Long Duration of Asymptomatic Mycoplasma genitalium Infection After Syndromic Treatment for Nongonococcal Urethritis

Abstract

Background

Although Mycoplasma genitalium (MG) is an acknowledged cause of nongonococcal urethritis (NGU), access to diagnostic testing is limited. Syndromic management is common, yet little is known about natural history.

Methods

Between August 2014 and April 2016, 13 heterosexual men aged ≥16 years with MG were identified within a cohort study of men with and without NGU attending an urban sexually transmitted diseases clinic. Men had 6–7 monthly visits. NGU was defined as ≥5 polymorphonuclear leukocytes per high-power field on urethral Gram stain plus either visible urethral discharge or urethral symptoms. Men with NGU received 1 g of azithromycin. Men with persistent NGU received moxifloxacin 400 mg for 14 days. First-void urine was retrospectively tested for MG using transcription-mediated amplification. Resistance-associated mutations were detected by polymerase chain reaction (PCR) and sequencing. Organism load was determined by quantitative PCR.

Results

Sixty-two percent of MG-positive men had macrolide resistance–mediating mutations (MRMM) at enrollment; 31% had parC mutations (all outside the quinolone resistance–determining region). MG persisted after azithromycin in 7 men, 6 of whom had MRMM. The median duration of persistence in the absence of curative therapy was 143 days (range, 21–228). Five men experienced symptom resolution after azithromycin, but MG persisted for another 89–186 days before moxifloxacin. Organism load was somewhat lower in MRMM than wild-type infections (P = .16)

Conclusions

The high prevalence of macrolide resistance and long duration of infection after symptom resolution highlights the need for diagnostic MG testing of men with NGU to direct therapy.

Among heterosexual men, Mycoplasma genitalium infection persisted asymptomatically for many months after routine therapy for nongonococcal urethritis. Macrolide resistance was high and explained most cases of persistence. Routine diagnostic testing for M. genitalium is needed to direct therapy.

Mycoplasma genitalium (MG) is a sexually transmitted bacterium associated with reproductive tract syndromes in men and women. Increasing evidence suggests that it is associated with cervicitis, pelvic inflammatory disease, preterm birth, spontaneous abortion, and infertility in women [1], although some debate remains regarding the causal nature of these associations [2]. In contrast, MG is an acknowledged cause of male urethritis [3] and is increasingly considered in the clinical management of nongonococcal urethritis (NGU). However, effective management of MG is complicated by the limited availability of diagnostic testing in many settings and increasing antimicrobial resistance.

Although several diagnostic tests for MG are approved for use in the European Union, no assay has received Food and Drug Administration (FDA) approval in the United States. European, Australian, and British guidelines for managing MG infections include recommendations for diagnostic testing, partner management, and tests of cure [4–6]. In contrast, the 2015 Centers for Disease Control and Prevention sexually transmitted disease (STD) treatment guidelines focus on syndromic management, recommending azithromycin for NGU when MG is suspected and moxifloxacin for persistent NGU. Given the absence of an FDA-approved assay, there are no testing guidelines and partner management recommendations are limited [7].

In addition to limited diagnostic testing, antimicrobial resistance is rising rapidly. In a 2015 meta-analysis, azithromycin efficacy was 85% prior to 2009, but only 67% after 2009 [8]. Similarly, a 2017 meta-analysis of moxifloxacin efficacy found 100% cure rates prior to 2010, but these declined to 89% after 2010 [9]. Coupled with increasing antimicrobial resistance are anecdotal reports of symptom resolution without concomitant eradication of MG, which could facilitate sustained transmission. However, the frequency of persistent asymptomatic infection is poorly understood and the contribution of antimicrobial resistance to persistence is unknown. A delay in initiating MG testing and treatment within a cohort study of heterosexual men with and without NGU enabled us to observe the natural history of MG infections and address these questions.

