Article Text

Download PDFPDF

Short report
Mycoplasma genitalium acquisition and macrolide resistance after initiation of HIV pre-exposure prophylaxis in men who have sex with men
  1. Jens Tomas Van Praet1,2,
  2. Sanne Steyaert1,
  3. Stefaan Vandecasteele1,
  4. Barbara Van Den Bergh1,
  5. Hilde Mahieu1,
  6. Stefanie De Buyser3,
  7. Veerle Grootaert4,
  8. Patrick Descheemaeker5,
  9. Marijke Reynders5
  1. 1 Nephrology and Infectious Diseases, AZ Sint-Jan AV, Brugge, Belgium
  2. 2 Internal Medicine and Pediatrics, Ghent University, Gent, Belgium
  3. 3 Biostatistics Unit, Ghent University, Gent, Belgium
  4. 4 Pharmacy, AZ Sint-Jan AV, Brugge, Belgium
  5. 5 Laboratory Medicine, AZ Sint-Jan AV, Brugge, Belgium
  1. Correspondence to Dr Jens Tomas Van Praet, Nephrology and Infectious Diseases, Algemeen Ziekenhuis Sint-Jan Brugge, 8000 Brugge, Belgium; jens.vanpraet{at}


Objectives Recent evidence shows that patients using HIV pre-exposure prophylaxis (PrEP) have an increased rate of bacterial STIs, including syphilis, chlamydia and gonorrhoea. Our study aimed to describe the acquisition and the susceptibility for macrolides of Mycoplasma genitalium in men who have sex with men (MSM) on PrEP.

Methods We studied all MSM who started PrEP in the AZ Sint-Jan Hospital Bruges from 1 June 2017 to 31 March 2019 with at least one follow-up visit. Patients were screened for M. genitalium and other STIs with pooled rectal swabs, pharyngeal swabs and first-voided urine, and blood samples at baseline and quarterly intervals after initiating PrEP. TaqMan Array Card technology was used to detect M. genitalium and determine macrolide-resistance mediating mutations in region V of the 23S rRNA gene (A2058G, A2059G, A2058C and others). Patients with an STI were treated based on a national guideline.

Results 131 MSM (median age 40 years, range 20–79) were included in the study. The median follow-up time was 12 months (IQR 6.1–17). Baseline prevalence of M. genitalium was 6.9% and incidence rate after PrEP initiation was 28.8 per 100 person-years (95% CI 21.7 to 37.2), without significant differences in proportions between the first four quarterly intervals. All but one acquisitions were asymptomatic. Younger age and positivity for M. genitalium at baseline were significantly associated with incident M. genitalium acquisition. The observed proportion of macrolide resistance increased not significantly from 44% at baseline to 57%–86% after PrEP initiation. None of the 27 macrolide-resistant M. genitalium acquisitions could be linked to azithromycin exposure in the three preceding months.

Conclusions After initiation of PrEP, we found a stable incidence of almost exclusively asymptomatic M. genitalium. However, a non-significant trend of an increased percentage of macrolide-resistant strains was observed.

  • mycoplasma
  • HIV pre-exposure prophylaxis
  • azithromycin

Statistics from

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.


Multiple randomised controlled trials have shown the use of pre-exposure prophylaxis (PrEP) to be highly effective in the prevention of HIV transmission. As a drawback, it was recently documented that people using PrEP have an increased rate of bacterial STIs, including syphilis, chlamydia and gonorrhoea.1

Mycoplasma genitalium is an emerging bacterial STI that was already recognised as a cause of non-gonococcal urethritis in the 1980s. Further clinical studies have linked M. genitalium to cervicitis and pelvic inflammatory disease,2 but most frequently it is found in the clinical context of asymptomatic carriership.3 4 Although a prolonged 5-day course of azithromycin is proposed by the European guidelines as first-line treatment, macrolide resistance in men who have sex with men (MSM) was reported to be common in recent studies.5 Importantly, one study found that recent use of azithromycin conferred an increased risk of obtaining macrolide-resistant M. genitalium urethritis.6

As MSM taking PrEP have a high exposure to antibiotics for the treatment of STI, we aimed to describe the incidence and macrolide resistance of M. genitalium in a cohort of MSM who started PrEP.


