual trends in macrolide-resistant Mycoplasma pneumoniae among children in three Asia–Pacific countries, 2000–2024. This graph, based on the results of the studies included in Supplementary Table 1, shows annual macrolide-resistant M. pneumoniae rates. For years with multiple studies, a case number–weighted arithmetic mean was used to calculate the annual proportion. The dotted lines indicate periods without available data, with values interpolated by connecting the rates from the preceding and following years. The yellow bar indicates the 2009 novel influenza pandemic period, and the gray bar represents the coronavirus disease 2019 pandemic period.

Antibiotic pressure and molecular epidemiology

The increase in MRMP might be attributed primarily to two factors, macrolide antibiotic pressure and the clonal expansion of resistant strains. In vitro studies have shown that exposure to macrolides can promote selection for MRMP strains.27 Macrolide-resistant strains have also emerged during treatment, with corresponding molecular mutations observed.28 Another finding supporting the role of antibiotic pressure is that Western Pacific countries, where macrolides are widely used for respiratory infections, have significantly higher MRMP rates than other regions do. Additionally, in Japan, increased use of alternative second-line agents for MRMP, such as tosufloxacin in children, have led to a reduction in macrolide use, corresponding to a decline in MRMP incidence.24,26,29,30

For genotyping M. pneumoniae, sequencing a ~400 bp region of the P1 gene and multiple-locus variable number tandem-repeat analysis have long been used; however, these methods have limited discriminatory power and weak correlation with MRMP. Consequently, multilocus sequence typing (MLST) has recently become the preferred approach.31,32,33 MLST analyses have shown that the most common genotype in Japan and South Korea is sequence type (ST) 3, which is especially prevalent among MRMP isolates.34,35 In South Korea, MRMP-ST3 strains appeared to have been introduced from approximately 2006–2007 and drove the increase in MRMP from 2019–2020.35,36 In contrast, during the decline in MRMP in Japan between 2015 and 2019, the prevalence of MRMP-ST3 decreased, whereas that of macrolide susceptible M. pneumoniae (MSMP)-ST7 increased markedly.37 In Taiwan, the MRMP rate has declined sharply in recent years following the COVID-19 pandemic, with evidence suggesting that the same ST3 genotype shifted from MRMP strains to MSMP strains.38 Collectively, these molecular epidemiology findings imply that factors beyond simple clonal expansion may contribute to the rise of MRMP or that distinct MRMP and MSMP strains may coexist within the ST3 genotype.

Clinical implications

With the increasing prevalence of MRMP, there is significant interest in its clinical impact and treatment strategies, especially in countries such as South Korea, China, and Japan, where MPP is relatively common. Consequently, most related studies have been conducted in these three countries. According to a meta-analysis that integrated these studies, MRMP infections showed no difference in clinical severity or chest radiographic findings compared with those of MSMP infections. However, in patients with MRMP, the duration of fever was on average 1.7 days longer, the length of hospital stay was 1.6 days longer, the duration of antibiotic treatment was 2.9 days longer, and the time to defervescence after macrolide administration was extended by 2.0 days.39 On the basis of Korean studies, it appears that, rather than MRMP status alone, the presence of lobar consolidation or pleural effusion on chest radiographs is a more important determinant of clinical severity and treatment response.40,41

THERAPEUTIC CHALLENGES IN M. PNEUMONIAE INFECTION

M. pneumoniae is a microorganism with characteristics that lie between those of pyogenic bacteria and viruses, and infections are generally known to be mostly self-limiting. Therefore, the effectiveness of antibiotic therapy has long been a subject of investigation. In a meta-analysis combining the results of 17 studies, including five randomized controlled trials (RCTs), antibiotic treatment for CAP caused by M. pneumoniae in children was associated with a reduction in fever duration but showed no statistically significant correlation with overall clinical improvement.42 To draw more reliable conclusions on this issue, a multicenter, double-blind RCT is currently being conducted in Switzerland.43