METHODS

Men in this case series were recruited from the Public Health–Seattle & King County (PHSKC) STD Clinic in Seattle, Washington, and had at least 1 MG-positive test during the retrospective testing period. Eligible men were ≥16 years of age, had valid contact information, were English speaking, had vaginal sex at least once in their lifetime, and had insertive oral and/or vaginal and/or anal sex with a female partner within 60 days before enrollment. Men reporting any male sex partners in the last year, receipt of antibiotics in the month before enrollment, known human immunodeficiency virus–infected status, recent sexual contact with someone with gonorrhea, or who had evidence of gonorrhea, were excluded.

Enrollment

After providing written informed consent, men underwent a genital examination and provided urethral swab and urine specimens. NGU was defined as ≥5 polymorphonuclear leukocytes (PMNs) per high-power field on a Gram-stained slide of urethral exudates plus either urethral symptoms (dysuria, pruritus, or tingling; reported urethral discharge) or visible urethral discharge on examination. Men also completed a computer-assisted self-interview reporting sociodemographic characteristics and sexual behavior. Those with NGU at enrollment were treated syndromically with 1 g of azithromycin.

Urine was tested for Chlamydia trachomatis and Neisseria gonorrhoeae using the Aptima Combo 2 assay (Hologic, Inc, San Diego, California). An additional 2 mL of urine was stored in Aptima transport medium, designed to stabilize RNA, at –80°C for future MG testing. From December 2015 to April 2016, transcription-mediated amplification (TMA) testing was conducted with the direct tube sampling system using analyte-specific reagents (ASR) on the SB100 instrument [10]. Subsequently, testing was conducted on the Panther instrument, using ASR. The 95% limit of detection for this assay is 0.02 genome equivalents (geq) per reaction (Damon Getman, personal communication).

Follow-up

Men without NGU at enrollment returned for 6 monthly visits. Men with NGU at enrollment had an additional visit 1 month after enrollment so that all began the 6-month follow-up period without NGU. Men with urethral symptoms between visits were asked to return for evaluation. At each visit, clinical examination and sexually transmitted infection (STI) testing employed the same procedures described for enrollment. New cases of NGU received 1 g of azithromycin. Persistent NGU was treated syndromically with 400 mg of moxifloxacin daily for 14 days.

With one exception, all MG testing was performed on stored specimens after men completed their study participation, withdrew, or were lost to follow-up (time from last visit to testing was 15 days–16.7 months). The exception was participant 4, whose MG-positive test result from month 4 was available prior to month 5. Any man with MG detected at his final study visit was asked to return for retesting. Men with MG-positive tests were prescribed moxifloxacin.

Duration of Infection

Duration of MG infection was calculated as the number of days from the first MG-positive visit to either the midpoint (~14 days) between the last MG-positive and first MG-negative visit, or to 14 days after the final study visit if no test of cure was available. Duration of infection after clinical cure was calculated in a similar manner, beginning 14 days after the last NGU-positive visit. Among men without NGU at enrollment, duration of asymptomatic MG infection prior to therapy was calculated as the number of days from the first MG-positive visit to 14 days after receipt of curative therapy. In sensitivity analyses, we assumed that infection persisted through the retesting timepoint for the 2 men who returned for a nonstudy test.

Markers of Antimicrobial Resistance and Organism Load

Aptima-positive specimens underwent testing for macrolide resistance–mediating mutations (MRMM) in the 23S ribosomal RNA gene, all of which correlate well with treatment failure [11] and mutations in the parC gene, some of which correlate with treatment failure (eg, mutations at positions S83 and D87) [12]. DNA extraction was performed using a MagnaPure96 (Roche, Pleasanton, California) with large volume (1 mL) universal pathogen extraction protocol and elution in 50 µL. Samples were subsequently tested by polymerase chain reaction (PCR) [13, 14], and MRMMs were detected with PyroMark96 sequencing [13]. The parC gene was amplified and sequenced by conventional Sanger sequencing [15]. Organism load was determined by quantitative PCR with a limit of detection of <5 geq per reaction [14]. We performed a Wilcoxon rank-sum test to test for differences in duration of infection by MRMM status. We performed a correlated t test to evaluate differences in organism load, plus a regression model of log₁₀ organism load vs resistance mutations, including a random-person effect.

All study procedures were approved by the University of Washington Institutional Review Board.

