Bacterial Vaginosis: Are We Coming Full Circle?

In 1975, a round table discussion was held on nonspecific vaginitis or Haemophilus vaginitis, which is now known as bacterial vaginosis (BV). Participants in this round table included Drs Herman Gardner, Sue Criswell, Wilbur Dunkelberg, Malcolm Fuller, Richard Hildebrandt, and William Josey, all from prestigious medical institutions. Also in attendance was Mukti Bhattacharyya, a consulting venereologist from Sheffield, England. During the discussion on the epidemiology of BV, Dr Josey stated the following: “In my opinion, H. v. vaginitis is a form of venereal disease. If we can identify an organism as being associated primarily with the genital system, I would automatically classify it as causing a venereal infection” [1]. Indeed, that was the teaching in textbooks of that era Despite the epidemiological evidence that BV is a sexually transmitted disease (STD), opinions changed over the next few decades, owing to several factors: (1) the lack of agreement on the role of Haemophilus vaginalis, now Gardnerella vaginalis as a pathogen causing BV; (2) failure of male partner treatment studies to demonstrate a clear benefit for prevention of recurrent BV in the female partner; (3) improved microbiological studies that showed that anaerobic vaginal flora were greatly increased in women with BV; (4) a few reports that found BV in virginal women; and (5) an overall lack of understanding of the pathogenesis of BV, including lack of an understanding of “optimal vaginal flora.” Thus, the currently accepted dogma is that BV is a polymicrobial syndrome [2] that appears to be associated with sexual activity but might somehow just “happen” to a woman due to an elusive cause. The possible explanations have included douching as a disruptor of the normal vaginal flora and “bad semen” as an alkaline enemy of the normally acidic vaginal pH Over the past several years, thoughts on the pathogenesis of BV have begun to shift because of research that has challenged the dogma and moved the field forward. These recent studies have led to important data on the role that condoms play in protection against BV, epidemiological distinctions of incident versus recurrent BV, a better appreciation for optimal vaginal health versus disrupted vaginal flora, and (as seen in this issue of the Journal) a reevaluation of the sexual practices of young women with BV [3] Several studies have recently investigated the role that condom use plays in the prevention of BV. In a study of BV involving 96 women who sought care at an STD clinic, incident BV was significantly associated with exposure to a new sexual partner (P=.02). Women who developed BV were more likely to have not used a condom with their occasional partner. Among women who developed BV, their male partners reported significantly more sexual partners within the previous 30 days than those women who did not develop BV (P=.03) [4]. Treatment studies for BV have shown that women who had unprotected sex during the follow-up period were more likely than those women who either abstained or used condoms to develop recurrent BV, and unprotected sex during the course of the study was associated with overall treatment failure [5, 6]. Hutchinson et al showed that consistent condom use was associated with a decrease in the risk for BV, with the strongest association being in women with lactobacillus-predominant normal flora as opposed to intermediate flora [7]. Most recently, Yotebieng et al [8] demonstrated that consistent condom use was associated with reduced rates of incident BV but had no effect in women who were already BV-positive at baseline In a study involving women with recurrent BV, Bradshaw et al [9] sought to identify risk factors associated with recurrence of BV over a 12-month period after oral metronidazole therapy. At 1 month follow-up, 23% of women had a recurrence of BV; by 12 months, 58% of women had recurrent BV, and 69% had recurrence of abnormal vaginal flora. Factors associated with recurrence were history of BV, lack of hormonal contraceptive use, having a regular sex partner throughout the study, and having a female sex partner. Having a new sex partner during the study period was not associated with recurrence. Women with a regular sex partner were more likely to have used condoms <50% of the time than were those with new sex partners, although condom use was not associated with recurrence in this cohort. The authors suggested that women treated for BV are reinfected by their regular sex partner but not by new partners, who would be less likely to carry the causative organisms for BV. Recurrence was associated with specific behavioral factors, which supports a role for sexual transmission in the pathogenesis of recurrent BV Although all would agree that BV is abnormal, the vaginal flora pattern that results in optimal vaginal health remains unclear. Intermediate vaginal flora—characterized by decreasing numbers of lactobacilli and increasing numbers of Gardnerella (Nugent scores of 4–6)—has previously been shown to be a pattern that tends to be transient and that can either revert to normal, lactobacillus-predominant flora (Nugent scores of 0–3) or develop into BV (Nugent scores of 7–10) [10]. Because variability in the vaginal flora with seemingly increased but transient levels of Gardnerella and other organisms has been commonly identified in women, it was dismissed in the past as being “normal” [11]. However, comparing the intermediate pattern of vaginal flora to vaginal flora that is persistently dominated by lactobacilli clearly