Studies have also been conducted to evaluate the actual clinical effectiveness of macrolides and alternative second-line antibiotics—tetracyclines and fluoroquinolones—in the treatment of MRMP. For example, during the 2011 epidemic in Japan, a study involving 258 children younger than 15 years reported that 87.1% had MRMP. The rates of defervescence within 24 hours of antibiotic initiation were significantly greater for minocycline (57.7%) and doxycycline (81.3%) than for macrolides (30.8%), whereas the rates for tosufloxacin (30.8%) were not different.44 During the nationwide epidemic in Japan between 2010 and 2013, when the MRMP proportion reached 80%, a large-scale study using national big data revealed no differences in the length of hospital stay or 30-day mortality among patients treated with macrolides, quinolones, or tetracyclines.45 In a similar period (2008–2012), a study in Japanese adolescents and adults aged 16 years and older demonstrated that, within 48 hours of antibiotic initiation, the proportion of patients who became afebrile did not differ significantly among the treatment groups in MSMP cases (71–91%) between the treatment groups. However, among MRMP patients, the defervescence rate was significantly higher with quinolones (77%) and minocycline (85%) than with macrolides (28%).46 A multicenter prospective study conducted in Japan between 2013 and 2015 also revealed that, among hospitalized children with MRMP, compared with macrolides and tosufloxacin, minocycline significantly reduced fever duration.47 In South Korea, a study analyzed cases of MPP collected during two epidemic periods, 2014–2015 and 2019–2020, and reported that the proportion of MRMP was 40.7%. In this study, a comparison between the macrolide and doxycycline groups revealed that doxycycline had a significantly greater effect, with improvement in fever and chest radiographic findings occurring approximately three days earlier than with macrolide treatment.48 A meta-analysis incorporating these key studies and other research conducted from 1990–2018 revealed that fever duration and length of hospital stay were shorter in the tetracycline group than those in the macrolide group. There were no significant differences in fever improvement within 24 hours between patients receiving fluoroquinolones and those receiving macrolides, although the defervescence rate was greater after 48 hours in the fluoroquinolone group.49

On the basis of the findings of these studies, treatment guidelines for severe macrolide-refractory MPP in children were established in Korea in 2019. In 2024, these guidelines were revised to include nonsevere cases as well.50,51 Both versions recommend the use of macrolides as the first-line treatment upon diagnosis of MPP, with a switch to second-line agents—specifically, tetracyclines (doxycycline and minocycline) or quinolones (levofloxacin and tosufloxacin)—if there is no clinical improvement within 48–72 hours, indicating macrolide-refractory disease. The 2024 revision, reflecting accumulated evidence, further recommends tetracyclines as the preferred second-line agents over quinolones. In addition, numerous studies have investigated the effectiveness of corticosteroids in cases of macrolide-refractory or severe MPP. Most of these studies have shown that corticosteroid use leads to a reduction in fever duration regardless of the antibiotic used; however, the evidence regarding its impact on overall clinical improvement remains inconclusive.52,53,54 Most other countries present antibiotic treatment guidelines similar to those of Korea, while the level of recommendation for corticosteroid use is likewise still limited.20,21

As a result of these trends, various treatments have been used for MPP that reemerged in South Korea from the second half of 2023 following the COVID-19 pandemic. Reported data show that macrolides (84.2%), tetracyclines (25.9%), fluoroquinolones (11.0%), and corticosteroids (55.6%) were administered for treatment, with the macrolide resistance rate estimated at approximately 87.0%.14 Notably, β-lactam antibiotics were also prescribed in 56.4% of cases, highlighting the need to reduce unnecessary β-lactam use in children with nonsevere pneumonia during MPP epidemics. An analysis of 389 MPP cases during the epidemic that continued through the end of 2024 examined treatment regimens and fever duration. Even in patients for whom no effective antibiotic therapy for Mycoplasma was provided (21.9%), the mean duration of fever was five days. Among patients treated only with macrolides (18.0%) and those receiving other regimens, including second-line agents such as tetracyclines or fluoroquinolones, alone or in combination with corticosteroids, the average time to defervescence after initiating treatment ranged from 0–2 days, with no substantial differences observed.55 These findings suggest that spontaneous recovery still occurs in MPP and that macrolides may remain clinically beneficial even when the MRMP rate is high. This finding supports the continued use of current guidelines, which recommend macrolides as the first-line treatment, with a switch to second-line agents if macrolide-refractory disease is suspected.51

THE MRMP PARADOX: WHY MACROLIDES SOMETIMES WORK

The fact that macrolides work sometimes may be attributable to M. pneumoniae being inherently a self-limiting pathogen, allowing for spontaneous recovery in many cases regardless of the treatment administered.42 At present, there is no way to predict which strains or factors may lead to spontaneous resolution or progression to severe pneumonia, and addressing this remains one of the most significant unmet needs in the field. Therefore, more research is needed to clarify the host, pathogen, and environmental factors that influence the clinical course of MRMP infections. For example, it is possible that strain-specific differences in key virulence factors—particularly the CARDS toxin gene—may contribute to variations in clinical outcomes.5 Large-scale analyses combining clinical data with whole-genome sequencing and quantitative assessments of virulence gene expression could help elucidate these differences. In addition, M. pneumoniae employs various immune evasion strategies, such as antigenic variation of surface proteins, which may influence persistence and disease severity. Exploring the genetic diversity among strains may help identify novel virulence determinants and clarify why certain strains are associated with more severe or refractory cases.16 On the host side, investigating whether certain genetic or immunological predispositions contribute to exaggerated inflammatory responses to M. pneumoniae infection is equally important.56 Advances in next-generation sequencing and transcriptomic studies offer promising tools for exploring these host factors.