RESULTS

Among 138 enrolled men, 16 had 1 or more MG-positive specimens between 8 August 2014 and 15 April 2016 (Figure 1). Of these, 3 had only 1 visit with a retrospective test and were excluded, leaving 13 men in this case series. Twelve of these 13 men had MG at enrollment, of whom 10 also had NGU. One man had a single asymptomatic infection at his final visit (Figure 2; Supplementary Table).

Figure 1.

Flowchart of enrollment and characteristics of men in the case series. Abbreviations: MG, Mycoplasma genitalium; MRMM, macrolide resistance–mediating mutation; NGU, nongonococcal urethritis; WT, wild-type. ^*Macrolide resistance mutation status could not be determined for the enrollment visit specimen but was determined to be wild-type at the subsequent visit.

Figure 2.

Clinical characteristics of heterosexual men and detection of urethral Mycoplasma genitalium infection, presence of macrolide resistance–mediating mutations, and organism load at monthly visits (no parC mutations previously associated with quinolone resistance were detected). Abbreviations: AZM, azithromycin; Enroll, enrollment; MG, Mycoplasma genitalium; MOX, moxifloxacin; MRM, macrolide resistance mutation; ND, not done; NGU, nongonococcal urethritis; Rx, treatment received. *Received azithromycin for Chlamydia trachomatis infection. ^†Patient declined antibiotic treatment. ^‡Received moxifloxacin for lower urinary tract symptoms. ^§Received azithromycin due to high polymorphonuclear leukocyte count, but did not meet criteria for nongonococcal urethritis diagnosis.

The mean age at enrollment was 28.1 years (range, 22–34 years). Eight (62%) participants were white, 3 (23%) were black, 1 (8%) was Asian, and 1 (8%) was multiracial (Table 1). Eleven participants (85%) were circumcised. Nine men reported an exact number of lifetime sexual partners (mean, 34 partners [range, 6–70]), whereas 4 men reported a range of 10–99. Nine men reported inconsistent condom use and multiple partners throughout the study, while 2 men reported no condom use and a single partner. Two men reported no sexual activity during follow-up (participants 3 and 10).

Table 1.

Sociobehavioral and Clinical Characteristics of Heterosexual Men With Urethral Mycoplasma genitalium Infection in Retrospectively Tested Urine Specimens From a Cohort Study

Participant No.	Age at Enrollment	Race	Lifetime No. of Sex Partners	No. of Sex Partners Throughout Follow-up	Condom Use Throughout Follow-up	Circumcision Status
1	29	White	15	1	None	Yes
2	27	Black	50–99	≥2	Inconsistent	No
3	22	White	9	0	NA	Yes
4	26	White	70	≥2	Inconsistent	Yes
5	23	Black	10–24	≥2	Inconsistent	Yes
6	34	White	60	≥2	Inconsistent	Yes
7	32	Asian	10–24	≥2	Inconsistent	Yes
8	29	White	50–99	≥2	Inconsistent	Yes
9	28	White	20	≥2	Inconsistent	No
10	30	White	6	0	NA	Yes
11	23	Asian & black	6	1	None	Yes
12	29	White	50	≥2	Inconsistent	Yes
13	33	Black	70	≥2	Inconsistent	Yes

Abbreviation: NA, not applicable - participant remained abstinent.

Resistance-associated Mutations

Testing for MRMM was successful in 37 of 48 (77%) MG-positive specimens. Eight men (62%) had MRMM at enrollment (participants 1–7 and 12), and 5 (participants 8–11 and 13) were infected with wild-type (WT) organisms (eg, no resistance mutations). All but 1 (participant 8) had the same mutation in all typeable specimens. The A2058C mutation was detected in 1 man (participant 5), 3 had A2058G (participants 4, 7, and 8), 1 had A2059C (participant 1), and 4 had A2059G (participants 2, 3, 6, and 12).

The parC gene was successfully amplified and sequenced in 36 of 48 (75%) MG-positive specimens. Four men had parC mutations (participants 2, 4, 5, and 13). All had the same mutation in all typeable specimens (Table 2), but none was in the quinolone resistance–determining region. Two men had a silent C234T mutation (participants 2 and 5), 1 had a P62S mutation (participant 13), and 1 had silent C234T/P62S mutations (participant 4).

