Mycoplasma genitalium (MG) is a sexually transmitted, fastidious bacterium that colonizes the female genital tract.1–6 There is increasing evidence that MG infection is associated with adverse reproductive health outcomes in women, including cervicitis, urethritis, pelvic inflammatory disease (PID), infertility, ectopic pregnancy, adverse birth outcomes, and HIV-1 acquisition.1,7–9 However, data on the natural history of MG infection, particularly with respect to duration of infection (persistence) and recurrence, are limited. Natural history studies are critical to improving our understanding of the contribution of MG to adverse health outcomes in women and informing guidelines for screening and treatment.
In addition, over the past several years, MG treatment has been complicated by the poor efficacy of doxycycline and increasing resistance to azithromycin in high-income countries.10 Azithromycin resistance in MG is mediated by macrolide resistance–mediating mutations (MRMMs) in the 23S rRNA gene.11 Most studies assessing MG macrolide resistance in women have been conducted in Europe and Australia,12–18 where circulating resistance greater than 50% has been reported in some settings. To date, only one published report has estimated macrolide-resistant MG in the United States. Among women with MG from 7 sites across the United States, MRMM prevalence was 51%.19 In East Africa, the prevalence of macrolide-resistant MG is unknown. Using data and stored specimens from women who previously participated in a randomized controlled trial of monthly periodic presumptive treatment to reduce vaginal infections, we conducted a retrospective cohort study to assess the natural history of MG infection and estimate the frequency of macrolide-resistant MG.
MATERIALS AND METHODS
Study Population and Procedures
The Preventing Vaginal Infections trial was a double-blinded, randomized controlled trial that assessed the effect of monthly metronidazole 750 mg + miconazole 200 mg intravaginal suppositories versus matching placebo to reduce rates of bacterial vaginosis (BV) and vulvovaginal candidiasis (VVC; No. NCT01230814). Detailed methods and results have been previously described.20 Briefly, 234 high-risk women from 3 sites in Kenya and 1 in the United States were enrolled between May 2011 and August 2012. Eligible women were 18 to 45 years of age, HIV-1 uninfected, not pregnant or breastfeeding, and sexually active, and had one or more vaginal infections at screening (BV, VVC, or Trichomonas vaginalis). All participants provided written informed consent, including a separate consent for the storage and future testing of biological specimens. The trial was approved by the human subjects research committees at Kenyatta National Hospital (Nairobi, Kenya), the University of Washington (Seattle, WA), and the University of Alabama at Birmingham (Birmingham, AL).
At enrollment, structured face-to-face interviews were conducted to collect data on demographic, clinical, and behavioral characteristics. At monthly follow-up visits, a urine pregnancy test was performed, and data on sexual behaviors, intravaginal practices, contraceptive use, product use, and genital tract symptoms (abnormal itching or discharge) were collected. Nonpregnant participants received a month's supply of study product and free male condoms. In addition, during follow-up visits at months 2, 4, 6, 8, 10, and 12, participants underwent a physical examination including pelvic speculum examination with collection of genital swabs for diagnosis of genital tract infections. If a participant missed an examination visit, a physical examination was performed at her next follow-up visit.
The Hologic APTIMA Combo 2 system (Hologic Inc, San Diego, CA) was used to obtain clinician-collected swabs containing cervicovaginal fluid at enrollment and months 2, 4, 6, 8, 10, and 12. Specimens collected at enrollment were tested for the presence of Neisseria gonorrhoeae and Chlamydia trachomatis. The remainder of the enrollment specimen and all follow-up specimens were stored at −80°C for future testing. At the completion of the study, stored specimens were tested for MG using a research use–only transcription-mediated amplification assay with reagents provided by Hologic as part of their ongoing research program. Specimens with a value of greater than 40,000 relative light units were considered positive.21 Samples that were MG positive underwent DNA extraction using the DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). Macrolide resistance–mediating mutations in region V of the MG 23S ribosomal RNA gene were detected by PyroMark (Qiagen) DNA sequencing of polymerase chain reaction amplicons.11,22
Vaginal Gram-stained slides were evaluated for BV using the Nugent score.23 Saline and KOH wet preparations were examined for the presence of motile trichomonads, clue cells, and yeast. Endocervical Gram-stained slides were scanned at low power to identify polymorphonuclear leukocytes (PMNs) in 3 nonadjacent oil immersion fields. Cervicitis defined as at least 30 PMNs on Gram stain. Yeast culture was performed on Sabouraud agar with a germ tube test to identify presumptive Candida albicans.
