|Home | About | Journals | Submit | Contact Us | Français|
The various microbiota normally associated with the human body have an important influence on human development, physiology, immunity, and nutrition. This is certainly true for the vagina wherein communities of mutualistic bacteria constitute the first line of defense for the host by excluding invasive, nonindigenous organisms that may cause disease. In recent years much has been learned about the bacterial species composition of these communities and how they differ between individuals of different ages and ethnicities. A deeper understanding of their origins and the interrelationships of constituent species is needed to understand how and why they change over time or in response to changes in the host environment. Moreover, there are few unifying theories to explain the ecological dynamics of vaginal ecosystems as they respond to disturbances caused by menses and human activities such as intercourse, douching, and other habits and practices. This fundamental knowledge is needed to diagnose and assess risk to disease. Here we summarize what is known about the species composition, structure, and function of bacterial communities in the human vagina and the applicability of ecological models of community structure and function to understanding the dynamics of this and other ecosystems that comprise the human microbiome.
Bacterial communities in the human vagina are thought to have a critical role in protecting the host against infectious disease. In reproductive age women, it is thought they do so through the production of lactic acid resulting in a low pH environment that restricts the growth of pathogens and other opportunistic organisms.1, 2 Thus, maintaining high numbers of lactic acid bacteria is a hallmark of healthy conditions. Although there are marked differences in the species composition and rank-abundances of populations in vaginal bacterial communities3, 4 among women, it appears that all are probably dominated by homofermentative lactic acid bacteria.5 This suggests the ecological function of various vaginal communities in reproductive age women – creating a low pH environment through the production of organic acids – is conserved despite differences in the bacterial species present.
The vaginal microbial ecosystem undergoes significant structural changes at various stages in a woman’s life that are directly linked to the level of estrogen in the body.6 Initial colonization occurs at birth, when the infant is first exposed to her mother’s vaginal tract if delivered vaginally, or by the skin bacteria of persons handling the infant in the case of a Caesarian-section delivery.7 While little is known about the importance of this initial colonization event, it is believed to establish the gut, skin and vaginal microbiota, and in the weeks and months following birth, these differentiate into communities distinct to each habitat.7-9 During the first 2-4 weeks following birth, maternal estrogen mediates thickening of the vaginal epithelium and the production of glycogen that is fermented by indigenous bacteria resulting in a lowering of the vaginal pH. This is transitory however, as subsequent metabolism of maternal estrogen is accompanied by thinning of the vaginal mucosa, a reduction of the level of glycogen, and a concomitant increase in vaginal pH.6
During childhood (Tanner stage 110), the pH of the vagina is nearly neutral and cultivation-dependent methods have shown it to be colonized by diverse assemblages of aerobic, strictly anaerobic, and enteric species of bacteria.11-14 Between the ages of 8 and 13 years of age, pubertal changes in the vulva and vagina occur that are induced by adrenal and gonadal maturation. During the maturation process, follicular development causes estrogen production to rise, and once again this is accompanied by a thickening of the vaginal epithelium and intracellular production of glycogen.6 These new environmental conditions select for microorganisms capable of fermenting glycogen to lactic acid and the concomitant acidification of the vaginal environment that is characteristic of reproductive age women.6, 11 Remarkably, the shifts in microbial community composition that occur during this transition have seldom been studied. Using cultivation-dependent methods, Alvarez-Olmos et al.13 found that the vaginal microbiota of many adolescent girls (14-18 y) resembled those of adult women with bacterial vaginosis. Yamamoto et al.15 assessed the vaginal microbiota of adolescent girls (13-18 y) using cultivation-independent methods and observed that the bacterial communities were comparable to those found in adults but remarked that this may not be the case for premenarcheal or perimenarcheal girls.
