BACTERIAL VAGINOSIS (BV) IS the most common cause of vaginitis worldwide,1–3 It has been repeatedly associated with a risk of preterm birth4,5 and is clearly a risk factor for the acquisition and perhaps transmission of sexually transmitted diseases (STDs), including human immunodeficiency virus (HIV) and subclinical and clinical pelvic inflammatory disease.4,6–9 There is general agreement that reducing rates of BV could lead to a decrease in these complications, but prospective studies addressing this issue have been thwarted by our lack of understanding of the pathogenesis of BV, which further limits treatment and prevention efforts. Previous studies have shown that 2 main behavioral factors, douching and sexual behavior, are associated with BV10–14; however, the majority of studies are cross-sectional in design. Moreover, there are few studies that have included data collected from the male sexual partner(s). It has been shown that male partners of women with BV may be colonized in the genital tract with BV-associated bacteria15 and that nongonococcal urethritis in the male is significantly correlated with BV in the female partner.16 Further, recent studies on the vaginal flora in women who have sex with women have provided compelling data that BV is sexually transmitted in these couples.17,18 Despite the available evidence, the hypothesis that BV is sexually transmitted in heterosexual couples has not been accepted. The most common reason given for this skepticism is the failure of studies in which the male partner was treated to show a decrease in BV recurrence rates in the female.19–21 However, these studies performed 2 decades ago were not optimally designed or executed in that they used a single dose of metronidazole and had no measures of compliance for the men. In order to gain additional insight into the pathogenesis of BV, we performed a prospective, longitudinal study, which examined risk factors for incident BV.
Materials and Methods
Women attending the Jefferson County Department of Health (JCDH) STD Clinic in Birmingham, AL, with a new problem or for screening without evidence of BV were eligible to participate in this longitudinal study on risk factors for the acquisition of BV. The study was approved by the institutional review boards of the University of Alabama at Birmingham and JCDH. All subjects provided informed consent. Inclusion criteria included age of at least 18 years and current heterosexual activity of at least 1 time per week on average. Women were excluded from the study if they had evidence of BV on vaginal Gram stain according to the criteria of Nugent et al.22 or evidence of STD.
Women agreeing to participate were asked detailed questions concerning symptomatology (specifically, vaginal discharge and odor), sexual history, sexual practices, hygienic practices, contraceptive history, and history of vaginitis and STDs. A pelvic examination was conducted, and diagnostic specimens were obtained for vaginitis and STD testing. Vaginal Gram stains were interpreted according to the method of Nugent et al.22
All women were counseled regarding safe sex practices and provided with condoms. Subjects were reinterviewed and reexamined at monthly intervals for 6 months, and then at 9 and 12 months. Laboratory testing was repeated as described above. Subjects were also asked to refer their male partners for enrollment into the male partner substudy. The latter consisted of a behavioral questionnaire, as well as screening for gonorrhea and chlamydia using urine ligase chain reaction (Abbott Laboratories, Abbott Park, IL). In addition, the clinical diagnosis of nongonococcal, nonchlamydial urethritis (NGU), defined by >5 polymorphonuclear leukocytes per oil immersion field on Gram stain, was recorded.
Subjects were treated according to protocol for any STD detected and were treated with oral metronidazole 500 mg twice a day for 7 days on development of symptomatic BV. Asymptomatic BV, defined as absence of vaginal discharge or odor, was not treated per CDC recommendations.23
Descriptive baseline characteristics for women classified as BV+ or BV− during the follow-up period were compared by the χ2 test for categorical variables or the t test for continuous variables. An episode of BV was defined as the development of BV as diagnosed by the Nugent criteria. Time to first episode was defined as the number of days from baseline to development of first episode of BV. BV rates per person-year were calculated as the sum of all new episodes of BV divided by the sum of all the follow-up days for the entire cohort, expressed as person-years. Persistent episodes of BV, defined as the presence of BV at a visit when the subject had received appropriate therapy for symptomatic BV at the prior visit or when the subject had had asymptomatic BV which was not treated at the prior visit, were not included in the analysis. The probability of remaining BV free was modeled using the Kaplan-Meier method. For calculation of the relative risks for incident BV during the follow-up period, repeated-measures models were used with the binomial distribution, with the outcome (BV) coded as 0 or 1. Risk factors were modeled univariately, and factors significant at P <0.05 were used to develop the multivariate model. Risk factors for women included information collected on the follow-up visits. For information collected on multiple sex partners since last visit, the exposure was totaled across all partners (such as number of times vaginal sex, number of times anal sex, number of different partners current partner had) and dichotomized at the median as 0 or 1. Risk factors examined in modeling included new partner since last visit, anal sex at least 1 time since last visit, sex with uncircumcised male at least once since last visit, given metronidazole since last visit, douched at least once since last visit, vaginal sex at least 6 times since last visit, more than 1 different partner since last visit, sexual partner had at least 1 other partner in the last 30 days, used condom always with regular partner, and used condom always with occasional partner. For all final analyses, a P value of <0.05 was deemed statistically significant.