The objectives of this analysis were to (1) estimate the prevalence, infection duration, and correlates of MG identified at enrollment; (2) assess the incidence (time to first incident infection), infection duration, and correlates of MG during follow-up; (3) assess the frequency of MG recurrence; and (4) determine MRMM frequency among women who tested positive for MG at any point during the trial. M. genitalium clearance was defined as testing negative after a previous positive result, and time of clearance was estimated as the midpoint between the last positive and the first negative result. If a participant had an MG+ visit, followed by a missed visit and then an MG− visit, the missed visit was assumed to be MG− and the expected data of the missed visit were used for the clearance estimate. For participants who were MG+ throughout the study, the duration of infection was calculated as the date of the final follow-up visit minus the date of enrollment. Recurrent MG infection required a previous positive result, with one or more interim negative results, followed by a new positive result. Time of recurrence was considered to be the visit at which recurrent MG was first detected. Strain typing was not performed, so it was not possible to determine whether recurrences represented infection with a new strain versus recurrence or reinfection with an identical strain.
The analysis population was restricted to participants who provided consent for future testing of stored specimens. Descriptive statistics were used to summarize participant characteristics, and χ2 and t tests were performed as appropriate. Logistic regression models were used to assess correlates of prevalent MG infection at enrollment. Age and study site were included in multivariable models based on a priori assumptions. In a prior analysis, we observed a trend toward a lower incidence of MG in the intervention arm compared with placebo24; therefore, initial analyses assessing correlates of incident infection were adjusted for study arm and stratified by site. Cox proportional hazards models were used to assess baseline and time-varying correlates of incident MG infection. Factors associated with MG infection with a P value of less than 0.10 in unadjusted analyses were considered for inclusion in multivariable models. Factors that continued to have a P value of less than 0.10 were retained in final multivariable models. Descriptive statistics were used to summarize the proportion of successfully tested specimens with MRMM. All statistical tests were assessed using a 2-sided α of 0.05. Analyses were conducted using Stata version 14.1 (StataCorp, Inc, College Station, TX).
Of 234 women enrolled, 221 (94%) returned for follow-up and provided consent for future testing. Participant characteristics are presented in Table 1. Median age was 29 years (interquartile range [IQR], 24–34 years), 24% of participants were from the United States, and 54% reported ever having sex in exchange for money, goods, or services. Overall, 77 (35%) participants tested positive for MG either at enrollment or during follow-up. A higher proportion of participants from Kenya tested positive for MG infection at any point in the study compared with those from the United States (64/168 [38%] vs. 13/53 [25%], respectively; P = 0.07).
Prevalence, Clearance, and Persistence
MG prevalence at enrollment was 11% and did not differ between Kenyan (11%) and US participants (11%). Of note, the prevalence of MG was higher than other curable sexually transmitted infections (STIs; C. trachomatis, 7%; N. gonorrhoeae, 1%; T. vaginalis, 7%), and no MG coinfections with C. trachomatis or N. gonorrhoeae were observed (Table 1). Associations between participant characteristics and prevalent MG are presented in Table 1. In unadjusted analysis, having a new sexual partner (odds ratio [OR], 2.50; 95% confidence interval [CI], 1.02–6.11) and having VVC (OR, 2.44; 95% CI, 1.02–5.83) were associated with prevalent MG. In multivariable analyses that adjusted for age and study site, the effect estimate was similar for VVC (adjusted OR, 2.45; 95% CI, 1.00–5.99) and attenuated slightly for having a new sexual partner (adjusted OR, 2.38; 95% CI, 0.87–6.49). Condom use and the prevalence of cervicitis did not differ between women with prevalent MG infection and those who were MG negative. Among 25 women with prevalent MG at enrollment, 22 participants cleared their MG infection before the end of the study, whereas 3 participants had MG detected at every study visit. Median time to clearance of prevalent infection was 3.9 months (IQR, 1.0–8.4 months; Fig. 1). Self-reported use of doxycycline was uncommon and did not seem to correlate with MG clearance (Fig. 1). No azithromycin was dispensed to women with MG infection at any point in the trial.
