Methods

We conducted a cross-sectional study of 521 MSM attending six SOPVs in Melbourne between October 2001 and September 2002, as previously described.10 The aims of that study were to determine the prevalence of C trachomatis and N gonorrhoeae in the pharynx, rectum and urine; to examine the feasibility of an STI screening programme in SOPVs; and to evaluate risk factors for acquisition of bacterial STIs in this population. The aim of this study is to describe the prevalence and clinical associations of M genitalium in urethral, rectal and pharyngeal specimens collected from this sample of MSM. The prevalence of C trachomatis and N gonorrhoeae in this sample of MSM has been previously reported10; some of these data are presented here for comparison to M genitalium.

In the original study,10 male sexual health nurses systematically approached SOPV patrons and invited them to participate. Participants completed a questionnaire on demographics, urethral and rectal symptoms in the previous week, and for the previous month: number of partners, unprotected oral and/or anal intercourse, and the number of visits to venues or beats that resulted in sexual contact. Men self-collected first-pass urine and rectal samples, and nurses collected throat swabs. All participants were given written participant information and consent was obtained. Participants were not tested for HIV infection. The original project was approved by the Victorian Department of Human Services human research ethics committee in 2001. All specimens were refrigerated and transported at 4°C to the molecular microbiology laboratory of the Royal Women’s Hospital in Melbourne, Australia, for nucleic acid amplification testing for C trachomatis and N gonorrhoeae. Samples were then stored, and ethics approval was obtained in 2008 from the Royal Women’s Hospital human research ethics committee, Melbourne, for analysis of samples for M genitalium.

Urine samples were stored at −80°C with DNA extraction occurring before use in this study. DNA was extracted from rectal and throat swabs after collection and stored in TE buffered solution (10 mM TRIS, 1 mM ETDA) at −30°C. DNA extraction was performed using the automated MagNA Pure LC (Roche Diagnostics, Mannheim, Germany) and DNA Isolation Kit I protocol. Detection of M genitalium in DNA samples was performed by amplification of a 516 base pair sequence of the 16S gene by rapid real-time PCR (LightCycler, RocheMolecular Biochemicals) in accordance with established methods.11 This assay has been shown to have analytical sensitivity of 10 copies per reaction,11 which is equivalent to 1000 copies per ml of urine and 400 copies per swab. It has undergone accreditation by our National Laboratory Accreditation Authority (NATA), and has been employed in a number of our previously published studies.12 13 14 15 16 The human β-globin gene was amplified by real-time PCR from all samples to act as an internal control to evaluate sample adequacy, inhibition and integrity following storage.17 Samples negative for both β-globin and M genitalium were deemed not assessable. No published data are available on the application of this real-time PCR assay11 to rectal and pharyngeal samples. However, assessment of the performance of this M genitalium PCR assay11 on urethral samples and urine was done before its use in our research studies. An exchange of a panel of M genitalium-positive and negative specimens occurred between our laboratory and colleagues in the field (David Martin, Louisana State University Health Sciences Centre, New Orleans, Louisiana, USA). This panel of specimens consisted of 31 samples (9 positive and 22 negative). Specimens were blindly tested with methods described above and compared to methods established elsewhere18 and a 100% concordance between the assays was found (data not shown).

Data analyses were conducted using SPSS (version 15.0 for Windows) and Stata (version 10). Proportions were compared using a test for equality of proportions, and χ² and Fisher’s exact tests where appropriate; 95% confidence intervals (CI) were calculated. Univariate analysis was used to calculate crude odds ratios (OR) and 95% CI. Logistic regression was used to investigate variables independently associated with M genitalium. Variables found to be significant on univariate analysis at p<0.05 were included in the regression model. MSM were excluded from the analysis where clinical information or specimens were not available.

Results

Of 2140 men approached, 521 (24%) agreed to participate in the study. Reasons for declining to participate have been previously described10 but no other data were available for those who declined to participate. All participants contributed at least one sample to the study: 511 (98%) collected a urine sample, 507 (97%) self-collected a rectal swab and all provided a throat swab.

