Since its initial description in the 1940s and eventual elucidation as a highly evolved pathogenic bacterium, Mycoplasma pneumoniae has come to be recognized as a worldwide cause of primary atypical pneumonia. Beyond its ability to cause severe lower respiratory illness and milder upper respiratory symptoms it has become apparent that a wide array of extrapulmonary infectious and postinfectious events may accompany the infections in humans caused by this organism. Autoimmune disorders and chronic diseases such as asthma and arthritis are increasingly being associated with this mycoplasma, which frequently persists in individuals for prolonged periods. The reductive evolutionary process that has led to the minimal genome of M. pneumoniae suggests that it exists as a highly specialized parasitic bacterium capable of residing in an intracellular state within the respiratory tissues, occasionally emerging to produce symptoms. This review includes discussion of some of the newer aspects of our knowledge on this pathogen, characteristics of clinical infections, how it causes disease, the recent emergence of macrolide resistance, and the status of laboratory diagnostic methods.
Biological characteristics of Mycoplasma pneumoniae
Mycoplasmas represent the smallest self-replicating organisms capable of cell-free existence, both in cellular dimensions and genome size. They are most closely related to the gram-positive bacterial group that includes streptococci, bacilli, and lactobacilli. The small genome typical of mycoplasmal species is believed to be the result of a gradual reduction in genome size from a common gram-positive ancestor by the process of degenerative evolution (Maniloff, 1992). More than 200 Mycoplasma species have been identified in humans, animals, plants, and arthropods, but only a few have been proven to produce human disease, the best known and most intensely studied being M. pneumoniae.
The organism eventually known as M. pneumoniae was first isolated from sputum in tissue culture from a patient with primary atypical pneumonia in 1944 (Eaton et al., 1944). It was soon apparent that this organism, initially called the Eaton Agent, caused respiratory infections in humans, but it was not until the 1960's that it was proven to be a Mycoplasma and not a virus (Waites & Talkington, 2004).
Individual spindle-shaped cells of M. pneumoniae are 1–2 μm long and 0.1–0.2 μm wide. Accordingly, the M. pneumoniae cell volume is <5% of that of a typical bacillus. Typical colonies of M. pneumoniae rarely exceed 100 μm in diameter and require examination under a stereomicroscope to visualize their morphologic features. The circular M. pneumoniae genome consists of 816 394 basepairs (bp) with 687 protein-coding genes (Himmelreich et al., 1996), about one sixth the size of Escherichia coli.
The small genome of M. pneumoniae and its limited biosynthetic capabilities are responsible for many of the biological characteristics and requirements for complex medium supplementation in order for the organism to be cultivated in vitro. Lack of a rigid cell wall makes these organisms pleomorphic and unable to be classified in the manner of conventional eubacteria. Mycoplasmas are not found freely living in nature because they depend on a host cell to supply the necessary nutrients. Another characteristic of the genus Mycoplasma is the requirement for sterols in artificial growth media, supplied by serum. Sterols are necessary components of the triple-layered mycoplasmal cell membrane providing structural support to the osmotically fragile organisms.
Although mycoplasmas can flourish within an osmotically stable environment in their eukaryotic host, their susceptibility to desiccation explains the need for close contact for transmission of infection from person to person by airborne droplets. Another structural component that is important for extracellular survival is a protein network that provides a cytoskeleton to support the cell membrane. Mycoplasma pneumoniae also produces capsular material that may have a role in cytadherence.
Mycoplasma pneumoniae possesses very limited metabolic and biosynthetic activities for proteins, carbohydrates, and lipids. It scavenges for nucleic acid precursors and apparently does not synthesize purines or pyrimidines de novo. Fermentation of glucose to lactic acid and acetic acid by means of substrate-level phosphorylation mediated by phosphoglyceric acid kinase, pyruvate kinase, and acetate kinase activities is a means of ATP generation; glycerol and some other small carbohydrates may also be metabolized to generate ATP. Mycoplasma pneumoniae carries all reactions of glycolysis, but the tricarboxylic acid cycle and a complete electron transport chain containing cytochromes are absent. Mycoplasma pneumoniae reduces tetrazolium and this property has been used historically to distinguish it from commensal oropharyngeal mycoplasmas. Mycoplasma pneumoniae cells exhibit a distinct, elongated shape with a tip structure, the attachment organelle, which is directly involved in several aspects of both the pathogenicity and life cycle of the organism (Fig. 1). Reproduction occurs by binary fission, during which the attachment organelle migrates to the opposite pole of the cell during replication and before nucleoid separation. Electron microscopy and genome analysis confirm the absence of flagella and pili.
The reductive evolutionary process that led to the minimal genome of M. pneumoniae suggests that it may exist as a facultative intracellular pathogen (Wernegreen, 2005). This possibility is supported by recent in vitro studies suggesting that the organism can take up residence and even replicate within cultured cells for prolonged periods, but this has not been proven to occur during natural infections (Dallo & Baseman, 2000; Meseguer et al., 2003; Yavlovich et al., 2004). Intracellular invasion was not observed after infection of a differentiated normal human bronchial epithelial cell culture (Krunkosky et al., 2007).
Molecular basis for cytadherence and production of disease
Like many other mycoplasmoses of humans and animals, the spectrum of respiratory conditions caused by M. pneumoniae stems from close association between the organism and the epithelial tissue of the host. This association is a prerequisite for the damage that results in disease. Factors that contribute to the disease state include ciliostasis and apoptosis, resulting in localized damage to the respiratory epithelium, and an immune response which, although often robust, is poorly efficacious in terms of clearing the organism or preventing subsequent reinfections.
The crucial association between M. pneumoniae cells and host cells is mediated through a polarized attachment organelle (Balish, 2006; Balish & Krause, 2006). Adhesin molecules, including P30 and P1, which are probably complexed with proteins B and C (Waldo & Krause, 2006), are localized to the attachment organelle through a complex series of events related to assembly of the organelle (Krause & Balish, 2004). Although direct interactions between these molecules and their putative targets on host cells have not been reported, M. pneumoniae cells interact specifically with nonproteinaceous molecules that are widely distributed on animal cells in general. The attachment organelle of M. pneumoniae itself is a cellular protrusion containing cytoplasm, in which a complex, detergent-insoluble, proteinaceous, electron-dense core is present (Krause & Balish, 2004; Balish, 2006; Balish & Krause, 2006). Recent studies (Henderson & Jensen, 2006; Seybert et al., 2006) reveal this core to consist of a pair of nonequivalent parallel rods with numerous cross-striations, connected to the tip of the attachment organelle through a terminal button. A complex at the base of the structure is also present, which might also be in contact with the cell membrane. Proteins P41 and P24, both of which are located in this region (Kenri et al., 2004), have recently been found to be required for normal relations between the attachment organelle and the cell body. In the absence of P41, the attachment organelle becomes separated from the cell body during gliding (Hasselbring & Krause, 2007a, b), while P41 and P24 are both crucial for proper timing and location of attachment organelle assembly during cell growth and development (Hasselbring & Krause, 2007a, b).
