Article Text

PDF

Original article
Oral and anal sex are key to sustaining gonorrhoea at endemic levels in MSM populations: a mathematical model
  1. B Hui1,
  2. C K Fairley2,3,4,
  3. M Chen2,3,4,
  4. A Grulich1,
  5. J Hocking5,
  6. G Prestage1,4,
  7. S Walker3,
  8. M Law1,
  9. D Regan1
  1. 1The Kirby Institute, UNSW Australia, Sydney, Australia
  2. 2School of Population and Global Health, University of Melbourne, Melbourne, Australia
  3. 3Melbourne Sexual Health Centre, Alfred Health, Melbourne, Australia
  4. 4Australian Research Centre in Sex, Health and Society, La Trobe University, Melbourne, Australia
  5. 5Centre for Women's Health, Gender and Society, University of Melbourne, Melbourne, Australia
  1. Correspondence to Dr Ben Hui, The Kirby Institute, UNSW Australia, Wallace Wurth Building, Sydney NSW 2052, Australia; b.hui{at}unsw.edu.au

Abstract

Objectives Despite early treatment of urethral infection, gonorrhoea is endemic in urban populations of men who have sex with men (MSM) in Australia. By contrast, gonorrhoea is not common in urban heterosexual populations. Sexual activities among MSM usually involve anal or oral sex, and as these behaviours are becoming increasingly common among heterosexuals, there is a need to investigate their roles in transmission of gonorrhoea.

Methods We developed individual-based models of transmission of gonorrhoea in MSM and heterosexuals that incorporate anatomical site-specific transmission of gonorrhoea. We estimated the probabilities of transmission for anal sex and oral sex by calibrating an MSM model against prevalence of gonorrhoea and sexual activity data. These probabilities were then applied to a heterosexual model in order to examine whether gonorrhoea can persist in a heterosexual population through the addition of anal sex and oral sex.

Results In the MSM model, gonorrhoea can persist despite prompt treatment of urethral infections. The probability of gonorrhoea persisting is reduced if use of condom for oral sex is increased to more than 15% of acts. Assuming that treatment of symptomatic infections is prompt, gonorrhoea is unlikely to persist in a heterosexual population even with the addition of anal and oral sex.

Conclusions Our models suggest that oral sex has an important role in sustaining gonorrhoea in a population of MSM by providing a pool of untreated asymptomatic infection. The importance of anal sex or oral sex in sustaining gonorrhoea in a heterosexual population remains uncertain due to the lack of information linking different types of sex acts and transmissibility.

  • GONORRHOEA
  • GAY MEN
  • MODELING
  • ORAL SEX

Statistics from Altmetric.com

Introduction

Gonorrhoea is a sexually transmitted infection (STI) caused by the bacterium Neisseria gonorrhoeae that commonly colonises the reproductive tract. It is associated with urethritis and proctitis in men, with ectopic pregnancy and tubal factor infertility in women, and increases the susceptibility and transmissibility of HIV in both men and women.1 Gonorrhoea infection of the male urethra usually results in noticeable symptoms and infected men tend to seek treatment promptly.1 ,2 In urban Australia, treatment for gonorrhoea is readily available and has effectively reduced the average duration of symptomatic urethral infection to a few days (personal communication with CK Fairley, based on a retrospective audit of 200 men diagnosed with urethral gonorrhoea at Melbourne Sexual Health Centre in 2014; unpublished). Furthermore, with at least 60% of urban populations of men who have sex with men (MSM) being tested annually,3 and untreated urethral infections persisting for only around 15–26 weeks,4 ,5 most gonorrhoea infections can be expected to resolve within considerably <1 year. With the relatively short period of infectivity (compared with other bacterial STIs such as chlamydia), it should be difficult for gonorrhoea to remain prevalent in the population. However, the gonorrhoea positivity among urban Australian MSM has remained high and stable for at least a decade and has even increased in recent years.6 ,7 While the prevalence of urethral gonorrhoea is generally not measured in surveillance studies due to the high likelihood of early treatment, prevalence of gonorrhoea at rectal and pharyngeal sites can be as high as 4%.8 Given that gonorrhoea infections at rectal and pharyngeal sites are usually asymptomatic, it is possible that a large proportion of rectal and pharyngeal infections remain untreated, allowing gonorrhoea to persist in the population despite the high rate of symptom-driven treatment.