The overall prevalence of STIs was high in the study population; 67 participants (12.9%; 95% CI 10.2% to 15.9%) had one or more infections with C trachomatis, N gonorrhoeae or M genitalium detected at a rectal, throat or urethral site. M genitalium (2.1%; 95% CI 1.1% to 3.6%) was less common than C trachomatis (8.1%; 95% CI 5.9% to 10.6%; p<0.001), and N gonorrhoeae (4.8%; 95% CI 3.2% to 6.9%; p = 0.02). M genitalium was detected in equal proportions in rectal (1.6%; 95% CI 0.8% to 3.0%) and urethral sites (0.6%; 95% CI 0.2% to 1.6%; p = 0.12), but was not detected in the pharynx (table 1). In comparison, C trachomatis was more commonly detected in the rectum (6.2%; 95% CI 4.3% to 8.6%) than the urethra (1.8%; 95% CI 0.9% to 3.2%, p = 0.004), but was also uncommon in the pharynx (0.6%; 95% CI 0.1% to 1.6%). Only one participant had urethral infection with N gonorrhoeae (0.2%; 95% CI 0.01% to 1.0%), but it was detected as often in the pharynx (2.5%; 95% CI 1.4% to 4.2%) as the rectum (2.2%; 95% CI 1.2% to 3.8%). All participants infected with M genitalium were infected in one site only and had no other STIs detected. In contrast, 13 (2.5%; 95% CI 1.4% to 4.1%) participants had C trachomatis and/or N gonorrhoeae detected at more than one site, four had anal N gonorrhoeae/C trachomatis co-infection, and one had urethral N gonorrhoeae/C trachomatis co-infection.

The majority of rectal infections were asymptomatic. None of the eight participants with rectal M genitalium infection reported anal symptoms (anal discharge, discomfort or bleeding); only two of 10 participants with anal N gonorrhoeae and five of 29 participants with anal C trachomatis reported anal symptoms. Anal symptoms were not more commonly reported in those with a rectal STI (7/44 (15.9%)) compared to those without a rectal STI (39/472 (8.3%)), (OR = 2.1; 95% CI 0.9% to 5.0, p = 0.09). Urethral discharge and/or dysuria were reported by only one of three patients with M genitalium urethral infection, one of eight participants with C trachomatis and the one patient with C trachomatis/N gonorrhoeae co-infection. Urethral symptoms were, however, more commonly reported in those with a urethral STI (3/12 (25%)) compared to those without urethral infection (22/494 (4.4%)), (OR = 7.2; 95% CI 1.8% to 28.3), p = 0.02.

The median age of all participants was 39 years (range 18–85 years). M genitalium was not associated with age, sexual identity, number of sexual partners or oral sex or unprotected anal sex in the previous month (data not shown); however, there were only 11 participants with M genitalium infection. Behavioural and demographic associations with detection of N gonorrhoeae and C trachomatis in this population were limited and have been previously described.10 Multivariate analysis found that only age <39 years was significantly associated with detection of any STI (adjusted OR = 0.4; 95% CI 0.2% to 0.7, p<0.01) in this population.

Discussion

This study found that M genitalium was less common than C trachomatis and N gonorrhoeae in MSM attending male-only saunas, and that it was most likely to be detected as an asymptomatic rectal or a urethral infection but was absent in the pharynx. M genitalium was not detected in individuals with gonococcal or chlamydial infections. No significant demographic or behavioural associations with M genitalium were identified in this population.