Mycoplasma pneumoniae exhibits gliding motility (Balish & Krause, 2006), a role for which has been suggested by data concerning a mutant strain lacking the attachment organelle protein P200 (Jordan et al., 2007). Although this strain has normal attachment organelle morphology and normal adherence characteristics with respect to red blood cells and cultured alveolar epithelial cells, its ability to interact productively with human bronchial epithelial cells and colonize them is profoundly reduced. As the only observable loss of function in these cells is an approximately threefold reduction in gliding speed, it appears that optimal gliding is necessary for colonization of fully differentiated, mucus-producing tissue.
The molecular mechanism underlying gliding motility in M. pneumoniae is unclear. Models have been proposed in which the electron-dense core is the principal source of motor activity, either through rotation (Hegermann et al., 2002) or vibration (Henderson & Jensen, 2006). Although no correlation between the morphology of the electron-dense core and motility properties was observed among M. pneumoniae relatives (Hatchel & Balish, 2008), it is possible that finer features of the core are involved in the gliding process. Additionally, direct roles in motility have been proposed for the major adhesin molecules, P1 (Seto et al., 2005) and P30 (Hasselbring et al., 2005). Finally, a large number of additional genes whose loss affect M. pneumoniae colony morphology and gliding motility have been identified (Himmelreich et al., 1996), providing an ample basis for future efforts in this area.
Variation in the primary structure of the P1 adhesin is believed to play a role in the epidemiology of M. pneumoniae disease. The two subtypes of M. pneumoniae most frequently isolated from clinical specimens differ to some extent in the amino acid and nucleotide sequences of P1 and its coding gene (Su et al., 1990), though numerous further variants have been identified in recent years (Kenri et al., 1999; Dorigo-Zetsma et al., 2000, 2001; Dumke et al., 2004; Pereyre et al., 2007). Although it has not been demonstrated experimentally, the source of these variant sequences is likely to be degenerate repeats of regions of the P1-coding gene present throughout the M. pneumoniae genome (Kenri et al., 1999); recent evidence in the related organism Mycoplasma genitalium supports a model in which these regions interchange with complementary regions, generating diversity in the P1 sequence (Iverson-Cabral et al., 2007; Ma et al., 2007). A recent report from Japan (Kenri et al., 2008) analyzed clinical isolates of M. pneumoniae collected over a several year period to characterize the P1 gene DNA and determined that the prevalent subtype of M. pneumoniae between 1995 and 2001 was group II but that group I strains became more prevalent between 2002 and 2005. Since the two main subtypes of M. pneumoniae appear to alternate over time, the variation in P1 might be associated with short-term immunity to individual subtypes, development of subtype-specific antibodies, and frequent reinfections with another subtype. Although recombination-mediated variation of P1 has not been demonstrated, evidence supports the occurrence of this process for a set of genes of unknown function (Dumke et al., 2004; Musatovova et al., 2008). A method for molecular subtyping of M. pneumoniae isolates directly in clinical samples has been described that is based on the amplification and sequencing of a repetitive region of the P1 gene (Dumke et al., 2006).
Subsequent to cytadherence, M. pneumoniae is believed to cause disease in part through generation of peroxide. Studies in the distantly related organism Mycoplasma mycoides ssp. mycoides SC have elucidated a correlation between pathogenicity and peroxide formation as a by-product of glycerol metabolism through l-α-glycerophosphate oxidase (Vilei & Frey, 2001; Pilo et al., 2005). As the genome of M. pneumoniae encodes the same enzyme, and metabolism of glycerol by M. pneumoniae is known to result in peroxide production (Low et al., 1971), it is reasonable to propose that the same metabolic pathway contributes to M. pneumoniae disease. HPr kinase (HprK) is central in the sensing of carbon metabolites and the regulation of carbon metabolism (Galinier et al., 1998). In M. pneumoniae, glycerol is required for activation of this enzyme (Halbedel et al., 2004). Its activity at low ATP concentrations (Merzbacher et al., 2004), unlike its counterpart in Bacillus subtilis, suggests strongly that M. pneumoniae, and probably other mycoplasmas, use HprK in a manner distinct from that of other organisms. However, the ultimate targets of this pathway, which in other organisms include a transcription factor absent in M. pneumoniae, remain unclear. Nonetheless, it appears that metabolism of glycerol, a molecule that is presumably readily available to M. pneumoniae as a component of the phospholipids of the membrane of the host cell to which they are so closely apposed, is linked tightly to peroxide production and virulence.
Superoxide anion produced by M. pneumoniae acts to inhibit catalase in host cells, thereby reducing the enzymatic breakdown of peroxides produced endogenously and by the Mycoplasma rendering the host cell more susceptible to oxidative damage (Almagor et al., 1984). Mycoplasma pneumoniae hemadsorption and lysis of guinea pig erythrocytes that are low in endogenous catalase, are also mediated by peroxide (Tryon & Baseman, 1992). This property was adapted for use as a diagnostic test to presumptively distinguish M. pneumoniae from other commensal mycoplasmas that are commonly found in the human respiratory tract that do not hemadsorb in this manner. Host cell lactoferrin acquisition by M. pneumoniae is yet another possible means by which local injury may occur through generation of highly reactive hydroxyl radicals resulting from the introduction of iron complexes in a microenvironment rendered locally acidic by cellular metabolism that also includes hydrogen peroxide (H2O2) and superoxide anion (Tryon & Baseman, 1987). In addition to direct oxidative damage, H2O2 elaborated by the organism may also activate tyrosine kinase-dependent-signaling pathways resulting in aberrant activation of the immune system as discussed below (Rhee et al., 2005).
