Article Text

Original article
Incidence and risk factors associated with chlamydia in men who have sex with men: a cohort analysis of Victorian Primary Care Network for Sentinel Surveillance data
  1. Anna Wilkinson1,2,
  2. Carol El-Hayek1,
  3. Christopher K Fairley2,3,
  4. David Leslie4,
  5. Norm Roth5,
  6. B K Tee6,
  7. Margaret E Hellard1,
  8. Mark Stoové1
  1. 1Centre for Population Health, Burnet Institute, Melbourne, Australia
  2. 2School of Population Health, University of Melbourne, Melbourne, Australia
  3. 3Melbourne Sexual Health Centre, Melbourne, Australia
  4. 4Victorian Infectious Diseases Reference Laboratory, Melbourne, Australia
  5. 5Prahran Market Clinic, Melbourne, Australia
  6. 6The Centre Clinic, Melbourne, Australia
  1. Correspondence to Anna Wilkinson, Burnet Institute, 85 Commercial Road, Melbourne, Victoria 3004, Australia; awilkinson{at}


Background Chlamydia is the most commonly notified sexually transmitted infection (STI) in Australia. Incidence studies of chlamydia in men who have sex with men (MSM) are rare and offer important public health information.

Objective To determine chlamydia incidence in MSM presenting at high caseload clinics and describe predictors of infection.

Methods The Victorian Primary Care Network for Sentinel Surveillance of bloodborne viruses and STIs (VPCNSS) links testing, demographic and behavioural data from individual testers at participating clinics. Data from MSM with more than one chlamydia test at the VPCNSS site between April 2006 and June 2010 were included. Chlamydia incidence per 100 person-years (PY) was calculated and Cox regression used to examine predictors of incidence.

Results 1206 positive tests for chlamydia were detected among 6333 MSM across 11 409 PY of follow-up. Overall chlamydia incidence was 10.6/100 PY (95% CI 10.0 to 11.2) and was highest among MSM aged 16–29 years (12.9/100 PY, 95% CI 11.7 to 14.1), presenting with STI symptoms (16.0/100 PY, 95% CI 14.2 to 18.0), HIV positive (18.5/100 PY, 95% CI 16.6 to 20.6) and self-identified sex workers (14.3/100 PY, 95% CI 10.0 to 20.6). Significant predictors of chlamydia infection among MSM were younger age (adjusted hazard ratio (aHR)=1.9, 95% CI 1.5 to 2.3), self-identifying as a sex worker (aHR=1.6, 95% CI 1.0 to 2.6), being HIV positive (aHR=2.6, 95% CI 1.8 to 3.8), presenting with STI symptoms (aHR=1.7, 95% CI 1.4 to 2.1) and reporting >10 sex partners in the past 6 months (aHR=2.5 95% CI 1.4 to 4.6).

Conclusion These results show that MSM represent a key risk population for chlamydia in Australia and identify a number of high-risk MSM subpopulations for whom clinical and public health interventions are warranted.

  • Chlamydia
  • incidence
  • men who have sex with men
  • risk factors
  • sexually transmitted infections
  • sexual health
  • epidemiology (general)
  • public health
  • chlamydia
  • anogenital cancer
  • AIDS
  • bacterial vaginosis
  • epidemiology (clinical)
  • syphilis
  • chancroid
  • chlamydia infection
  • genital ulcers
  • genitourinary medicine

Statistics from


Chlamydia is the most commonly notified infectious disease in Australia.1 In Victoria, the second most populous jurisdiction in Australia, notification rates for chlamydial infection have increased significantly over the past 10 years, from 69.8/100 000 population in 2000 to 296.7/100 000 population in 2010.2 The known sequelae and serious complications of untreated male urethral chlamydia infection include epididymitis, infertility, Reiters syndrome and conjunctivitis. For men who engage in receptive anal intercourse, chlamydial proctitis can occur.3

