Mycoplasma genitalium: An Emerging Cause of Sexually Transmitted Disease in Women

Mycoplasma genitalium is an emerging sexually transmitted pathogen implicated in urethritis in men and several inflammatory reproductive tract syndromes in women including cervicitis, pelvic inflammatory disease (PID), and infertility. This comprehensive review critically examines epidemiologic studies of M. genitalium infections in women with the goal of assessing the associations with reproductive tract disease and enhancing awareness of this emerging pathogen. Over 27,000 women from 48 published reports have been screened for M. genitalium urogenital infection in high- or low-risk populations worldwide with an overall prevalence of 7.3% and 2.0%, respectively. M. genitalium was present in the general population at rates between those of Chlamydia trachomatis and Neisseria gonorrhoeae. Considering more than 20 studies of lower tract inflammation, M. genitalium has been positively associated with urethritis, vaginal discharge, and microscopic signs of cervicitis and/or mucopurulent cervical discharge in seven of 14 studies. A consistent case definition of cervicitis is lacking and will be required for comprehensive understanding of these associations. Importantly, evidence for M. genitalium PID and infertility are quite convincing and indicate that a significant proportion of upper tract inflammation may be attributed to this elusive pathogen. Collectively, M. genitalium is highly prevalent in high- and low-risk populations, and should be considered an etiologic agent of select reproductive tract disease syndromes in women.


Introduction
An estimated 340 million new curable cases of sexually transmitted infections (STIs) are acquired annually throughout the world [1], making these infections an important public health and economic concern. Mycoplasma genitalium is an emerging cause of STIs in the United States [2] and has been implicated in urogenital infections of men and women around the world. More than 25 years after its initial isolation from men with non-gonococcal urethritis (NGU; [3]), M. genitalium is now recognized as an independent etiologic agent of acute and persistent male NGU and is responsible for approximately 20%-35% of non-chlamydial NGU cases [4,5]. Implicating this organism in male urogenital disease was a significant advancement in our knowledge of STIs, but it has been less clear whether M. genitalium is also a cause of inflammatory reproductive tract disease in women. This comprehensive literature review (PubMed database; MeSH ''Mycoplasma genitalium'' with no restrictions on publication year) addresses the overall population prevalence and associations of M. genitalium with inflammatory syndromes of the female reproductive tract.

Epidemiology and Prevalence of M. genitalium Infections
After the initial isolation in 1980 [3], few epidemiologic studies of M. genitalium infection were undertaken largely because of difficulties in cultivation of this fastidious organism. Some 10 years later, the polymerase chain reaction (PCR) was first employed for detection of M. genitalium in patient specimens [6,7], thereby facilitating larger investigations of prevalence and associations with urogenital disease. Using such molecular methods, sexual transmission of the organism has been suggested by high concordance rates among sexual partners [8][9][10][11][12][13][14][15] and documented specifically in infected couples with concordant M. genitalium genotypes [11,15]. In addition, sexual transmission of M. genitalium infection can be inferred from increased prevalence values in cohorts reporting sexual intercourse and the association with number of sex partners [2,16]. For the purpose of this review, high-risk individuals were defined as those attending an STI clinic, those enrolled in a study where inclusion criteria included signs of urogenital disease, patients presenting to family planning clinics for termination of pregnancy, or those individuals classified as sex workers. Low-risk enrollees were those not attending an STI clinic, fertility clinic attendees, those chosen randomly from an otherwise healthy population, and all women enrolled in studies of adverse pregnancy outcomes.
Using the above definition of high-risk populations, 18,838 women have been tested for M. genitalium urogenital infection [6][7][8]10,13,with a substantially higher prevalence than low-risk groups (7.3%; Table S1). Among studies of high-risk individuals, the population prevalence values ranged from 0% to 42% and can be explained by several factors, including the clinical setting, specificity of the employed NAAT assay, participant enrollment criteria (e.g., specific symptoms or signs), geographic location of study site, high-risk behavior (e.g., commercial sex workers), coinfection with other STIs, and the reporting of point or cumulative, multi-sampling values. Importantly, we considered only point prevalence values in the overall prevalence calculations because cumulative, multi-sampling values were not directly comparable. In conclusion, considering sexual transmission and the high prevalence worldwide, the public health significance of M. genitalium infections in women is potentially very large.