When vaginal epithelial cells are sloughed in reproductive age women the glycogen present is then presumably metabolized by bacterial populations to produce organic acids; however, as widely cited as this mechanism is it is backed by very little evidence collected from in vitro analyses.16 The resulting low pH (4.0-4.5) of the vagina creates an environment that restricts or precludes the growth of many pathogenic organisms. However, with the onset of menopause, estrogen levels again decrease, and menstruation ceases. This is accompanied by atrophy of the vaginal epithelium and reduced cervico-vaginal secretions.6 In most menopausal women the vaginal microbiota is thought to shift from populations of lactic acid producing bacteria to an assortment of species that include strictly anaerobic and enteric bacteria.17, 18 The dynamic nature of this ecosystem underscores the importance of resolving its microbial constituents at different stages of human development and the prominent influence of estrogen levels in the host on the vaginal environment.
In general, the presence of high numbers of lactic acid bacteria in the vagina is often equated with “healthy” and low numbers, or absence thereof, as being “abnormal.”19-21 This has historically focused attention on members of the genus Lactobacillus as keystone species because of their well-known ability to produce lactic acid through the fermentation of sugars. This view originates from the earliest studies of the vaginal microbiota over a century ago, when Professor Albert Döderlein first reported culturing the bacteria from vaginal secretions. He found they produced lactic acid, which in turn inhibited growth of pathogens both in vitro and in vivo.22 “Döderlein’s bacillus” was later classified in 1928 as Lactobacillus acidophilus.23 Several decades later in the 1980s, it was determined that L. acidophilus was not a single species, but rather a group of closely related, obligately homofermentative species collectively known as the Lactobacillus acidophilus complex.24 Because species within this complex are difficult to distinguish phenotypically or biochemically,25 they were differentiated on the basis of DNA homology.26, 27 All of the Lactobacillus spp. found to be prevalent in the vagina today are members of this complex.
Following Döderlein’s discovery of what later came to be known as Lactobacillus, cultivation-dependent studies eventually revealed that a diverse array of facultative and strictly anaerobic bacteria, and sometimes the yeast Candida, can be present in the healthy vagina but typically in much lower numbers.28-30 Furthermore several species of Lactobacillus belonging to the Lactobacillus acidophilus complex were identified, including L. jensenii, L. casei, L. gasseri, L. crispatus, L. plantarum, L. fermentum, L. cellobiosus, L. brevis, L. minutus, and L. salivarius.31-34 It is interesting to note that L. minutus was reclassified in 1992 as a member of a new genus, Atopobium, and subsequently renamed A. minutum.35 Given the uncertainty and controversy surrounding the potential role of Atopobium species in bacterial vaginosis,36-38 future studies might seek to better understand the traits that distinguish these genera and the species within them.
Efforts to characterize vaginal microbial communities using cultivation methods undoubtedly led to significant improvements in understanding the role of microbes in vaginal health, but they were limited due to the inherent biases in cultivation methods. Today it is well known that most host-associated and environmental microbes resist cultivation in the laboratory using traditional techniques.39 Undoubtedly cultivation of microorganisms is fundamental to understanding their physiology and phenotypic characteristics, and it remains a very useful tool in studies of microbial ecology. Promising developments in cultivation of fastidious bacteria using state-of-the-art techniques40-43 are likely to enable the cultivation of many previously inaccessible microbes. However, studies aimed at assessing fine-scale variation in host-associated microbial communities within and among individuals or exploring ecological relationships within these communities require methods that provide detailed information about microbial diversity while also being cost-effective and scalable to high-throughput sample processing. In response to this need, cultivation-independent methods have in recent years become the standard approach to characterizing the diversity of microbes residing in and on the human body.44-48
Major advances in DNA sequencing technology over the last decade have fundamentally changed the way we assess microbial community structure and composition. For investigations of bacterial diversity, these methods commonly utilize 16S rRNA gene sequences as a means to compare and classify taxa. This approach circumvents the need for cultivation by analyzing DNA sequences extracted directly from samples. Typically, partial 16S rRNA gene sequences are amplified using primers that anneal to highly conserved sequences in the gene, and the resulting amplicons are sequenced. Phylogenetic analyses of the sequences allows for classification of phylotypes and determination of the numerically dominant taxa in a community. Other methods that rely on other conserved genes (cpn6049, rpoC, uvrB or recA50) have also been developed but are not as widely used.