Young and predominantly heterosexual populations make up the majority of chlamydia notifications in Australia, with 81% of chlamydia notifications being among those aged 15–29 years, of which 63% are females.2 However, prioritising men who have sex with men (MSM) in clinical and public health prevention practice is important due to a high prevalence of sexually transmitted infections (STIs) and sexual risk practices and the known contribution of STIs to HIV transmission.4 ,5 The Second National Sexually Transmissible Infections Strategy 2010–2013 identifies MSM as a priority population for reducing chlamydia incidence.6 While Australian4 and international7 studies have identified a number of key risk factors for chlamydia in MSM, including younger age, high numbers of casual sex partners and high-risk sexual behaviours, authors also note the rarity of chlamydia incidence data among MSM from which to reliably determine predictors of infection.4

Although passive notifications of chlamydia can provide useful outcomes with regard to time trends and burden of disease, data from passive surveillance rely heavily on testing practices and access to services8 ,9 and cannot provide estimates of incidence or examine predictors of incidence. Passive notifications are assumed to underestimate chlamydia transmission due to limited asymptomatic testing and presumptive treatment of patients and sexual partners.10 An increase in screening patients by general practitioners has been reported,10 however, significant under screening and therefore under-reporting of chlamydia occurs, especially among males.8 ,11

In the past decade, Australian governments have invested millions of dollars on health promotion and clinical strategies to reduce chlamydia transmission, including a national pilot programme for chlamydia screening in general practice planned for 2011–2013.12 ,13 It is important that such initiatives are informed by data that identifies population subgroups and risk factors that contribute to chlamydia transmission.9

In this study, we estimate the incidence of chlamydia in MSM presenting to high caseload sexual health and gay community clinics and identify risk factors for chlamydia infection among MSM.


The Victorian Primary Care Network for Sentinel Surveillance of bloodborne viruses and STIs (VPCNSS) was established by the Burnet Institute in collaboration with the Victorian Infectious Diseases Reference Laboratory, Melbourne Sexual Health Centre and the Victorian Government Department of Health. In 2006, the VPCNSS established four disease networks: HIV, syphilis, chlamydia and hepatitis C infection.14 The chlamydia network consists of 11 sentinel sites selected to cover at-risk populations, young people (aged 16–24 years) and MSM. The MSM network, from which data presented in this study were collected, includes three general practices that specialise in gay men's health, a metropolitan sexual health centre and a regional sexual health centre.14 The system methods have been described in detail previously,14 so only select methodological details follow.

Demographic and sexual risk behaviour data across a 6-month recall period were collected from MSM undergoing chlamydia testing as part of standard clinical practice using a paper questionnaire at the general practice sites. The regional sexual health centre moved from paper to electronic surveys in 2009 and one general practice did not contribute behavioural data from questionnaires. The metropolitan sexual health clinic used a computerised medical records system to collect pre-test risk assessment and sexual behaviour questions using a 12-month recall period. All questionnaire items were completed by patients at the metropolitan sexual health centre. For other sites, behavioural items were predominantly patient completed, and clinical and demographic items were completed by the clinician in consultation with patients. Patients' country of birth, Aboriginal or Torres Strait Islander status and sexworker status were self-reported. The presence of STI symptoms was reported by MSM at the time of the consultation; however, these were not defined and recorded based on the clinician's judgement; symptoms may have been unrelated to chlamydia. Questions regarding condom use related specifically to anal sex at all participating sites. At the metropolitan sexual health centre, questions about number of partners for sex workers are specific to non-paying partners.

Chlamydia trachomatis pathology results were obtained from the laboratories that conducted chlamydia testing for the sentinel sites and linked with questionnaires at the Burnet Institute. Indeterminate chlamydia results and positive tests within 30 days of a previous positive were excluded. If individuals had samples taken from multiple anatomical sites during the same visit, they were classified as chlamydia positive if a test from any anatomical site was positive. Men were classified as MSM if they self-reported as MSM or reported male sexual partners. Where behavioural data were unavailable, men were assumed to be MSM if tested for chlamydia using an anal or throat swab (510 test records) or were HIV positive (246 test records). HIV-positive status in MSM were determined by matching viral load testing data to chlamydia test data, a positive HIV result within the system or self-reported in questionnaire data.