Risk factors for males that were examined for the relationship with BV included age at first sex, number of partners in last 30 days, new partners in last 30 days (≥1 vs. 0), condom use (always in last 30 months vs. less than always), STD diagnosis (NGU, gonorrhea, or chlamydia), circumcision status, and number of days since last sex.
Incidence of BV
Ninety-six women were enrolled into the study. The mean age of the participants was 24.2 ± 6.0 years, and 92% were black. Table 1 shows the relationship between baseline factors and the development of BV. Women who developed BV were more likely to have a greater number of partners in the last 30 days (2.5) compared to women who did not develop BV (1.0), but not significantly so. Women who developed BV also reported more vaginal sex episodes in the last 30 days compared to women who did not develop BV. Women with intermediate flora (Nugent scores of 4–6) at enrollment were more likely to develop BV compared to women with Nugent scores of 0 to 3 at baseline (P = 0.09).
The median time to develop BV was 81 days (95% CI, 56–113) (Fig. 1). Overall, 67 women (69.8%) developed BV at some point during follow-up, with a total of 227 episodes reported during the follow-up period. Of the 67 women who developed BV during the follow-up period, a majority (79.1%) had a recurrent or persistent episode (Fig. 2). Forty women had at least 1 episode of BV that was defined as persistent, with a total of 92 persistent episodes among these women. The incidence rate of new episodes of BV was 2.33/ person-year (95% CI, 1.63–2.50). Among those women who completed at least 6 months of follow-up, 20/72 (20.8%) never developed BV.
Ongoing Behaviors Associated with BV: Repeated-Measures Model
Analysis of single risk factors over time associated with episodes of BV is presented in Table 2. Women who developed BV were significantly more likely to have had a new sex partner since their previous visit (RR, 1.13; 95% CI, 1.02–1.25) and were significantly more likely to have had more frequent vaginal sex since the last visit (RR, 1.07; 95% CI, 1.01–1.15). Among women who developed BV, those who always used a condom with their occasional partner were significantly less likely to develop BV (RR, 0.80; 95% CI, 0.67–0.98). Having clinical symptoms of vaginal odor or discharge was associated with BV, as expected (P ≤0.0001). In multivariate analysis of behavioral risk factors associated with acquisition of BV, the only risk factor which remained significant was sex with a new partner since the prior visit (RR, 1.74; 95% CI, 1.05–2.87; P = 0.03), with those with a new partner nearly twice as likely to develop BV.
Male Sexual Partner Behaviors Associated With Development of BV
Among normal women followed in this study, 37 male partners were evaluated with a baseline questionnaire and laboratory examination (Table 3). Among women who developed BV, their male partners reported significantly more partners in the last 30 days (1.8) versus those who did not develop BV (1.2 partners). Other factors were not related to the acquisition of BV.
The etiology of BV remains unproven; however, evidence continues to accumulate in favor of sexual transmission. In the 1950s, Gardner and Dukes24 showed that BV could be induced in a woman by introduction into the vagina of vaginal secretions from an infected woman. More recent studies of women who have sex with women have shown similar findings with high concordance rates of abnormal vaginal flora between partners, again suggesting transmission via infected vaginal secretions.17,18
The epidemiology of BV mirrors that of an STD rather than that of vaginal candidiasis. Among women enrolled in a study of a topical microbicide, the number of sexual partners, but not the frequency of intercourse, was significantly associated with the occurrence of BV and trichomoniasis,13 suggesting that exposure to an infected male and not the act of intercourse itself was the causative factor. A cross-sectional study of women with BV found that women with BV were more likely to report a new sexual partner within the preceding month than women without BV.25 In a prospective longitudinal study of factors associated with acquisition of BV conducted in Seattle, sex with a new partner since the last visit was, similar to our study, the strongest predictor of incident BV (OR, 2.5; 95% CI, 1.3–4.6).11 Further, Keane et al.16 reported a highly significant relationship between NGU in males and BV in their sexual partners.