Although toxins had not been thought to be part of the M. pneumoniae repertoire in previous years, recent evidence suggests otherwise. A protein with significant sequence homology to the S1 subunit of pertussis toxin, which carries out the ADP-ribosylating activity crucial to the toxin's function (Locht & Antoine, 1995), was identified as specifically binding to surfactant protein A (Kannan et al., 2005). Subsequently, a recombinantly expressed form of this protein was found to function as a protein: ADP-ribosyltransferase; it further caused vacuolation and ciliostasis in cultured host cells provided its ADP-ribosyltransferase activity was intact, suggesting its potential to act as an exotoxin during M. pneumoniae infection (Kannan & Baseman, 2006). Thus, the protein was dubbed community-acquired respiratory distress syndrome toxin (CARDS TX). As M. pneumoniae lacks homologs of proteins associated with the pertussis toxin S1 subunit that confer its ability to be translocated from the pathogen to the host cell cytoplasm, it is unclear how this protein reaches the host cell. However, the C-terminal moiety of CARDS TX is novel and, the protein itself is immunodominant (Kannan & Baseman, 2006), suggesting that novel features of this protein confer the ability to be translocated. Its targets in the host cytoplasm remain unidentified. Additionally, it is unclear how a toxin function might relate to its specific binding to surfactant protein A, although it has become clear that numerous proteins of M. pneumoniae and its relatives double as adhesins to host extracellular proteins like fibronectin (Dallo et al., 2002; May et al., 2006) and mucin (Alvarez et al., 2003).
Human lung alveolar type II pneumocytes infected with M. pneumoniae show an increase in IL-8, tumor necrosis factor-α (TNF-α), and IL-1-β mRNA production (Yang et al., 2002), supporting the idea that the adherence to human airway epithelial cells leads to production of cytokines and recruitment of lymphocytes and other inflammatory cells, and that these cytokines subsequently modulate the activity of the inflammatory infiltrates. Mammalian cells parasitized by M. pneumoniae can exhibit a number of cytopathic effects as a result of the local damage. Host cells may lose their cilia, appear vacuolated, and reduce their oxygen consumption, glucose utilization, amino acid uptake, and macromolecular synthesis, ultimately resulting in exfoliation. These subcellular events result in some of the clinical manifestations of respiratory tract infection such as the persistent, hacking cough (Waites et al., 2007).
Interaction with host immune cells and autoimmunity
The observation that atypical pneumonias were often associated with cold agglutinins was made as early as 1918 (Clough & Richter, 1918), and these autoantibodies were later characterized as recognizing the I antigen of human red cells, a carbohydrate antigen of surface glycolipids and proteins (Feizi & Taylor-Robinson, 1967; Yu et al., 2001). When considering the potential pathophysiology of M. pneumoniae-induced human diseases, a confounding factor lies in the possibility that either or both active infection and infection-induced autoimmune mechanisms may play a role in the ensuing inflammatory process. Clearly, autoimmune phenomena, in some cases due to antigenic mimicry, can occur, but identification and characterization of mycoplasmal antigens involved in such processes are difficult.
Perhaps the most commonly associated significant autoimmune phenomena associated with M. pneumoniae infections involve the peripheral nervous system. Strong clinical data now exist linking peripheral neurologic syndromes to pathologic antibodies against carbohydrate moieties expressed on a variety of gangliosides, especially GM1 (Willison & Yuki, 2002). About 5–15% of Guillain–Barré syndrome (GBS) cases have been associated with a preceding M. pneumoniae infection (Goldschmidt et al., 1980; Hughes et al., 1999). Patients with GBS who had evidence of a recent M. pneumoniae infection as assessed by the presence of IgM antibodies were also much more likely than other GBS patients to have antibodies against galactocerebrosides (GalC), and the patients' sera could be depleted of GalC reactivity by preincubation with M. pneumoniae antigens (Ang et al., 2002). Patients with Campylobacter jejuni-associated GBS were uniformly negative for anti-GalC antibodies. Other studies have suggested that anti-GM1 antibodies are also produced during M. pneumoniae infection and that the anti-GalC antibodies are an epiphenomenon (Susuki et al., 2004; Christie et al., 2007a, b).
Antecedent M. pneumoniae infections have also been associated with encephalitis, acute demyelinating encephalomyelitis (ADEM) and optic neuritis (Biberfeld, 1971; Pellegrini et al., 1996; Yamamoto et al., 1996; Milla et al., 1998; Bitnun et al., 2001, 2003; Riedel et al., 2001; Lin et al., 2002; Candler & Dale, 2004; Daxboeck et al., 2004). Evidence linking these cases with autoantibodies is weaker than the association with GBS but some data exist (Nishimura et al., 1996; Komatsu et al., 1998). Anti-Gal-cerebroside antibodies were found in one study to be more common among patients with M. pneumoniae-associated encephalitis compared with M. pneumoniae patients without central nervous system involvement or normal controls (Nishimura et al., 1996). In one study of almost 2000 persons with encephalitis, M. pneumoniae was the most common infectious agent identified (Christie et al., 2007a, b). Approximately 5% of cases showed evidence of acute M. pneumoniae infection, most of whom were children. Unfortunately, as with GBS cases, most cases have been diagnosed only by detection of the presence of antibody to M. pneumoniae with a preceding respiratory illness, so establishment of a causal relationship is difficult. The organism has rarely been cultured from spinal fluid in patients with encephalitis, but detection of M. pneumoniae by PCR brings up the possibility that some cases may be due to active infection, not an autoimmune phenomenon. Clusters of encephalitis may also occasionally occur during epidemics of M. pneumoniae respiratory illness. In late 2006–early 2007 three cases of encephalitis, one of them fatal, occurred in Rhode Island among elementary school children during an ongoing Mycoplasma epidemic. The cases were investigated by epidemiologists from the Centers for Disease Control and Prevention (CDC), and were felt to be Mycoplasma related, though no formal report has been issued. Additional evidence for molecular mimicry by M. pneumoniae comes from a study in which adherence-inhibiting anti-P1 adhesin monoclonal antibodies showed cross-reactions with intracellular antigens of eukaryotic cell lines in immunofluorescence microscopy experiments. The cross-reacting antigens were isolated and characterized as glyceraldehyde-3-phosphate dehydrogenase and 2-phospho-d-glycerate hydrolyase (Jacobs et al., 1995).