Table 2.

Detection of Mutations in the Quinolone Resistance–determining Region of parC at Monthly Visits Among Heterosexual Men With Urethral Mycoplasma genitalium Infection

Participant No.	Enrollment	Month 1		Month 2	Month 3	Month 4	Month 5		Month 6	Month 7	Poststudy MG Test
1	WT	WT		WT	NT	WT	WT		WT	WT	WT
2	Silent C234T	NT		NT	NT	Silent C234T	–		Silent C234T	Silent C234T	NT
3	NT	WT		NT	WT	ND	WT		–	–	NA
4	P62S, Silent C234T	NT		NT	a	P62S, Silent C234T	P62S, Silent C234T		–	–	NA
5	Silent C234T	Silent C234T		Silent C234T	Silent C234T	Silent C234T	NDb	–	–	–	NA
6	WT	WT		–	–	–	–		–	–	NA
7	WT	–		–	NT	–	–		–	–	NA
8	WT	WTc	–	a	–	–	d		d	d	NA
9	WT	–		–	d	d	d		d	d	NA
10	WT	–		d	d	d	d		d	d	NA
11	NT	WT		NT	WT	WT	–		–	–	NA
12	WT	WT		WT	d	d	d		d	d	NA
13	–	–		–	–	–	–		–	P62S	NA

Abbreviations: –, participant was M. genitalium negative at this visit; MG, Mycoplasma genitalium; NA, not applicable (participant was not recalled for a poststudy test); ND, not done; NT, not typeable; WT, wild-type.

^aParticipant missed this visit.

^bParticipant had a symptom visit preceding month 5.

^cParticipant had a symptom visit preceding month 1.

^dParticipant withdrew or was lost to follow-up.

Duration of Infection After Azithromycin for NGU

Ten men with MG at enrollment had NGU and received 1 g of azithromycin (participants 1–10). MG was eradicated in 3 of these by month 1 (participants 7, 9, and 10). Two men had macrolide-sensitive infections (participants 9 and 10) and no subsequent positive tests. One had MRMM at enrollment (participant 7) and asymptomatic MG at month 3.

MG persisted after initial azithromycin therapy in 7 (70%) of the men with NGU (participants 1–6 and 8), of whom 6 had MRMM. The seventh man had a WT infection at enrollment but MRMM at month 1 (participant 8). One man with persistent infection had a negative test at month 5 but was MG positive again at month 6 (participant 2). All others remained persistently infected through their last positive test. Five of the 7 (71%) men in whom MG was detected after azithromycin had persistent NGU (participants 1–3, 6, and 8): 2 received moxifloxacin (participants 6 and 8), 2 received azithromycin again because they were considered reexposed (participants 1 and 2), and 1 was not treated (participant 3). Of the 4 men who received additional therapy at month 2, MG was only eradicated in the 2 who received moxifloxacin.

The median duration of MG infection after azithromycin among the 2 men with macrolide-sensitive infections (participants 9 and 10) was 14.75 days (range, 12.5–17 days). In contrast, the median duration of infection in the absence of curative therapy among the 7 men with MRMM (participants 1–6 and 8) was 143 days (range, 21–228 days) (P = .04).

Five of the 6 men with persistent infections experienced clinical cure after 1–2 doses of azithromycin but remained infected for another 89–186 days (participants 1–5). One man received moxifloxacin at month 1 for persistent NGU (participant 6) and 3 received moxifloxacin at month 5 for recurrent NGU (participants 3–5). All experienced clinical and microbiological cure after moxifloxacin.

Sensitivity Analysis

Participants 1 and 2 remained asymptomatically infected at their final study visits. When recalled 9–10 months later, both still had MG. One had MRMM. One had an interim negative test and nontypeable poststudy MG but MRMM at month 7. Both were treated with moxifloxacin, but neither had a test of cure. Assuming these 2 men remained infected from enrollment to receipt of moxifloxacin, the maximum duration of infection was 542 days; the maximum duration of persistence after clinical cure was 500 days.