Vaginal Discharge
In more than 20 independent clinical studies, M. genitalium has been evaluated as a cause of inflammatory lower genital tract syndromes including urethritis, cervicitis, and vaginal discharge. From 2002 to 2010, six studies addressed the relationship between M. genitalium infection and vaginal discharge (n = 3,059; Table 1) with significant associations observed in three [14,43,51]. Vaginal discharge was measured either as a symptom or sign among the cited studies and the criteria varied from any sign of discharge to defined pathologic symptoms such as discharge characterized as heavy, yellow, or green-gray with mucous-like or creamy consistency (see Table 1 for specific diagnostic criteria for each study). No clear trend was evident as to whether signs or symptoms were better predictors of M. genitalium infection, as only two studies used vaginal symptoms in their diagnostic criteria.
Etiologies of vaginal discharge are extremely diverse, can be either microbial or non-microbial, may be normal or abnormal, and can be attributed to inflammation in other parts of the reproductive tract (reviewed in [57]). Only one study adjusted for the presence of bacterial vaginosis (BV; [43]) and, despite a high rate of BV in the cohort, demonstrated significant associations of M. genitalium with vaginal discharge. However, evaluation of a similar population 4 years later did not reproduce the finding [13]. Further, the case definition of pathologic discharge was variable between studies and, most importantly, patient-reported symptoms are a highly subjective measure. Therefore, the disparity among studies is not surprising. Future studies with defined and/or quantitative signs that control for co-infection with other STIs and concurrent inflammatory syndromes (e.g., bacterial vaginosis) will be necessary to determine whether M. genitalium is independently associated with vaginal discharge.

Urethritis
Considering only microscopic signs of urethral inflammation (.4-5 or .10 polymorphonuclear leukocytes per high-powered microscope field [PMNL/hpf]), positive associations with M. genitalium infection were found in three of four studies [8,40,49]. One study, the largest of Scandinavian women ( [49]; n = 7,604; Table 1), found a significant association between M. genitalium and microscopic urethritis. Three other Scandinavian studies of urethritis failed to show a significant association with M. genitalium infection even when patients co-infected with C. trachomatis [10] were removed or after adjusting for concurrent cervicitis [8,9]. However, Anagrius and colleagues showed a significant association with microscopic signs of urethritis and/or cervicitis in Swedish women [8]. This study exemplifies that exclusion of women with concurrent cervicitis or other inflammatory syndromes is important because inflammation from other sites may contaminate the urethra leading to a false diagnosis of urethritis. Two of the studies that failed to show a significant association between M. genitalium and urethritis, but did control for concurrent cervicitis [8,40], showed a strong trend towards association with lower bounds of their respective 95% CIs close to the null. Importantly, M. genitalium is a recognized cause of sexually acquired acute and persistent urethritis [4,5] in men. However, considering the disparate results of the cited studies, we cannot conclusively implicate M. genitalium as a cause of female urethritis. Additional investigations of M. genitalium urethritis are warranted, especially in populations outside of Scandinavia.

Cervicitis
Cervicitis, often termed mucopurulent cervicitis [58], is characterized by the presence of clinical signs such as mucopurulent discharge, friability at the external os (easily induced bleeding), elevated counts of PMNL detected by Gram staining of endocervical swab material, or a combination of these signs [58]. However, there is no generally accepted case definition of cervicitis. Among epidemiologic studies of cervicitis in high-and low-risk populations ( Table 2; n = 13,000 women), M. genitalium has been positively associated with cervical inflammation in all studies where microscopic signs were considered independent of nonmicroscopic signs [8,10,30,40,47,49]. Of these, only two studies showed significant correlations [10,49]. Considering non-microscopic criteria (see study by Pepin et al. [51] for diversity of nonmicroscopic signs), cervical discharge was the most consistently measured among the retained studies. Four of eight studies showed positive associations between cervical discharge and M. genitalium infection, all of which were significant relative to women without this sign (Table 2; [25,36,47,51]).
Two studies have addressed whether microscopic or nonmicroscopic signs are better predictors of M. genitalium cervicitis within the same patient population. Casin et al. found no significant associations using either cervical discharge or microscopic signs (.10 pmn/hpf) [30], but all women in this study had vaginal discharge. However, Manhart and colleagues found no significant association between M. genitalium infection and cervicitis defined by .30 PMNL/hpf, but statistical significance was observed with abnormal cervical discharge ( [47]; Table 2). Further, considering all studies of M. genitalium cervicitis, those where a high threshold of microscopic cervicitis (.20 or .30 PMNL/hpf, or more PMNL than epithelial cells) was employed, only three of seven studies showed a significant correlation between microscopic signs and M. genitalium infection ( [10,46,49]; Table 2). This suggests that a high microscopic threshold of inflammation is not a more specific sign of M. genitalium cervicitis and might also fail to detect less severe inflammation. Collectively, it is clear that discrepancies among these studies can be attributed to the variable case definition of cervicitis and studies with uniform criteria will be required to address which sign(s) best predict M. genitalium cervicitis.
C. trachomatis is a common cause of cervicitis and a potentially confounding variable for implicating M. genitalium as an independent etiologic agent. Where possible, we excluded subjects with C. trachomatis co-infection for all OR calculations (see Table 2). Six of Cases are those individuals with defined signs or symptoms; controls are individuals without signs or symptoms. 2 M. genitalium prevalence and OR calculated at enrollment. 3 Cumulative, multi-sampling prevalence over a 5-y study that included M. genitalium PCR-and culture-positive individuals; only PCR-positive individuals considered for prevalence in cases vs. controls and OR calculations. 4 Excluded women co-infected with NG, CT, and TV. 5 Cumulative, multi-sampling prevalence. 6 Excluded patients co-infected with M. genitalium and CT. 7 OR calculated from symptoms reported at each sampling and excluded patients co-infected with M. genitalium and CT or NG. 8 Excluded patients with concurrent cervicitis. 9 For Cases are those individuals with defined signs or symptoms; controls are individuals without signs or symptoms. 2 Excluded patients with concurrent urethritis and those co-infected with M. genitalium and CT. 3 Population prevalence and prevalence of M. genitalium in cases vs. controls excluded patients with human papilloma virus and CT. 4 Each enrolled patient sampled at 2-3 sites; prevalence calculated from positive result at any site.
nine studies where C. trachomatis co-infection was either excluded or adjusted for in multivariate analyses found significant associations between M. genitalium and cervicitis [10,25,36,46,49,51], whereas only a single study [47] showed significant associations without this adjustment. Studies controlling for N. gonorrhoeae infection are lacking. Despite these differences, several clinical investigations of urogenital disease in women indeed indicate that M. genitalium should be considered an independent risk factor for cervicitis, particularly when urogenital specimens are negative for other known pathogens. Importantly, this magnitude of increased risk is similar to those of other known causes of cervicitis, including C. trachomatis and N. gonorrhoeae [46].

