A Silent Epidemic: The Prevalence, Incidence and Persistence of Mycoplasma genitalium Among Young, Asymptomatic High-Risk Women in the United States

Abstract

Background

Mycoplasma genitalium can result in pelvic inflammatory disease and adverse pregnancy outcomes. We analyzed data collected from a prospective study of asymptomatic bacterial vaginosis (BV) to determine the natural history of M. genitalium.

Methods

Women aged 15–25 years, with asymptomatic BV and ≥2 risk factors for sexually transmitted infection were recruited from 10 sites throughout the United States. Vaginal swab samples were collected at enrollment and through home-based testing every 2 months over 12 months. M. genitalium nucleic acid amplification testing was performed for M. genitalium using transcription-mediated assays (Hologic). The prevalence, incidence, and persistence of M. genitalium, defined as all positive specimens during follow-up, were estimated with 95% confidence intervals (CIs). Adjusted odds ratios (AOR) were calculated using logistic and Poisson regression to evaluate participant characteristics associated with M. genitalium infection.

Results

Among 1139 women, 233 were M. genitalium positive, for a prevalence of 20.5% (95% CI, 18.2%–22.9%); 42 of 204 had persistent M. genitalium (20.6%). Among 801 M. genitalium–negative women at baseline, the M. genitalium incidence was 36.6 per 100 person-years (95% CI, 32.4–41.3). Black race (AOR, 1.92; 95% CI, 1.09–3.38), age ≤21 years (1.40; 1.03–1.91), and prior pregnancy (1.36; 1.00–1.85) were associated with prevalent M. genitalium; only black race was associated with incident M. genitalium (P = .03).

Conclusions

We identified high rates of prevalent, incident, and persistent M. genitalium infections among young, high-risk women with asymptomatic BV, supporting the need for clinical trials to evaluate the impact of M. genitalium screening on female reproductive health outcomes.

Young, asymptomatic, high-risk women in the United States have high rates of prevalent, incident and persistent Mycoplasma genitalium infections, highlighting the need for clinical trials to assess the impact of M. genitalium screening programs on female reproductive health outcomes.

Decades since its discovery as one of the smallest free-living bacteria affecting humans, Mycoplasma genitalium has earned attention as a sexually transmitted infection (STI) resulting in both male and female genitourinary syndromes. A meta-analysis of studies in women has demonstrated a significant association of M. genitalium with cervicitis, pelvic inflammatory disease, spontaneous abortion, and preterm birth [1]. Furthermore, longitudinal studies have noted the association of M. genitalium with an increased likelihood of human immunodeficiency virus (HIV) acquisition [2, 3]. Treatment guidelines for sexually transmitted diseases (STDs) [4] in the United States reflect the increased importance of M. genitalium, which is important for healthcare providers to recognize as STI care shifts from public STD clinics to primary care [5].

The diagnosis of M. genitalium infections is limited by the absence of Food and Drug Administration approved detection methods; however, research use and validation of M. genitalium nucleic acid amplification tests (NAATs) in some clinical laboratories have improved estimation of its prevalence worldwide. Among low-risk women in the United States and the United Kingdom [6, 7], the population prevalence of M. genitalium infection ranges from 1% to 2% and is highest among women 16–19 years of age. The burden of M. genitalium is considerably higher among female patients in STD clinics, with prevalences ranging from 7.0% to 19.2% [8–10]. Several studies have identified high coinfection rates with Neisseria gonorrhoeae and/or Chlamydia trachomatis among women with M. genitalium infection [8–10].

Presently, there is little information regarding the natural history of M. genitalium. Earlier studies noted its persistence among women in the United Kingdom and female sex workers in Africa [11–13], but the persistence rate among sexually active women in the United States is unknown. The persistence of M. genitalium may be due to lack of detection in asymptomatic persons or antimicrobial resistance to therapy (eg, azithromycin and doxycycline), potentially leading to chronic infections and long-term complications, such as infertility.

Schwebke et al [14] conducted a multicenter clinical trial involving women 15–25 years of age with asymptomatic bacterial vaginosis (BV) followed up over 12 months to determine whether treatment of asymptomatic BV would result in reduced acquisition of gonorrhea and chlamydia. We conducted analyses of the specimens and data collected from this study to determine the prevalence, incidence, natural history of M. genitalium, and factors associated with infection in young high-risk women.

