ORGANISMS SUCH AS GARDNERELLA VAGINALIS, Mycoplasma hominis, Ureaplasma urealyticum, and other anaerobes (eg, Prevotella, Mobiluncus sp, Bacteriodes) are implicated in bacterial vaginosis (BV), with a decrease in normal vaginal lactobacilli. This change in vaginal flora may be asymptomatic or may result in a vaginal discharge with an odor and/or vaginal inflammation. BV is common in women who have sex with women (WSW),1,2 but there is debate about whether BV can be sexually transmitted between women.1,3,4
The etiology of BV is unclear. BV is associated with smoking, having a new sexual partner, increasing age, black and African Caribbean ethnicity, and vaginal douching, but many of these associations may be confounded.5 BV is associated with complications such as second-trimester miscarriage,6 premature rupture of membranes, and premature delivery of low birth weight infants,7 and with infection after obstetric or gynecologic surgery.8
This study uses multivariate analysis to investigate whether BV is associated with sexual activity between women.
Materials and Methods
A questionnaire was offered to new patients at 2 sexual health clinics for lesbians and bisexual women in London, U.K. Eight hundred three completed the questionnaire out of an estimated 1000 attenders between 1992 and 1995. The questionnaire was self-completed before consultation with clinic staff and gathered data about demographic features, sexual history, and sexual practice. Seven hundred eight women (88%) had a full genitourinary screen, including testing for Neisseria gonorrhoeae, Chlamydia trachomatis, Trichomonas vaginalis, and Candida sp.9 Diagnosis of BV was made using 3 of 4 of Amsel’s criteria10 (thin vaginal discharge, malodorous reaction to potassium hydroxide, vaginal pH ≥4.7, and clue cells on vaginal wet preparation).
Univariate analysis explored the associations of BV diagnosis with sexual history, sexual practice, and demographic variables. Most variables were examined as categorical variables. Those reaching statistical significance at the level of P <0.1 were included in a multivariate model, as well as variables found to be related to BV in other studies (eg, vaginal douching). A logistic regression model was built up in a stepwise fashion by using the likelihood ratio (LR) test and by observing change in odds ratios.
The demographic features of the 708 clinic attenders are shown in Table 1.
BV was diagnosed in 222 women (prevalence 31.4%; confidence interval [CI], 27.9–34.8). A total of 31.8% of women with BV were asymptomatic, whereas 34.6% reported vaginal discharge, 14.5% reported vulvovaginal irritation, and 8.9% reported menstrual symptoms. Coinfection was uncommon: 6.8% of those with BV also had Candida sp. (15 women), 1.4% (3) had genital herpes, 1.4% (3) genital warts, 0.9% (2) chlamydia, 0.5% (1) trichomoniasis, and no one was coinfected with gonorrhea.
Eight variables were positively associated with BV diagnosis at a level of P <0.05 in univariate analysis: number of female partners ever, number of female partners in the last year, vaginal penetration with fingers, vaginal penetration with sex toys, oroanal sexual contact, lesbian sexual identity, Asian ethnicity, and smoking (Table 2).
Vaginal penetration with fingers was highly correlated with vaginal penetration with sex toys and oroanal sex; the latter variables did not make an independent contribution to the multivariate model and so therefore were not retained. Similarly, total number of female sexual partners was highly correlated with number in the last year, so only the former was retained. Sexual activity with men, use of lubricant and vaginal douching were not associated with a diagnosis of bacterial vaginosis, so these were not retained in the final model (Table 3).
Multivariate analysis showed that diagnosis of BV was significantly associated with larger numbers of female sexual partners (60% increase in odds ratio with 11 or more lifetime female partners compared with 1–5 female partners) (Table 3). More frequent vaginal penetration with fingers was associated with a 30% increase in BV, although this was not significant. Smoking and Asian ethnicity significantly increased the odds ratio for BV.
We found a higher prevalence of BV (31.4%) than in a U.K. population with a comparable distribution of ethnicity (15.7% in a subsample of WSW).2 The prevalence of BV was found to be higher in WSW than in heterosexual women in 2 studies2,11 but not another.12 Epidemiologic studies of risk factors for BV present contradictory findings.5 The high prevalence we describe is not explained by age, sex with men, vaginal douching, the use of lubricants, a history of pregnancy, or employment status.