Using these methods numerous studies have been done to characterize the vaginal microbial communities of healthy, asymptomatic, reproductive age women.3-5, 37, 51-55 Although these studies have relied on various analytical methodologies and study designs – sampling different regions of the vagina, women from various ethnic backgrounds, different geographical locations of populations, sampling times in relation to the menstrual cycle, and so on – their findings consistently demonstrate that vaginal bacterial community composition differs both within and between women, and several types of communities are known to exist. Together, these findings paint a much more complicated picture of the vaginal microbiota than had been considered in the past.
Previous studies performed in our laboratory have shown that several distinct kinds of vaginal communities with markedly different species composition occur in white, black, Hispanic and Asian women in North America3, 4 and Japanese women in Tokyo, Japan.5 Since vaginal bacterial communities differ in species composition they are likely to differ in how they respond to disturbances. Conceptually this is important since vaginal communities continually experience various kinds of chronic and acute disturbances caused by human behaviors such as the use of antibiotics, hormonal contraceptives and other methods of birth control, sexual intercourse, vaginal lubricants, douching and many others. In addition, the structure and composition of vaginal microbial communities are known to be influenced by natural changes in normal healthy women including aging,17, 56 time in the menstrual cycle,57 menstruation,58-61 pregnancy,62 and stress.63-65
In a previous study we analyzed 144 vaginal samples from healthy Caucasian and black women in North America.4 The results showed that 80% of the women had microorganisms phylogenetically related to Lactobacillus iners, L. crispatus, L. jensenii, or L. gasseri as a numerically dominant member of the vaginal microbiota. Overall, L. iners was the most common species of Lactobacillus in women of both ethnic groups having been recovered in 66% of the women sampled, and other groups have reported this organism as being highly common in the vaginas of reproductive age women as well.51, 52, 54, 66-76 Surprisingly L. iners was only first described in 199977 as it does not grow on the media typically used to isolate and enumerate Lactobacillus, so it was absent from earlier cultivation-dependent studies of the vaginal microbiota. The remainder of communities in our study contained a relatively low proportion of lactobacilli, exhibited greater species evenness and included high numbers of clones most closely related to Atopobium and genera of the order Clostridiales, including Megasphaera, Dialister, Anaerococcus, Finegoldia, Peptostreptococcus, and Eubacterium. Additionally, 20-30% of the clones from these communities were from novel clades in the phylum Firmicutes. Comparable results were obtained in a recent study of healthy, reproductive age Japanese women.5 The findings of these studies indicate there are a limited number of different kinds of vaginal microbial communities in asymptomatic, apparently healthy women. Moreover, from studies of adolescent women (13-15 y),15 it appears that these communities are established in puberty and may reside in women until menopause.
Recently, we completed a more detailed and expansive study to characterize vaginal microbiota using high-throughput methods based on pyrosequencing of barcoded 16S rRNA genes.4 The subjects were a cohort of 396 North American asymptomatic, reproductive age women equally representing four ethnic backgrounds (Asian, white, black, and Hispanic). We found a total 282 phylotypes among these women. Their vaginal bacterial communities were characterized into five groups, four of which were dominated by Lactobacillus iners, L. crispatus, L. gasseri, or L. jensenii, and the fifth which had lower proportions of lactic acid bacteria and higher proportions of strict and facultative anaerobes. The latter community type accounted for about 25% of the women sampled, a notable finding considering the prevailing view that high numbers of Lactobacillus are necessary for a healthy vaginal tract. Furthermore we observed high bacterial species diversity in all vaginal communities, even those in which the phylotype abundance distribution was highly skewed toward one or very few numerically dominant phylotypes.