In addition to antigenic mimicry as a mechanism for autoimmunity, M. pneumoniae has been shown to directly activate cells of the immune system for cytokine production. Hoek et al. (2002, 2005) demonstrated that the organism induces production of IL-4 by mast cells in coculture experiments. Mast cell cytokine production is dependent upon the presence of sialic acid residues on the target cell membrane and the P1 adhesin. More recent studies have revealed that the activation process requires expression of the heavily sialylated FcɛRI α chain by the mast cell (Luo et al., 2007). A model of cellular activation has been developed in which adherence by the bacterium to surface sialoglycoproteins results in cellular activation through the normal receptor-signaling mechanisms. This process is augmented in mast cells by H2O2 elaborated by the organism (Atkinson et al., 2007) specifically inhibits protein tyrosine phosphatases and the lipid phosphatase PTEN by reversibly oxidizing an essential cysteine in the catalytic site (Rhee et al., 2005).
Mycoplasma pneumoniae also directly activates and induces cytokine production from unsorted peripheral blood leukocytes (Kita et al., 1992), lymphocytes (Arai et al., 1983; Simecka et al., 1993), respiratory epithelial cells (Yang et al., 2002; Dakhama et al., 2003; Kraft et al., 2008), and monocyte/macrophages (Yang et al., 2003; Broaders et al., 2006). Unlike the results obtained with mast cells, activation of phagocytes for TNF-α production does not appear to require cell contact and can be induced with Mycoplasma-culture supernatants (Luo et al., 2008). Activation of this cell type likely proceeds through the activation of toll-like receptors 1 and 2 (TLR1 and TLR2) by Mycoplasma-derived lipopeptides (Shimizu et al., 2007).
The frequent occurrence of outbreaks of mycoplasmal infections and lack of protective immunity against reinfection has stimulated research since the early 1960's into vaccine development using animal models and human volunteers. Thus far, these efforts have been disappointing and not much has been performed in this field in recent years largely due to, first, the HIV epidemic and, more recently, preoccupation of governmental funding agencies with select agents more useful in bioterrorism. The earlier research in mycoplasma vaccines has been summarized previously (Waites & Talkington, 2004).
Epidemiology of mycoplasmal respiratory disease
Mycoplasma pneumoniae infections can involve both the upper and lower respiratory tract and occur both endemically and epidemically worldwide in persons of all ages. Climate, seasonality, and geography are not thought to be of major significance, although most outbreaks in the United States. tend to occur in the late summer and early fall. Mycoplasma pneumoniae was believed to be responsible for 15–20% of all cases of community-acquired pneumonia (CAP) between 1962 and 1975 in Seattle, Washington (Foy et al., 1993). Serological studies performed in Denmark showed a pattern of M. pneumoniae infections over a 50-year period from 1946 through 1995 with endemic disease transmission punctuated with cyclic epidemics every 3–5 years (Lind et al., 1997). The long incubation period, relatively low transmission rate, and persistence of the organisms in the respiratory tract for variable periods following infections may explain in the prolonged duration of epidemics of M. pneumoniae infections. A number of well-described outbreaks of M. pneumoniae respiratory infections in the community and in closed or semi-closed settings such as military bases, hospitals, religious communities, schools, and institutions for the mentally or developmentally disabled have been reported. Even though long-term morbidity is uncommon, the acute illnesses are often disruptive and can consume significant resources (Waites & Talkington, 2004).
A Finnish study (Korppi et al., 2004) reported that M. pneumoniae was detected in 30% of pediatric CAP, and in over 50% among children aged 5 years or older, making it the single most common pathogen detected. A study performed in the United States during the 1990s detected M. pneumoniae in 23% of CAP in children 3–4 years of age (Block et al., 1995). A French study (Layani-Milon et al., 1999) documented its occurrence in children <4 years of age without significant differences in infection rates for other children or adults. These findings may reflect the greater number of young children who attend day care centers on a regular basis than in previous years, and the ease with which young children share respiratory secretions with older household members or contacts. Marston (Marston et al., 1997) reported that M. pneumoniae was definitely responsible for 5.4% and possibly responsible for 32.5% of 2776 cases of CAP in hospitalized adults in Ohio. Extrapolation of these data nationally provides an estimated number of CAP cases due to M. pneumoniae in hospitalized adults that exceeds 100 000 on an annual basis. Because the majority of cases of CAP are treated as outpatients, the total number of pneumonias due to M. pneumoniae is many times greater. An additional striking finding was their observation that the incidence of mycoplasmal pneumonias in hospitalized adults increased with age and it was second only to Streptococcus pneumoniae in elderly persons. Taken together, these investigations prove that M. pneumoniae must be considered as a possible etiologic agent of CAP in persons of all ages who present with variable severity of illness.
Clinical manifestations of mycoplasmal respiratory disease
Like the mucosal pathogen Helicobacter pylori, M. pneumoniae seems to operate primarily as a ‘stealth pathogen’, engaging the innate immune system while subverting or avoiding activation of the adaptive elements (Merrell & Falkow, 2004). The typical respiratory infection caused by M. pneumoniae is a slowly developing syndrome presenting with pharyngitis, sinus congestion, occasionally otitis media, and eventually prolonged lower respiratory involvement up to and including primary atypical pneumonia with fever and bibasilar pulmonary infiltrates. The incubation period may be as long as 3 weeks. Although this picture has been presented as typical, in actuality, family studies have revealed that many individuals never progress to the severe lower respiratory phase of the infection and up to 20% may be asymptomatic (Clyde, 1983). The organism can be detected using PCR in the lungs and respiratory tissues of apparently asymptomatic individuals in relatively high frequency, especially those with underlying asthma (Martin et al., 2001a, b).
The clinical symptoms manifested can be quite diverse with the most severe respiratory effects being CAP and occasionally abscesses (Cherry & Welliver, 1976). During the slowly developing respiratory illness, a dry cough that later becomes mucoid manifests after 3–4 days accompanied by dry rales on auscultation of the chest. The cough represents an evolving tracheobronchitis, the most common form of the infection, and is associated during the early stages with upper respiratory congestion, flu-like symptoms, and pharyngitis. If CAP develops, fever in the 101–103° F range is commonly seen. Respiratory symptoms in the most severe cases can precipitate admission to the hospital with decreased blood oxygen saturations and increased work of breathing. Chest radiographs may reveal bibasilar streaky infiltrates.
