Objectives There are limited data on the prevalence of Mycoplasma genitalium (Mgen) coinfection with rectal chlamydia (Chlamydia trachomatis (CT)) and rectal gonorrhoea (Neisseria gonorrhoeae (NG)) infections and few studies examining the prevalence of pharyngeal Mgen in men who have sex with men (MSM). Using transcription-mediated amplification assay, this study aimed to determine the proportion of rectal CT and rectal NG infections in MSM who are coinfected with rectal Mgen, and the proportion of MSM with Mgen detected in the pharynx in order to inform clinical practice.
Methods This was a cross-sectional study conducted at Melbourne Sexual Health Centre in Australia. Consecutively collected rectal swabs from MSM that tested positive for CT (n=212) or NG (n=212), and consecutively collected pharyngeal samples (n=480) from MSM were tested for Mgen using the Aptima Mycoplasma genitalium Assay (Hologic, San Diego). Samples were linked to demographic data and symptom status.
Results Rectal Mgen was codetected in 27 of 212 rectal CT (13%, 95% CI 9 to 18) and in 29 of 212 rectal NG (14%, 95% CI 9 to 19) samples, with no difference in the proportion positive for Mgen. MSM with rectal CT/Mgen coinfection had more sexual partners than those with rectal CT monoinfection (mean 6 vs 11, p=0.06). MSM with rectal NG/Mgen coinfection were more likely to be HIV-positive than those with rectal NG monoinfection (OR=2.96, 95% CI 1.21 to 7.26, p=0.023). MSM with rectal CT/Mgen coinfection were more likely to be using pre-exposure prophylaxis than MSM with rectal NG/Mgen coinfection (OR 0.25, 95% CI 0.10 to 0.65, p=0.002). Pharyngeal Mgen was uncommon and detected in 8 of 464 samples (2%, 95% CI 1% to 3%). Pharyngeal Mgen was associated with having a rectal STI (OR=10.61, 95% CI 2.30 to 48.87, p=0.002), and there was a borderline association with being HIV-positive (p=0.079).
Conclusion These data indicate one in seven MSM treated for rectal CT or rectal NG will have undiagnosed Mgen that is potentially exposed to azithromycin during treatment of these STIs. Rectal Mgen coinfection was associated with specific risk factors which may inform testing practices. Pharyngeal Mgen was uncommon.
- mycoplasma genitalium
- men who have sex with men
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Mycoplasma genitalium (Mgen) is an established cause of urethritis1; however, there are limited data on the prevalence of Mgen at anatomical sites other than the urethra. A recent meta-analysis estimated the prevalence of Mgen in men who have sex with men (MSM) in the community at 3.2% (95% CI 2.1 to 5.1).2 Studies in urban STI clinics in Melbourne and Sydney estimate the overall prevalence of Mgen in MSM between 9.5% and 13.4%, respectively, with a prevalence of 7.0%–8.9% in the rectum and 2.7%–4.9% in the urethra.3 4 There are limited data on the prevalence of Mgen coinfection with the two most common STIs, Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG). In a study of clinical samples collected at STI clinics in London, Mgen was found to be a coinfection in 13.0% of CT and 2.4% of NG infections in all people at the clinic, including women, using PCR.5 Screening and/or testing for CT and NG at extragenital sites is commonly performed in MSM in line with international guidelines6; however, this is not currently recommended for Mgen.