An important finding from these studies is that the distribution of community types varies significantly among women from different ethnic backgrounds (Fig. 1). For example, in the study by Ravel et al.4 vaginal bacterial communities dominated by Lactobacillus spp. were found in 80.2% and 89.7% of Asian and white women, respectively, but just 59.6% and 61.9% of black and Hispanic women, respectively. On the other hand, occurrence of communities with low proportions or no detectable Lactobacillus species community type were elevated in Hispanic (38.1%) and black (40.4%) women compared to Asian (19.8%) and white (10.3%) women. These findings are in accordance with results obtained by Zhou et al.,3, 5, 53 who assessed the vaginal bacterial communities of white, black and Japanese women. Moreover, vaginal pH was found to differ among ethnic groups as well (Table 1), with the overall median vaginal pH of black (4.7 ± 1.04) and Hispanic (5.0 ± 0.74) women being slightly elevated over what is typically considered to be healthy (4.0-4.5). Vaginal pH was elevated (5.3 ± 0.6) among women of all racial groups in the “diversity group,” and it was above 4.5 for the community types dominated by L. gasseri (5.0 ± 0.7) and L. jensenii (4.7 ± 0.4). L. crispatus and L. iners-dominated community types had median pH values of 4.0 ± 0.3 and 4.4 ± 0.6, respectively.4
There is compelling evidence to suggest Lactobacillus spp., the production of lactic acid, and the resulting low pH are important for preventing the proliferation of nonindigenous organisms in the vagina.1, 28, 78-81 However, these observations have been over-interpreted and through faulty logic have led to the assertion and common wisdom that Lactobacillus spp. must be present for health to be maintained. This has been extended and some claim that women whose vaginal communities are depauperate of Lactobacillus spp. are somehow abnormal. Unfortunately, this fallacy is the premise of the commonly used Nugent criteria used for the diagnosis of bacterial vaginosis wherein the degree of “healthiness” is in part assessed by scoring the abundance of Lactobacillus morphotypes, ignoring the possibility that their ecological function could be supplanted by bacteria with other morphotypes. It might be more reasonable to postulate that a key ecological function of vaginal communities, namely the production of lactic acid, might be accomplished by a variety of taxa capable of homolactic and heterolactic fermentation of substrates. Atopobium, Streptococcus, Staphylococcus, Megasphaera and Leptotrichia are among the genera found in the vagina that possess this capability in addition to the better known Lactobacillus. If the maintenance of a low environmental pH is indeed a key function of the vaginal microbial community then perhaps it may be more appropriate to consider ‘lactic acid bacteria’ in toto to be members of the same ecological guild82 because they use the same resource pool to accomplish the same ecological function. If this is the case, then the prevailing view that species of Lactobacillus are both necessary and sufficient for maintaining health, may well be overly simplistic because functionally equivalent species may in fact ‘substitute’ for species of Lactobacillus.
Vaginal bacterial communities reside in an ecosystem that is strongly influenced by characteristics of the host, local environment, and constituent populations (Fig. 2). The human microbiome is often referred to as a commensal relationship in which one member derives benefit from the other member without providing benefit or causing harm in return. This is almost certainly not the case for the vaginal microbiome wherein the bacteria are entirely dependent on the host for nutrients, and in turn the bacterial communities play a role in protecting against disease-causing organisms. Consequently, it should be viewed as a mutualism in which an understanding of community composition, function, and dynamics requires that the vagina be viewed as an ecosystem and not simply the sum of its parts. For example, it is now clear that bacterial communities in the vaginas of reproductive age women are reasonably complex and include diverse species from several bacterial lineages. Hence, although vaginal bacterial communities are dominated by lactic acid producing bacteria, they coexist and interact with a wide array of other bacterial species by competition for space and resources. Through metabolic activities these populations modify their environment in ways that either facilitate or preclude colonization by other species through resource competition, predation by bacteriophage and the production of antimicrobial substances. Likewise the host influences the composition of communities by determining the quantity and composition of vaginal transudates that constitute an important source of nutrients for resident bacterial populations. Moreover, it is likely though not proven that vaginal mucus and epithelial cell receptors play an important role in colonization by certain bacterial species. Further, innate and local immune systems may work to exclude or select community members (reviewed by Linhares et al.2).
