While a cervicovaginal mucosa covered with lactobacilli is still considered the optimal environment, molecular studies have shown that not all lactobacilli are equal [1]. Lactobacillus crispatus only occasionally co-occurs with organisms other than lactobacilli, has been associated with an anti-inflammatory cervicovaginal immune profile, and seems to protect women from developing anaerobic dysbiosis and from the above-mentioned adverse outcomes [1–5]. In contrast, L. iners does not seem to protect women from developing anaerobic dysbiosis and often co-occurs with BV-associated anaerobes, pathobionts (streptococci, staphylococci, or Enterobacteriaceae), or pathogens [1–6]. A vaginal microbiome dominated by L. iners is, however, not associated with a proinflammatory profile, and data on whether it increases the risk of adverse outcomes are conflicting. Vaginal microbiomes with a high relative abundance of the other vaginal lactobacilli are much less prevalent and less well studied [1,7].

Vaginal dysbiosis is often defined as a prolonged deviation from a low-diversity, lactobacilli-dominated vaginal microbiome. Molecular studies have identified different types of vaginal dysbiosis [1,7]. The most common type is high-diversity anaerobic dysbiosis, almost always including Gardnerella vaginalis and Atopobium vaginae as well as multiple other anaerobes, with or without a low relative abundance of L. iners. Low-diversity anaerobic dysbiosis, characterized by G. vaginalis or A. vaginae domination, also occurs, albeit less commonly. Studies employing multiple methods of vaginal microbiome characterization have shown high correlations between BV by Gram stain Nugent scoring and anaerobic dysbiosis (high and low diversity combined) [1]. Another type of vaginal dysbiosis that is likely important from a clinical point of view is a high relative abundance of pathobionts, also referred to as pathobiont carriage. Anaerobic dysbiosis and pathobiont carriage have been associated with proinflammatory immune profiles and with the above-mentioned adverse outcomes, and anaerobic dysbiosis has also been associated with cervicomucosal barrier disruption [3]. The roles of Bifidobacteriaceae (other than G. vaginalis) and Corynebacterium in the vaginal microbiome have not yet been studied, and domination by these bacteria is rare [1,7].

Many epidemiological studies have found that BV, VVC, vaginal pathobiont carriage, and STIs are interrelated and that many of the associations are bidirectional [8,9]. The interrelationships could be explained by both behavioral and biological factors. First, many of these conditions share risk factors related to sexual transmission. While BV, VVC, and vaginal pathobiont carriage were never considered STIs and can indeed occur in the absence of sexual activity, it is now clear that sexual transmission of the implicated organisms does play a role, especially in sex with uncircumcised male partners [2,10]. Second, most dysbiosis types and VVC cause mucosal barrier disruption, which decreases the ability of mucus and vaginal secretions to trap or inactivate pathogens and creates epithelial portals of entry, and cervicovaginal inflammation, which increases the concentration of target cells for HIV at the mucosal sites where HIV exposure takes place [3]. Interestingly, STIs and anaerobic dysbiosis often overlap, but VVC seems to occur more often in the presence of lactobacilli-domination than in the presence of anaerobic dysbiosis (Fig 1) [8]. A recent study showed positive associations between vaginal S. agalactiae carriage and vaginal Escherichia coli and Candida albicans carriage, but a negative association with anaerobic dysbiosis [9]. Many explanations for these associations have been hypothesized, such as vaginal pH (C. albicans, S. agalactiae, and E. coli are not inhibited by the low vaginal pH produced by lactobacilli), competition between micro-organisms for nutrients, microbial defense mechanisms against one another, biofilms that include some micro-organisms but not others, and attachment of some bacteria to Candida hyphae. Further in-depth characterization of these mechanisms is important because they may lead to new targets for drug development and an increased understanding of how intervening in one pathway might influence other pathways.

While first-line treatment of BV with oral or vaginal metronidazole or clindamycin is typically efficacious in the short term (as defined by Nugent or Amsel criteria), recurrence rates are high [10,11]. Clinical studies have shown that BV recurrence rates can be reduced by longer duration and/or prophylactic use of first-line antibiotics, by (estrogen-containing) hormonal contraception, and by circumcision of male sexual partners, but not by adding other antibiotics (azithromycin or moxifloxacin) to first-line antibiotics or by metronidazole/clindamycin treatment of male sexual partners [11]. Some argue that recurrence is particularly likely when a mucosal biofilm is present. In vitro and in vivo studies have shown that such a biofilm is damaged and suppressed by metronidazole but not completely eliminated [12]. The interrelationships between various urogenital conditions also pose challenges. For example, treatment of anaerobic dysbiosis often leads to VVC [13]. Treatments might be more efficacious in the longer term when they specifically target dysbiosis-associated anaerobes or pathobionts while sparing lactobacilli and are combined with biofilm disrupting agents, systemic or topical estrogen, and/or Lactobacillus-containing vaginal pro- or synbiotics. Estrogen-containing hormonal contraception, and Lactobacillus-containing vaginal pro- or synbiotics if found to be clinically effective, could also be implemented for routine use on a larger scale to prevent vaginal dysbiosis in women at risk.

While current knowledge suggests that maintaining lactobacilli-dominant, inflammation-free vaginal environments could advance the prevention of HIV, STIs, and adverse outcomes, this would not be an easy task, and many research questions remain. At the moment, the vaginal health of women is seldom routinely checked, even in pregnancy. HIV/STI screening programs targeting at-risk populations do exist, but otherwise, only women who seek medical care for urogenital symptoms are likely to be evaluated and treated, often in the absence of any diagnostic laboratory testing. Further, there is ample evidence that presumptive and syndromic management of urogenital conditions in the absence of any diagnostic testing have low sensitivity and specificity compared to diagnostic testing followed by treatment [14]. Even if diagnostic testing were to be introduced in order to optimize vaginal health and minimize complications, we currently do not know which women to target, when intervention would be required (i.e., which relative abundances or concentrations of G. vaginalis, A. vaginae, and pathobionts or which levels of cervicovaginal inflammation should be considered harmful), and which interventions, or combinations of interventions, would be optimal. However, the progress made in recent years has made it possible to start contemplating these issues and work towards solutions.