Both CT and NG have been commonly treated with macrolides, in accordance with international treatment guidelines.6 7 Britain has recently moved away from azithromycin for first-line treatment of CT and NG due to growing concerns over macrolide resistance.8 9 One gram (1g) of azithromycin has also commonly been recommended for treatment of Mgen6 7; however, macrolide resistance is now detected in at least 40%–60% of Mgen infections in many countries,10 11 and treatment failure following azithromycin currently exceeds 60% in Melbourne, Australia. Treatment with 1g azithromycin leads to the selection of macrolide resistance in at least 12% of infections.12 13 The inadvertent exposure of Mgen to azithromycin during treatment of CT or NG infections may lead to the selection of macrolide resistance and contribute to the rising rates of macrolide resistance seen in Mgen globally over the past decade.12
Screening for NG and CT in the pharynx is widely recommended in MSM.14 There are few studies on the prevalence of Mgen in the pharynx, with conflicting data from published studies. In two Australian studies Mgen was not detected in the pharynx of MSM.3 15 However another study presented at the STI & HIV World Congress in 2015 showed a high prevalence of pharyngeal Mgen (13.5%) among 388 MSM recruited from gay bars in five cities across China.16 All studies to date have used PCR-based assays. We used a highly sensitive transcription-mediated amplification (TMA) assay to determine the prevalence of Mgen in MSM attending Melbourne Sexual Health Centre (MSHC), given the high background prevalence of rectal (7%) and urethral (3%) Mgen infection in asymptomatic MSM at MSHC.4
This study aimed to determine the proportion of rectal chlamydial and gonococcal infections in MSM coinfected with rectal Mgen and predictors of coinfection, as well as the proportion of MSM infected with Mgen in the pharynx and the risk factors associated with pharyngeal Mgen.
Materials and methods
This was a cross-sectional study among MSM attending MSHC in Victoria, Australia, between May 2017 and January 2018. Patients presenting to MSHC routinely complete a computer-assisted self-interview (CASI) about their sexual history, including gender, number of partners and condom use over the last 3 months, HIV status, and use of pre-exposure prophylaxis (PrEP) for HIV. MSM were identified as men attending MSHC who reported having anal sex with another man in the previous year via CASI. MSHC is the largest public sexual health service in Victoria, Australia. The centre provides around 50 000 consultations every year, with 36% of consultations for MSM.
We examined the prevalence of Mgen rectal coinfection with either CT or NG, and the prevalence of Mgen pharyngeal infection, in serially collected specimens. Clients attending MSHC are offered free screening for STIs, including rectal and pharyngeal CT and NG, with swabs tested using Aptima Combo 2 TMA Assay (Hologic, San Diego, California, USA).
Consecutive swabs that tested positive for CT or NG were stored at room temperature in Aptima specimen transport tubes until further testing, and samples obtained from MSM were identified. Rectal samples that tested positive for CT from HIV-positive patients at MSHC were sent to an external laboratory for Lymphogranuloma venereum (LGV) testing and were therefore not available for this study. Samples from MSM were stored until batches of 100 samples were formed. Samples were stored and tested within 60 days. Samples were assessed prior to further testing to determine if there was sufficient remaining buffer; a minimum of 1.7 mL of buffer was required for Mgen testing and samples without sufficient buffer were discarded. The remaining samples were tested for Mgen using the Aptima Mycoplasma genitalium Assay (Hologic). Equal numbers of sequential CT and NG positive samples were selected.
All consecutive pharyngeal swabs with sufficient remaining buffer following testing for CT and NG were stored at room temperature, and swabs from MSM were identified. Samples from MSM were stored until batches of 100 samples were formed. Samples were stored and tested within 60 days. Samples determined to have adequate buffer remaining were tested for Mgen using the Aptima Mycoplasma genitalium Assay (Hologic).
Samples were linked to demographic and epidemiological data (eg, age, number of male partners, condom and PrEP use, and HIV status), as well as clinical diagnosis, presence of any symptoms and reason for presentation, and then irreversibly deidentified.
All analyses were performed using Stata IC V.14 (StataCorp, College Station, Texas). We estimated that if 5% of 225 CT or NG samples were positive for Mgen, this would yield CIs of 2.5% to 8.6%. We used univariable and multivariable logistic regression analyses to determine risk factors associated with rectal coinfection with CT/Mgen and rectal coinfection with NG/Mgen, and pharyngeal Mgen monoinfection. Variables were included in multivariable models if the p value was ≤0.05; strongly correlated variables were excluded from multivariable analyses. Ninety-five per cent binomial CIs were calculated for all proportions.
Characteristics of rectal samples examined for coinfection
Four hundred and eighty rectal samples, positive for either CT or NG, were selected for testing for Mgen. Fifty-five (11%) samples yielded invalid results due to insufficient buffer despite careful selection. Valid results were obtained for 212 rectal samples that were positive for CT and 212 rectal samples that were positive for NG. One sample which was positive for both CT and NG and excluded from analyses tested negative for Mgen.