METHODS

The initial prospective, multicenter study was conducted from July 2008 to January 2015. Women were recruited from 10 different clinical sites (ie, STD, family planning, obstetrics-gynecology, and clinical research clinics) located in Alabama, Maryland, North Carolina, Pennsylvania, and California [14]. Young women 15–25 years of age were eligible if they had asymptomatic BV meeting Amsel’s criteria (based on vaginal pH >4.5 and >20% clue cells on wet mount microscopy), reported vaginal intercourse within the past 3 months, and had ≥2 of the following STI risk factors: age ≤20 years, black race, Hispanic ethnicity, douching at least once per month, ≥2 sex partners in the past 12 months, and STI diagnosis in the past year. Exclusion criteria included pregnancy, use of oral antibiotics, and trichomoniasis at baseline. All sites received approval to conduct the study by their local institutional review boards.

STI Testing and Treatment

After giving written consent, participants provided either clinician-collected or self-collected vaginal swab samples for BV and STI testing; they also completed a baseline questionnaire assessing demographic factors, sexual risk behaviors, and sexual history. A home testing kit was then mailed to participants every 2 months over a 12-month period; the kit included 3 vaginal swabs, a microscope slide, and specimen containers.

Women randomized to the treatment arm received oral metronidazole (500 mg twice daily for 7 days) for asymptomatic BV identified at baseline and from vaginal specimens collected during the study period. Women in both groups diagnosed with chlamydial and/or gonococcal infections were referred for standard evaluation and treatment. All participants were encouraged to seek evaluation for genitourinary symptoms and were allowed to receive treatment outside the study protocol. Because M. genitalium testing was not conducted in real time and the NAAT was used for research only, participants with M. genitalium identified did not receive directed therapy.

Laboratory Procedures

Vaginal swab samples were collected every 2 months over 12 months. Women were instructed to roll a vaginal swab over a microscope slide, which was subsequently tested for BV using Nugent criteria [15]. A second vaginal swab sample was tested for N. gonorrhoeae and C. trachomatis, using the manufacturer’s instructions for the BDProbeTec NAAT (Becton Dickinson), at 4, 8, and 12 months after enrollment. The third vaginal swab sample was processed using the APTIMA (Hologic) specimen collection kit and frozen at −70°C for future testing. M. genitalium testing was performed using frozen specimens collected at enrollment and at 2-, 4-, 6-, 8-, 10-, and 12-month follow-up from participants who provided consent for specimen storage and future testing. M. genitalium NAATs were performed according to the manufacturer’s instructions for APTIMA (Hologic), using M. genitalium–specific reagents for target capture, transcription-mediated amplification, hybrid protection assay, and amplicon detection using a manual direct tube sampling system or an automated Tigris instrument. Purified M. genitalium nucleic acid and nuclease-free water were included as positive and negative controls, respectively, in each test run. Specimens were categorized according to their output in relative light units (RLU); outputs of 0–10000 RLU identified specimens as negative for M. genitalium, 10000–30000 RLU as equivocal, and >30000 RLU as positive. Specimens with initial positive or equivocal results were retested, and any specimen with 2 separate results showing an output >10000 RLU was considered positive for M. genitalium.

Definitions

Prevalent M. genitalium infection was defined as a positive NAAT result from a vaginal specimen collected at enrollment. Incident M. genitalium infection was defined as a prior negative result followed by a positive result from any specimen collected 2–12 months after enrollment, and participants could have multiple incident infections. Clearance of infection was defined as negative results on all subsequent specimens after a positive NAAT result. Time of clearance was estimated as the midpoint between the last positive and the first negative visit. Persistent infection was defined as repeated positive M. genitalium results from all available specimens for each participant during follow-up.