Gonorrhoea is known to be transmitted through penile–vaginal sex, anal sex and from the male urethra to the pharynx through oral sex (fellatio).2 ,9 Transmission through cunnilingus appears to be less common, but studies have suggested an increased chance of pharyngeal gonorrhoea infection in heterosexual men who perform cunnilingus.10 The number of rectal and pharyngeal infections in heterosexuals is expected to increase as the number of heterosexuals who ever experienced anal sex and fellatio increases.11 However, quantitative evaluation of the role of rectal and pharyngeal infections in transmission of gonorrhoea is difficult because the probabilities of transmission for specific sex acts, and between different anatomical sites in men and women, are largely unknown.

By contrast with urban MSM populations in Australia, diagnosis and notification rates of gonorrhoea in urban heterosexual populations are below endemic levels.6 ,12 This can likely be attributed to differences in partner acquisition rates and sexual behaviours between the two population groups. Previous modelling studies of transmission of gonorrhoea, such as that conducted by Garnett et al,4 have established the role of partner acquisition in the persistence of STIs at a population level through metrics such as the basic reproduction number.13 Such studies, however, have generally neglected the role of different modes of transmission due to the added complexity this involves and a lack of data on site-specific transmissibility. In this study, therefore, we have used a mathematical transmission model to examine the role of anal sex and fellatio in the persistence of gonorrhoea in an urban MSM population.

The most recent survey of sexual behaviour conducted in Australia indicated a slight decrease in the frequency of penile–vaginal sex within regular heterosexual partnerships, an increase in the number of individuals who have ever experienced anal sex and an increase in the frequency of fellatio.14 However, the implications of these trends for incidence and prevalence of STIs have not been studied. In this study, we also developed a model of heterosexual gonorrhoea transmission with an analogous structure to the MSM model in order to assess whether gonorrhoea can persist in a general heterosexual population through the inclusion of anal sex and fellatio. We assumed that the transmission probabilities for anal sex and fellatio among heterosexuals are the same as for these acts among MSM. We then use the heterosexual model to explore the likelihood of gonorrhoea persisting in an urban heterosexual population under various assumptions, including an increase in the frequency of anal sex and fellatio.

Methods

A summary of the modelling methodology follows. A more detailed description of the model is provided in the online supplementary technical appendix.

For each member of the population, we assume that there are three sites of infection: urethral, rectal and pharyngeal. We assume that infections are localised, and the infection status at one site has no effect on the natural history or susceptibility and transmissibility of infection at the other sites. The infection parameters are listed in the online supplementary table A1 in the technical appendix.

MSM model

We developed an individual-based mathematical model of gonorrhoea transmission in a sexually active population of MSM. The model tracks sexual activity and status of infection on a daily basis. The sexual behaviour in the MSM model is based on published modelling studies and reports on MSM in Australia.3 ,7 ,15–18 Online supplementary table A2 in the technical appendix lists the behavioural parameters used in the model.

We assume that gonorrhoea can be transmitted bidirectionally between urethral and rectal sites through anal sex19 and between urethral and pharyngeal sites through fellatio.10 Apart from anal sex and fellatio, oral–anal sex acts (such as anilingus or ‘rimming’) are also known to be common among MSM,18 and it has been speculated that transmission from the pharynx to the rectum may also occur.10 We, therefore, also include the case where transmission from pharynx to rectum can occur directly without urethral infection. We assume oral–anal sex acts occur in 23% of casual partnerships, with a frequency of 1–3 acts per partnership, based on the frequency of rimming reported in Phang et al.18 For simplicity, we assume there are no other transmission pathways.

Heterosexual model

For comparison, a model was also developed to investigate the transmission of gonorrhoea in a hypothetical heterosexual population. We assume that this population has equal numbers of men and women, and only heterosexual partnerships are allowed.

