What is known about the occurrence of sexually transmitted infections (STI) in women who have sex with women (WSW), and should it matter? Demographically, this is not a trivial issue: estimates of lifetime same sex behaviour among women range from 8% to 20%, and between 1.4% and 4.3% of all women may be WSW on the basis either of behaviour or self defined identity.^1–3 WSW have traditionally been viewed as “low risk” for STI, including HIV, and data from several small studies seem to support this belief.^4–9 However, as is often the case when one attempts to categorise any group by a descriptive “measurement” as complex as sexual behaviour, the real situation is of course more complicated.

As several studies have reported, the sexual practices of WSW present a reasonable means for vagina to vagina transmission of infected cervicovaginal secretions,^10–12 a concept most directly supported by documentation of trichomoniasis being sexually transmitted between women.¹³ There is strong evidence that transmission of human papillomavirus (HPV) between female sex partners occurs, as HPV and associated squamous intraepithelial lesions (SIL) have been detected in WSW who reported no previous sex with men.^10,14,15 Bacterial vaginosis (BV), a condition associated with pelvic inflammatory disease and adverse outcomes of pregnancy,¹⁶ occurs in 24% to 51% of WSW,^5,9,12,17,18 and sexual transmission of some responsible factor has been debated.^12,17 Although uncommon, transmission of HIV and hepatitis B between female sex partners has been reported.^19–23 Many WSW are also at risk for STI acquisition from male partners. Even when surveyed outside STI clinic settings, most WSW report having had sex with men, and many (20–30%) continue to have sex with men as well as women.²⁴ Female adolescents who have sex with other females may be especially likely to engage in unprotected sex with both male and female partners.²⁵

In this issue of Sexually Transmitted Infections, Fethers and colleagues report (p 345) a case-control study of 1432 WSW attending a public STI and HIV clinic in Sydney for routine sexual health screen.²⁶ Almost one in 10 women attending these clinics were WSW; only two thirds of these were screened for STI. Why those unscreened sought care at the clinic, and what determinants went into the decision not to screen them, is not described. More common among WSW relative to matched heterosexual controls were BV, previous diagnosis of STI, and seropositivity to hepatitis B and C; less common was a report of previous genital warts. Equally prevalent were gonorrhoea, chlamydia, HIV, and Papanicolaou smear evidence of SIL. Trichomonas vaginalis, Chlamydia trachomatis, and HIV infections were detected in WSW reporting sex exclusively with women in the previous year. WSW more commonly reported having had sex with men who have sex with men (MSM) and injecting drug users (IDU), as well as a higher number of lifetime partners, ever having exchanged sex for money, and being IDU themselves. As the authors note, the prevalence of BV (8%) was low relative to that seen in other studies of WSW, possibly because one third of WSW attending the clinic were not screened for BV.

This report is important for two major reasons. Firstly, while case-control studies have their own set of limitations, they offer one approach to circumvent some of the methodological challenges inherent in studying WSW. As eloquently reviewed in a discussion on research in sexual minorities by a multidisciplinary task force report on lesbian, “gay,” bisexual, and transgendered health,²⁷ the methodological considerations underlying population selection, subject sampling, and recruitment are complex. Precisely defining a reproducible study group, while simultaneously acknowledging and accounting for the inherent heterogeneity in most populations, is especially challenging when studying sexual behaviour and its attendant risks and outcomes. Well controlled studies in reproducible populations can surmount some of these challenges; the case-control methodology employed by Fethers and colleagues in studying their STI clinic population is a good example.

Secondly, and more importantly, this is now the third study to show an alarming prevalence of HIV related risk behaviours in WSW who report sex with men during a visit to an STI clinic. These risks include sex with homosexual or bisexual men, use of injection drugs and of crack cocaine, and exchange of sex for drugs or money.^28,29 In our analysis of WSW attending our STI clinic in Seattle, Washington, WSW who reported sex with both male and female partners in the preceding 2 months had a high prevalence of risk behaviour for HIV acquisition, and women reporting sex only with other women in the preceding 2 months more commonly reported ever having had sex with a homosexual or bisexual man.²⁹ Further, among 550 WSW queried in a community survey done in San Francisco, 40% reported unprotected vaginal or anal sex with men during the past 3 years, including men who had sex with men and male IDU.³⁰ A large, cross sectional, community based study of sociometric networks among IDU found that same sex behaviour among women was independently associated with a twofold increase in the likelihood of HIV infection.³¹ Thus, the proportion of HIV infected women who have sex with women may be substantial, while use of barrier methods to prevent STI transmission in such encounters may be low.³² In summary, these WSW could theoretically function as a “bridge” population: one with sexual links to men (possibly men at higher risk for STI/HIV by virtue of either being IDU or having sex with other men) and to WSW who are sexually active only with women. Such exposures could prove significant for WSW who may be classified as “low risk” when they later report exclusive same sex behaviour.

Despite this provocative and diverse evidence, we have only a limited understanding of the frequency and range of behaviours that put WSW at risk for STI acquisition, and no generalisable estimates of the epidemiology of STI in WSW. Why don't we know more? Attempts to use national or local surveillance data to estimate the prevalence of STI among WSW are limited in that many risk classifications have either excluded same gender sex among women or subsumed it under a hierarchy of other behaviours viewed as conferring greater risk.³³ Few STI reporting systems routinely collect information on same sex behaviour among women. The situation is further complicated by the question of whether sexual minority women receive appropriate preventive health care,^34,35 and thus are even likely to access systems which might capture incident STI or consequent syndromes. For example, WSW probably do not receive Papanicolaou smear screening according to recommended guidelines; whether this relates to erroneous beliefs about personal risk for HPV and cervical cancer, or because providers do not appropriately screen WSW, is not known.^10,36,37

What are the challenges and tasks for those working in this area and those responsible for directing STI research and policy? As noted by Fethers and colleagues, traditional assumptions about the sexual practices between women have generally implied that such activities confer low or no risk. Such generalisations are often made categorically, without specific knowledge of sexual practices. At best, they are informed by disinterest, or by a lack of willingness to believe that the area is worthy of further study; at worst, homophobia and sexism contribute to these views. Such premature conclusions about STI risk among WSW adversely affect our ability to gather data that would illuminate our lack of knowledge. This “early closure” is deleterious not only for research directly related to STI in WSW; in large studies of more “traditional” health outcomes, such as breast cancer and heart disease, data on sexual orientation are not routinely collected despite the importance of measurements related to women's reproductive health histories. A tendency to dismiss the value of studying STI in WSW may certainly adversely affect the likelihood that research proposals are funded, and that research findings are published in journals with relatively wide readership in the scientific community.

Why study STI in WSW? Reasons for testing any scientific hypothesis should not be because it is politically mandated, because it is politically expedient, or because “it hasn't been studied.” Research and funding priorities should be based on sound hypotheses and solid data. In the United States, the Institute of Medicine report on research priorities for lesbian health confirmed the need for more extensive data on sexual practices and healthcare seeking behaviours that put women who have sex with women at risk for STI.³⁸ The work of Fethers and colleagues adds to the growing body of evidence that supports this line of scientific inquiry. It also reminds us that our attempts to categorise people by “sexual behaviour” into risk groups that neatly predict STI epidemiology and transmission must account for the complexity and subtleties that characterise human sexual behaviour.