Duration of Infection in the Absence of NGU and Associated Antibiotic Therapy

Two men with MG at enrollment did not have NGU. Participant 11 had an asymptomatic macrolide-sensitive infection through month 4 when he received azithromycin for urethral inflammation. MG was eradicated. Participant 12 had an asymptomatic MRMM infection, developed NGU at month 2, received azithromycin, and was lost to follow-up. The median duration of asymptomatic infection prior to azithromycin was 100 days (range, 70–130 days).

Sporadic Asymptomatic Infection

Participant 13 had a new MG infection at his final visit accompanied by elevated PMNs but no symptoms and no visible discharge. He was not treated and could not be reached for retesting.

Treatment Efficacy

MG was eradicated in 3 of the 4 (75%) men with WT infections after azithromycin (participants 9–11) and a test of cure, while MRMM emerged in 1 man (participant 8). In contrast, MG was undetectable after azithromycin in only 1 of the 7 (14%) men with MRMM (participant 7). MG was eradicated in all 5 (100%) who received moxifloxacin and had a test of cure (participants 3–6 and 8), none of whom had resistance-associated parC mutations.

Organism Load

The organism load assay yielded results in 38 of 49 (78%) specimens (Figure 2). In WT infections, the median load was 97 geq/5 µL template (range, 4–108403). In MRMM infections, it was 39 geq/5 µL template (range, 8–11829). In the regression model, men with MRMM had a 0.15-fold lower organism load than those with WT infections (95% confidence interval, .01–2.1; P = .16).

DISCUSSION

In this case series of 13 men with urethral MG infection, mutations consistent with macrolide resistance were detected in 62% at enrollment and organism load was somewhat lower in MRMM than in WT infections. Only 1 man appeared to clear his infection in the absence of antimicrobial therapy to which MG was susceptible. Six of the 7 men with MRMM experienced persistent infections following azithromycin, a currently recommended treatment for NGU in the United States, and overall MG infections persisted for a median of 143 days (range, 21–228 days) from initial clinical presentation to receipt of curative treatment or study end. Five of 6 men with persistent infections experienced symptom resolution after azithromycin and would have been considered cured in the absence of testing.

Our estimated median duration of infection was significantly longer in MRMM than in WT infections. The true duration of infection was likely longer. We estimated duration conservatively, defining the end of infection as the first negative test. If an MG-negative test between 2 positive tests represents a drop below the threshold of detection followed by reemergence rather than clearance and reinfection or flare-up, the median duration of infection would be longer. Additionally, the duration of infection prior to enrollment was unknown and not included in our estimates. Although persistence was primarily due to antibiotic resistance–associated treatment failure, we also observed 1 man with an asymptomatic WT infection that persisted for 5 months prior to receiving azithromycin, suggesting that long duration infections can also occur in the absence of treatment failure or MRMM. While few other data are available on the duration of MG infection in men, persistence in women from various settings ranges widely from 1.5 to 21 months [16–21]. As estimates of duration depend on length of follow-up time and frequency of testing, larger studies with longer follow-up periods and frequent sampling will be needed to more accurately determine the duration of MG infection.

Syndromic therapy for NGU with 1 g of azithromycin was associated with clinical cure in many cases but was mostly ineffective in eradicating the organism. In contrast, moxifloxacin was uniformly effective. Prior observations of azithromycin resistance range from 40% to 100% [13, 22–26] and the 62% prevalence of MRMM we observed is consistent with increasing rates of treatment failure [22]. The resolution of symptoms after azithromycin despite persistent infection may be related to anti-inflammatory properties inherent in azithromycin, although the long duration of infection in the absence of symptoms suggests that this does not entirely explain the disparate clinical and microbiologic outcomes we observed. Organism load decreased after initial administration of azithromycin, and this may have been below the threshold required to promote clinical symptoms. Alternatively, patients’ NGU may have been caused by an organism other than MG that was eradicated with azithromycin.

