Current Treatment of Bacterial Vaginosis—Limitations and Need for Innovation

Abstract

Practitioners and patients alike widely recognize the limitations of current therapeutic approaches to the treatment of bacterial vaginosis (BV). Options remain extremely limited, and our inability to prevent the frequently, often relentless symptomatic recurrences of BV and to reduce serious sequelae such as preterm delivery, remains an acknowledged but unresolved shortcoming. Our incomplete understanding of the pathophysiology of this unique form of vaginal dysbiosis has been a significant impediment to developing optimal treatment and prevention approaches. New drugs have not been forthcoming and are not likely to be available in the immediate future; hence, reliance on the optimal use of available agents has become essential as improvised often unproven regimens are implemented. In this review, we will explore the limitations of currently recommended therapies, with a particular focus on the contribution of reinfection and pathogen persistence to BV recurrence, and the development of interventions that target these mechanisms. Ultimately, to achieve sustained cure and effectiveness against BV-associated sequelae, it is possible that we will need approaches that combine antimicrobials with biofilm-disrupting agents and partner treatments in those at risk of reinfection.

Bacterial vaginosis (BV) represents a profound shift in the vaginal microbiota and is characterized by high bacterial species diversity; increased loads of facultative anaerobes, including Gardnerella vaginalis, Atopobium vaginae, and other fastidious BV-associated bacteria, such as Megasphaera, Sneathia, and Clostridiales species; increased production of volatile amines; and a rise in vaginal pH to > 4.5 [1]. This change is accompanied by marked depletion of key Lactobacillus species such as Lactobacillus crispatus, which produces lactic acid, bacteriocins, and other antimicrobial molecules, and is thought to play an important role in host defense against pathogens [2, 3]. Recent studies have identified a polymicrobial biofilm dominated by G. vaginalis and A. vaginae that is adherent to vaginal epithelial cells in women with BV and appears to be absent in healthy controls [4, 5]. The initiating event that results in this adverse shift in the vaginal microbiota remains unclear and has been a significant impediment to understanding the pathogenesis of this common condition and to optimizing treatment and prevention approaches. In this review, we explore the limitations of currently recommended therapies, with a particular focus on the contribution of reinfection and pathogen persistence to BV recurrence.

BV is vaginal condition with a high global burden in women of reproductive age. Prevalence estimates range from 12% in Australian women [6] and 29% in North American women [7, 8] to >50% of women affected in East/Southern Africa [9]. Over 50% of women with BV experience an unpleasant vaginal malodor and discharge, with qualitative studies showing that BV is associated with a significant negative impact on self-esteem, sexual relationships, and quality of life [10]. This dysbiotic state is associated with an approximate 2-fold increased risk of acquiring a broad range of sexually transmitted infections, including chlamydial infection, gonorrhea, herpes simplex type 2 infection, and human immunodeficiency virus (HIV) infection [11–14], and increases the risk of HIV-infected women transmitting HIV to male partners [15]. BV is associated with important reproductive and obstetric sequelae, increasing women's risk of pelvic inflammatory disease, spontaneous abortion, preterm delivery, low birth weight, and postpartum endometritis [16–18].