demonstrates that the intermediate pattern is not optimal for vaginal health. Although studies on the complications of abnormal vaginal flora have focused on women with BV, it is possible that women with intermediate flora may also be at risk for complications, especially those associated with absence of lactobacilli, such as STD acquisition and perhaps preterm birth [12, 13] The presence of abnormal vaginal flora is not typically associated with clinical evidence of an inflammatory process [14]. However, multiple studies involving nonpregnant and pregnant women have demonstrated elevated levels of proinflammatory cytokines in cervical and vaginal secretions of women with altered flora, compared with those of women with normal vaginal flora [15–22]. This inflammation may contribute to the risk for adverse outcome associated with BV. Yudin et al [19] measured cervical cytokines before and after treatment with oral or topical metronidazole in pregnant women with BV. They noted a statistically significant decrease in cervical cytokine levels in those women who achieved a therapeutic cure (Gram stain score 0–3 and absence of all 4 Amsel clinical criteria). However, no change was noted in women who were not cured, including those women with intermediate Nugent scores. Vaginal proinflammatory cytokine (interleukin 1β [IL-1β], interleukin 8, and interleukin 6) levels are elevated in women with intermediate flora, as well as in women with BV [22]. Furthermore, elevated vaginal cytokines appear to correlate predominantly with higher scores for G. vaginalis/Prevotella rather than with other components of vaginal flora identified on vaginal Gram stain [22] Vaginal mucosal inflammation may be an important contributing factor for abnormal vaginal flora as a biological risk factor for human immunodeficiency virus (HIV) acquisition and transmission. For example, proinflammatory cytokines have been associated with increased HIV replication and spread [17, 23]. Elevated levels of these same cytokines associated with BV may explain the association between abnormal vaginal flora and increased risk for transmission of HIV. Mitchell et al [24] showed that in women with HIV and BV, BV was the factor responsible for the increase in proinflammatory cytokines, as opposed to HIV. Among pregnant women colonized with anaerobic gram-negative rods and/or G. vaginalis, Genc et al [18] found that elevated concentrations of vaginal IL-1β or a diminished interleukin 1RN to IL-1β ratio were associated with an increased risk for preterm delivery. This localized inflammation may induce cervical ripening, weaken amniotic membranes, or precipitate contractions that result in preterm delivery Small antimicrobial peptides known as defensins are important for controlling the growth of normal bacteria on human mucosal surfaces and are the body’s first line of defense against microbial invasion. In studies involving nonpregnant women without clinical evidence of pelvic inflammatory disease, elevated defensin levels in vaginal secretions were strongly associated with upper genital tract infection or endometritis, as well as with elevated numbers of vaginal neutrophils [25, 26]. Elevated maternal serum defensin levels in the mid-trimester have also been associated with subsequent preterm birth [27]. In other studies of pregnant women, elevated vaginal defensin levels were associated with intermediate vaginal flora (Gram stain score 4–6), the presence of vaginal neutrophils, and early preterm birth at <32 weeks (odds ratio, 3.1; 95% confidence interval, 1.0–9.8) [28, 29]. Thus, it appears that the source of intermediate vaginal flora, although perhaps common, is indeed not conducive to optimal vaginal health In this issue of the Journal, Fethers et al [3] reexamine risk factors for BV among young sexually experienced and inexperienced women. Five hundred forty college-aged women were recruited into a cross-sectional study of the association of behaviors and BV. Of 528 evaluable women, 25 (4.7%) had BV. Among the 83 women who reported no history of any sexual contact, the prevalence of BV was 0%. Among the 78 women who reported engaging in only noncoital sexual contact, 3 had BV. The noncoital sexual practices were digital-vulvovaginal contact, oral sex, and sex toy use. Thus, no BV was found among truly sexually inexperienced women. Previous studies that reported BV in virgins defined virginity as lack of history of penile-vaginal sex [30, 31]. These studies collected no data on other sexual practices. Both studies were conducted in settings that may have led to underreporting of sexual intercourse by the young women, namely, American military recruits and a gynecology clinic in which the participants were the classmates of the daughters of the clinical staff Bacterial vaginosis is an important public health problem that, until lately, has been the subject of little innovative research. Although their study is straightforward and relatively simple in its design, Fethers et al [3] should be commended for their willingness to challenge the prevailing dogma in part created by the data from studies such as those of “virgins” mentioned above. It is time to accept (again) that BV is sexually transmitted—as was generally acknowledged in 1975—and begin to focus on studies to elucidate the STD organism responsible for its pathogenesis. Without progress in determining the microbial origin of this disease, we will not be able to move forward on preventing the costly public health complications associated with BV I thank Dr William Josey for his thoughtful advice and mentoring