Data Analyses

Clinical and laboratory data collected from baseline and follow-up were analyzed using SAS software (SAS). The proportions of unique participants with prevalent, incident and persistent M. genitalium infections were estimated with 95% confidence intervals (CI), and incidence rates were calculated in person-years. Associations between demographic (ie, age, race/ethnicity, education level), behavioral (ie, number of sexual partners in the past year; new sexual partner in the past year; history of pregnancy) and clinical characteristics (ie, coinfection with C. trachomatis or N. gonorrhoeae) at baseline were initially investigated with χ² tests. No methods were used to impute missing data. Factors significantly associated with prevalent M. genitalium infections at the .10 level from the univariate analysis were analyzed using logistic regression, and adjusted odds ratios (AORs) were estimated with 95% CIs to determine independent associations with prevalent M. genitalium. Factors associated with incident M. genitalium were analyzed using Poisson regression, and statistical significance was assessed at the .05 level.

RESULTS

From the primary study, there were 1365 asymptomatic women included in the intent-to-treat population, of whom 1139 (83.4%) had specimens available for M. genitalium testing; 226 participants (16.6%) did not consent to additional specimen testing and were excluded from further analyses (Figure 1). Among the 1139 women, the median age was 21 years (range, 15–25 years), and 84.1% were black in race/ethnicity (Table 1). Most (47.9%) reported having vaginal sex with only 1 sexual partner in the past year, but more than half of participants reported having a new sexual partner and having receptive oral sex with >2 partners in the past year (50.6% and 52.5% respectively). The majority (54.8%) reported a history of pregnancy.

Figure 1.

Flow diagram of study. The initial study [14] enrolled 1365 asymptomatic high-risk women 15–25 years of age, of whom 1139 were screened for Mycoplasma genitalium at baseline. M. genitalium testing was conducted at 2-, 4-, 6-, 8-, 10-, and 12-month follow-up, using self-collected vaginal swab samples.

Table 1.

Baseline Characteristics of Women Enrolled in the Study and Associations With Prevalent Mycoplasma genitalium Infection

Characteristic	Women Enrolled, No. (N = 1139)	Women With Prevalent M. genitalium Infection, No. (%) (n = 233)	Odds Ratio (95% CI)
Age
15–21 y	642	145 (22.6)	Reference
22–25 y	497	88 (17.7)	0.74 (.55–.99)
Race/ethnicitya
White	141	23 (16.3)	Reference
Black	958	214 (22.3)	2.45 (1.49–4.04)
Other	135	17 (12.6)	0.53 (.31–.89)
Educational level
High school or trade school	585	138 (23.6)	Reference
Some college education	554	95 (17.1)	0.67 (.50–.90)
Vaginal sex: No. of partners in past year
0	70	16 (22.9)	Reference
1	546	111 (20.3)	0.86 (.48–1.56)
≥2	523	106 (20.3)	0.86 (.47–1.56)
Receptive oral sex: No. of partners in past year
0	138	33 (23.9)	Reference
1	403	91 (22.6)	0.93 (.59–1.46)
≥2	598	109 (18.2)	0.71 (.46–1.10)
Receptive anal sex: No. of partners in past year
0	917	187 (20.4)	Reference
1	192	36 (18.8)	0.90 (.61–1.34)
≥2	30	10 (33.3)	1.95 (.90–4.24)
New sexual partners in past year
No	563	131 (23.3)	Reference
Yes	576	102 (17.7)	0.71 (.53–.95)
Sexual activity with other women in past year
No	983	203 (20.6)	REF
Yes	156	30 (19.2)	0.92 (.60–1.40)
History of pregnancy
No	514	89 (17.3)	Reference
Yes	625	144 (23.0)	1.43 (1.06–1.92)
Coinfectionb
Chlamydia trachomatis	167	50 (29.9)	1.84 (1.28–2.66)
Neisseria gonorrhoeae	55	13 (23.6)	1.22 (.64–2.30)
Treatment for bacterial vaginosisc
No	627	132 (21.0)	Reference
Yes	495	100 (20.2)	0.95 (.71–1.27)

Abbreviation: CI confidence interval.

^aPatients can be in multiple race categories.

^bPatients can have more >1 coinfection.

^cAssignment to treatment for bacterial vaginosis, which is not expected to have an effect on the likelihood of prevalent M. genitalium infection at baseline.

Prevalent M. genitalium Infection

Of the 1139 participants with specimens for analyses, 233 were M. genitalium positive at baseline, for a prevalence of 20.5% (95% CI, 18.2%–22.9%). In comparison, 167 women (14.7%) had C. trachomatis and 46 (4.8%) had N. gonorrhoeae infection at baseline. The prevalence of M. genitalium was higher among women aged 15–21 than among those aged 22–25 years (22.6% vs 17.7%, respectively) (Table 1). The prevalence of M. genitalium coinfection was 29.9% among women with chlamydial infection and 23.6% among those with gonococcal infection.