The available evidence suggests vaginal communities are in a state of dynamic equilibrium, in which short term fluctuations (at least in reproductive age women) occur in response to changes driven by hormonal changes that are a part of normal menstrual cycles (reviewed by Farage and Sobel83). It is unknown whether any stage in a menstrual cycle should be thought of as a “disturbed” state and therefore more susceptible to invasion. However, it is known that many normal human activities are associated with destabilization of vaginal communities. For example, frequent sexual intercourse, having multiple sex partners, and frequent episodes of receptive oral sex84-87 cause destabilization of vaginal microbial communities, as do douching88 and use of spermicides.89-92 Each of these has the potential to destabilize vaginal bacterial communities and increase their invasibility. Here we discuss an ecological model as it may apply to the vaginal microbiota and their stability. First we must emphasize that a relevant and meaningful ecological theory for the vaginal microbiome is currently not feasible until additional research on community dynamics has been done. However, while the model presented here is conceptual in nature it does provide a useful framework for evaluating the possible importance and roles of community members and guiding future studies in this field.
The strong linkage between high numbers of lactic acid bacteria and a “healthy” vaginal microbial community is consistent with Walker’s Driver-Passenger model of community structure and function.93 Under this model, species of lactic acid bacteria would be considered ‘drivers’ that strongly influence the function or structure of the ecosystem by producing lactic acid and maintaining a low pH. The environment thus created would be a strong determinant of community species composition and activity because they would all have to flourish or at least tolerate an environmental pH of 4.0-4.5. The non-lactic acid bacteria would be considered ‘passengers’ that are typically present at lower numerical abundance, may have little influence on the ecology of the system, and might be lost from the community or change over time without markedly affecting community function. Vaginal communities seem to conform to this model at least from a numerical perspective since the rank abundance of species is highly skewed and lactic acid bacteria often outnumber others by two orders of magnitude, with diverse kinds of organisms present in lower numbers.
The occurrence of functional redundancy among ‘drivers’ (i.e., multiple species of lactic acid bacteria) may impact the stability and resilience of a community in the face of disturbances that lead to changes in community structure or function. Such disturbances will differ in magnitude, frequency and source. Those disturbances that are small in scale or infrequent may not affect communities that have functional redundancy in driver species, because if one driver is disadvantaged or lost, another can compensate and thereby maintain community function. If, however, the disturbance is overwhelming or the community is driven by a single species, it may be more susceptible to change in function. This change in function could be harmful if it is necessary for health, or it can provide opportunities for colonization by pathogens.
Ecological theory and empirical data indicate that communities are not equally resilient or equally stable. The stability (or resistance) of a community reflects its capacity to resist change in structure or function in response to a disturbance event94 (Fig. 3), while resilience reflects the ability of a community to recover from a disturbance and return to a ‘quasi-stable’ equilibrium state94-96 (Fig. 4). The resistance of a community is reflected in the magnitude of change that can occur without having an impact on community function, whereas resilience is a measure of disturbance frequency or intensity that alters community function. Both the resistance and resilience of communities are largely determined by the ability of ecological networks of indigenous species to tolerate stresses and disturbance events, the physical, chemical and metabolic interactions among the species present, and the degree of functional redundancy present. A disturbance event is an environmental change that causes shifts in population densities, the gain or loss of species, and concomitant changes to community function.97 The response of any community to one or more disturbance events (or to a disturbance regime characterized by a distribution of disturbance sizes, frequencies, intensities, and timing) is determined by the attributes of component species.98 Disturbed communities may or may not return to their previous state.97 The ability of an ecosystem to buffer against perturbations and resist species invasions is dependent on the redundancy of species that have important stabilizing roles as well as their ability to differentially respond to perturbations. This ‘insurance hypothesis’ posits that increasing diversity increases the odds that at least some species will respond differentially to variable conditions and perturbations, and that greater diversity increases the odds that an ecosystem has functional redundancy by containing species that are capable of functionally replacing important species.94
Events that destabilize microbial communities are not equal in terms of intensity or duration. Communities with low resistance and resilience may be disturbed by a single intense event of short duration, or multiple events of moderate to low intensity. This can result in transitory changes to the structure of these communities rendering them more susceptible to invasion by species that are not indigenous to the human vagina including transient species of fecal origin and opportunistic pathogens. In contrast, robust vaginal communities will exhibit stability in the face of more frequent events of low to moderate intensity and retain ecological functions that are characteristic of healthy communities. Given this, we postulate that stability and resilience of vaginal bacterial communities is likely to vary widely since the species composition and structure of these communities differs among women and this in turn may account for differences in the susceptibility of individuals to urogenital infectious diseases. Nonetheless, this hypothesis has yet to be empirically tested. To date, most studies of vaginal microbiology have relied on cross-sectional study designs in which individuals are sampled at one time point or over regular intervals of a few weeks or months.57, 99-101 As a result, very little is known about the dynamics of vaginal microbiota over short time-scales and in response to various host-associated perturbations.