Microbiologic studies also support sexual transmission of BV. In their original report on Haemophilus vaginalis (now Gardnerella vaginalis), Gardner and Dukes24 believed the disease was sexually transmitted based on high recurrence rates and the isolation of H vaginalis from over 90% of the male partners. G vaginalis and anaerobic bacteria similar to those isolated from BV patients have been associated with balanoposthitis in males, an infection which clinically and microbiologically resembles BV.26,27 Further, G vaginalis and anaerobes have been frequently isolated from the semen of asymptomatic males.28 In a small study of BV-associated organisms in males, male partners of women with BV were more likely to have Mobiluncus, Gardnerella, and Mycoplasma hominis isolated from genital sites than partners of women without BV. Further, specimens that were obtained from the male less than 20 hours after intercourse were more likely to be positive than specimens collected from the same men 2 weeks later, suggesting that the colonization was transient.15
Although douching has been associated with BV in cross-sectional studies, no prospective data are available. Further, studies that have shown an association between BV and douching have frequently also shown an association of BV with number of sexual partners or acquisition of a new sexual partner.10,11,14 Douching is highly interrelated to sexual activity, and thus it is difficult to determine the true relationship of BV to douching, especially in cross-sectional studies. Further, in a study of vaginal flora among women who have sex with women, douching was not associated with BV.18
Cited as evidence against the sexual transmission of BV is a study that found BV among a group of sexually inexperienced (based on self-report) adolescent females.29 Additionally, the failure of partner studies to show reduction in BV recurrence rates among women whose male sexual partners were treated with a 2-g dose of metronidazole is seen as nonsupportive of sexual transmission.19,20,30 However, there was no measure of compliance with the medication in these studies. Further, single-dose therapy is only marginally effective for females31 and thus may be ineffective for males.
Our study confirms that the most significant risk factor for acquisition of BV among women at high risk for STDs is exposure to a new male sexual partner. Further, use of condoms with the occasional partner was protective against development of BV. With regard to the male data, it is also of interest that males with greater numbers of sexual partners were more likely to have a partner who developed BV and that women who thought their partner had other recent partners were at greater risk of acquiring BV. Although the actual mechanism of transmission is unclear, the available data concerning transient colonization of the male genital tract with BV-associated bacteria suggests that recent sexual exposure of a male to a female with BV could enhance or allow transmission to an uninfected female partner. Future studies should focus on the mechanism of transmission, as well as the potential for appropriate treatment of the male partner to prevent transmission.
1. Fleury FJ. Adult vaginitis. Clin Obstet Gynecol 1981; 24:407–438.
2. Amsel R, Totten PA, Spiegel CA, Chen KCS, Eschenbach D, Holmes KK. Non-specific vaginitis: diagnostic and microbial and epidemiological associations. Am J Med 1983; 74:14–22.
3. Rein MF, Holmes KK. “Non-specific vaginitis,” vulvovaginal candidiasis, and trichomoniasis clinical features, diagnosis and management. Curr Clin Top Infect Dis 1983; 4:281–315.
4. Eschenbach DA. Bacterial vaginosis and anaerobes in obstetric-gynecologic infection. Clin Infec Dis 1993; 16:S282–S287.
5. Hillier SL, Nugent RP, Eschenbach DA, et al. Association between bacterial vaginosis and preterm delivery of a low-birth weight infant. N Engl J Med 1995; 333:1737–1742.
6. Wiesenfeld H, Hillier L, Krohn MA, et al. Lower genital tract infection and endometritis: insight into subclinical pelvic inflammatory disease. Obstet Gynecol 2002; 100:456–463.
7. Martin H, Richardson B, Nyange P, et al. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition. J Infect Dis 1999; 180:1863–1868.
8. Wiesenfeld H, Hillier S, Krohn MA, Landers D, Sweet R. Bacterial vaginosis is a strong predictor of Neisseria gonorrhoeae
and Chlamydia trachomatis
infection. Clin Infect Dis 2003; 36:663–668.
9. Taha TE, Hoover DR, Dallabetta GA, et al. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV. AIDS 1998; 12:1699–1706.