Univariate analyses demonstrated that women who reported black race/ethnicity had increased odds of prevalent M. genitalium infection compared with nonblack women (Table 1). Prevalent infection was less likely among women who reported race/ethnicity other than black or white, and among college-educated women compared with those with a high school or trade school education. Women with a history of pregnancy and those with C. trachomatis infection also had an increased likelihood of prevalent M. genitalium, compared with those without prior history or infection (Table 1).

In multivariate analysis, black race remained significantly associated with prevalent M. genitalium infection (AOR, 1.92; 95% CI, 1.09–3.38), compared with white race (Table 2). Age ≤21 years was also associated with increased odds of prevalent infection compared with age >21 years (AOR, 1.40; 95% CI, 1.03–1.91). Women who reported a history of pregnancy also had a higher likelihood of prevalent M. genitalium (AOR, 1.36; 95% CI, 1.00–1.85).

Table 2.

Results of Logistic Regression Modeling, Including Characteristics Associated With Prevalent Mycoplasma genitalium Infection in Univariate Analyses

Characteristic	AOR (95% CI)	P Value
Black race	1.92 (1.09–3.38)	.03
Other race	0.74 (.41–1.36)	.33
Age (≤21 vs >21 y)	1.40 (1.03–1.91)	.03
New sexual partner(s) in past year	0.79 (.58–1.06)	.12
History of pregnancy	1.36 (1.00–1.85)	.05
Educational level (high school, trade school, or less vs some college, college degree, or postgraduate study)	0.80 (.59–1.09)	.15

Abbreviations: AOR, adjusted odds ratio; CI, confidence interval.

Incident M. genitalium Infection

Among 801 M. genitalium–negative participants at baseline who had ≥1 follow-up specimen over 12 months, there were 260 M. genitalium–positive results during 711.2 person-years of follow-up, for an incidence rate of 36.6 per 100 person-years (95% CI, 32.4–41.3). The M. genitalium incidence was similar among age groups; however, black women had a higher incidence than nonblack women (39.6 [95% CI, 34.7–45.1] vs 24.0 [17.1–33.8] per 100 person-years, respectively). Women who reported a history of prior pregnancy had a higher M. genitalium incidence than those without (40.8 vs 32.0 per 100 person-years). Among participants identified with chlamydial or gonococcal coinfections at baseline, M. genitalium incidence rates were 47.6 and 38.1 per 100 person-years, respectively.

In univariate analysis involving the same demographic, behavioral and clinical factors listed in Table 1, incident M. genitalium infection was associated with black race (P = .01), educational level (P = .02), and asymptomatic BV at baseline (P = .03). Incident M. genitalium infection, however, was not associated with the number of new sexual partners in the past year (P = .11), history of prior pregnancy (P = .06), or chlamydial (P = .07) or gonococcal (P = .89) coinfections at baseline. Treatment with metronidazole for asymptomatic BV identified throughout the study period did not affect the likelihood of incident M. genitalium infection (P = .73). When race, educational level and asymptomatic BV were included in multivariate analysis, only black race remained significantly associated with incident M. genitalium (P = .03).

Clearance and Persistence of M. genitalium Infection

Of the 233 women identified with prevalent M. genitalium at enrollment, 204 (87.5%) had follow-up specimens for analysis of clearance or persistence (Table 3). Of the 204 women, 50/ (24.5%) had complete clearance of their M. genitalium infection after baseline; however, 13 of the 50 (26.0%) also had chlamydia and/or gonorrhea coinfection at enrollment and probably received therapy with some activity against M. genitalium. When participants coinfected at baseline were excluded from analysis, 61 (41%) of 147 M. genitalium–infected women who provided specimens at 2-month follow-up had evidence of clearance. Of 77 women with mixed infections (both positive and negative results), 71% had M. genitalium detected from ≥4 specimens over 12 months. Overall, 42 women (20.6%) had persistent infection with M. genitalium–positive NAAT results throughout their entire follow-up period.