Upper Reproductive Tract Infection by M. genitalium
Pelvic Inflammatory Disease  Table 3). In the first of these, of 58 Kenyan women with histologically confirmed endometritis, M. genitalium was found significantly more often in women with endometritis compared to women without this condition (16% versus 2%; [31]). Similarly, in a sub-study from the US PID Evaluation of Clinical Health cohort, women with M. genitalium were three times more likely to have endometritis at enrollment compared to women without M. genitalium [39].
In two cross-sectional studies where the endometrium was sampled directly to measure the associations of current infection and upper tract disease [31,33,39], M. genitalium was associated significantly with endometritis [31,39]. In contrast, one prospective study of commercial sex workers in Kenya failed to find an association of M. genitalium infection with PID [33] over 36 months. Considering the persistent nature of M. genitalium, as with other STIs, it is possible that the follow-up period and high percentage of loss to follow-up was not adequate to detect incident PID. Importantly, clinical diagnosis of PID includes several variable signs (see Table 3) that often do not correlate with laparoscopic findings [54]; this undoubtedly contributes to variability among PID studies and could impact the associations with M. genitalium infection. Further, no clear trend was observed when comparing studies that removed co-infections with C. trachomatis or controlled for co-infections in multivariate analyses. Overall, M. genitalium has been associated with microscopic endometritis and PID, and confirmatory studies are clearly necessary to establish an independent role and investigate the mechanisms for upper reproductive tract inflammation.

Pregnancy-Related Complications and Infertility
PID can be a pre-cursor to several significant upper tract complications, including ectopic pregnancy, chronic pelvic pain and tubal factor infertility [60]. No association between M. genitalium and ectopic pregnancy was observed in a single study using serological testing for M. genitalium exposure ( [42]; Table 4). Considering other adverse pregnancy outcomes including preterm birth, spontaneous abortion or miscarriage, stillbirth, and small for gestational age, of five independent studies ( [20,21,42,44,53]; Table 4), two studies have indeed shown an independent Table 2. Cont. 5 Exluded women co-infected with M. genitalium and CT. 6 Cumulative, multi-sampling prevalence over a 5-y study that included M. genitalium PCR-and culture-positive individuals; only PCR-positive individuals considered for prevalence in cases vs. controls and OR calculations. 7 Excluded women co-infected with M. genitalium and NG, CT, or TV. 8 Excluded women with CT and women co-infected with M. genitalium and CT. doi:10.1371/journal.ppat.1001324.t002  Cases are those individuals with defined signs or symptoms; controls are individuals without signs or symptoms. 2 M. genitalium prevalence calculated at enrollment; PID HR calculated from incident infections in prospective study. 3 M. genitalium prevalence and association with endometritis calculated from infection status and signs at enrollment. 4 M. genitalium prevalence and association with ectopic pregnancy were evaluated using a serological assay rather than with a NAAT. 5 Excluded patients co-infected with M. genitalium and CT. 6 Excluded women with CT and women co-infected with M. genitalium and CT. association of M. genitalium with preterm birth [18,20], but no other syndromes have been linked to this infection. In contrast, investigations into M. genitalium as a cause of infertility have consistently shown a strong correlation. Two Danish studies have found a significant association between women with M. genitalium-specific serum antibodies and laparoscopically confirmed tubal infertility ( [17,24]; Table 4). Exclusion of women with prior C. trachomatis infection still resulted in significant association between M. genitalium and infertility ( [24]; Table 4). In one study, PCR detection of M. genitalium was attempted from endocervical swabs with no success, suggesting that previous M. genitalium infection might cause permanent damage to the oviduct, or that endocervical swabs are ineffective for detecting upper genital tract infection. However, in a recent study of Polish women by Grzesko and colleagues, M. genitalium was detected by PCR more often in cervical swabs from infertile patients compared to healthy, fertile women [37], suggesting that endocervical swabs can predict upper tract infection. It is important to note that NAAT studies make associations between current infection and infertility, while serological studies determine associations with prior M. genitalium exposure. Because tubal scarring can result in long-term infertility, serological studies probably best address whether M. genitalium is a cause of tubalfactor infertility and can be useful in determining recent or longterm infections (e.g., IgM versus IgG antibodies).