Another hypothesis is that a considerable proportion of cases diagnosed as MPP in previous studies—and even in current clinical practice—may in fact represent viral pneumonias that are self-limiting. As mentioned earlier, diagnostic limitations contribute to a substantial risk of false positives, since M. pneumoniae DNA can be detected for extended periods after infection or even in colonizers. In fact, numerous studies have shown that a significant proportion of CAP patients diagnosed with MPP also have respiratory viruses detected concurrently.10,14,41 Future studies should therefore employ more accurate diagnostic methods to minimize this possibility. Indeed, a recent RCT on MPP treatment initiated in Switzerland has incorporated the acute-phase ASC ELISpot test, which is currently considered the most reliable method for distinguishing true acute infection.43

A final hypothesis is that this phenomenon may be due to the immunomodulatory effects of macrolides.57 This property is already well established, and numerous studies have demonstrated that corticosteroids, as representative anti-inflammatory agents, are effective in reducing fever in patients with MPP.52,53,54 Therefore, macrolides may facilitate the self-limiting course of M. pneumoniae infection through their immunomodulatory, rather than antimicrobial, action. This area also requires further well-designed experimental and clinical studies to be adequately validated.

CONCLUSIONS

Since the COVID-19 pandemic, heightened public awareness and concern regarding respiratory infections have led to greater focus on MPP as a significant public health issue. In addition, the recent marked increase in MRMP in Korea has drawn considerable clinical and research interest. However, the quantity and quality of clinical and basic research data on M. pneumoniae remain insufficient. Given the inherent challenges of conducting well-controlled studies on CAP in children and ensuring accurate diagnosis, careful interpretation and application of available evidence are essential. In particular, key unanswered questions include the clinical significance of M. pneumoniae detection, the true impact of macrolide resistance, and the effectiveness of second-line therapies (Table 1). These topics represent urgent priorities for future research. Clinicians should also remain attentive to emerging data in these areas and apply new evidence cautiously in daily practice.

Table 1. Current knowledge and unmet needs in macrolide-resistant Mycoplasma pneumoniae .

Aspect	What we know	What we need
Epidemiology	High and increasing MRMP rates in Western Pacific countries (esp. Korea, Japan, China); periodic epidemics; clonal expansions (e.g., ST3)	Better real-time surveillance and genotype monitoring globally; understanding local transmission dynamics
Resistance mechanism	23S rRNA mutations (mainly A2063G); clonal spread of resistant strains	Mechanistic studies to explain persistence and spread; exploration of other possible resistance pathways
Diagnosis	PCR and serology are commonly used but have limitations (false positives/negatives, persistent colonization)	Rapid, accurate tests distinguishing colonization from active infection; validated biomarkers (e.g., ASC tests)
Clinical impact	MRMP prolongs fever and hospital stay slightly; not clearly linked to more severe radiologic findings	Predictive markers to identify patients at risk of severe outcomes; large-scale prospective studies
Treatment	Macrolide is first-line; tetracyclines and fluoroquinolones effective as second-line; steroids may help reduce fever	Better evidence for optimal second-line therapy in children; safety data on alternative antibiotics; more robust RCTs
Natural history	Many cases resolve spontaneously; macrolide benefit partly due to anti-inflammatory effect	Clear criteria to identify when antibiotics are truly needed; trials comparing antibiotic vs. placebo arms
Public health and guidelines	Updated local guidelines exist (e.g., Korea’s 2024 update); but practice still varies	Harmonized, evidence-based global guidelines; stewardship to avoid unnecessary macrolide use

MRMP = macrolide-resistant Mycoplasma pneumoniae, ST = sequence type, PCR = polymerase chain reaction, ASC = antigen-secreting cell, RCT = randomized controlled trial.

SUPPLEMENTARY MATERIAL

Supplementary Table 1

Prevalence of macrolide-resistant Mycoplasma pneumoniae in children across three Asia-Pacific Nations, 2000–2024