Table 3.

Clearance and Persistence of Mycoplasma genitalium Among 204 Women With Prevalent Infection and Follow-up Testing Over 12 Months

Follow-up Status	Women, No. (%)
Clearance after baselinea	50 (24.5)
Clearance after 2 mo	13 (6.4)
Clearance after 4 mo	11 (5.4)
Clearance after 6 mo	11 (5.4)
Mixed (both positive and negative results throughout follow-up)	77 (37.8)
Persistent M. genitalium (all follow-up results positive)	42 (20.6)

^aClearance is defined as no positive tests after a given point.

DISCUSSION

Adolescents and young adults bear the largest burden of STIs in the United States and account for half of the roughly 20 million new STIs that occur annually in this country [16]. We identified an M. genitalium prevalence of 20.5% and an incidence rate of 36.6 per 100 person-years in a multicenter, prospective study of young, high-risk women with asymptomatic BV. To our knowledge, these represent some of the highest prevalence and incidence rates reported for any STI in the United States. The prevalence of M. genitalium was almost twice that of C. trachomatis and 5 times that of N. gonorrhoeae in our population, and M. genitalium coinfected 24%–30% of women who had either chlamydial or gonococcal infections. The striking incidence rate of M. genitalium among women 15–24 years of age with multiple STI risk factors indicates its widespread transmission within their sexual networks. Given the association of M. genitalium with reproductive tract complications [1, 13, 17–21], these data support the consideration of targeted screening programs for the detection of this organism in similar populations of women and their sexual partners.

A history of pregnancy was significantly associated with prevalent M. genitalium infection, and women with prior pregnancy also had a higher M. genitalium incidence rate than nulliparous women. These findings may be due, in part, to the influence of hormonal changes during menstruation and pregnancy on the colonization of the human female genital tract with M. genitalium [22]. Animal studies have demonstrated that progesterone affects genital tract colonization with different mycoplasmas, including M. genitalium, Mycoplasma hominis, and Ureaplasma sp. [23]. Another possibility is that prior pregnancy may serve as a marker for increased sexual risk, considering that more than half of our participants reported this history. However, other characteristics of high-risk activity, such as number of sexual partners and new partners in the past year, were not associated with M. genitalium infection in our study, suggesting that this explanation is less likely.

Other variables, such as age <21 years and black race, were significant predictors of prevalent and incident M. genitalium infections. In an earlier study involving M. genitalium and cervicitis, Manhart et al [24] reported that the average age of infected women was 20 years, compared with 23 years for uninfected women, indicating that adolescents have increased risk for M. genitalium infection. Other investigators have identified associations of age <20 years and nonwhite race with an increased risk for M. genitalium, C. trachomatis, and N. gonorrhoeae [25]. Nationally, young minority women have higher STI rates than their nonminority counterparts [16], which may be attributed to multiple factors, including behaviors, access to healthcare, and sexual networks [26, 27].

Investigators have suggested that BV and alterations in the normal vaginal flora increase the risk for STIs [28, 29]. Lokken et al [30] reported in 2017 that BV among Kenyan female sex workers was associated with a 3.5-fold increased odds of incident M. genitalium. Our primary study was conducted with the hypothesis that screening and treatment of asymptomatic BV would restore optimal vaginal flora, thus reducing susceptibility to STIs [14]. However, no statistically significantly associations remained in our population between incident M. genitalium and asymptomatic BV or its treatment. This is consistent with the findings reported by Schwebke et al [14] for gonorrhea and chlamydia and suggests that either asymptomatic BV is not causally linked to a woman’s risk for M. genitalium or that the rate or the duration of BV treatment did not shift the vaginal microbiota of these women in a meaningful way.

After excluding women who probably received treatment for gonorrhea or chlamydia that had some activity against M. genitalium, we found a clearance rate of 41% at 2 months. Vandepitte et al [13] reported that 93% of their sex workers had clearance of their M. genitalium infection within 12 months of follow-up, with a median time to clearance of 2.1 months (interquartile range, 1.5–4.8 months). In that study, the overall clearance rate was 25. 7 per 100 person-years, but the authors’ analysis did not account for potential M. genitalium treatment among participants who received antibiotics for other STIs [13, 31].