A variety of extrapulmonary invasive manifestations of M. pneumoniae infection have been described, and patients who have compromised immunity, including humoral immunodeficiencies are more likely to suffer these complications (Johnston et al., 1983; Roifman et al., 1986; Gelfand et al., 1993; O'Sullivan et al., 2004). It is noteworthy that complement-deficient patients have not been described as having undue susceptibility to M. pneumoniae, suggesting that protective antibody plays a unique role in preventing dissemination of the organism. As discussed above, there is some evidence that invasive central nervous system infection takes place in some cases and is responsible for some cases of encephalitis. Many of these neuroinvasive cases lack the typical pulmonary symptoms and antibody, unlike patients with M. pneumoniae-associated GBS (Christie et al., 2007a, b). Pneumonia due to M. pneumoniae may uncommonly be complicated by empyema (Shuvy et al., 2006). Pericarditis has been frequently associated with M. pneumoniae infection and may be underdiagnosed (Kenney et al., 1993; Szymanski et al., 2002; Levy et al., 2003). Finally, acute arthritis has occasionally be identified as being due to dissemination of the organism in patients with apparently normal immunity (Davis et al., 1988; Dionisio et al., 2001). Other cases of M. pneumoniae infection-associated arthritis may represent a reactive process, possibly due to the development of autoantibodies resulting in synovial inflammation (Lambert, 1968; Cimolai et al., 1989; Poggio et al., 1998; Dionisio et al., 2001; Natarajan et al., 2001; Perez & Artola, 2001; Harjacek et al., 2006). Other nonspecific extrapulmonary manifestations such as hemolytic anemia, renal dysfunction, gastrointestinal complaints and other systemic involvement have been extensively reviewed elsewhere (Waites & Talkington, 2004).
Role of M. pneumoniae in chronic asthma
Infection by M. pneumoniae has been suspected to play a role in some chronic human diseases including adult rheumatoid arthritis (RA) (Ramirez et al., 2005; Johnson et al., 2007), juvenile idiopathic arthritis (JIA) (Oen et al., 1995), and Crohn's disease (Chen et al., 2001; Roediger, 2004). The evidence is particularly strong in the case of asthma, implicating M. pneumoniae both in the pathogenesis as well as in exacerbations of acute attacks (Sutherland et al., 2004). Given the prolonged nature of M. pneumoniae infections,it seems likely that a subset of asthmatics may have a chronic infection that induces a Th2-dominant inflammatory response.
Mycoplasma pneumoniae has long been appreciated as a trigger for acute asthmatic attacks (Seggev et al., 1986; Leibowitz et al., 1988; Lieberman et al., 2003; Biscardi et al., 2004) and some studies have shown that the organism can be isolated in higher prevalence from asthmatics (Esposito et al., 2000; Teig et al., 2005). In addition, some studies and case reports document the initial onset of asthma following M. pneumoniae infection (Petrovsky, 1990; Yano et al., 1994; Wilsher & Kolbe, 1995; Biscardi et al., 2004). Follow-up studies in children have demonstrated prolonged airway dysfunction that can persist for years, perhaps consistent with a persistent infection.
Martin et al. (2001a, b) conducted a study in 55 chronic stable asthmatics and identified M. pneumoniae using PCR in bronchial washings and biopsies in 23 of the group (42%). Only one of 11 normal control subjects was PCR-positive for M. pneumoniae. A further six asthmatic individuals were positive for Chlamydophila pneumoniae and two were positive for other pathogenic Mycoplasma species (M. genitalium and M. fermentans) A 6-week treatment trial of clarithromycin resulted in a statistically and clinically significant improvement in the subset of patients with positive PCR findings but not in the PCR-negative subjects (Kraft et al., 2002). Interestingly, all of the subjects and controls were negative for M. pneumoniae antibody suggesting that, in the setting of chronic airway colonization, an effective immune response is evaded by the organism.
Although M. pneumoniae exclusively parasitizes humans, animal models of chronic respiratory infection in mice have been developed and are providing insight into the role of this organism in asthma pathogenesis (Wubbel et al., 1998). The organisms can be cultured from rodent lung over one year after initial infection permitting long-term studies on effects of infection on lung anatomy and physiology and the examination of the effects of infection in combination with allergic sensitization (Hardy et al., 2002). Two different laboratories have demonstrated airway hyperresponsiveness resembling chronic asthma using this murine model (Martin et al., 2001a, b; Hardy et al., 2002). Interestingly, Th2-dominant airway inflammation seems to potentiate the survival of the organism in lung. Prior allergic sensitization of mice to hen egg ovalbumin is associated with downregulation of TLR2 expression and decreased clearance of M. pneumoniae in mouse lung (Chaplin et al., 2007; Wu et al., 2008). In contrast, a Th17-dominant inflammatory response appears to be important in clearance of the organism. Depletion of pulmonary macrophages (Lai et al., 2008) and neutralization of IL-23-mediated IL-17F production (Wu et al., 2007) both prolong the course of the infection. Thus, human subjects with allergic sensitization of the lung may be at more risk to develop chronic airway colonization by M. pneumoniae with resultant augmentation of airway hyperreactivity.
Pathology and laboratory detection
Descriptive pathological findings and general laboratory features of M. pneumoniae infection have been reviewed elsewhere (Waites & Talkington, 2004) and need not be reiterated in depth here. The main aspects to emphasize are that clinical chemistry and hematological laboratory findings are seldom diagnostic. Specific microbiological, molecular, and serological assays must be used in order to confirm a suspected clinical diagnosis, but each of these approaches has its own inherent limitations.
Histopathological examination of tissues from persons with acute mycoplasmal respiratory infection, animal models, and tracheal organ cultures demonstrate ulceration and destruction of ciliated epithelium of bronchi and bronchioles, edema, bronchiolar and alveolar infiltrates of macrophages, lymphocytes, neutrophils, plasma cells, and fibrin. Type II pneumocyte hyperplasia and diffuse alveolar damage have also been reported. Pleura may contain patches of fibrinous exudates. Pleural effusions and diffuse alveolar damage sometimes occur in association with more severe cases and long-term sequelae such as pleural scarring, bronchiectasis, and pulmonary fibrosis in some cases. Lung abscesses may also occur. Immunosuppressed persons with M. pneumoniae infection may lack pulmonary infiltrates, further attesting to the importance of the host immune response in lesion development (Waites & Talkington, 2004).
Serology has historically been the most common laboratory means for diagnosis of M. pneumoniae infections. A fourfold increase in antibody titer in acute and convalescent sera is still considered the ‘gold standard’ for diagnosis of acute M. pneumoniae respiratory infection (Gavranich & Chang, 2005). Although culture and PCR are also used, persistence of the organism for variable lengths of time following acute infection makes it difficult to assess the significance of a positive culture or PCR assay without additional confirmatory seroconversion (Foy et al., 1993). The length of time necessary for culture ranging from a few days to up to 3 weeks or more makes it impractical for patient management and it is not widely available except in specialized reference laboratories.