Incidence, Clearance, Persistence, and Recurrence
Among 196 women without MG at enrollment, there were 52 incident MG infections during 155.7 person-years of follow-up (overall MG incidence [excluding recurrences], 33.4 per 100 person-years). M. genitalium incidence was higher among Kenyan participants compared with US participants (38.4 per 100 person-years [95% CI, 28.7–51.4] vs. 18.2 per 100 person-years [95% CI, 8.7–38.1]; P = 0.06). In univariate analysis, being a current smoker was associated with incident infection, whereas younger age and concurrent BV trended toward an association (Table 2). In a multivariable model that included these 3 factors in addition to study arm and stratified by site, smoking was independently associated with increased risk of incident MG (adjusted hazard ratio [HR], 3.02; 95% CI, 1.32–6.93), whereas the effect estimates for concurrent BV (adjusted HR, 1.59; 95% CI, 0.87–2.88) and age less than 25 years (adjusted HR, 1.70; 95% CI, 0.95–3.06) continued to trend toward an association.
The duration and patterns of incident MG are presented in Figure 2. Median time to MG clearance was 1.5 months (IQR, 1.4–3.0 months) and did not differ by study arm (P = 0.65). Of 52 women who experienced an incident infection during follow-up, 43 (83%) experienced one incident infection (i.e., no recurrence) and the majority had MG detected at only one study visit (26/43[60%]). Again, reported use of doxycycline was uncommon and did not seem to correlate with clearance of incident or recurrent infection (Fig. 2). Of 61 participants who had either prevalent (n = 22) or incident (n = 39) MG and cleared their infection, 20 (34%) had a recurrence of MG during study follow-up, with a median time to recurrence of 4.1 months (IQR, 3.7–7.3 months).
Macrolide Resistance-Mediating Mutations
Among 77 women who tested positive for MG during the trial, specimens from 26 women were unevaluable for MRMM testing. Fifty-one women had a total of 120 evaluable MG samples that were tested for MRMM. Sixteen specimens (13.3%) from 15 different women tested positive for MRMM; mutations A2059G and A2058G were detected. The proportion of specimens with MRMM did not differ by country (Kenya, 12/82 [14.6%]; United States, 4/38 [10.5%]; P = 0.54). Patterns of MG infection among women with MRMM are displayed in Figure 3. Ten of 15 women with MRMM were MG negative at the prior visit, suggesting that macrolide-resistant MG was acquired from a sex partner. In most cases of MRMM, infection was cleared by the next assessment. Mean duration of infection differed by MRMM status. Women with wild-type MG infection had a longer median duration of infection (48 days; IQR, 43–106 days) compared with those with MG with MRMM (42 days; IQR, 40–44 days; P = 0.01).
In this prospective cohort of Kenyan and US women, we observed a high prevalence of MG infection (11.3%), which was detected more frequently than other curable STIs. We observed a similar prevalence of MG between Kenyan and US women, with an overall prevalence that was consistent with observations from studies of high-risk women in Uganda (14%),25 Kenya (12.9%–16%),26–28 South Africa (8.7%),29 and the United States (7.7%–19.2%).19,30,31 We also observed a high incidence of MG infection (33.4 per 100 person-years), which again was substantially higher than the incidence of other curable STIs in this cohort reported previously (C. trachomatis incidence, 11.7 per 100 person-years; N. gonorrhoeae incidence, 7.2 per 100 person-years).24 Data on MG incidence are sparse, but our estimates are similar to those reported by other Kenyan studies (22.7–34.6 per 100 person-years).27,28 The high prevalence and incidence of MG highlight the significant burden of this pathogen among reproductive-aged women.