In addition to anal sex and fellatio, gonorrhoea can also be transmitted bidirectionally between men and women through penile–vaginal sex.20 The parameters of infection for this model are listed as italicised entries in the online supplementary table A1 in the technical appendix, while the behavioural parameters used in the heterosexual model are listed in the online supplementary table A3 in the technical appendix.

Results

MSM model

Gonorrhoea can persist in the modelled MSM population when the partner acquisition rate and sexual behaviour (which includes permitted partnership type and maximum number of partners that can be sought in 6 months) is calibrated to the published data.3 ,15 ,18 ,21 The median estimates of prevalence at equilibrium (defined as the average prevalence over the last 10 years of simulations run to 60 years) are summarised in table 1, and the probabilities of transmission inferred through model calibration are summarised in table 2. The modelled estimates of prevalence of gonorrhoea at the rectal and pharyngeal sites correspond well with observed estimates of prevalence. In order to calibrate the model such that the prevalence of rectal and pharyngeal infection can be maintained in the vicinity of 4% as reported, prevalence at the urethral site has to be more than 1%.

Table 1

The prevalence at equilibrium for all sites in the men who have sex with men model

Table 2

Calibrated transmission probability per unprotected sex act

The inclusion of pharyngeal-to-rectal transmission (eg, through the practice of rimming) allows some transmission to occur without involvement of the urethral site. However, with <23% of casual partnerships engaging in rimming, and with less than 1–3 acts per casual partnership, the influence of rimming on urethral prevalence in this population is limited.

Prevalence can be reduced by increasing usage of condoms during anal sex. However, as the results in table 3 show, even if condom usage for anal sex is increased to 100% (thereby effectively eliminating transmission through anal sex, assuming 100% prophylactic efficacy of condoms) gonorrhoea cannot be completely eliminated in the model. By contrast, the results in table 4 show that increasing condom usage for fellatio, while maintaining existing condom usage during anal sex, can lead to a large reduction in prevalence of gonorrhoea. Gonorrhoea does not persist in most (85%) simulation runs if condoms are used in 20% of acts of fellatio. At condom usage of >25% for acts of fellatio, gonorrhoea cannot persist in the MSM model at all.

Table 3

Gonorrhoea prevalence at equilibrium for all sites in the men who have sex with men model under increased condom usage for anal sex

Table 4

The prevalence at equilibrium for all sites in the men who have sex with men model, under increased condom usage for oral sex

Heterosexual model

Even if the frequency of fellatio and non-fellatio (ie, penile–vaginal sex and anal sex) within partnerships are assumed to be the same as for the MSM model, gonorrhoea cannot persist in the modelled heterosexual population unless the probability of transmission for cunnilingus is at least 15% of the value assumed for fellatio (ie, probability of transmission from vagina to pharynx >0.0943 and from infected pharynx to vagina >0.0130). Gonorrhoea persists in the modelled population in <1.5% of simulation runs if the probability of transmission for cunnilingus is increased to 25% of the value assumed for fellatio (ie, probability of transmission from vagina to pharynx of 0.1572, and from pharynx to vagina of 0.0216). However, a clear link between pharyngeal infection and cunnilingus remains unclear due to the small number of reported cases and, therefore, this level of probability of transmission cannot be substantiated.10

If the probabilities of bidirectional transmission for cunnilingus are again fixed to 15% of the value assumed for fellatio, gonorrhoea can persist in the modelled heterosexual population if every act of penile–vaginal sex is accompanied by an act of anal sex, although the likelihood of persistence remains very low (occurs in 2% of simulation runs, with prevalence well below 0.05%). Likewise, increasing the frequency of fellatio also increases the likelihood of gonorrhoea persisting in the modelled population. Gonorrhoea cannot persist unless the per-encounter probability of fellatio exceeds 0.5, in which case there is a small probability of gonorrhoea persisting in the population (persistence occurs in 4% of simulation runs, with prevalence below 0.6%). Note also that the per-encounter probability of fellatio used (>0.5) is already higher than the value of 0.3 quoted in the most recent surveys of sexual behaviour in Australian heterosexual populations.11 ,14

Alternatively, increasing the duration of symptomatic infection also allows gonorrhoea to persist in the modelled heterosexual population. Gonorrhoea persisted in 50% of simulation runs if the average duration of symptomatic urethral infection in men is increased to 3.6 days. While this may be plausible (heterosexual men may, eg, be slower in seeking treatment due to lesser awareness regarding STIs and their symptoms than MSM), data are not available to support this.