Current clinical approaches to the management of BV are somewhat empirical and involve the use of either nitroimidazoles or clindamycin, antimicrobials with broad-spectrum anaerobic coverage. Recommended first-line regimens include oral metronidazole 500 mg twice daily for 7 days, intravaginal 2% clindamycin cream once daily for 7 days, or intravaginal metronidazole gel once daily for 5 days [19]. Single doses and short courses of metronidazole, tinidazole, and intravaginal clindamycin are less effective [19]. While short-term cure rates following first-line recommended regimens are equivalent and approach 80% [20], studies with extended follow-up show that recurrence rates in excess of 50% occur within 6–12 months [21, 22]. These high rates of recurrence have led investigators to evaluate a range of alternative therapeutic approaches, including extended and suppressive antimicrobial regimens [23–27], combination first-line regimens [28], and adjunctive intravaginal and oral probiotic therapies [28, 29]. While some of these approaches appear promising and are under further evaluation, overall there has been limited progress in achieving sustained long-term cure following cessation of these regimens. This lack of success is a reflection of our poor understanding of the pathogenesis of both recurrent and incident BV. Higher baseline loads of some BV-associated bacteria have been associated with an increased risk of recurrence in one study [30], and 2 studies showed that BV-associated biofilm re-accumulates following antibiotic therapy [31, 32]. These data indicate that persistence of certain BV-associated bacteria and biofilm may be a determinant of recurrence following antimicrobials and raises the question as to whether antimicrobial resistance also plays a role. While there have been data indicating that clindamycin use can result in the emergence of clindamycin-resistant anaerobic gram-negative rods [33], in a panel of 865 anaerobic species obtained from women with BV, resistance to metronidazole was rare (0.3%) [34]. Despite evidence that there may be differences in antimicrobial susceptibility between the 2 first-line agents, overall BV cure rates following clindamycin and metronidazole have been equivalent [19, 35]. Intriguingly, however, recent whole-metagenome sequencing studies have identified at least 4 clades of G. vaginalis [36, 37], with preliminary studies showing that 2 of these clades may be intrinsically resistant to metronidazole, providing one possible mechanism for BV persistence after treatment [38]. There are also tantalizing data suggesting that reintroduction of BV-associated bacteria through sexual intercourse may be contributing to BV recurrence and that sexual transmission or exchange of BV-associated bacteria may play an integral role in the pathogenesis of BV and recurrence. Making significant progress in achieving sustained long-term cure of BV, in reducing BV-associated sequelae, and in developing prevention strategies necessitates investigating the contribution of reinfection, organism/biofilm persistence, and antimicrobial resistance to the pathogenesis of BV recurrence.

EVIDENCE FOR SEXUAL TRANSMISSION OF BV-ASSOCIATED BACTERIA AND IMPLICATIONS FOR TREATMENT AND PREVENTION STRATEGIES FOR BV RECURRENCE

The proposition that BV may be sexually transmitted has been evident in published literature for several decades [39–44]. However, confusion regarding the etiology of BV, difficulties in identifying a single causative agent, and absence of a clear disease counterpart in males have all been significant impediments in determining whether BV is sexually transmitted. In the 1950s, a highly unethical study demonstrated that women inoculated with secretions from women with BV but not a pure culture of G. vaginalis developed BV [45]. This led to the belief that BV was likely to be sexually transmitted, and the study was followed by 6 randomized, controlled male-partner-treatment trials that were conducted from the mid-1980s to the 1990s [39] that have recently been systematically reviewed [46]. Three trials found male partner treatment to be associated with a reduction in either BV recurrence or BV symptoms in women [47–49], although this was statistically significant in only 1 trial, by Mengel et al [46, 48]. Mengel et al reported a significant increase in symptom resolution and BV cure by Gram stain in women whose partners received either a 7-day course or a 2-g dose of metronidazole, but data were presented graphically, and no effect sizes were reported [48]. Vutyanavic et al reported a 13% increase in clinical cure of BV among women at 4 weeks whose male partners had been treated with 2 g of tinidazole [47], and Vejtorp et al found a 15% reduction in BV and a 36% reduction in culture of G. vaginalis at 5 weeks in women whose male partners were treated with two 2-g doses of metronidazole; neither finding was statistically significant [49]. Two trials found no effect on BV recurrence from treating male partners with a single dose or 7-day course of metronidazole [50] or with oral clindamycin for 1 week [51]. One trial found BV recurrence to be higher among women whose partners were treated with two 2-g doses of metronidazole, compared with those who were not [52].

The apparent incongruity between the strong evidence for sexual transmission and the failure of these trials now appears likely to be due to issues in trial design that have been detailed in a rigorous systematic review by Mehta [46]. Importantly, Mehta found that none of the trials had sufficient power to detect reasonable effect sizes, that randomization methods were deficient or insufficiently reported, and that adherence to therapy was not reported in women and was only reported in men in 2 trials. Five trials used treatment regimens in women that are now considered to be less effective, such as single-dose regimens, and 5 trials used regimens in men that are now known to be suboptimal in women. Mehta concluded that the trial findings were inconclusive by the current clinical trial standards and recommended that sufficiently powered trials using recommended therapies be conducted to determine whether antibiotic treatment in men can reduce BV and to generate an accurate evidence base for guidelines [46].