10. Schwebke J, Desmond RA, OH MK. Predictors of bacterial vaginosis in adolescent women who douche. Sex Transm Dis 2004; 31:433–436.
11. Hawes S, Hillier S, Benedetti J, Stevens C, et al. Hydrogen peroxide-producing lactobacilli and acquisition of vaginal infections. J Infect Dis 1996; 174:1058–1063.
12. Ness R, Hillier S, Richter H, et al. Douching in relation to bacterial vaginosis, lactobacilli, and facultative bacteria in the vagina. Obstet Gynecol 2002; 100:765–772.
13. Barbone F, Austin H, Louv WC, Alexander WJ. A follow-up study of methods of contraception, sexual activity and rates of trichomoniasis, candidiasis and bacterial vaginosis. Am J Obstet Gynecol 1990; 163:510–514.
14. Holzman C, Leventhal J, Qiu H, Jones N, Wang J, Group BS. Factors linked to bacterial vaginosis in nonpregnant women. Am J Public Health 2001; 91:1664–1670.
15. Holst E. Reservoir of four organisms associated with bacterial vaginosis suggests lack of sexual transmission. J Clin Microbiol 1990; 28:2035–2039.
16. Keane FE, Thomas BJ, Whitaker L, Renton A, Taylor-Tobinson D. An association between non-gonococcal urethritis and bacterial vaginosis and the implications for patients and their sexual partners. Genitourin Med 1997; 73:373–377.
17. Berger BJ, Kolton S, Zenilman JM, Cumminep MC, Feldman J, McCormack WM. Bacterial vaginosis in lesbians: a sexually transmitted disease. Clin Infect Dis 1995; 21:1402–1405.
18. Marrazzo J, Koutsky L, Eschenbach D, Agnew K, Stine K, Hillier S. Characterization of vaginal flora and bacterial vaginosis in women who have sex with women. J Infect Dis 2002; 185:1307–1313.
19. Mengel MB, Berg AO, Weaver CH, et al. The effectiveness of single-dose metronidazole therapy for patients and their partners with bacterial vaginosis. J Fam Pract 1989; 28:163–171.
20. Swedberg J, Steiner JF, Deiss F, Steiner S, Driggers DA. Comparison of single-dose vs. one-week course of metronidazole for symptomatic bacterial vaginosis. JAMA 1985; 254:1046–1049.
21. Vejtorp M, Bollerup AC, Vejtorp L, et al. Bacterial vaginosis: a double-blind randomized trial of the effect of treatment of the sexual partner. Br J Obstet Gynecol 1988; 95:920–926.
22. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation. J Clin Microbiol 1991; 29:297–301.
23. Centers for Disease Control and Prevention. 2002 Sexually transmitted disease treatment guidelines. MMWR Morb Mortal Wkly Rep 2002; 51:RR-6.
24. Gardner HL, Dukes CD. Haemophilus vaginalis
vaginitis. Am J Obstet Gynecol 1955; 69:962–976.
25. Paavonen J, Miettinen A, Stevens CE, Chen KCS, Holmes KK. Mycoplasma hominis in non-specific vaginitis. Sex Transm Dis 1983; 45:271–275.
26. Kinghorn GR, Jones BM, Chowdhury FH, Geary I. Balanoposthitis associated with Gardnerella vaginalis
infection in men. Br J Vener Dis 1982; 58:127–129.
27. Masfari AN, Kinghorn GR, Duerden BI. Anaerobes in genitourinary infections in men. Br J Vener Dis 1983; 59:255–259.
28. Hillier SL, Rabe LK, Muller CH, Zarutskie P, Kuzan FB, Stenchever MA. Relationship of bacteriologic characteristics to semen indices in men attending an infertility clinic. Obstet Gynecol 1990; 75:800–804.
29. Bump RC, Buesching WJ. Bacterial vaginosis in virginal and sexually active adolescent females: evidence against exclusive sexual transmission. Am J Obstet Gynecol 1988; 158:935–939.
30. Moi H, Erkkola R, Jerve F, et al. Should male consorts of women with bacterial vaginosis be treated? Genitourin Med 1989; 65:263–268.
31. Eschenbach DA, Critchlow CW, Watkins H, et al. A dose duration study of metronidazole for the treatment of non-specific vaginosis. Scand J Infect Dis 1983; S40:73–80.