In contrast, M. genitalium persisted in up to a fifth of our participants with asymptomatic infections at baseline. Vandepitte et al [13] reported a much higher rate of M. genitalium persistence, 45%, among female sex workers from Uganda at 3 months, and other investigations involving sex workers in Kenya [11] found that 17% had persistent M. genitalium at 3 months and 21% at ≥7 months. Oakeshott et al [12] previously identified M. genitalium infections among female university students in the United Kingdom, demonstrating that 26% with M. genitalium infection had persistence after 12–21 months. However, their population had a low M. genitalium prevalence of 3.3%, and a small number of participants with follow-up (n = 27) [12].

We identified an M. genitalium persistence rate of 21% up to 12 months, which is disconcerting given the high prevalence rate in our population and the morbidity associated with M. genitalium in women. This could still be an underestimate, however, because another 38% had mixed infections, and intermittent negative results could have represented fluctuations in bacterial loads below the limit of detection [32]. Although pathogen persistence could indicate long-term colonization of the female genital tract, McGowin et al [33] demonstrated that persistent M. genitalium elicits chronic inflammatory cytokine secretion in endocervical epithelial cells, even when there are low organism burdens associated with asymptomatic infections. M. genitalium has mechanisms to evade the local immune response and persist over time [34]; thus, chronic inflammation in the upper female genital tract can result in pelvic inflammatory disease, endometritis, and tubal-factor infertility.

Our study design was limited by focusing on young, high-risk women with asymptomatic BV, so our findings cannot be extrapolated to other women >25 years of age, with symptoms or lower risk for STIs. We did not collect data regarding the HIV status of study participants; therefore, the natural history of M. genitalium among HIV-infected women remains unclear. We had limited information regarding specific treatment for other STIs but accounted for coinfected participants who presumably received antibiotics with some M. genitalium activity. Finally, we conducted M. genitalium testing from stored vaginal swab samples, and participants did not undergo clinical evaluations after the enrollment visit unless they had new symptoms or were identified as having gonorrhea or chlamydia. Therefore, women with persistent M. genitalium were not evaluated for pelvic inflammatory disease or other related complications.

A recent technical consultation with the National Institutes of Health reported by Martin et al [35] identified knowledge gaps and key research priorities that could advance our understanding of M. genitalium, its effect on reproductive tract tissues, and optimal therapeutic interventions. Clinicians should be aware of the gaps in M. genitalium knowledge in order to advocate for increased availability of diagnostic testing and research into better treatment approaches. In the current study, we addressed one of the research priorities for this organism—the natural history of M. genitalium in asymptomatic high-risk women. Prior investigations had indicated that the low prevalence of M. genitalium among asymptomatic women would result in low-yield screening efforts [7, 36]. This was the premise for the argument against the development of national screening programs for M. genitalium infections. Our findings in the context of the published literature suggest that a targeted M. genitalium screening and treatment program among certain populations (ie, high-risk women aged ≤25 years in the United States with asymptomatic BV) could be a highly effective public health intervention; however, there is an urgent need for research designed to test this hypothesis.

Notes

Acknowledgments. We gratefully acknowledge the support of Ned Hook, University of Alabama at Birmingham; Linda McNeil, FHI 360; David Martin, Louisiana State University; and Jonathan Glock, National Institute of Allergy and Infectious Diseases clinical project manager for all of their support with implementation of the study and editing the manuscript. We thank Dana Lapple at the University of North Carolina, Chapel Hill, and Eric Munson at Wheaton Franciscan Laboratory for M. genitalium testing, and Hologic for providing research-use-only reagents for M. genitalium testing.

Financial support. This work was supported by the Division of Microbiology and Infectious Diseases, National Institute of Allergy and Infectious Diseases (contract HHSN26620040073C to the University of Alabama at Birmingham Sexually Transmitted Infections Clinical Trials Group) and by the Microbiology Core Laboratory of the Southeastern Sexually Transmitted Infections Cooperative Research Center (through grant U19-AI031496 from the National Institutes of Health).

Potential conflicts of interest. A. C. S. and A. M .R. receive royalties from UptoDate. A. C. S receives FOCUS grant funding from Gilead. J. S. and M. M. H. have received consulting fees and research support from Hologic. A. M. R. is on the Biofire advisory board. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.