Differences were observed in factors associated with prevalent MG infection versus incident infection. Detection of VVC and having a new sex partner were associated with prevalent MG, but were not associated with incident infection. Conversely, being a current smoker was associated with increased risk of incident infection, whereas younger age and concurrent BV trended toward an association. For smoking, age, and concurrent BV, no association was observed for prevalent MG. These differences may be due to the fact that risk factor assessment was performed most likely after MG infection occurred (i.e., prevalent infections could have been recent or persisted for months), whereas risk factor assessment for incident infections occurred closer to the time of actual infection. The association of smoking, age, and concurrent BV with incident, but not prevalent infection, is likely due to limited study power and the smaller number of prevalent infections at baseline. Of note, the effect estimates for the association of smoking, age, and concurrent BV with prevalent MG were in the same direction as those in the incidence analysis and of generally similar magnitude. Our findings regarding BV and incident MG differed from a study by Lokken et al.28 that reported an association with BV at the preceding visit and incident infection. Both cohorts were from Kenya and had sampling every 2 months. However, the inclusion of women receiving a vaginal health intervention as part of present study decreased the overall prevalence of BV and could have impacted the relationship between BV and MG susceptibility. Lastly, in contrast to other reports,5,30,32,33 no association was observed between cervicitis by Gram stain and prevalent MG.
Among women with prevalent MG, median time to clearance was approximately 4 months. This represents the lower bound for infection duration among those with prevalent infection, because we were unable to account for person-time before enrollment. Median infection duration among women with incident MG was 1.5 months. This infection duration was similar to the median clearance time reported in a study of Ugandan women (2.1 months),25 but slightly shorter than those observed in another cohort of Kenyan women (2.9 months).28 Overall, there was little reported use of azithromycin and doxycycline in our cohort, and no women who tested positive for MG during follow-up ever received azithromycin. In all cases where doxycycline was dispensed during the trial, it was either received at MG-negative time points or did not result in eradication of the organism in infected women, suggesting that most infections were cleared spontaneously. Recurrent MG infection occurred in approximately one third of women, which is also consistent with the recurrence rate among women in the Ugandan study (39%).25
The proportion of MG infections with MRMM was relatively low (13.3% of MG-positive specimens had MRMM detected) compared with reports from other regions, where MRMM prevalence has approached 60%.16–19 The frequency of MRMM in our study population was similar to that observed in a cohort of women in South Africa (MRMM prevalence, 9.8%).29 Azithromycin-resistant infections emerge after azithromycin therapy in approximately 10% of MG-positive individuals.34 The lower frequency of MRMM observed in these African studies may be due to lower rates of background azithromycin use, reducing selective pressure. Despite the low overall frequency of MRMM detection, most MG infections with MRMM were detected as incident infections, demonstrating that macrolide-resistant MG is circulating in these regions. Interestingly, the duration of infection with macrolide-resistant MG was shorter than wild-type MG, which suggests that macrolide-resistant MG may be less fit or more easily cleared than wild-type strains. Given the rapid emergence of macrolide resistance,35 it is important to characterize the prevalence of MRMM in populations at high risk for MG infection to better inform treatment guidelines and potentially minimize treatment failure.
Our study should be interpreted in the context of several limitations. Specimens for MG testing were collected every other month, which may have inflated estimates of infection duration and may have also resulted in failure to detect infections of short duration. More frequent, monthly sampling would improve the precision of infection duration estimates and incidence. A proportion of MG-positive specimens were unevaluable for MRMM testing, limiting our ability to assess MRMM among all women with MG. Women participating in the trial were considered to be at higher risk for STIs because approximately half reported past or present transactional sex. As a result, the observed prevalence and incidence in our study population may be higher than that in the general population. Lastly, data on MG strain were not available. Therefore, we were unable to determine if women who experienced a second MG infection were infected with a new strain or if the MG infection actually persisted, with intervening negative visits being false negatives due to low organism concentration rather than true negatives.