Mycoplasma pneumoniae has both lipid and protein antigens which elicit antibody responses that can be detected after about 1 week of illness, peaking at 3–6 weeks, followed by a gradual decline, allowing several different types of serological assays, based on different antigens and technologies. Serology is a useful epidemiologic tool in circumstances where the likelihood of mycoplasmal disease is high, but it is less suited for assessment of individual patients. Its main disadvantage is the need for both acute and convalescent paired sera collected 2–3 weeks apart that are tested simultaneously for IgM and IgG to confirm seroconversion. This is especially important in adults over 40 years of age who may not mount an IgM response, presumably because of reinfection. The percentage of persons with acute infection that demonstrate a positive IgG response in the acute phase was <50% in a recent study (Talkington et al., 2004), perhaps due to the presence of IgG from previous infections However, when convalescent sera were tested, the number of IgG-positive specimens rose to 82%. A single measurement of IgM may detect an acute infection if the test is performed after at least 7 days following onset; but the result may be negative if the test is performed sooner than this. This same study (Talkington et al., 2004) found that only 14 of 27 (52%) acute-phase sera-tested positive by various IgM assays, but this number rose to 39 (88%) when convalescent sera were tested. IgM antibodies can sometimes persist for several weeks to months. One recent study (Ozaki et al., 2007) found that a single assay using the IgM ImmunoCard (Meridian Bioscience) had a sensitivity of only 31.8% for detection of acute M. pneumoniae infection in seropositive children with pneumonia, but this increased to 88.6% when paired sera were analyzed. These findings suggest it is risky to base diagnosis of acute mycoplasmal respiratory infection on a single assay for IgM alone. Antibody production may also be delayed in some infections, or even absent if the patient is immunosuppressed. False-negative tests can also occur if serum is collected after antibiotics are administered. Because M. pneumoniae is a mucosal pathogen, IgA is typically produced early in the course of infection. Measurement of serum IgA alone or in combination with IgM may therefore be an alternative approach for diagnosis of acute infection, but very few assays include reagents for IgA detection and no commercial IgA EIAs are currently available in the United States. Several studies have been performed in Europe using a variety of IgA assay types (Sillis, 1990; Granstrom et al., 1994; Seggev et al., 1996; Watkins-Riedel et al., 2001; Lieberman et al., 2002; Abele-Horn et al., 2006; Daxboeck et al., 2006) that generally support its use to detect acute infection, especially in older persons. In some cases, IgA is the only antibody class that is positive (Lieberman et al., 2002). However, (Csángó et al., 2004) measured IgG, IgM, and IgA antibodies against M. pneumoniae in healthy blood donors and in patients with various infections caused by microorganisms other than M. pneumoniae using various commercial EIA assays. They found that 22.8% blood donors and 53.8% of patients with various non-Mycoplasma infections were positive for IgA, raising doubts about its value to support serodiagnosis of a current M. pneumoniae infection.
The best-commercial serological test for individual patient diagnosis depends on the age of the patient, timing of serum collection, whether paired sera are obtained, availability of appropriate, equipment, and experience of the laboratory personnel.
Complement fixation (CF) was the first method developed for serological testing for M. pneumoniae. CF measures mainly the early IgM response and does not differentiate among antibody classes, which is desirable to differentiate acute from remote infection. CF suffers from low sensitivity and specificity because the glycolipid antigen mixture used may be found in other microorganisms, as well as human tissues, and even plants. Cross-reactions with M. genitalium are well recognized. Owing to these significant limitations most clinical laboratories have replaced CF by alternative techniques with greater sensitivity and specificity, many of which have been developed and sold as commercial kits. However, the consideration of a single 1: 64 CF titer as an indication of recent M. pneumoniae infection is still considered useful by some microbiologists. Immunofluorescent antibody (IFA) assays, direct and indirect hemagglutination using IgM capture, and other particle agglutination antibody assays (PAs) have been developed to detect antibody to M. pneumoniae. Enzyme immunoassays (EIAs) have become the most widely used commercial methods for detection of M. pneumoniae. EIAs are more sensitive for detecting acute infection than culture, and can be comparable in sensitivity to PCR, providing a sufficient time has elapsed since infection for antibody to develop and the patient has a functional immune system. These assays may be qualitative or quantitative, may or may not require specialized equipment for performing the assay and reading the results, and can be performed with very small volumes of serum. The need for acute and convalescent sera has remained the obvious limitation for prompt point-of-care diagnosis. However, qualitative, rapid point-of care serologic assays that detect both IgM and IgG or IgM alone in an easy-to-read format without the need for any instrumentation have been developed (Waites & Talkington, 2004). Talkington et al. (2004) found that single use point-of-care EIAs produced by Remel (Remel, Lenexa KS) measuring IgG and IgM simultaneously and the IgM ImmunoCard (Meridian Biosciences, Cincinnati, OH) were better able to identify seropositive samples than several plate-type EIAs. However, plate-type EIAs may be more efficient and cost effective in laboratories that need to measure larger numbers of specimens at the same time. The variability of results from comparative studies and concerns for basing diagnosis of acute M. pneumoniae infection on a single serum specimen underscores the need for improved sensitivities and specificities among serological reagents used commercially for detecting acute M. pneumoniae infection (Talkington et al., 2004; Beersma et al., 2005).
Owing to the insensitivity and prolonged time needed for detection of M. pneumoniae by culture, the need for acute and convalescent sera collected 2–3 weeks apart for optimum serological diagnosis, and other problems inherent with serological assays as described above, PCR gained considerable interest very soon after the early methodologies were developed in the late 1980s. The first reports of PCR for detection of M. pneumoniae in clinical samples appeared in 1989 (Bernet et al., 1989; Jensen et al., 1989). Since then, there have been more than 200 publications describing the use of PCR for detection of M. pneumoniae in human infections and characterization of its basic biological features. Gene targets used in various types of PCR assays for M. pneumoniae include 16S rRNA gene, P1 adhesin, an ATPase operon gene, the tuf gene, and repetitive element repMp1, among others (Waites & Talkington, 2004). The sensitivity of PCR is very high, theoretically corresponding to a single organism when purified DNA is used. Other advantages are the potential ability to complete the procedure in one day, the possibility of obtaining a positive result sooner after onset of illness than serology, the requirement of only one specimen containing organisms that do not have to be viable, as well as the ability to detect nucleic acid in preserved tissues. Use of PCR may also be valuable in identifying a mycoplasmal etiology in persons with a variety of extrapulmonary syndromes in which an obvious contribution of respiratory infection may not be readily apparent. Detection of the organism using PCR is possible in body fluids such as blood and cerebrospinal fluid. This represents an important improvement in diagnosis because, in our experience, cultures of these sites are very rarely positive.