Characteristics of cases with rectal CT compared with rectal NG
Compared with MSM with rectal NG, MSM with rectal CT were more likely to be presenting to MSHC as a contact with partners diagnosed with an STI (p=0.07) and to be using PrEP (p=0.084), although these did not reach significance (table 1). MSM with rectal NG were more likely to be symptomatic compared with those with rectal CT (OR=2.54, 95% CI 1.71 to 3.77, p<0.001; table 1). MSM with rectal NG were more likely to be HIV-positive compared with men with rectal CT (OR=5.40, 95% CI 2.32 to 12.52, p<0.001) (table 1), although this finding is likely to have been influenced by the removal of a significant proportion of rectal CT samples from HIV-positive men for LGV testing. Both groups were similar in terms of age, number of sexual partners and condom use.
A relatively high proportion of MSM with rectal CT also had urethral CT (41/211 (19%, 95% CI 14 to 25)) and pharyngeal CT (25/212 (12%, 95% CI 8 to 17)) (table 1). In patients who had rectal NG, 66 of 208 (32%, 95% CI 25 to 39) also had urethral NG, and 85 of 209 (41%, 95% CI 34 to 48) had pharyngeal NG (table 1).
Characteristics of rectal CT and rectal NG cases that were coinfected with rectal Mgen
Overall, Mgen was detected in 27 of 212 rectal CT samples (13%, 95% CI 9 to 18) and in 29 of 212 rectal NG samples (14%, 95% CI 9 to 19) (table 1), with no difference in the proportion of cases with Mgen coinfection (p=0.774) (table 1).
MSM with rectal CT/Mgen coinfection had more sexual partners than those who had rectal CT monoinfection (mean 6 vs 11, p=0.06); however, this was of borderline significance (table 2). There were no significant differences in terms of age, condom use in the last 3 months, symptom status, whether they were an STI contact or STIs at other anatomical sites (table 2).
MSM with rectal NG/Mgen coinfection were more likely to be HIV-positive than those with rectal NG monoinfection (OR=2.96, 95% CI 1.21 to 7.26, p=0.023) (table 2). There were no significant differences in terms of age, number of sexual partners, PrEP usage, symptom status or reason for presentation (table 2).
Overall MSM with rectal NG/Mgen coinfection were less likely to be using PrEP than MSM with rectal CT/Mgen coinfection (OR 0.25, 95% CI 0.10 to 0.65, p=0.002; table 3).
Pharyngeal Mgen infection
Four hundred and eighty pharyngeal samples from MSM were tested for Mgen during the course of the study. Fourteen samples yielded invalid results due to insufficient buffer, with valid results obtained for 464 pharyngeal samples. Eight of 464 samples were positive for Mgen (2%, 95% CI 1% to 3%), 7 of 464 were positive for CT (2%, 95% CI 1% to 3%), and 23 of 464 were positive for NG (5%, 95% CI 3% to 7%).
Pharyngeal Mgen was associated by univariable analyses with being HIV-positive (OR=5.13, 95% CI 1.19 to 22.12), having Mgen detected in the rectum (OR=60.14, 95% CI 3.41 to 1061.47) and having either rectal CT or rectal NG (OR=12.99, 95% CI 3.01 to 56.06) (table 4). Using a composite variable of any rectal STI, in multivariable analyses, MSM with pharyngeal Mgen were more likely to have a rectal STI detected (OR=10.61, 95% CI 2.30 to 48.87; table 4) and have a borderline association with being HIV-positive. Although MSM with pharyngeal Mgen were more likely to have either rectal CT and/or rectal NG, none of the eight patients had an STI detected in their urethra or a coinfection detected in their pharynx. Of note, no male patient with pharyngeal Mgen had pharyngeal symptoms.
As for pharyngeal Mgen, men who tested positive for CT or NG in the pharynx were also more likely to have CT or NG detected in the rectum. Five of seven pharyngeal CT infections were associated with concurrent rectal CT (71%, 95% CI 29 to 96, OR=34.14, 95% CI 6.35 to 183.65), and 10 of 23 pharyngeal NG infections were associated with concurrent rectal NG (43%, 95% CI 23 to 66, OR=28.29, 95% CI 10.14 to 78.90).