As discussed above, acidification of the vaginal milieu via lactic acid production is probably important to preventing the invasion of vaginal communities by nonindigenous organisms. It has been suggested that hydrogen peroxide may also help prevent such invasions since some vaginal species of Lactobacillus produce H2O2 in vitro.34, 85, 102-106 However recent studies have shown this is not likely the case in vivo since dissolved oxygen levels in the vagina are exceptionally low.1, 2, 107, 108 Other mechanisms such as production of bacteriocins have also been suggested80 but their importance has yet to be documented.
The establishment of a pathogen in a community closely mirrors the process of exotic species invasion in plant and animal communities. One concept from invasive species research that applies most readily is the relationship between disturbance and invasion. A review by Didham et al. (2005)109 examined whether invasive species were drivers of change in community composition or passengers that took advantage of a disturbance that had changed the community’s structure. The implication for medical microbiology is that without disturbance a pathogen maybe be able to invade some community types but not others. The second concept is the role of redundancy in driver species to maintain function and prevent invasion. Understanding the extent of functional redundancy and the contribution of specific community members to a healthy vaginal microbiota could help define risk factors for infection.
Ecologists have long known that the biological communities of disturbed ecosystems are more susceptible to invasion by non-indigenous, ‘weedy’ species.98 This is likely true for the bacterial communities of the human vagina too. If the proliferation of these invasive species proceeds unchecked, it could lead to clinically significant symptoms and disease. If the resilience of a vaginal community is low, then transitory changes to the structure of these communities may occur more readily in response to disturbances and these disturbed communities may be more susceptible to invasion by species that are not indigenous to the human vagina. These might include transient species of fecal origin and opportunistic pathogens.98 Moreover, we speculate that the disturbed state may itself constitute the clinical syndrome of BV where there is a reduction in the presence of lactic acid bacteria.
Evidence that host specific characteristics influence the species composition and dynamics of microbial communities that colonize the vagina is accruing, though direct evidence is lacking. As described previously, recent studies have shown that the vaginal communities of women can be classified into several types based on similarities in bacterial community composition.3-5 This can be viewed in two ways. One is that vaginal communities can show marked differences in the composition and rank abundance of species present,55, 110, 111 and these differences may be potentially important in terms of ecosystem resilience and resistance to infectious agents. On the other hand, it also demonstrates that the differences among women are apparently not boundless, and therefore colonization of the host and vaginal community composition are probably not random events. Said plainly, there appear to be host factors that facilitate or select for bacterial species with particular characteristics. These might be linked to the presence of epithelial cell surface receptors, variation in the composition or amount of vaginal secretions, the host immune system, or other factors (reviewed in Linhares et al.2 and Wira et al.112).
The notion that there is selection for particular bacterial species by host-determined characteristics is also supported by the observation that a rather limited number of different Lactobacillus species are found to dominate vaginal communities. Recent reports3-5, 15, 51, 53 suggest only four species of Lactobacillus, namely, L. crispatus, L. jensenii, L. gasseri and L. iners, are found as dominant members of vaginal communities. This is surprising given the plethora of different species of lactobacilli that are recognized26 and suggests that species of lactobacilli found in the vagina possess characteristics that allow them to compete and be successful under the environmental conditions of the vagina. If there is selection within an individual for species (or strains of species) that possess a suite of specific characteristics, then this could have important implications for efforts to develop prebiotics and probiotics for maintaining or re-establishing normal, healthy vaginal communities, because ideally they would be tailored to reflect differences in the species composition of an individual’s normal community.