Discussion

Gonorrhoea infection can occur at urethral, rectal or pharyngeal sites. To our knowledge, previous gonorrhoea modelling studies have not considered site-specific infection, most likely because of the increased complexity required of models to capture this and the additional pathways of transmission, and a lack of data on site-specific probabilities of transmission. The individual-based model developed for this study allows transmission between anatomical sites to be captured and provides insights to the importance of different transmission pathways on persistence and prevalence of gonorrhoea in different populations having different sexual behaviour.

Our study shows that gonorrhoea can persist in an MSM population despite the short average duration of symptomatic urethral infection brought about by prompt treatment. The inclusion of rectal and pharyngeal infections, which are usually asymptomatic, allows a reservoir of untreated infection to be established, thereby increasing the likelihood of persistence of gonorrhoea. On the other hand, gonorrhoea does not persist in our modelled heterosexual population even if anal sex and fellatio are included. Without changes to assumptions regarding sexual behaviour, such as increasing the number of sexual partners, gonorrhoea can only persist in the heterosexual population if transmission probabilities or the duration of urethral infection are increased to values that are likely unrealistic.

Furthermore, the model suggests that fellatio has a greater influence on persistence of gonorrhoea than anal sex, even if the per-act probability of transmission for fellatio is lower than for anal sex. This is largely due to the high frequency of fellatio compared with that of anal sex between casual partners, and the assumption that condoms are not used during fellatio.

To reduce prevalence of gonorrhoea at rectal and pharyngeal sites, the most direct approach is to reduce the duration of rectal and pharyngeal infection, for example, by increasing screening coverage for rectal and pharyngeal gonorrhoea. An alternative approach is to reduce transmission to and from rectal and pharyngeal sites. Our results suggest that prevalence of gonorrhoea at all sites can be reduced by increasing condom usage during anal sex as well as fellatio. However, achieving a substantial increase in condom usage for fellatio may be difficult given that fellatio is usually perceived to be safe,22 ,23 and healthcare practitioners may be less enthusiastic about promoting condom use for fellatio when compared with anal sex.24

In the MSM model, prevalence of gonorrhoea can be sustained at observed levels at the rectal and pharyngeal sites, but is possibly overestimated for the urethral site. While the true prevalence of urethral infection for the urban MSM population in Australia is not known, it is unlikely to be as high as 1.5% if most urethral infections are treated within a few days. This discrepancy is likely due to limitations in model structure and assumptions. In the model, transmission occurs primarily through anal sex and fellatio, and the urethral site is always either the source or target of infection. It is known that MSM do engage in sexual practices other than anal sex and fellatio, and many of these practices have the potential to be pathways for transmission of gonorrhoea.18 As the results in table 1 demonstrate, the inclusion of urethral–rectal and urethral–pharyngeal pathways of transmission, such as that may occur through the practice of rimming, has the potential to reduce the prevalence of urethral infection required to sustain prevalence of rectal and pharyngeal infection. However, more detailed data on the frequency and associated probabilities of transmission of urethral–rectal and urethral–pharyngeal transmission pathways are needed before the importance of these can be established.

By contrast with the MSM model, gonorrhoea cannot be sustained at an endemic level in our modelled heterosexual population without imposing additional (and likely unrealistic) assumptions on probabilities of transmission or aspects of sexual behaviour. While gonorrhoea can be made to persist in the heterosexual model by increasing the frequency of anal sex or fellatio, the likelihood of persistence remains small (occurs in <5% of simulation runs) and the resulting prevalence of well <1% is likely to be too low for gonorrhoea to persist long-term in a real population. This finding is consistent with the observed absence of endemic infection of gonorrhoea in urban heterosexual populations in Australia. However, more detailed data, such as the duration of untreated gonorrhoea infection in heterosexuals and the possible changes to sexual behaviour (eg, frequency of sex, condom usage and sexual repertoire) over long-term partnerships is required to substantiate this finding.