Data collected between April 2006 and June 2010 were used in the analysis. To estimate chlamydia incidence, MSM who only had one test for chlamydia within the follow-up period were excluded from the analysis. χ2 tests were used to assess differences in the characteristics of single versus multiple testers. Person-years (PY) of follow-up were calculated from the first chlamydia test to the end of follow-up (30 June 2010). In addition, statistical methods used excluded outcomes from the first record of participants including records that indicated a positive chlamydia result. Participants had at least two tests for chlamydia and may have had multiple positive tests.

Incidence was calculated using standard formulae and reported as events per 100 PY overall and stratified by age, Australian born, self-identified sex worker, HIV status and STI symptoms present at the time of testing. Log-rank tests for equality were used to assess differences in incidence between groups. Participants' HIV status was reported as a baseline variable and taken from the first test record. Over the follow-up period, there were 524 (9.8%) participants who were HIV negative on their first test record and then subsequently HIV positive on their last test record.

The Anderson and Gill method for analysis of multiple failure-times data with ordered events was chosen and included robust SEs to adjust for clustering (ie, multiple tests and re-infections in the same individual).15 Cox regression was used to examine associations between exposure variables and chlamydia incidence. The final multivariable model was constructed using backward selection.

Six Human Research Ethics Committees including the Victorian Department of Health approved the project (online appendix A). All collaborators and health services involved had individual ethics committees and processes.14


A total of 6333 MSM with more than one test for chlamydia were included in the analysis, contributing 11 409 PY of follow-up between April 2006 and June 2010. The median follow-up was 1.7 PY. Excluding their first record, the 6333 MSM contributed 17 050 test records of which 72.3% (n=12 372) were matched to a behavioural questionnaire. Survival analysis included 1206 positive tests for chlamydia. The median time between tests was 6 months.

Baseline characteristics

A comparison was made of the baseline characteristics of participants who had only one test (excluded from subsequent analyses) with those who had multiple tests (using their first test record). Those with multiple test records were more likely to be Australian born, a self-identified sex worker, report no STI symptoms at this first test and HIV positive (χ2, p<0.05).

MSM tested for chlamydia were typically aged in their mid-30s and Australian born. Most were asymptomatic at the time of testing and most commonly reported >10 sex partners and inconsistent condom use with any sex partners in the past 12 months. MSM attending general practices most commonly reported between two and five partners in the past 6 months and most commonly reported condom use with casual sex partners. Approximately one in seven participants were HIV positive at baseline and of those with regular partners that were asked; a small proportion reported an HIV-positive regular partner (table 1).

Table 1

Description of baseline demographics, clinical characteristics and sexual risk practices of MSM

Crude incidence rates

The overall rate of chlamydia infection was 10.6 per 100 PY (95% CI 10.0 to 11.2). Chlamydia incidence was highest among those aged 16–29 years, those who reported sex work, those who were HIV positive at baseline and those presenting with STI symptoms (table 2).

Table 2

Crude incidence rates of men who have sex with men by demographic and clinical characteristics

Cox regression model

After adjusting for other variables in the model, younger age, self-reported sex work, being HIV positive, presenting with STI symptoms, greater numbers of anal sex partners and inconsistent condom use with casual partners in the past 6 months were associated with chlamydia incidence (table 3).