MG,
Experimental animal models have also provided evidence that M. genitalium can colonize upper reproductive tract tissues, leading to salpingitis or endometritis [61][62][63]. Thus, it is evident that M. genitalium could be an independent cause of tubal factor infertility. Importantly, however, the few studies to date have been relatively small in size and longitudinal studies would be of tremendous benefit for understanding this complex condition whereby prior infection can lead to long-term sequelae.

Recommended Treatment of M. genitalium and Important Considerations
Evaluation of M. genitalium treatment efficacy has been a subject of obvious importance but conclusive recommendations are lacking largely because, to date, only a single randomized controlled clinical trial has been reported [64]. In this trial, a single 1-g dose of azithromycin was more effective than 7-day, multi-dose doxycycline for eradication of M. genitalium infection in men. In patients diagnosed with PID, the US Centers for Disease Control and Prevention guidelines recommend therapy with ceftriaxone plus doxycycline or cefoxitin and probenecid plus doxycycline [65]. These treatments are primarily targeted towards C. trachomatis and N. gonorrhoeae, to which less than half of PID cases can be attributed [66]. Importantly, several reports suggest these treatment regimens would be ineffective for eradicating M. genitalium, as male and female genital infections persist in a significant proportion of patients treated with tetracyclines [67][68][69][70][71] or levofloxacin [68,72].
Using azithromycin (1 g; single dose), clinical cure rates are only between 79% and 87% for M. genitalium-positive male and female patients, leaving a significant subset of patients with persistent urogenital tract infections [29,64,67,73]. Extended 5-day regimens of azithromycin therapy increase cure rates to 96% after doxycycline treatment failure [67], and additional randomized trials are now required to determine the optimum dosage and regimen. If patients fail extended azithromycin therapy, moxifloxacin is the only available antibiotic with a successful rate of cure [73] and should be used only with patients failing other therapies. Successful treatment of M. genitalium infection in female patients is of particular importance because prolonged inflammation at upper genital tract sites might lead to significant reproductive tract morbidity and infertility [74]. In women, treatment must be effective for both lower and upper genital tract infection.

Conclusions and Implications for Future Research
Following the firm establishment that M. genitalium causes NGU in men and is a cause of STI, many studies have now found significant associations with lower and upper reproductive tract disease in women. Taken together, M. genitalium should be considered an etiologic agent of cervical inflammation and upper tract disease syndromes, including PID and infertility. Importantly, additional studies with defined diagnostic criteria will be required to fully understand the relationship between M. genitalium and cervicitis. A systematic review and meta-analysis would be of significant benefit for defining the associations of M. genitalium infection with reproductive tract disease in women.
Although not addressed in this review, M. genitalium likely maintains persistent infection through intracellular survival in mucosal epithelial cells [75,76] resulting in inflammation [75,77]. The observed correlations between M. genitalium reproductive tract infection and HIV-1 (reviewed in [78]) may be explained by longterm inflammation elicited by M. genitalium infection; these associations are likely of particular importance considering the enormous public health burden of HIV infections worldwide. Therefore, continued research will be important to understand the dynamics of persistent HIV-1 and M. genitalium co-infections of vaginal and cervical tissues, particularly when dissecting clinical correlates with disease.
We still have much to learn about reproductive tract infections from both a clinical and basic science standpoint. Overall, M. genitalium appears to be a highly prevalent sexually transmitted bacterial pathogen that, if not diagnosed and the patient treated appropriately, can cause persistent urogenital inflammation in men and women and increase the risk of HIV transmission and infection. Continued investigation of M. genitalium's role in sexually transmitted disease will be pivotal for understanding the complex and dynamic inflammatory syndromes of the female reproductive tract.