METHOD

Ethical permission was obtained, and all new patients at two lesbian sexual health clinics in London were offered a questionnaire to gather data about potential risk factors for STI: demographic features, sexual history, and sexual practice. Between 1992 and 1995, 803 patients completed questionnaires out of an estimated 1000 attenders. No record of non-responders was kept. A total of 708 out of 803 had a GU examination and results from this were linked with questionnaire responses. This study includes 199 women in an initial prevalence study of STI in WSW³; we present here the results from the complete study population focusing on sexual history and other risk factors for STI.

A GU examination comprised cervical and urethral Gram stained smear and culture for Neisseria gonorrhoeae, cervical swab for Chlamydia trachomatis by ELISA, and high vaginal samples for Trichomonas vaginalis by wet preparation. Gram stained vaginal preparations were taken for Candida sp identification and culture. Bacterial vaginosis (BV) was diagnosed using Amsel’s criteria. Screening for treponemal disease used the Venereal Disease Research Laboratory (VDRL) test and the Treponema pallidum haemagglutination assay (TPHA). Herpes simplex cultures were performed as clinically indicated and genital warts were diagnosed by visual inspection. Pelvic inflammatory disease (PID) was diagnosed on clinical examination with a raised white cell count.

Statistical analysis

Diagnosis of one or more STIs (excluding candida and BV) in women who had had a GU screen was the outcome of interest. Variables reaching statistical significance at the level of p<0.1 in univariate analysis were included in a multivariate model, as well as variables related to STI in other studies. A logistic regression model was built up in a stepwise fashion, using the likelihood ratio test and observing change in odds ratios.

RESULTS

Description of the sample (n = 708)

The majority of the sample (92%) identified as lesbian, 7% as bisexual; 65% were employed, 21% were unemployed, and 10% were students. Most (90%) came into the category of white ethnicity, and the sample had a mean age of 31 years (SD 6.4). Fifty five per cent were asymptomatic.

Prevalence of sexually transmitted infections

There were no significant differences in prevalence of STIs when stratified by history of sex with men (table 1). However PID, chlamydia and gonorrhoea were diagnosed only in women with a history of sex with men.

Three women who had not been sexually active with men were diagnosed with one of the following: genital warts, primary genital herpes, and trichomoniasis. All three women reported engaging in a large variety of sexual practices with female partners so it is not possible to draw conclusions about modes of transmission.

Two women were diagnosed with hepatitis C; both had histories of injecting drug use. There were no cases of HIV, syphilis, or hepatitis B. However, while most subjects were tested for syphilis, a minority were tested for HIV or hepatitis, so we cannot estimate prevalence rates.

Univariate and multivariate analysis

The distribution of sexual history with male and female partners, sexual practice with women, pregnancy, alcohol intake, and smoking is shown in tables on the STI website (www.stijournal/supplemental/ tables A and B). Past or current smoking was the only variable significantly related to diagnosis of any STI at a level of p<0.05. Lifetime number of female partners, use of dental dams and employment status were significant at a level of p<0.1 and so were also included in multivariate analysis (table 2). None of the other risk factor variables were significantly associated with diagnosis of any STI.

A non-significant trend was observed with increasing numbers of female sexual partners (the odds ratio for any STI increased by about 40% for each category increase in numbers of lifetime female partners) (table 2; further tables are available online at http://www.stijournal.com/supplemental/). Odds of any STI increased with a history of sex with men, although this was not statistically significant. There was a non-significant reduction in the odds of any STI with employment and with the use of dental dams.

DISCUSSION

Main findings

Candida and bacterial vaginosis were common. Trichomoniasis, genital warts, and genital herpes were infrequently diagnosed, and pelvic inflammatory disease, chlamydia, and gonorrhoea infections were rare. It appears that trichomoniasis, genital herpes, and genital warts may be sexually transmitted between women, but there is no evidence for woman to woman transmission of gonorrhoea, chlamydia, or PID.

Risk factors

Heterosexual sexual contact would be expected to increase the prevalence of STI. However, only 6.7% of WSW had been sexually active with men within the past year. Employment and smoking status are likely to be markers for other factors (for example, choice of higher risk sexual partners). It is plausible that a latex dental dam would reduce the sexual transmission of infection, and that increasing numbers of female partners would increase the risk. These trends need confirming in larger studies.

Limitations of the study

Incomplete response rates may introduce selection bias. Recall bias and social desirability bias are well recognised in sexual health surveys⁴ and we used convenience sampling, so our results cannot be generalised to all WSW. However, this sample is the largest from the United Kingdom to date and includes a large proportion of asymptomatic women. STI diagnoses were infrequent, so our study lacked the power to analyse all risk factors since some categories contained few subjects.

Comparison with other studies

No STIs except PID were found in 148 WSW in San Francisco in 1978.⁵ Two much larger studies conducted in Washington State⁶ and Sydney⁷ in the 1990s describe higher prevalence than our results for all STIs except BV and candida. These comparison populations included a high percentage of women who were sex workers (14.3% and 22%, respectively). Only 1.0% of our sample reported more than 50 lifetime male sexual partners and 1.7% reported more than 50 lifetime female sexual partners.

Trichomoniasis, genital warts, and genital herpes

Others have described transmission of trichomoniasis,⁸ genital warts,⁹ and genital herpes³ between women. Cervical human papillomavirus (HPV) and cervical smear abnormalities have also been noted in women who have no history of sex with men.¹⁰ Genital herpes, genital warts, and trichomoniasis may have non-sexual modes of transmission. Despite this, sexual transmission between women remains plausible.

Gonorrhoea, chlamydia, and pelvic inflammatory disease

In common with others^3,5–7 we found that infection with gonorrhoea, chlamydia, or PID was present in WSW, but was diagnosed only in those who reported sexual histories with men. These results suggest that routine or repeated screening for gonorrhoea and chlamydia in women who have no history of sex with men may not be necessary.

Comparison with heterosexual clinic attenders

Studies which included heterosexual female clinic attenders show a lower prevalence of most STIs (except BV) in WSW.^3,6,7,9 However, in two of these studies^6,7 WSW reported injecting drug use and higher risk partnerships more frequently than heterosexual women, and in the latter the prevalence of hepatitis B and C was significantly higher in WSW.

CONCLUSIONS

Although we have demonstrated a low prevalence of STI in this sample of WSW, lesbians and bisexual women may have sexual histories with men as well as women or histories of injecting drug use, and are therefore vulnerable to sexually transmitted or blood borne infections. These factors outline the importance of taking a thorough history and offering health advice based on behaviour rather than sexual identity.¹¹