Table 3

Predictors of chlamydia diagnosis among men who have sex with men


The incidence of chlamydia in MSM in the VPCNSS between April 2006 and June 2010 was 10.6/100 PY. This is more than double the incidence of chlamydia infection in women aged 16–24 years (the group thought to be most affected by chlamydia), which was recently reported as 4.4/100 PY in Victoria (Walker J, Tabrizi S, Fairley C et al., unpublished data, 2012).16 This is the first Australian study to have estimated chlamydia incidence among HIV-positive MSM. HIV-positive MSM in this study were two and a half times more likely to test positive for chlamydia compared with HIV-negative MSM.

The key limitations of this study arise from participants self-referring to the VPCNSS sites, thus potentially limiting the generalisation of chlamydia incidence rates and predictors to the general MSM population. Participants in this study may over-represent MSM at higher risk of STIs. First, because MSM actively sought or presented with characteristics that prompted STI testing and participants included in analyses had more than one test for chlamydia. Lower STI risk MSM may be under-represented because sexual history taking did not prompt a chlamydia test. Second, the selection of sites in the VPCNSS was based on a caseload profile showing high rates of MSM presentations, and there is potential that MSM who are at higher risk of STIs, and HIV-positive MSM presenting for routine care, may be attracted to such clinical settings. Although HIV infection has been associated with enhanced risk of being diagnosed with other STIs,7 the results of this study are also likely to reflect enhanced opportunities for timely chlamydia diagnoses among MSM presenting for routine HIV care. In comparing participants with only one test with multiple testers, a greater proportion (59%) of HIV-positive participants were multiple testers. Results may also be influenced by the first record for each participant not being included in the survival analysis as no PY were contributed and positive chlamydia tests on the first record excluded. A sensitivity analysis was undertaken to examine the impact of missing data, and the conclusion was unchanged (online appendix B). Finally, participants tested at the VPCNSS sites may also have received testing at alternate services during the follow-up period and thus additional chlamydia tests and diagnoses may have been missed.

The incidence rate in HIV-negative MSM (9.0/100 PY) is comparable to the 11.5/100 PY reported in the only previous study of chlamydia incidence in HIV-negative MSM in Australia.4 The Sydney Health in Men (HIM) Study used community-based sampling as opposed to this study, which recruited through General Practitioners and Sexual Health Centres. The aforementioned study also included a large proportion (90%) of incident urethral chlamydia infections that were self-reported and diagnosis occurred between study visits.4 Although this study presents data from a cohort of participants proactively seeking and engaging in healthcare from high caseload MSM clinics, the comparable incident rate reported here with that reported in the community recruited HIM Study affords some reliability to the estimation of chlamydia incidence among Australian MSM reported in this study.

This study reported largely similar chlamydia risk findings to the HIM Study4 and other international studies of MSM,7 ,17 with age, number of sex partners and unprotected anal intercourse being predictive. Number of anal sex partners was a stronger risk factor for a positive test for chlamydia than unprotected anal intercourse in this study and a stronger association than reported in the HIM Study in the comparable category (6–10 partners in the previous 6 months). The outcome in this study was a positive test for chlamydia from any site, including pharyngeal, differing to the HIM Study, which specifically reported on urethral chlamydia. It has previously been reported that a proportion of urethral infections may be attributable to pharyngeal chlamydia.18 Oral sex may be occurring with anal sex partners and the capturing of pharyngeal chlamydia in this study may explain why the number of anal sex partners more sensitively predicted a positive test for chlamydia. There may also be the potential for reporting bias of sexual risk practices to clinicians by participants in this study and the potential for under-reporting practices like inconsistent condom use.19