Real-time PCR assays have also been described (Hardegger et al., 2000; Templeton et al., 2003; Ursi et al., 2003; Waites & Talkington, 2004; Pitcher et al., 2006; Dumke et al., 2007; Gullsby et al., 2008). Advantages of real-time PCR over traditional PCR include more rapid turnaround time and less handling of PCR products using electrophoretic analysis (Saito et al., 2005). The advantage of real-time PCR over traditional PCR in detection of systemic spread of M. pneumoniae was demonstrated in a recent study in which 15 of 29 (52%) of patients with serologically proven M. pneumoniae infection had a positive PCR on their sera, while conventional PCR was uniformly negative (Daxboeck et al., 2005).
Comparison of PCR with culture and/or serology has yielded varied results that are not always in agreement (Loens et al., 2003a, b). As would be expected, molecular-based assays often demonstrate equivalent or superior sensitivity for detection of acute infection over serology as well as culture (Abele-Horn et al., 1998; Templeton et al., 2003), but this is not always the case (Michelow et al., 2004; Pitcher et al., 2006). Positive PCR results in culture-negative persons without evidence of respiratory disease suggests inadequate assay specificity, persistence of the organism after infection, or asymptomatic carriage, perhaps in an intracellular compartment that does not yield culturable organisms. Quantitative studies may be useful in drawing conclusions. Positive PCR results in serologically negative persons may be due to an inadequate immune response, early successful antibiotic treatment, or to the collection of specimens before specific antibody synthesis could occur. Negative PCR results in culture or serologically proven infections increase the possibility of inhibitors or other technical problems with the assay and its gene target. If antibiotics have been administered, PCR results may be negative even though serology is positive.
Combined use of PCR with IgM serology may be a useful approach for diagnosis of M. pneumoniae respiratory infection in children, but potentially less useful in adults who may not mount an IgM response and would add to the expense of laboratory testing. A possible alternative, especially in older adults may be a combination of PCR with IgA serology. Combining serology with PCR may also provide some interpretive guidance in distinguishing colonization from active disease. We suggest that positive PCR assays for M. pneumoniae, especially in specimens from normally sterile sites such as spinal fluid or blood should ideally be confirmed by a second unrelated target gene, but there are no universal recommendations for this practice. Thus so far, there are no commercial PCR kits for detection of M. pneumoniae available for use in the United States, but some companies market various products in several European countries. Reference laboratories may also offer assays they have validated themselves using various gene targets.
Clinical samples suitable for M. pneumoniae PCR include nasopharyngeal and oropharyngeal secretions, sputa, bronchoalveolar lavage, and lung tissue obtained by biopsy. Many persons with mycoplasmal respiratory disease do not produce significant amounts of sputum, especially children. Disease severity may not justify invasive procedures to obtain lower respiratory tract specimens. Michelow et al. (2004) evaluated nasopharyngeal and oropharyngeal samples obtained from children with serologically proven M. pneumoniae pneumonia and reported that either specimen type was equally effective for detection of the organism using PCR, but combining results form both sites provided the greatest diagnostic yield. One group of investigators (Raty et al., 2005) found sputa to be superior to nasopharyngeal aspirates and throat swabs in young adults with serologically proven M. pneumoniae infection. However, others found no difference in detection of M. pneumoniae using PCR in adults (Gnarpe et al., 1997) or children (Reznikov et al., 1995) in these anatomic sites.
Because M. pneumoniae is only one of a variety of fastidious and/or slow-growing pathogenic microorganisms that can cause respiratory tract infections that may produce clinically similar manifestations, there has been considerable interest and effort to develop multiplex PCR assays for their detection. Most commonly these assays have included gene targets for M. pneumoniae, C. pneumoniae, Legionella pneumophila (Maltezou et al., 2004), and occasionally other organisms (Corsaro et al., 1999; Grondahl et al., 1999; Tong et al., 1999; Welti et al., 2003; Ginevra et al., 2005; Khanna et al., 2005; McDonough et al., 2005; Raggam et al., 2005; Stralin et al., 2005, 2006a, b; Morozumi et al., 2006; Geertsen et al., 2007; Wang et al., 2008).
Nucleic acid sequence-based amplification (NASBA), which is an isotheramal RNA amplification technique has been applied for the detection of M. pneumoniae (Loens et al., 2002, 2003a, b, 2007, 2008) Initial studies have shown it can be comparable to PCR assays in terms of sensitivity. Multiplex NASBA assays that detect M. pneumoniae as well as other respiratory pathogens have also been described (Loens et al., 2007, 2008).
Antimicrobial susceptibility testing and treatment
Administration of antimicrobials to patients with M. pneumoniae infections will generally produce satisfactory results with a marked reduction in duration of respiratory symptoms. Management has been guided primarily by well known and consistent susceptibilities to a variety of drugs. Macrolides are the treatments of choice, but tetracyclines and fluoroquinolones are also effective (Waites & Talkington, 2004). Most clinical trials evaluating treatments for CAP identified small numbers of cases proven to be due to M. pneumoniae by serologic diagnosis, though some recent studies incorporated culture and/or PCR. Use of serology alone precludes determination as to whether a treatment regimen actually eradicates the organism, thus very few data are available regarding microbiological efficacy of any regimen. High-dose steroids have been reported to be useful in treatment of encephalitis in children with complicated M. pneumoniae infection. Plasmapheresis and intravenous immunoglobulin therapy might also be considered if steroid therapy is ineffective in these settings.
Antimicrobial management of systemic infections caused by M. pneumoniae that extend beyond the respiratory tract can be difficult, especially in persons who have antibody deficiencies or are otherwise immunocompromised. Despite in vitro susceptibility to logical treatment alternatives, some infections can be expected to fail to respond clinically, even when treatment is administered for prolonged periods, accompanied by failure to eradicate the organisms or their rapid return when antimicrobials are discontinued.