In the absence of any commercially available assay for M genitalium, efforts to examine its prevalence and contribution to STI syndromes and to obtain comparative data across populations are hampered. A number of research assays for detection of M genitalium have been described.9 11 19 20 21 We employed a PCR assay11 which has undergone accreditation by NATA, and has been utilised in a number of our previous studies12 13 14 15 16; it has been in routine diagnostic use at Melbourne Sexual Health Centre for 3 years. The prevalence of M genitalium (2%) was low, but not unexpected, given that N gonorrhoeae and C trachomatis were detected in 4% and 8% of samples, respectively. Although factors such as degradation of DNA during storage (to which M genitalium is particularly susceptible) could have resulted in lower detection levels, this is likely to be minimal as samples were stored optimally at −80°C and extracted DNA was stored in TE buffer at −30°C. The sensitivity of the amplification assay utilised in this study could also contribute to reduced levels of detection, but based on our comparison to other assays this difference is not likely to result in a significant increase in detection of M genitalium in this population.

Another important consideration is that very limited data are available on the application and performance of any of these assays on rectal and pharyngeal samples. Francis et al recently showed high concordance (96%) between TMA9 and a multi-target real-time PCR on rectal samples,7 but no published comparative data on the performance of other PCR assays on rectal samples, or any data on their application to pharyngeal samples are available. Only one published study has employed PCR to detect M genitalium in the pharynx and they found no infections in 28 MSM 8 The absence of M genitalium in a larger sample of 521 MSM in this study supports this finding. Despite the limited available data to validate the use of research PCRs for the testing of non-urethral specimens, these findings may indicate that M genitalium, like C trachomatis, may not achieve sustained infections in the pharynx, and that the bulk of transmission of M genitalium in the community may be occurring predominantly through unprotected vaginal and anal intercourse, with oral sexual transmission having a lesser role. Clearly further research is needed to support this premise. However, a study of heterosexual and MSM STD clinic attendees found urethral M genitalium and C trachomatis were both significantly associated with unprotected vaginal sex and these infections were more common in heterosexual males compared to MSM.13 In contrast with these findings, adenovirus and herpes simplex type one (HSV-1) were significantly associated with oral sex and were more common in MSM.13

In this study M genitalium was not significantly associated with anal symptoms; however there were only three cases of rectal infection, which limited our ability to examine clinical associations. The only other sizable study of rectal infection by Francis et al7 also did not report a significant association with M genitalium and anal symptoms (p = 0.35) and proctitis (p = 0.70), in their adjusted analysis. Of their confirmed cases of proctitis (n = 27), three men had M genitalium, but they were co-infected with C trachomatis or HSV-1.7 These findings suggest that M genitalium may be only weakly associated with anal symptoms and proctitis in MSM, although more data are needed.

Limited demographic and behavioural data were collected in this study to minimise barriers to recruitment and facilitate rapid screening in SOPVs. The number of cases with M genitalium limited our ability to examine associations with this infection, but apart from younger age, there were no significant behavioural associations with detection of any STI in this population. It is possible that behaviour over the last month may not have predicted infection as participants could have under-reported high-risk sexual behaviour, or a substantial number of infections may have been acquired more than one month earlier, particularly as a high proportion were asymptomatic. There is likely to be some selection bias associated with our study, as the participation rate was 24% of those approached, and participants could only be selected from certain areas of the sauna.

This study shows M genitalium was less common than C trachomatis and N gonorrhoeae in MSM attending male-only saunas. It accounted for a small, but substantial proportion of asymptomatic rectal infections and urethral infections. M genitalium was not detected in the pharynx, which may indicate that this infection is not sustained at this site, and that oral sex is not a significant mode of transmission of M genitalium; however further studies in other populations employing a range of assays would be required to validate this finding. The lack of a commercial assay for detection of M genitalium poses a significant barrier to testing for this organism in many clinical settings, even in syndromes such as non-gonococcal urethritis, in which its aetiological role is clearly established. Further research to understand the contribution of M genitalium to clinical syndromes such as proctitis, and its role in rectal, pharyngeal and urethral infections in large studies of MSM is clearly needed before advocating incorporation of M genitalium testing into routine screening practices in asymptomatic MSM.