Participants who reported sex work were at a 60% higher risk of chlamydia infection. This result contrasts with other Australian findings that report a lower risk for chlamydia infection in females who report sex work.20 Recent findings from a Melbourne study showed a low chlamydia incidence among female sex workers and infections were almost entirely due to transmissions from private partners.21 There are a number of potential explanations for the association of reported sex work with chlamydia incidence among MSM in this study. While research findings suggest that MSM sex workers tend to practice safe sex with their clients,22–25 they also tend to report more frequent engagement in sexual risk behaviours with non-clients26 ,27 and increased illicit drug use compared with other MSM.28–31 Others have suggested male sex workers engage in more risky sexual behaviour in their private sex lives and have recommended the targeting of this population for HIV prevention with a priority given to sex lives and relationships in general, in addition to their status as sex workers.32 In the VPCNSS data set, sex workers who presented at the metropolitan sexual health clinic (n=121) were asked specifically about their non-paying partners and reported a median number of partners of 5.0 from April 2006 to June 2010. These data suggest that among these MSM sex workers, a core group reported a high number of private partners and these specific MSM may be driving the high overall incidence in sex workers in this study. This finding is supported by the significant association of high numbers of partners with chlamydia incidence, rather than sexual risk practices per se. In addition, similar to HIV-positive MSM, MSM reporting sex work tested more frequently for chlamydia. The mean number of chlamydia tests among MSM reporting sex work was more than twice that of MSM who did not report sex work (9.4 vs 3.5), while the proportion of sex workers testing positive was approximately 40% lower than non-sex workers.

The high chlamydia incidence rate among MSM in this study, alongside the facilitative role of chlamydia and other STIs for HIV transmission, justify the inclusion of MSM as a priority population for reducing chlamydia incidence in national STI strategies6 and clinical guidelines recommending high-frequency STI screening for MSM.33 Although chlamydia incidence was higher among STI symptomatic MSM, incidence was also high in asymptomatic men, similar to a recent study of HIV-positive MSM.34 More than one in 10 HIV-positive MSM who were asymptomatic at the time of testing in this study returned a positive test for chlamydia. Again, these findings support both general clinical guidelines35 and those targeting MSM33 that recommend chlamydia testing regardless of symptoms.

The VPCNSS is unique in the context of Australian STI surveillance and provides the opportunity to link testing, clinical, demographic and behavioural data from at-risk populations, allowing the calculation of incidence and the identification of risk factors for STIs. The VPCNSS is ongoing, and future analysis will allow comparisons of chlamydia incidence estimates over time, providing a valuable tool to evaluate and inform targeted STI prevention initiatives. As more data become available, future analysis should also compare HIV and other STI incidence to examine STI infection history and assess the impact of bacterial STIs on HIV acquisition and transmission.

Key messages

  • Chlamydia notifications in Australia have increased significantly in recent years and men who have sex with men (MSM) are a priority population for reducing chlamydia incidence.

  • In this study, overall chlamydia incidence in MSM was 10.6/100 person-years (95% CI 10.0 to 11.2) and higher than other risk populations in Victoria, Australia.

  • MSM populations at higher risk of chlamydia infection in this study included younger MSM, HIV-positive MSM and self-identified sex workers.

  • The results show that MSM represent a key risk population for chlamydia in Australia and a priority for public health and clinical prevention practices.


The authors would like to gratefully acknowledge the patients attending sentinel sites, the ongoing contribution of the surveillance officers at the Burnet Institute and notifying medical practitioners. The collection of data for the VCPNSS is made possible by the ongoing participation and contribution of sentinel sites and the laboratories that undertake pathology services for these clinics. The authors would like to acknowledge the NHMRC who provide funding to MEH as a senior research fellow. The authors gratefully acknowledge the contribution to this work of the Victorian Operational Infrastructure Support Program.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Funding Victorian Government Department of Health; NHMRC provide funding to MEH.

  • Competing interests None.

  • Ethics approval Department of Health Human Research Ethics Committee; Department of Justice, Human Research Ethics Committee; Family Planning Victoria, Human Research Ethics Committee; The Royal Children's Hospital, Melbourne, Human Research Ethics Committee; The Royal Women's Hospital, Human Research Ethics Committee; The Alfred, Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement All authors, external and internal had full access to all of the data in this study and take responsibility for the integrity of that data and the accuracy of the data analysis.

Request Permissions

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.