This study examined coinfection of Mgen with CT and NG in the rectum of MSM, as well as the prevalence of pharyngeal Mgen in MSM being screened for CT and NG at the largest urban sexual health centre in Australia. We found high rates of coinfection, with Mgen present in 13%–14% of MSM with rectal CT or rectal NG. Pharyngeal Mgen was uncommon and detected in 2% of MSM, and most cases were associated with rectal CT or rectal NG infection. These findings indicate one in seven MSM infected with CT or NG in the rectum are coinfected with Mgen and that a substantial number of undiagnosed rectal Mgen infections are being inadvertently exposed to macrolide antibiotics, which is likely to be contributing to increasing macrolide resistance in Mgen.
A recent meta-analysis, including five studies mostly testing urine, estimated the prevalence of Mgen in MSM in the community at 3.2% (95% CI 2.1 to 5.1) at any site.2 Studies have shown Mgen to be most commonly detected in the rectum of MSM, with 40.7% of contacts of the infection positive in the rectal sites.17 Mgen was detected in 7% of rectal samples from asymptomatic MSM in a recent study of 1001 asymptomatic MSM at MSHC,4 and 9% of consecutive asymptomatic and symptomatic anorectal samples from 505 MSM in Sydney.3 The Melbourne study also found that of 89 MSM with Mgen at any site, 17% were coinfected with CT or NG, and in 143 MSM diagnosed with CT or NG at MSHC over the same period, 11% were coinfected with Mgen.4 Notably a recent Sydney study reported rectal CT to be independently associated with anorectal rectal Mgen (OR=5.0, 95% CI 2.1 to 11.8, p<0.001).3 A study at our centre did not show this association in the rectum, but that CT and NG were associated with urethral Mgen (p=0.03 and p=0.002, respectively).4
Globally for over a decade CT and NG have commonly been treated with regimens that include 1g azithromycin. Recent studies in Australia have shown resistance to azithromycin in Mgen in MSM now exceeds 80%.3 13 18 A meta-analysis and published data from our group have shown de novo resistance occurs in 12% of Mgen infections following use of 1g azithromycin.12 19 Data on the effect of extended azithromycin regimens on de novo resistance are less clear. A recent meta-analysis containing 82 patients who received 1.5g azithromycin (without preceding doxycycline) found de novo resistance to be 3.7% (95% CI 0.8% to 10.3%), while a prospective study of 106 Mgen-infected patients treated with 1.5g azithromycin found de novo resistance to be similar to 1g (12%, 95% CI 3% to 27%).12 19 The rising levels of macrolide resistance seen in Mgen is likely to be due to azithromycin use, particularly 1g in the treatment of Mgen, as well as inadvertent exposure during treatment of CT and NG. Macrolide resistance is also rising in gonorrhoea and syphilis.20 21 Collectively, these data add to a growing body of evidence that suggests azithromycin use should be reduced in the STI field.9 However, while limited safe and available treatment options exist for macrolide-susceptible Mgen, it seems sensible to use an extended azithromycin regimen with preceding doxycycline as published data suggest this may be associated with low levels of de novo resistance (2.6%, 95% CI 0.3% to 9.2%).22
Coinfection with Mgen in MSM with either rectal CT or rectal NG was associated with specific risk factors. Compared with having rectal NG alone, rectal Mgen/NG coinfection was more common among individuals with HIV. This association could not be assessed among patients with CT due to the removal of samples for LGV testing. A recent meta-analysis showed an association between HIV and Mgen in MSM, predominantly reflecting cross-sectional observational data.23 Prospective studies have suggested an association between Mgen and HIV infection in women in Africa,23 24 but there are no prospective studies examining this relationship in MSM, and the nature of this relationship remains to be established. Coinfection with Mgen in individuals with either rectal CT or rectal NG was associated with factors that largely reflect increased risk—current use of PrEP, HIV infection and having more male partners. While individually not particularly useful, collectively these risk factors may indicate which individuals are more likely to be rectally coinfected with Mgen. Mgen has been associated with current use of PrEP in previous studies,3 and the majority of studies have found younger age and increased number of sexual partners to be associated with Mgen.25 26 Neither coinfection with Mgen in the rectum or infection in the pharynx was associated with site-specific symptoms.