In summary, our data show that MG infection is very common among sexually active women in Nairobi and Mombasa, Kenya, and Birmingham, Alabama. Given its association with adverse reproductive health outcomes in women, a high prevalence and incidence of MG could contribute to substantial morbidity. Additional work is needed to definitively demonstrate that MG causes adverse reproductive health outcomes.36,37 Studies of MG treatment and prevention will be critical to improving our understanding of this pathogen's contribution to adverse reproductive outcomes and developing appropriate public health control strategies.
1. McGowin CL, Anderson-Smits C. Mycoplasma genitalium
: An emerging cause of sexually transmitted disease in women. PLoS Pathog 2011; 7:e1001324.
2. Thurman AR, Musatovova O, Perdue S, et al. Mycoplasma genitalium
symptoms, concordance and treatment in high-risk sexual dyads. Int J STD AIDS 2010; 21:177–183.
3. Tosh AK, Van Der Pol B, Fortenberry JD, et al. Mycoplasma genitalium
among adolescent women and their partners. J Adolesc Health 2007; 40:412–417.
4. Hjorth SV, Björnelius E, Lidbrink P, et al. Sequence-based typing of Mycoplasma genitalium
reveals sexual transmission. J Clin Microbiol 2006; 44:2078–2083.
5. Anagrius C, Lore B, Jensen JS. Mycoplasma genitalium
: Prevalence, clinical significance, and transmission. Sex Transm Infect 2005; 81:458–462.
6. Tully JG, Taylor-Robinson D, Cole RM, et al. A newly discovered mycoplasma in the human urogenital tract. Lancet 1981; 1:1288–1291.
7. Manhart LE, Broad JM, Golden MR. Mycoplasma genitalium
: Should we treat and how? Clin Infect Dis 2011; 53(Suppl 3):S129–S142.
8. Mavedzenge SN, Van Der Pol B, Weiss HA, et al. The association between Mycoplasma genitalium
and HIV-1 acquisition in African women. AIDS 2012; 26:617–624.
9. Lis R, Rowhani-Rahbar A, Manhart LE. Mycoplasma genitalium
infection and female reproductive tract disease: A meta-analysis. Clin Infect Dis 2015; 61:418–426.
10. Manhart LE, Jensen JS, Bradshaw CS, et al. Efficacy of antimicrobial therapy for Mycoplasma genitalium
infections. Clin Infect Dis 2015; 61(Suppl 8):S802–S817.
11. Jensen JS. Protocol for the detection of Mycoplasma genitalium
by PCR from clinical specimens and subsequent detection of macrolide resistance–mediating mutations in region V of the 23S rRNA gene. Methods Mol Biol 2012; 903:129–139.
12. Walker J, Fairley CK, Bradshaw CS, et al. Mycoplasma genitalium
incidence, organism load, and treatment failure in a cohort of young Australian women. Clin Infect Dis 2013; 56:1094–1100.
13. Gesink DC, Mulvad G, Montgomery-Andersen R, et al. Mycoplasma genitalium
presence, resistance and epidemiology in Greenland. Int J Circumpolar Health 2012; 71:1–8.
14. Chrisment D, Charron A, Cazanave C, et al. Detection of macrolide resistance in Mycoplasma genitalium
in France. J Antimicrob Chemother 2012; 67:2598–2601.
15. Twin J, Jensen JS, Bradshaw CS, et al. Transmission and selection of macrolide resistant Mycoplasma genitalium
infections detected by rapid high resolution melt analysis. PLoS One 2012; 7:e35593.
16. Salado-Rasmussen K, Jensen JS. Mycoplasma genitalium
testing pattern and macrolide resistance: A Danish nationwide retrospective survey. Clin Infect Dis 2014; 59:24–30.
17. Dumke R, Thurmer A, Jacobs E. Emergence of Mycoplasma genitalium
strains showing mutations associated with macrolide and fluoroquinolone resistance in the region Dresden, Germany. Diagn Microbiol Infect Dis 2016; 86:221–223.
18. Gesink D, Racey CS, Seah C, et al. Mycoplasma genitalium
in Toronto, Ont: Estimates of prevalence and macrolide resistance. Can Fam Physician 2016; 62:e96–e101.