Given the worldwide emergence of moxifloxacin treatment failures [27, 28] and parC mutations linked to treatment failure [29, 30], it was surprising that we observed neither of these. This may be due to the small number of men in this case series, particularly since moxifloxacin treatment failures were previously observed in Seattle men (unpublished data). The PHSKC STD Clinic does not routinely perform MG testing and only prescribes moxifloxacin for persistent/recurrent NGU. There may be less selective pressure than in areas with routine MG testing and more frequent administration of moxifloxacin. While this suggests that moxifloxacin is still a viable option in this population, this may change with the introduction of more widespread MG testing and should be monitored.

The large proportion of men who experienced asymptomatic infections that persisted for months after therapy is cause for concern. These men likely assumed they had been effectively treated, and nearly all engaged in condomless sex, potentially infecting their sex partners. Given the high prevalence of MRMMs and the high frequency of multiple partners among these heterosexual men, MRMMs are probably also common among women with MG in the Seattle area, potentially rendering treatment of female syndromes equally challenging.

Strengths of this study include the prospective design with 6 months of follow-up and the use of sensitive nucleic acid amplification tests that captured low organism load infections [31]. Identification of MRMMs and parC mutations through sequencing is highly specific. There are also limitations. The number of men in these analyses was relatively small. We may have detected residual nucleic acids in some rather than viable organisms, although the long duration of asymptomatic infection in many suggests that most were active infections. We were unable to differentiate persistent infections from new infections or reinfections, and this likely occurred in some cases. However, the long duration of infection from the man who reported abstinence throughout his study participation (participant 3) and the stable MRMM and parC mutations, suggest that some cases truly represented persistence. Finally, some TMA-positive samples contained insufficient DNA for sequencing to identify resistance-associated mutations, likely due to the lower limit of detection of the TMA assay [31].

In conclusion, we found that MG often persists for many months and that this persistence can occur despite resolution of signs and symptoms after syndromic therapy for NGU. The low microbiologic cure rates after azithromycin and the high prevalence of MRMMs highlights the increasingly diminished utility of this antibiotic as routine therapy for NGU. Indeed, widespread use of azithromycin to treat STIs has likely contributed to the development of MG resistance and is one factor prompting recent changes in European, Australian, and British guidelines that no longer recommend azithromycin for NGU. While current data do not support widespread screening for MG, the cases of persistent MG after resolution of NGU that we describe highlight the need for MG testing among men with NGU and tests of cure after azithromycin. Where macrolide resistance is not already widespread, initial MG testing should ideally include testing for MRMMs to direct therapy. Beyond this, our findings highlight the need to better understand the natural history of MG, which appears to be a chronic and increasingly difficult-to-treat STI.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We gratefully acknowledge the contributions of the staff at the Public Health Seattle & King County STD Clinic for facilitating enrollment; Sean Proll for creation of figures; Sabina Astete for M. genitalium testing; and Hologic, Inc, for donated test kits and reagents.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases (grant number R01 AI110666). Study data were collected and managed using REDCap electronic data capture tools hosted at the Institute of Translational Health Sciences through a grant from the National Center for Advancing Translational Sciences (grant number UL1 TR002319). Partial support for this research (computing software) came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant to the Center for Studies in Demography and Ecology at the University of Washington (grant number P2C HD042828).

Potential conflicts of interest. L. E. M. has received honoraria, reagents, and test kits for diagnostic assays from Hologic, Inc; J. S. J. has received grants, personal fees, and speaker’s fees from Hologic; serves on the scientific advisory board of Roche Molecular Systems; serves on the Cepheid scientific advisory board; and has received travel support and personal fees from Cepheid. The Statens Serum Institut has received remuneration for contract work from SpeeDx, Hologic, NYTor, Diagenode, Nabriva, and GlaxoSmithKline (GSK). P. A. T. has received speaker’s fees from Hologic and has consulted with SpeeDx for their diagnostic assays. L. C. C. has received trainee support from the Institute of Translational Health Sciences (grant number TL1 TR002318). M. G. has received grants from the National Institutes of Health, Hologic, and GSK. The University of Washington has received funding for contract work from Hologic, SpeeDx, Abbott, and Cepheid. 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.