Mgen has been rarely reported in the oropharynx of MSM. Previous studies in Sydney and Melbourne, of 508 and 515 men, respectively, both failed to detect pharyngeal Mgen using PCR. TMA has a higher analytical sensitivity than PCR for Mgen,27 although the Aptima Mycoplasma genitalium Assay (Hologic) has yet to be validated for detection of pharyngeal Mgen. Given the organism load of Mgen is up to 100 times lower than that of CT,28 and may be at particularly low loads in the pharynx, TMA may be more likely to detect pharyngeal infections. We detected Mgen in 2% of pharyngeal samples, which is similar to the prevalence of CT in the pharynx in MSM attending our clinic.29 Pharyngeal Mgen was associated with having a concomitant rectal STI (either CT or NG), although a recent study at our centre found only 1.9% of 54 rectal Mgen infections had pharyngeal Mgen by PCR.4 Other pharyngeal STIs commonly occur concurrently with rectal infections, with 71% and 43% of CT and NG pharyngeal infections found to have a concurrent rectal infection, respectively. MSM who were HIV-positive were also more likely to have pharyngeal Mgen detected than HIV-negative men. As mentioned previously, there is a known association between HIV and Mgen,23 and the clearance rate of Mgen appears to be slower among HIV-positive people30; however, this association with HIV has not been previously reported with pharyngeal Mgen. There was no difference in the proportion of MSM reporting to be STI contacts between those with Mgen in the pharynx and those without. The transmissibility of pharyngeal Mgen is not known as it is not clear if this low positivity in the pharynx reflects passive infection/deposition, rather than active infection as has been hypothesised for pharyngeal CT. Overall these and other data indicate the pharynx is unlikely to be a significant source of Mgen transmission.
There were several limitations to this study. The Aptima buffer evaporates from the stored rectal and pharyngeal samples when stored at room temperature, which left insufficient remnant buffer for testing for many specimens. Subsequently, the samples were resealed with parafilm, which reduced evaporation. Rectal and pharyngeal samples that were positive for both chlamydia and gonorrhoea were more likely to have insufficient buffer remaining for additional Mgen testing, impacting on our ability in this study to examine cases with triple (Mgen, CT and NG) infections. Removal of CT-positive rectal samples from HIV-positive men being tested for LGV impacted on our ability to examine the association between HIV and rectal CT. This study was conducted at a single sexual health clinic, and so is likely to reflect a higher risk population that may impact on prevalence estimates.
We have demonstrated a high prevalence of coinfection of Mgen with CT and NG in the rectum of MSM attending our service, which highlights how commonly Mgen is being inadvertently exposed to antibiotics in the treatment of other STIs. While the prevalence of Mgen in men with other rectal STIs is high, lack of clarity around the natural history of Mgen in the rectum and concerns around issues of cost, toxicity and antimicrobial resistance in the treatment of macrolide-resistant Mgen make the issue of screening for Mgen in MSM far more complex than it is for CT and NG. Few papers have investigated pharyngeal Mgen in MSM. Our findings are in line with other publications and show Mgen is not commonly detected in the pharynx of MSM being screened for CT and NG, indicating it is not a common site of infection.
One in seven men who have sex with men treated for rectal Chlamydia trachomatis or rectal Neisseria gonorrhoeae will have undiagnosed Mycoplasma genitalium (Mgen).
Undiagnosed rectal Mgen appears to be commonly exposed to azithromycin during screening and treatment of rectal chlamydia and gonorrhoea.
Pharyngeal Mgen is uncommon.
We would like to acknowledge Catherine Flowers and Kate Paoli for assistance with testing samples for this project.
Handling editor Henry John Christiaan de Vries
Contributors All authors have contributed significantly to the work and approved the manuscript.
Funding This study received support from Hologic. RLL is supported by an Australian Government Research Training Program (RTP) Scholarship. TRHR was supported by NHMRC Early Career Fellowship (no 1091536).
Competing interests The Melbourne Sexual Health Centre receives funding from SpeeDx (Australia) for research projects on Mycoplasma genitalium; however, no funding from SpeeDx was received or used to support this project.
Patient consent for publication Not required.
Ethics approval The Alfred Hospital Research Ethics Committee approved this study (project number 178/17).
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request.