19. Getman D, Jiang A, O'Donnell M, et al. Mycoplasma genitalium
prevalence, coinfection, and macrolide antibiotic resistance frequency in a multicenter clinical study cohort in the United States. J Clin Microbiol 2016; 54:2278–2283.
20. McClelland RS, Balkus JE, Lee J, et al. Randomized trial of periodic presumptive treatment with high-dose intravaginal metronidazole and miconazole to prevent vaginal infections in HIV-negative women. J Infect Dis 2015; 211:1875–1882.
21. Hardick J, Giles J, Hardick A, et al. Performance of the gen-probe transcription-mediated [corrected] amplification research assay compared to that of a multitarget real-time PCR for Mycoplasma genitalium
detection. J Clin Microbiol 2006; 44:1236–1240.
22. Jensen JS, Bradshaw CS, Tabrizi SN, et al. Azithromycin treatment failure in Mycoplasma genitalium
–positive patients with nongonococcal urethritis is associated with induced macrolide resistance. Clin Infect Dis 2008; 47:1546–1553.
23. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991; 29:297–301.
24. Balkus JE, Manhart LE, Lee J, et al. Periodic presumptive treatment for vaginal infections may reduce the incidence of sexually transmitted bacterial infections. J Infect Dis 2016; 213:1932–1937.
25. Vandepitte J, Weiss HA, Kyakuwa N, et al. Natural history of Mycoplasma genitalium
infection in a cohort of female sex workers in Kampala, Uganda. Sex Transm Dis 2013; 40:422–427.
26. Gomih-Alakija A, Ting J, Mugo N, et al. Clinical characteristics associated with Mycoplasma genitalium
among female sex workers in Nairobi, Kenya. J Clin Microbiol 2014; 52:3660–3666.
27. Cohen CR, Nosek M, Meier A, et al. Mycoplasma genitalium
infection and persistence in a cohort of female sex workers in Nairobi, Kenya. Sex Transm Dis 2007; 34:274–279.
28. Lokken EM, et al. Recent bacterial vaginosis is associated with acquisition of Mycoplasma genitalium
. Am J Epidemiol 2017; 186:194–201.
29. Hay B, Dubbink JH, Ouburg S, et al. Prevalence and macrolide resistance of Mycoplasma genitalium
in South African women. Sex Transm Dis 2015; 42:140–142.
30. Gaydos C, Maldeis NE, Hardick A, et al. Mycoplasma genitalium
as a contributor to the multiple etiologies of cervicitis in women attending sexually transmitted disease clinics. Sex Transm Dis 2009; 36:598–606.
31. Hancock EB, Manhart LE, Nelson SJ, et al. Comprehensive assessment of sociodemographic and behavioral risk factors for Mycoplasma genitalium
infection in women. Sex Transm Dis 2010; 37:777–783.
32. Bjartling C, Osser S, Persson K. Mycoplasma genitalium
in cervicitis and pelvic inflammatory disease among women at a gynecologic outpatient service. Am J Obstet Gynecol 2012; 206:476.e1–476.e8.
33. Manhart LE, Critchlow CW, Holmes KK, et al. Mucopurulent cervicitis and Mycoplasma genitalium
. J Infect Dis 2003; 187:650–657.
34. Jensen JS, Bradshaw C. Management of Mycoplasma genitalium
infections—Can we hit a moving target? BMC Infect Dis 2015; 15:343.
35. Horner P, et al. Which azithromycin regimen should be used for treating Mycoplasma genitalium
? A meta-analysis. Sex Transm Infect 2017; 94:10–20.
36. Wiesenfeld HC, Manhart LE. Mycoplasma genitalium
in women: Current knowledge and research priorities for this recently emerged pathogen. J Infect Dis 2017; 216:S389–S395.
37. Martin DH, Manhart LE, Workowski KA. Mycoplasma genitalium
from basic science to public health: Summary of the results from a National Institute of Allergy and Infectious Diseases Technical Consultation and Consensus Recommendations for Future Research Priorities. J Infect Dis 2017; 216:S427–S430.