Bacterial vaginosis (BV) is the most frequently cited cause of vaginal discharge and malodor and the most common vaginal disorder of reproductive age women, resulting in millions of health care visits annually in the United States alone.113, 114 Moreover, in non-pregnant women BV is associated with serious adverse sequelae including infertility,115 endometritis,116 and pelvic inflammatory disease,117 as well as an increased risk of acquiring HIV, Neisseria gonorrhoeae, and other sexually transmitted diseases.118-122 During pregnancy, BV is associated with several adverse outcomes including preterm delivery of low birth weight infants,123 spontaneous abortion,124, 125 premature rupture of membranes,126 preterm birth,126 amniotic fluid infections,127 postpartum endometritis,125 and endometritis following Caesarian section.128 The exact etiology of BV remains elusive, but different pathogenicity models have been proposed involving either the depletion or displacement of lactobacilli in the development of BV (Fig. 5).36 The prevalence of BV among women varies widely and depends on the subject population. It is present in 10% to 20% of white, non-Hispanic women and 30%-50% of African-American women, and it may occur in up to 85% of sex workers in Africa.113, 129-131 It has a prevalence of 5-26% in pregnant women worldwide.110
Despite over a century of work, attempts to find a single causative agent have failed. Since Duke and Gardner implicated Gardnerella vaginalis as a causative agent of BV (1955),132 numerous efforts have been made to associate BV with the presence of certain bacteria in hopes of identifying an infectious agent. In recent years investigators have used cultivation-independent methods to continue the search for organisms that might cause BV. Fredricks et al.133 used broad-range PCR and sequencing of 16S rRNA genes to find that women with BV had a relatively high prevalence abundance of bacteria such as Atopobium vaginae, Leptotrichia amnionii, Sneathia sanguinegens, Porphyromonas asaccharolytica, G. vaginalis, and novel members of the Clostridiales referred to as BV-associated bacteria (BVAB). Moreover, in a study by Ferris et al. broad range PCR assays were also used to characterize the vaginal microbiota before and after metronidazole treatment to find that the diversity of anaerobic bacterial types of BV flora was shifted to a predominantly L. iners flora in cured patients and that unresponsive patients had the highest concentrations of A. vaginae.71 While these studies corroborate the co-occurrence of BV with so-called BV-associated bacteria it is unclear if these bacterial species are causally related to the symptoms of BV or whether the association is contingent on the criteria used to diagnose BV. Studies in our lab and others have demonstrated these very same suspected agents and other closely related bacteria are found in asymptomatic women though perhaps at somewhat lower number (Fig. 6). This argues that these organisms might not be ‘infectious agents’ in the strict sense, but when present in high number they might elicit some or all of the symptoms classically associated with bacterial vaginosis. That said, it is risky to deduce that certain organisms cause BV simply because they are abundant when symptoms are present.
Others have pursued the question of whether BV is a sexually transmitted disease. Evidence supporting this notion comes from the observed concordance of BV status in monogamous lesbian couples ranges up to 95%.30, 134, 135 Also, women with BV have more sex partners and an earlier age of sexual debut than women without BV,84, 86 and an association between receptive oral sex and BV has been suggested.84 However, BV has been diagnosed in asymptomatic virginal women,136-138 and this begs the question of what, if any, role sexual behaviors may have in the acquisition of BV. On balance the data available support the hypothesis that BV is not an infectious disease.
Over the years, the definition of BV and the diagnostic criteria commonly used have been conflated, and they remain mired in controversy. The Amsel test, which is often used for the clinical diagnosis of BV, is based on four criteria: (a) a vaginal pH of >4.5, (b) the presence of clue cells, (c) a fishy odor upon addition of 10% KOH to vaginal discharge, and (d) a white, thin, homogenous vaginal discharge.139 The diagnosis of BV is made if at least three of these criteria are confirmed. The gold-standard for the diagnosis of BV in research and laboratory settings has been the Nugent score.140 This diagnostic test is a scored scale based on (1) the presence of Gram-positive rods (Lactobacillus morphotypes) (2) the presence of Gram-variable rods and cocci (Gardnerella vaginalis, Prevotella, Porphyromonas, and peptostreptococci morphotypes) and (3) the presence of curved Gram-variable rods (Mobiluncus spp. morphotypes). In a formal sense, an obvious potential problem is the logic of the Nugent score premise that high numbers of Lactobacillus spp. define “health,” and this imposes a bias against normal vaginal microbial communities that lack appreciable numbers of lactobacilli, yet maintain a low pH. The Amsel test, on the other hand, may lack sensitivity due to the subjectivity of the clinician’s interpretation. Reports comparing the two diagnostic measures arrive at opposing conclusions,141, 142 which has led many to suspect the accuracy of these tests.
