Article Text

Download PDFPDF

When the presumption of innocence is not beneficial: periodic presumptive treatment for STIs and HIV in female sex workers
  1. John M Murray1,2,
  2. Andrew Vallely3,
  3. Andrew Page4
  1. 1School of Mathematics and Statistics, University of New South Wales, Sydney, Australia
  2. 2National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Darlinghurst, Australia
  3. 3Australian Centre for International and Tropical Health, School of Population Health, University of Queensland, Herston, Brisbane, Australia
  4. 4Discipline of Epidemiology and Biostatistics, School of Population Health, University of Queensland, Herston, Brisbane, Australia
  1. Correspondence to Dr John M Murray, School of Mathematics and Statistics, University of New South Wales, Sydney, NSW 2052, Australia; j.murray{at}

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.

The recent announcement that a prime-boost combination of two HIV vaccine candidates had a partial protective efficacy of around 31% in phase III clinical trials in Thailand is encouraging,1 2 but for the foreseeable future the prevention of HIV acquisition will continue to require a multifactorial response.3 In resource-poor countries where heterosexual transmission is the dominant mode of infection, the promotion of evidence-based behavioural and biomedical prevention strategies that effectively reach vulnerable individuals most at risk of HIV and sexually transmitted infections are paramount. Female sex workers (FSWs) and their clients may represent a significant source of new HIV infections, particularly in the early stages of the epidemic.4 5 Although 100% condom usage is the gold standard for HIV and STI risk reduction for FSWs and their clients, condoms are not always available, nor are they always used, due to women's limited ability to negotiate safer sexual practices in commercial or transactional relationships.6

Sexually transmitted infections (STI) such as Neisseria gonorrhoeae (Ng) and Chlamydia trachomatis (Ct) can increase cervico-vaginal HIV shedding in women, while Ng results in greater shedding in the semen of men infected with HIV.7 This can lead to a higher likelihood of HIV transmission with an RR of about 3 for a number of STIs in both men and women.7 The presence of an STI in an HIV-uninfected partner can also increase their risk of HIV acquisition,8 as observed in a recent herpes simplex virus type-2 (HSV-2) suppressive therapy trial among female bar workers in Tanzania where HIV acquisition was associated with recent gonorrhoea infection (HR=3.62, 95% CI 1.62 to 8.08)9. The treatment of STIs in FSWs should theoretically therefore provide some measure of protection against HIV transmission and acquisition among FSWs and their clients.

The true measure of benefit of an intervention targeting STIs in FSWs as a means to reduce HIV incidence entails determining incidence not only among FSWs but also among their clients and possibly their partners. Randomised controlled trials of this type of intervention have been conducted in the past; however, the measure of effectiveness has been limited to the FSWs involved in the trial and not any associated groups.10 Moreover there are other questions of interest associated with the optimal implementation of such programmes. In this complex environment, mathematical modelling can play a useful role.

In this edition, Vickerman et al investigate the effectiveness and optimal implementation of periodic presumptive treatment (PPT) for STIs in FSWs (see page 163). Their assessment covers a number of situations and is based on data from randomised clinical trials and pre- and postintervention analyses in Johannesburg and Free State in South Africa, Cotonou in Benin and Laos. Previous mathematical modelling has indicated the benefits of PPT compared with syndromic treatment where STI prevalences are high, given the asymptomatic nature of many of the infections and the inability to deliver treatment at the onset of each infection.10 The modelling of Vickerman et al reproduces the significant reduction in Ng and Ct prevalence in these settings, as well as in lowering levels of infection with Haemophilus ducreyi (Hd) in Free State.

Of particular interest is how the optimal coverage and timing of azithromycin-based PPT varies with the setting. The administration of PPT at a frequency greater than once a month appears to add little benefit in reducing STIs and, correspondingly, HIV incidence. Moreover, in a region such as Free State where Hd acts as a high-risk cofactor for HIV acquisition, and where azithromycin provides extended protection against reinfection, PPT remains effective if administered less frequently. Similarly, PPT in Laos requires less frequent administration due to the relatively low HIV prevalence. Also of interest is the extent of PPT coverage required to achieve maximum effect. PPT directed to Ct in an area such as Johannesburg requires around 75% coverage, and for Free State, where Hd is the dominant cofactor, only 50% coverage is needed. In each of these cases, the optimal frequency is a complicated calculation dependent on the dominant STI, their HIV cofactors and the efficacy of the PPT.

The authors raise a note of caution, suggesting that the impact of PPT may have been overestimated due to the model's assumption of homogeneous risk behaviour and patterns of sexual mixing. Incorporating data derived from sexual and social network analyses may be a valuable direction for future generations of HIV/STI projection models.

Mathematical modelling suggests that the effectiveness of PPT in reducing HIV incidence among FSWs can exceed 40% in some settings. Although these reductions can be small in absolute number where HIV prevalence is relatively low,11 these findings suggest that where a significant proportion of new HIV infections in the general population continue to be generated in vulnerable groups such as FSWs,5 PPT may have a significant role to play in improving sexual health both among so-called ‘core’ and ‘bridging groups’ and in the wider population at risk. The effectiveness of both syndromic management and PPT among FSWs and their clients in terms of population level impact is likely to be modulated by HSV-2 both in vulnerable groups and in the broader population. As Cherisch and Rees point out, up to 50% of new HIV infections in populations that have a high prevalence of HSV-2 could be attributed to HSV-2 enhancement of susceptibility.12 In future estimates of the potential impact of targeted PPT on broader population level outcomes, it will therefore be vital for mathematical models to incorporate not only information on sexual mixing patterns but also robust data on the epidemiology of key STI cofactors, particularly HSV-2, in different subgroups within a given population.

The authors' recommendation that PPT be implemented in settings where FSWs have high rates of STIs and HIV but limited access to STI treatment is relevant in a variety of settings outside the study population areas of Asia and sub-Saharan Africa. For example, in Papua New Guinea, high prevalences of chlamydia (19.4–32%), gonorrhoea (21–36%), trichomoniasis (22–33%), syphilis (24–33%) and HIV (7–13%) have been reported among sex workers in Port Moresby, Lae and Goroka.13 14 Investigating the effectiveness of PPT through a combination of highly focused acceptability studies, supported by locally relevant, robust mathematical projection modelling, represents an important opportunity to inform the development of comprehensive evidence-based national HIV prevention strategies in such settings.



  • Linked articles 034678.

  • Competing interests None.

  • Patient consent Obtained.

  • Provenance and peer review Commissioned; not externally peer reviewed.

Linked Articles