A mounting body of epidemiologic and microbiologic data suggest that sexual transmission is likely to be integral to the pathogenesis of BV. Detection of BV has been associated with inconsistent condom use and exposure to increased numbers of recent and lifetime sex partners, by meta-analysis [43]. Women with BV have an earlier median age of sexual debut than women without BV [53]. While 2 studies identified BV in “virgins,” both restricted the definition of ‘virgin’ to absence of a past history of penile-vaginal sex [54, 55]. In contrast, a study of young female students that collected detailed data on sexual behaviors found that BV was not detected in women without a history of sexual activity with others, was uncommon in women who reported having only ever engaged in noncoital sexual activities, and was significantly associated with the practice of penile-vaginal sex [56]. Consistent epidemiological data show high rates of concordance of BV within female partnerships [40, 57–60], and BV has been associated with practices that implicate sexual transmission between women, including increased number of female partners, a female partner with BV, and receptive oral sex [58, 60–62]. Marrazzo et al showed that women who have sex with women, in monogamous relationships, shared Lactobacillus strain types [63], and Vodstrcil et al found that incident BV was associated with exposure to a new female sex partner and a female partner with symptoms of BV in a 2-year cohort [60]. These data suggest that dynamic exchange of both protective and detrimental vaginal bacterial species is occurring between women in sexual relationships.

Published data also support the fact that reintroduction of BV-associated bacteria through sexual intercourse is likely to be contributing to BV recurrence after treatment. Treated women with ongoing exposure to an untreated male partner had twice the risk of recurrence, after adjustment for sex frequency, condom use, and contraception, in 2 studies [22, 64]. Several studies have found that inconsistent condom use during penile-vaginal sex increased the risk of recurrence following treatment [64–66]. Deep-sequencing studies have shown that the subprepucial space and distal urethra of males can harbor a broad range of BV-associated bacteria and that these bacteria are more prevalent among the male partners of females with BV than among those of females without BV [67]. The composition of the sulcus microbiota also appears to be strongly influenced by circumcision and sexual activity [68], with male circumcision prospectively associated with a significant reduction in BV-associated genera [69, 70] and a 40%–60% reduction in the development of BV in female partners [71]. BV-associated biofilm has recently been detected in male urine and semen and is more commonly found in the male partners of females with BV than in healthy controls [4, 72, 73].

Collectively, these findings provide evidence of carriage of BV-associated bacteria in males and broad support for the exchange of BV-associated bacteria within sexual partnerships between men and women, and women and women. They strongly suggest that sexual transmission is integrally involved in the pathogenesis of both incident and recurrent BV and highlight the need to fund and conduct sufficiently powered partner-treatment trials to determine whether this strategy reduces BV recurrence and associated sequelae. They also provide compelling evidence for an unforeseen benefit from male circumcision and for promotion of condom use in strategies to reduce the risk of BV acquisition and recurrence. There is 1 registered male-partner-treatment trial enrolling couples in North America, in which men are randomly assigned to receive a 7-day course of oral metronidazole versus oral placebo, and these data are eagerly awaited (ClinicalTrials.gov identifier NCT02209519). Trials involving the use of topical antimicrobials in addition to oral agents are planned and may be required to eradicate cutaneous carriage of BV-associated bacteria from the penile skin. Female-partner-treatment trials are clearly needed, and because they will be logistically challenging, the design of such studies will require innovative thinking and careful planning.

BIOFILM DISRUPTION

As summarized above, there is compelling evidence for the existence of pathogen-containing biofilm in the vagina of women and the urethra and subpreputial surface of the glans penis of males, which provides a rational explanation for microbial persistence after therapy [32, 73–76]. Moreover, biofilm has been shown to persist after putatively successful antimicrobial therapy in females [32, 77]. Biofilm may serve to reduce antimicrobial penetration, allowing antibiotic-susceptible microbes to persist, creating a carrier state. Biofilm sanctuary function may well explain the failure of effective antimicrobial therapy. Understanding biofilm production provides insight into expression of bacterial virulence factors [78–80]. It may be necessary to break down biofilm to achieve optimal efficacy of antimicrobial therapies, and novel strategies to eliminate biofilm have emerged that target women with recurrent BV.