Two fallacies permeate thinking about the diagnosis and treatment of BV. The first is that BV is an infectious disease. This seems to arise from classical thinking about infectious diseases in the framework of Koch’s postulates, wherein a single species is both necessary and sufficient to cause infection. Although this is certainly true for a number of pathogens, it may be inadequate as an explanation for other diseases caused by mixtures of organisms. These so-called polymicrobial infections may not be infections in the strict sense of the word, wherein a nonindigenous organism invades a community. Instead they may be caused by indigenous populations that are typically rare but become abundant due to changes to ecologically important characteristics modulated by the host (e.g., nutrient levels) or disturbances that alter the competitive dynamics of bacterial populations. In other words, the incidence of infectious disease may often depend not only on the competitiveness of an invasive species (i.e., the infectious agent) but also on the ecological dynamics of the habitat (i.e., anatomical site). To us it seems there is much to be gained from borrowing and testing ecological theory that has been developed over decades by plant and animal ecologists who have studied invasive species in a wide variety of circumstances.
A second fallacy is directly tied to the notion that the vaginas of normal healthy women are populated by high numbers of Lactobacillus spp. This statement is accurate so far as it goes. However, the converse statement – that women whose vaginal communities have few or no Lactobacillus spp. – are abnormal is unsupported by data. We postulate that because of this logical fallacy, BV is often over-diagnosed. This could partly account for the reported high incidence of so-called ‘asymptomatic’ BV in reproductive age women,143, 144 and also explain a proportion of BV treatment failures and apparent recurrences of BV in women. Acknowledging that not all vaginal communities of healthy women are dominated by Lactobacillus spp. would also be in accordance with the observation that the vaginal communities of post-menopausal women (not taking hormone replacement therapy) often lack appreciable numbers of Lactobacillus spp., yet these individuals do not exhibit other untoward symptoms. We suspect that the causes and cures of BV will continue to be enigmatic until it is recognized that while ‘normal and healthy’ can be equated with high numbers of lactobacilli, the converse statement (“unhealthy” is equated with low numbers of or no lactobacilli) is not necessarily true. We must be vigilant and realize that for a significant proportion of women ‘normal and healthy’ can also occur in the absence of appreciable numbers of Lactobacillus spp.
Multiple kinds of normal vaginal microbial communities are found in healthy women. The data provide strong evidence that more species than lactobacilli can dominate the vaginal microbial ecosystem of healthy women. The community function of maintaining low pH is highly conserved among women despite the difference in their vaginal microbial community composition and structure. These communities are postulated to provide different levels of protection against disease and infection, and their ability to offer protection may be lessened if the communities are disrupted. We propose that all vaginal microbial communities are not equally resilient and their stabilities differ in the face of disturbances. Moreover, differences in the resilience of various vaginal microbial communities may account for the differential susceptibility of races to HIV, BV and other urogenital infectious diseases. Obviously our knowledge of the factors that affect and control the vaginal microbiota is still incomplete and increasingly we should view the vagina as a microbial ecosystem so that we can better understand the full range of factors that affect risk to disease.
This work was supported by grants UO1 AI070921 from the National Institute of Allergy and Infectious Diseases and UH2 AI083264 from the Human Genome Research Institute of the National Institutes of Health. We also wish to acknowledge the contributions of past and present members of the Forney and Ravel research teams who have contributed to our research on the ecology of the vaginal microbiome, especially Zaid Abdo, Stephen Bent, Rebecca Brotman, Pawel Gajer, and Sam Ma.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure of Potential Conflicts of Interest
The authors have reviewed the journal’s policy on disclosure of potential conflicts of interest and confirm they have none to declare.
Hickey and Forney wrote the manuscript, with intellectual contributions and editorial suggestions from Pierson, Zhou and Ravel.