It has been known for many years that persons with mycoplasmal respiratory infections may continue to shed the organisms for following clinical resolution of the illness and antimicrobial therapy (Smith et al., 1967) and that erythromycin-resistant strains can occur and are sometimes isolates from patients who have received prior macrolide therapy (Niitu et al., 1970; Stopler et al., 1980; Stopler & Branski, 1986). Because the organism is very rarely isolated from clinical specimens, and performance of in vitro susceptibility tests is an even less common procedure, whether naturally occurring resistance to antimicrobial agents occurs to any significant extent is virtually unknown in most countries. Naturally occurring resistance to tetracyclines or fluoroquinolones has never been reported in M. pneumoniae, but in vitro selection of mutants resistant to both drug classes through serial subcultures with increasing concentrations has been successful (Gruson et al., 2005; Degrange et al., 2008). Mutations in the quinolone resistance determining regions resulted in minimum inhibitory concentrations (MICs) for ciprofloxacin up to 32 μg mL−1 (Gruson et al., 2005).
A report from Japan published in 2001 described the isolation from children with pneumonia and bronchitis of macrolide-resistant M. pneumoniae possessing a 23S rRNA gene mutation (Okazaki et al., 2001). Investigations of strains isolated in Japan before 1999 did not reveal any resistance (Matsuoka et al., 2004; Suzuki et al., 2006). However, azithromycin use for respiratory infections has increased in Japan over the past several years as it has in many other countries. Among 76 M. pneumoniae strains isolated between 2000 and 2003 in the northern, central, and southern regions of Japan, 13 (17%) were resistant to erythromycin, 12 of which had MICs >256 μg mL−1 (Matsuoka et al., 2004). Nucleotide sequencing of 23S rRNA gene domains II and V and ribosomal proteins L4 and L22 showed that 10 strains had an A-to-G transition at position 2063 (M. pneumoniae numbering equivalent to 2058 in Escherichia coli numbering), one strain had A-to-C transversion at position 2063, 1 strain showed A-to-G transition at position 2064 and one a C-to-G transversion at position 2617 (M. pneumoniae numbering, or 2611 in E. coli numbering). The C2617G substitution was associated with low-level resistance. Domain II and ribosomal proteins L4 and L22 were not involved with this resistance. In addition, using PCR, these investigators detected the A-to-G mutation in 23 of 94 (24%) PCR-positive oral samples taken from children with respiratory infections. In 2005, a second report (Morozumi et al., 2005) identified 12 of 183 (6.6%) M. pneumoniae isolates from Japanese children with respiratory tract infections collected between 2002 and 2004 that were resistant to erythromycin with MICs of 32 to >64 μg mL−1. They found an A-to-G transition at position 2063 in domain V of the 23S rRNA gene in nine strains and an A-to-G transition at position 2064 in two strains. All 12 strains had a C-to-T 785 transition in domain II. One strain did not have a transition in the 23S rRNA gene. The 12 erythromycin-resistant strains were subjected to pulsed field gel electrophoresis which classified seven as group I and five as group IIb as defined previously (Cousin-Allery et al., 2000). Additional reports from Japan (Suzuki et al., 2006; Morozumi et al., 2008) found macrolide resistance in 10–33% of M. pneumoniae strains obtained between 2001 and 2006, all of which had mutations in domain V of 23S rRNA gene. It seems likely that the increase in macrolide resistance since 2000 in Japan is due to antibiotic pressure from increased macrolide use, which has occurred in that country during this time period. Another recent report has confirmed the presence of macrolide-resistant M. pneumoniae in France with the Domain V 23S rRNA gene mutations (Pereyre et al., 2007). However, M. pneumoniae macrolide-resistant mutants selected in vitro harbored mutations in other positions of 23S rRNA gene (A2062, E. coli numbering) or in ribosomal proteins L4 and L22 and could be predictive for mutations that will be eventually be observed in clinical strains (Pereyre et al., 2004).
Perhaps the most important question arising from the emergence of macrolide-resistant M. pneumoniae in Japan is whether there is an associated clinical implication. In one report, patients with macrolide-resistant M. pneumoniae who received macrolide treatment experienced more febrile days than patients with macrolide-susceptible isolates, but there were no apparent treatment failures or serious illnesses reported (Suzuki et al., 2006). Another recent study (Morozumi et al., 2008) reported that children with macrolide-resistant M. pneumoniae frequently had to be changed to treatment with minocycline or levofloxacin because of persistent fever, cough, no resolution or worsening of chest radiographs.
Because macrolides, especially azithromycin, have been widely used in many countries worldwide, it is prudent to perform global surveillance for the occurrence of macrolide-resistant M. pneumoniae to identify whether these trends in macrolide resistance are emerging elsewhere and whether they affect adults. It is also apparent that the impact of macrolide resistance on the outcomes of respiratory infections is not clear because the number of cases described thus far is very small and the data obtained from them are largely descriptive and difficult to quantify.
Evidence of naturally occurring macrolide resistance, the need to be able to quantify antimicrobial susceptibilities for currently available antibiotics in some cases for clinical purposes, and the need to evaluate investigational antibiotics to determine their spectrum of activities means it is desirable to have a standardized method for performance, interpretation, and quality control of in vitro antimicrobial susceptibility tests. A Subcommittee on Mycoplasma Antimicrobial Susceptibility Testing was established in 2002 by the Clinical and Laboratory Standards Institute (CLSI). Though the report from a multicenter study has not yet been finalized and accepted by the CLSI, preliminary results have been communicated that summarize test methodology for agar and broth dilution techniques and establish recommended quality control reference strains for each method (Waites et al., 2006a, b).
Over the past several years, sophisticated molecular-based techniques such as PCR, along with older technology such as serology, and culture, augmented by knowledge obtained from the complete genome sequence, have been applied in epidemiologic investigations, animal models of disease, evaluation of diagnostic tests, and clinical trials of antimicrobial agents. As a result, our understanding of M. pneumoniae's cell biology, mechanisms of cytadherence, disease production, evasion of host defenses, disease transmission, contribution to chronic lung diseases, emergence of antimicrobial resistance, and efficacy of new antimicrobial treatments have improved. Despite these many advances, much is still unknown about this microorganism, which is among the smallest of all bacteria. Most Mycoplasma infections never have a microbiological diagnosis because rapid, sensitive, specific, and reasonably priced methods for its direct detection are not readily available in physician offices or hospital laboratories. A reliable and user-friendly nucleic acid amplification method for Mycoplasma detection in clinical specimens adapted for performance in clinical diagnostic laboratories would be of immense importance both for patient diagnosis and management and for epidemiological research. Development of a safe vaccine that offers protective immunity might also go a long way towards reducing the extent of M. pneumoniae infections, particularly in high-risk populations such as the military, schools, hospitals, and other institutions where large numbers of people dwell in close proximity, but this seems unlikely to be developed in the foreseeable future.
T.P.A. is supported by NHLBI P01 HL073907-04.