An experiment by Gardner and Dukes supports the hypothesis that BV may be sexually transmitted between women; introduction of the vaginal secretions of a woman with BV into the vagina of another woman resulted in the development of BV.13 Others have found that the female partners of women with BV were also more likely to have BV.1,3
Our results are consistent with 2 multivariate analyses, which found associations between BV and numbers of female sexual partners.1,14 Thirty-eight percent of our sample reported 2 or more female partners within the last year, and 57% reported 6 or more lifetime female partners. It may be that exposure to female partners over time is important, especially because recurrence of BV after treatment is common.15
Others describe associations with receptive oroanal sex (odds ratio [OR], 2.4), and failure to clean an insertive sex toy between uses on the subject and her partner (OR, 2.7).1 We were unable to demonstrate significant associations with particular sexual activities. However, we did not collect data on the time elapsed from each specific sexual behavior, and some sexual activities were almost universally reported (Table 2), so association is difficult to demonstrate. Heterosexual contact did not increase the odds of BV, although only 6.7% of the sample reported sexual activity with men within the last year. Less than 10% of the sample reported vaginal douching.
Smoking status is unlikely to contribute directly to the etiology of infection but is more likely to be a marker for other factors (eg, background prevalence or choice of sexual partner).5 The association of BV with Asian ethnicity should be interpreted with caution because the numbers in this subgroup were small.
Sexual health surveys are vulnerable to inaccuracies such as recall bias and social desirability bias. Because this clinic-based study relies on convenience sampling, results cannot be generalized to all women who have sex with women. However, this sample is the largest from the U.K. to date and includes a large proportion of asymptomatic women.
The high prevalence of BV in this population of WSW has implications for obstetric or gynecologic treatment. Screening for BV and/or prophylactic antibiotics is recommended before gynecologic surgery.15 Lesbians may choose to become pregnant9; routine screening for BV in low-risk asymptomatic pregnant women does not lead to improved maternal or fetal outcomes, although antibiotic treatment of BV in women at high risk of premature delivery is recommended.16
This multivariate study strengthens the evidence for an association of BV with sexual activity between women. The demonstration of a trend with increasing numbers of female partners supports the hypothesis that BV may be sexually transmitted between women. This suggests that female partners of women being treated for BV should also be screened and treated for BV.
1. Marrazzo JM, Koutsky LA, Eschenbach DA, et al. Characterization of vaginal flora and bacterial vaginosis in women who have sex with women. J Infect Dis 2002; 185:1307–1313.
2. Morris MC, Rogers PA, Kinghorn GR. Is bacterial vaginosis a sexually transmitted infection? Sex Trans Infect 2001; 77:63–68.
3. Berger BJ, Kolton S, Zenilman JM, et al. Bacterial vaginosis in lesbians: A sexually transmitted disease. Clin Infect Dis 1995; 21:1402–1405.
4. McCaffrey M, Varney P, Evans B, et al. Bacterial vaginosis in lesbians: Evidence for lack of sexual transmission. Int J STD AIDS 1999; 10:305–308.
5. Morris M, Nicoll A, Simms I, et al. Bacterial vaginosis: A public health review. Br J Obstet Gynecol 2001; 108:439–450.
6. Oakshott P, Hay P, Hay S, et al. Association between bacterial vaginosis or chlamydial infection and miscarriage before 16 weeks' gestation: prospective community based cohort study. BMJ 2002; 325:1334–1336.
7. Flynn CA, Helwig AL, Meurer LN. Bacterial vaginosis in pregnancy and the risk of prematurity: A meta-analysis. J Fam Pract 1999; 48:885–892.
8. Koumans EH, Kendrick JS, CDC Bacterial Vaginosis Working Group. Preventing adverse sequelae of bacterial vaginosis: A public health program and research agenda. Sex Transm Dis 2001; 28:292–297.
9. Bailey JV, Farquhar C, Owen C, et al. Sexually transmitted infections in women who have sex with women. Sex Transm Infect 2004; 80:244–246.
10. Amsel R, Totten PA, Spiegel CA, et al. Non-specific vaginitis; diagnostic criteria and microbial and epidemiologic associations. Am J Med 1983; 74:14–22.
11. Skinner CJ, Stokes J, Kavanagh J, et al. A case-controlled study of the sexual health needs of lesbians. Genitourin Med 1996; 72:277–280.
12. Marrazzo JM, Koutsky LA, Handsfield HH. Characteristics of female sexually transmitted disease clinic clients who report same-sex behaviour. Int J STD AIDS 2001; 12:41–46.
13. Gardner HL, Dukes CD. Haemophilus vaginalis
vaginitis. Am J Obstet Gynecol 1955; 69:962–967.
14. Smart S, Singal A, Mindel A. Social and sexual risk factors for bacterial vaginosis. Sex Transm Infect 2004; 80:58–62.
15. Management of bacterial vaginosis. Drug Ther Bull 1998; 36:33–35.
16. McDonald H, Brocklehurst P, Parsons J, et al. Antibiotics for treating bacterial vaginosis in pregnancy (Cochrane Review). Cochrane Library, issue 4. Chichester, UK: John Wiley and Sons Ltd; 2003.