Between 1 June 2017 and 31 March 2019, we studied all MSM who started PrEP in the AZ Sint-Jan Hospital Bruges and had at least one follow-up visit. Patients were screened for STIs at baseline and quarterly intervals after initiating PrEP. Testing for syphilis, HIV and hepatitis infections was performed by means of standard serological assays. Depending on their sexual modus, patients were instructed to collect a rectal swab, pharyngeal swab and/or first-voided urine and pool these in one container. A sample of this container was tested by a laboratory-developed highly sensitive and specific molecular screening assay based on the Taqman Array Card technology detecting Chlamydia trachomatis (including lymphogranuloma venereum (LGV) strains), Neisseria gonorrhoeae, Mycoplasma genitalium, Treponema pallidum, Haemophilus ducreyi and Trichomonas vaginalis. This method detects macrolide-resistance mediating mutations for M. genitalium in region V of the 23S rRNA gene (A2058G, A2059G, A2058C and others). Patients with an STI were treated regardless of their symptoms based on a national guideline (IGGI guide Belgium). A prolonged 5-day regimen of azithromycin was used to treat wild-type M. genitalium and a 10-day course of moxifloxacin 400 mg in case of macrolide resistance. No tests of cure were performed.

Using the ‘PropCIs’ package in R V.3.6.1, Wilson’s score 95% CI was calculated for single proportions and incidence rates and Tango’s score CI was calculated for differences in paired proportions of M. genitalium acquisition between quarterly intervals. Using SPSS V.25, several regression models were fitted for binary outcomes with a binomial distribution and logit link. Simple (unadjusted) logistic regression models were used to find associations with M. genitalium acquisition after PrEP initiation. Age at baseline (in years), positive screening for an STI at baseline and mode of PrEP use were entered as variables. A Generalised Estimating Equations (GEE) model with a robust independent covariance matrix for repeated measurements within a patient and with quarterly interval as fixed effect was used to estimate the proportion of macrolide resistance at each interval.


From 1 June 2017 until 31 March 2019, 150 patients attended the STI polyclinic of AZ Sint-Jan Hospital Bruges asking for a first PrEP prescription. Seventeen patients were left out of the analysis because there was no follow-up visit and two were omitted because they were women. A total of 131 men (median age of cohort at baseline 40 years (range, 20–79)) were included in the study. All were gay or bisexual. All except for one African patient were Caucasians. The median follow-up time was 12 months (IQR 6.1–17). In total, 109 patients (83%) used PrEP daily at their last visit.

At baseline, 11% (15/131) of the cohort tested positive for chlamydia, 9.2% (12/131) for gonorrhoea, 6.9% (9/130) for M. genitalium, 3.1% (4/131) for syphilis and 0.8% (1/130) for T. vaginalis. After PrEP initiation, the incidence rate in the cohort was 35.0 per 100 person-years (PY) for chlamydia (95% CI 27.3 to 43.6), 29.6 per 100 PY for gonorrhoea (95% CI 22.4 to 38.0), 28.8 per 100 PY for M. genitalium (95% CI 21.7 to 37.2), 14.8 per 100 PY for syphilis (95% CI 9.7 to 21.9), 1.6 per 100 PY for LGV (95% CI 0.43 to 5.5) and 0.77 per 100 PY for chancroid (95% CI 0.14 to 4.3). The proportion of M. genitalium acquisition in the first four quarters after PrEP initiation in patients without M. genitalium at baseline was 11.6%, 6.7%, 8.8% and 5.9%, respectively. The 95% CIs for the differences in these paired proportions did not exceed the range of (−14%, +14%). The 37 positive tests for M. genitalium after PrEP initiation were documented in 27 patients (21% of the cohort). Moreover, 76% were mono-infections, 8.1% were coinfections with gonorrhoea, 5.4% coinfections with chlamydia, and 5.4% coinfections with chlamydia and syphilis. In only one patient with a mono-infection, there were symptoms of urethritis.