Summary data from the latest survey of sexual behaviour conducted in Australia suggest that the partner acquisition rate for heterosexuals has remained considerably lower than for MSM, with a median of eight lifetime partners for heterosexual men versus 22 for MSM.14 ,25 This latest finding is consistent with our assumption that partner acquisition rates of heterosexual men are lower than those of MSM, and would not change the conclusions drawn from our study, as persistence of gonorrhoea in the modelled heterosexual population is unlikely even if the high partner acquisition rate assumed for the MSM model is applied to the heterosexual model. The survey also found that anal sex remains rare among heterosexual partnerships, that there has been a slight decrease in the frequency of penile–vaginal sex and only a slight increase in the frequency of fellatio in regular partnerships.14 If our assumptions regarding sex-act-specific transmissibility (in particular, the assumption that transmissibility for cunnilingus is much lower than for fellatio) are reasonable, then the latest data on sexual behaviour in Australia should have minimal impact on the conclusions drawn from our study.11 ,14 ,25 ,26

The aim of this study was to investigate the impact of site-specific infection on transmission of gonorrhoea and persistence. Due to a lack of data on site-specific probabilities of transmission, we relied on the model calibration process to infer values for most of these parameters. The study by Bissessor et al8 found a significant difference in bacterial load across infection sites, and high bacterial load is thought to be associated with increased transmissibility of STIs and symptomatic infection,27 but a link has not been firmly established for gonorrhoea. Given that urethral prevalence of infection is assumed to be much lower than the prevalence at non-urethral sites, the calibration process suggests that transmission to the urethral site must occur with much lower probability than transmission from the urethra. However, low urethral prevalence of infection can also be caused by the shortening of the duration of infection through treatment, such that the true probability of transmission to the urethra cannot be ascertained by calibration to prevalence alone. This uncertainty remains in the absence of accurate site-specific data on incidence and duration of infection for model calibration.

We make a number of necessary simplifying assumptions in our model. We assume that all symptomatic infections will be treated and that treatment is effective for infection at all anatomic sites. In reality, however, not all urethral infections are symptomatic or treated promptly, and some treatment failures are known to occur, especially for pharyngeal infections.28 These assumptions mean that we likely underestimate the duration of infection and, thereby, overestimate the probabilities of transmission required to sustain prevalence of gonorrhoea at the observed levels.

In this study, the assumed duration of asymptomatic infection at non-urethral sites is based on a study from an urban area where screening and treatment is already available.29 The probabilities of transmission required to sustain the observed prevalence will likely change should the rate of asymptomatic treatment be different. For example, the probabilities of transmission required to maintain prevalence will be higher if asymptomatic infection is detected and treated more promptly (eg, through increased screening coverage or frequency, and partner notification). The probabilities of transmission required will also increase if the appearance of symptoms leads to changes in sexual behaviour, such as abstinence or increased condom usage.30

Key messages

  • Untreated extragenital infections have an important role in sustaining gonorrhoea in an urban population of men who have sex with men (MSM).

  • The impact of oral sex is notable due to its common occurrence and the likelihood that condoms will not be used.

  • Increasing the use of condoms for oral sex could lead to a substantial reduction in the prevalence of gonorrhoea among MSM.

  • Gonorrhoea is unlikely to persist in a heterosexual population with access to treatment unless there is an increase in partner acquisition rates or the frequency of sex.

References

View Abstract
  • 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:

Footnotes

  • Handling editor Jackie Cassell

  • Contributors BH and DR collaborated in the writing of the manuscript. BH was also involved in the design and implementation of the model. CKF, MC, SW and ML were involved in providing laboratory data and expert opinions on transmission pathways. AG, GP and JH were involved in providing expert options on sexual behaviours. All coauthors revised the manuscript before submission.

  • Funding This study was funded, in part, by National Health and Medical Research Council Project Grants Program Grant 568971. The Kirby Institute is funded by the Australian Government Department of Health and Ageing, and is affiliated with the Faculty of Medicine, University of New South Wales. The views expressed in this publication do not necessarily represent the position of the Australian Government.

  • Competing interests None.

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

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.