The first evidence of a potential therapeutic benefit from biofilm disruption emerged in a multicenter, long-term study of maintenance suppressive therapy for the prevention of recurrent BV, in which Reichman et al used topical boric acid 600 mg daily following a 1-week course of systemic nitroimidazole therapy [81]. Previous unpublished studies performed by Sobel et al had demonstrated that boric acid alone was inadequate in even achieving a satisfactory clinical response in BV, reflecting its weak antimicrobial potency. In the study by Reichman et al, after a 1-month course of daily boric acid treatment, asymptomatic women were additionally prescribed suppressive twice weekly metronidazole for 4 months. The overall combination regimen dramatically reduced BV recurrence on treatment, although recurrence late after treatment was common [81]. Unfortunately, study design precluded objective evaluation of the unique contribution of boric acid. Ongoing research is also evaluating boric acid enhanced with an ethylenediaminetetraacetic acid excipient, which boosts antimicrobial activity and retains activity against vaginal biofilm [82].

These clinical studies, together with the biofilm studies by Swidsinski et al, have enhanced interest in the identification of more-potent agents capable of biofilm disruption, including octenidine, DNAses, retrocyclin, and the naturally occurring antimicrobials subtilosin, ploy-L-lysine, and lauramide arginine ethyl ester. There is also interest in quorum-sensing inhibitors [77, 83–86]. Octenidine and boric acid are the only agents to have been evaluated in human studies. Octenidine, an antiseptic with broad-spectrum antimicrobial activity, is effective against biofilms involved in oral, wound, and orthopedic-implant infections. Swidsinski et al recently reported on its use in 24 women with recurrent BV [77]. With application of octenidine daily for 7 days and, in women with further recurrence, daily for 28 days and weekly for 60 days, initial cure rates looked promising, but the efficacy of prolonged and repeated treatment was poor, and BV recurrence was observed in a significant proportion of women [77].

G. vaginalis biofilms contain extracellular DNA (eDNA), which is integral to their structural integrity, and enzymatic disruption of this eDNA by DNase has been shown to inhibit G. vaginalis biofilm formation and to disrupt established biofilms in vitro [84]. Hymes et al recently reported that low concentrations of DNase and metronidazole were more efficacious against G. vaginalis biofilm, compared with either agent alone, in vitro, perhaps due to increased susceptibility of G. vaginalis to the antibiotic when liberated from the biofilm [84]. The synthetic agent, retrocyclin (RC-101), an antimicrobial peptide with antiviral activity, has also been shown to inhibit the formation of G. vaginalis biofilms in vitro [83, 87]. RC101 is a potent inhibitor of vaginolysin, a toxin produced by G. vaginalis. Vaginolysin inhibition has been proposed as a potential strategy for BV treatment [83, 87], and while it is not clear whether the action of RC101 is entirely mediated through inhibition of vaginolysin, it may be another potential candidate for studies of BV in humans. A final, highly novel and emergent area of biofilm research involves inhibition of quorum sensing, a strategy that some bacterial species use to coordinate expression of genes involved in virulence, biofilm formation, and pathogenicity [88]. While quorum-sensing inhibitors have not yet been evaluated in human studies, they have been shown to be active in vitro against biofilms produced by Pseudomonas aeruginosa and Staphylococcus species [88, 89]. There has been much successful recent research on biofilm prevention and removal in the field of intravascular catheter and prosthetic device infections, using a variety of in vitro models. BV research lags behind, and one significant impediment has been the lack of a suitable animal model, with vaginal primate and other animal models being less optimal owing to significant differences in their microbiota and pH compared to humans [90, 91]. There is urgent need for progress in the development and evaluation of safe and effective topical vaginal agents capable of disrupting and eliminating genitourinary tract biofilm that could be included in combination therapy with antimicrobials.