Using simple logistic regression models, we explored factors associated with incident M. genitalium acquisition after PrEP initiation. A positive screening at baseline for M. genitalium was predictive for incident M. genitalium acquisition (odds 6.0 (95% CI 1.5 to 24.0), p=0.012). From the nine patients who were treated at baseline for M. genitalium, four tested positive again 3 months later whereas five were negative. One of these five patients had a recurrence at 6 months. A follow-up screening was available in 18 of the 22 patients who were negative at baseline and acquired M. genitalium after PrEP. Only three of these 18 tested positive again 3 months later. The age of the patients with incident M. genitalium varied between 25 and 65 years. Entering age as a continuous variable in a binary logistic regression model for the group older than 25 years, we found a decrease in the odds of 6% for every year increase in age (95% CI 0.8% to 10.5%, p=0.024).

The observed percentage of macrolide resistance in our cohort was 44% at baseline, 71% after 3 months, 57% after 6 months, 86% after 9 months and 75% after 12 months. Using GEE, we found no significant differences in the different estimated proportion of macrolide resistance between baseline and the first four quarterly follow-up visits (p>0.05) (see figure 1). At baseline, 6.8% of the cohort was exposed to azithromycin. After PrEP initiation, the percentages of exposure were 4.8%, 3.8%, 1.3% and 1.4% at the first four quarterly follow-up visits, respectively. In none of the 27 positive tests for macrolide-resistant M. genitalium documented after PrEP initiation, a patient had been exposed to azithromycin in the three preceding months.

Figure 1

Estimated proportion ±95% CI of macrolide-resistant M. genitalium strains at baseline and four quarterly follow-up visits (GEE, p=NS).


Recent evidence points to an increased rate of bacterial STI after PrEP initiation1 and a high percentage of macrolide-resistant M. genitalium strains at baseline screening of MSM patients enrolled in PrEP programmes.7 Moreover, PrEP users are frequently exposed to a single 1 g dose of azithromycin, which was recently shown to be associated with an increased risk of future macrolide-resistant M. genitalium infection.6 This exposure presumably is one of the driving factors of increasing resistance in the community.8 We therefore determined the M. genitalium incidence and resistance in our local PrEP cohort, and explored predictors of M. genitalium acquisition.

Baseline screening showed that 6.9% of our cohort was positive for M. genitalium. Other studies which screened asymptomatic HIV-negative MSM reported a prevalence between 2.1% and 10%.7 9 10 After PrEP initiation, M. genitalium was the third most frequently observed STI, with the rates of chlamydia, gonorrhoea and syphilis being comparable with another large series of patients on PrEP.1 Furthermore, we found that younger age and baseline positivity for M. genitalium were significantly associated with incident M. genitalium acquisition. Traeger et al also found that younger age was associated with a greater STI risk after PrEP commencement.1 The percentage of asymptomatic infections (97%) exceeded the number (61%) reported in a series of male STI clinic attendees tested by urethral swabs.4

The baseline resistance for macrolides was 44% in our cohort, whereas others reported resistance rates in asymptomatic screened MSM between 58% and 70%.7 10 We observed a non-significant trend of increased macrolide resistance after PrEP initiation. Since none of the patients who acquired a resistant strain were exposed to azithromycin in the preceding 3 months, we could not link the use of this antibiotic with increased resistance. With the widespread macrolide resistance in Europe, it is strongly recommended that all positive tests should be followed up with detection of macrolide-resistance mediating mutations.

Our study is the first to provide longitudinal data on M. genitalium acquisition and resistance in a cohort of PrEP users. The pooling of samples prevented us from reporting on the anatomical site of M. genitalium carriership and could have influenced the analytical test performance. The incidence of M. genitalium might be overestimated since we have not performed tests of cure. We were not able to collect information on antibiotic exposure before PrEP initiation. Finally, it should be noted that most guidelines nowadays advise against screening for M. genitalium and that treatment of asymptomatic patients is debatable.

Abstract translation

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.



  • Handling editor Federico Garcia

  • Contributors JTVP, SV and BVDB collected the patient data. MR and PD collected the STI data and supervised the laboratory testing. JTVP and SDB analysed the data. All authors contributed to the preparation of the manuscript.

  • Funding This work was supported by an unrestricted grant from Gilead Sciences.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval This study was approved by the local ethical committee (advice no. 1905).

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

  • Data availability statement Data are available upon request