ANTIMICROBIAL RESISTANCE

There remains scant knowledge regarding the role of antimicrobial resistance in contributing to both initial clinical and bacteriologic failures and recurrence [92]. Clinical experience indicates that the former is uncommon, whereas relapse after initial improvement is much more common. Beigi et al performed one of the few studies examining in vitro resistance to cultivatable BV-associated bacteria, concluding that, in contrast to clindamycin, anaerobic gram-negative bacterial resistance to metronidazole was rare [33, 34]. Similar in vitro studies that focus specifically on women with refractory or recurrent BV have not been performed. A major limitation of such studies in general is that the majority of BV-associated bacteria cannot be grown in vitro and thus cannot be tested for antimicrobial susceptibility. It should also be emphasized that BV represents a functional dysbiosis with a community of contributing pathogens, reflecting a model that is difficult if not impossible to evaluate in vitro by use of traditional methods of antimicrobial susceptibility testing.

Both G. vaginalis and A. vaginae demonstrate well-described intrinsic resistance to nitroimidazole agents. Nevertheless, in vivo these organisms are dramatically reduced in population numbers following conventional nitroimidazole therapy, indicating an effective in situ effect, possibly reflecting the effect of more-potent antimicrobial intermediate or degradation products [93]. It has always been assumed that the currently recommended regimens of oral and topical nitroimidazoles achieve concentrations in vaginal secretions far above concentrations needed to eradicate BV pathogens [94]. Nevertheless, some investigators have opined that higher concentrations may be more effective and essential, especially given the possibility of microbial pathogen persistence in biofilm. The use of available topical formulations of metronidazole result in vaginal fluid drug concentrations 10–30-fold higher than achievable with the oral drug, and several uncontrolled noncomparative pilot studies using higher-dose topical metronidazole have reported achieving improved cure and control rates [95, 96]. There is a need to perform studies comparing high-dose topical metronidazole with conventional dose commercially available metronidazole, especially given the availability and documented safety of the high-concentration/dose preparations.

CONCLUSION

Practitioners and patients alike widely recognize the current limitations in achieving successful therapy of BV. Therapeutic options remain extremely limited, and our inability to prevent frequent symptomatic recurrences of BV remains an acknowledged but unresolved shortcoming. New drugs have not been forthcoming and are not likely to be available in the near future. Thus, despite the limitations, we recommend adhering to current treatment guidelines rather than relying on unproven alternative regimens. Probiotics, although attractive as supplementary agents, have not consistently been shown to be advantageous. Our lack of progress is the direct result of our continuing incomplete understanding of the pathophysiology of this unique form of vaginal dysbiosis. On the other hand, increasing recognition of the potential role of pathogen-rich biofilm in facilitating disease persistence, together with the future introduction of effective antibiofilm agents, results in hope for improved therapeutic success in the future. It is also important to acknowledge the likelihood that reinfection from partners may be contributing to recurrence and may obscure the benefits of new therapeutic approaches. Rigorous partner-treatment trials, conducted in accordance with current clinical trial standards, are essential to not only determine the contribution of reinfection to recurrence, but also to provide an accurate evidence base for developing new treatment guidelines. Ultimately optimization of future BV treatment strategies may require combination approaches, such as antibiotics given along with biofilm-disrupting agents and in conjunction with partner treatment. Future research also should focus on the possibility that BV is a heterogeneous condition involving subtly different vaginal microbiomes. Such putative BV subtypes might respond differently to treatment regimens. In the field of HIV, ambitious goals have been set and result in intense coordinated efforts between clinicians and researchers, accompanied by advocacy from the public sector, to attempt to meet these targets. BV is a condition that is strikingly common but largely hidden from public view. Our collective goal needs to be the achievement of sustained cure for women if we are to reduce the global burden of this common and distressing condition and impact on its serious sequelae such as preterm delivery and HIV transmission.

Notes

Supplement sponsorship. This article appears as part of the supplement “Proceedings of the 2015 NIH/NIAID Bacterial Vaginosis Expert Consultation,” sponsored by the Division of Microbiology and Infectious Diseases of the National Institute of Allergy and Infectious Diseases in partnership with the University of Alabama at Birmingham Sexually Transmitted Infections Clinical Trials Group; contract HHSN272201300012I.

Acknowledgments. We thank the University of Alabama Clinical Trials Group for funding to attend a meeting at which some of this content was presented.

Potential conflict of interest. Both authors: No reported conflicts. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes