# Understanding Bacterial Vaginosis: How One Bacterium (Gardnerella vaginalis) Starts the Infection

Bacterial vaginosis (BV) is the most common cause of vaginal discharge and is linked to serious reproductive health complications, including preterm birth and increased risk of HIV and other sexually transmitted infections — yet its exact cause has been controversial for decades. This review article proposes a conceptual model showing that BV is actually initiated by the sexual transmission of a single pathogen, *Gardnerella vaginalis*, which has the unique ability to attach to vaginal cells, form a protective biofilm, and outcompete the healthy lactobacilli that normally keep the vagina healthy. The researchers present decades of evidence, including fulfilled Koch's postulates, demonstrating that *G. vaginalis* is not part of the normal vaginal flora but rather a sexually transmitted pathogen that paves the way for other anaerobic bacteria to overgrow and cause the symptoms of BV.

# Understanding Bacterial Vaginosis: How One Bacterium (*Gardnerella vaginalis*) Starts the Infection

## Table of Contents

- Key Points
- What Is Bacterial Vaginosis (BV)?
- How the Researchers Developed This Model
- The Healthy Vaginal Microbiome: How It Normally Develops
- G. vaginalis in the Female Genital Tract
- G. vaginalis in the Male Genital Tract
- How G. vaginalis Attaches to the Body's Cells
- The Mechanisms of G. vaginalis Infection
- The BV Syndrome: Symbiosis in the Vaginal Microbiome
- Genetic Diversity: Are There "Good" and "Bad" Strains?
- Koch's Postulates: Proof That G. vaginalis Causes BV
- What This Means for Patients and Treatment
- Study Limitations
- Recommendations for Patients
- Frequently Asked Questions
- Source Information

## Key Points

- BV is initiated by sexual transmission of Gardnerella vaginalis, not by poor hygiene.
- G. vaginalis forms a protective biofilm and outcompetes healthy lactobacilli, allowing other anaerobes to overgrow.
- BV is linked to preterm birth and increased risk of HIV and other sexually transmitted infections.
- Recurrent BV may be due to reinfection from a partner or biofilm resistance to standard antibiotics.
- Treatment strategies that specifically target G. vaginalis and its biofilm should be tested to improve prevention and treatment.

## What Is Bacterial Vaginosis (BV)?

Bacterial vaginosis is the most prevalent cause of symptomatic vaginal discharge worldwide, affecting millions of women each year. It's not just an uncomfortable nuisance, though. BV is associated with serious complications of reproductive health, including preterm birth, as well as the acquisition or transmission of sexually transmitted infections (STIs), including human immunodeficiency virus (HIV) infection.

Despite its prevalence and impact, the precise cause of BV has remained unknown for over 60 years. This lack of understanding has had real-world consequences: current treatment regimens and prevention strategies are inadequate, and recurrence rates remain frustratingly high.

The mystery dates back to 1955, when researchers Gardner and Dukes first wrote: "We are prepared to present evidence that the vast majority of so-called 'non-specific' bacterial vaginitides constitute a specific infectious entity caused by a single etiological agent... We have assigned the name *Haemophilus vaginalis* to this newly isolated bacillus." This organism was later renamed *Gardnerella vaginalis*.

At the heart of the controversy is a simple observation: BV is characterized by dramatic changes in the vaginal flora. A healthy vagina is dominated by **lactobacilli** (beneficial bacteria that produce hydrogen peroxide), but in BV, these protective bacteria drop dramatically and are replaced by *Gardnerella vaginalis* and a marked increase in strictly anaerobic bacteria (bacteria that grow without oxygen).

This review, published in *The Journal of Infectious Diseases*, aims to settle the debate by presenting a clear conceptual model for how BV develops — a model that puts *G. vaginalis* at the center as the initiating pathogen.

## How the Researchers Developed This Model

This was not a traditional clinical trial or laboratory experiment. Instead, the authors conducted a comprehensive review of past and current research spanning more than 50 years of published studies. They examined evidence from epidemiology (how BV spreads in populations), microbiology (how bacteria behave in the lab), genomics (the genetic makeup of bacteria), and clinical studies (how BV presents in patients).

By synthesizing this vast body of evidence, they developed a **conceptual model** for the pathogenesis (disease development) of BV. Their model integrates findings on how the vaginal microbiome develops from birth, how *G. vaginalis* is transmitted sexually, how it adheres to and invades vaginal tissue, how it forms biofilms, and how it interacts with other vaginal bacteria to produce the full syndrome of BV.

The epidemiology of BV strongly indicates that it is acquired via sexual transmission — a key pillar of the model presented here.

## The Healthy Vaginal Microbiome: How It Normally Develops

To understand how BV happens, it helps to understand what a healthy vagina looks like. The vaginal flora is acquired shortly after birth from maternal and environmental sources.

In a landmark study of prepubertal girls, Hill and colleagues demonstrated the presence of anaerobic organisms in the vaginal fluid of the majority of girls, yet they were unable to demonstrate the presence of *G. vaginalis* by culture — even though it was specifically sought. This finding is important: if *G. vaginalis* were a normal part of the vaginal flora, it should be present in girls before puberty too, but it is not.

At puberty, with the production of **estrogen**, the vaginal flora changes to lactobacillus predominance. Estrogen promotes the deposition of **glycogen** (a stored form of sugar) in the vaginal epithelium (the lining of the vagina). This glycogen is used as a food source by the saccharolytic (sugar-fermenting) lactobacilli. The subsequent creation of **lactic acid** as their metabolic end product lowers the vaginal pH to below 4.5, creating an acidic environment that is hostile to many harmful bacteria.

In addition, the growth of lactobacilli increases the **reduction-oxidation (redox) potential** of the vagina — a measure of the electron activity in the environment — which inhibits the growth of indigenous anaerobes (bacteria that thrive without oxygen).

In short: the healthy vagina is acidic, dominated by lactobacilli, and actively suppresses anaerobic bacteria. *G. vaginalis* is not part of this healthy picture.

## G. vaginalis in the Female Genital Tract

Epidemiologic data strongly support the sexual transmission of *G. vaginalis*. The bacterium has been recovered from vaginal fluid in close to **100% of women with clinically diagnosed BV**.

Other frequently identified BV-associated bacteria, recovered at variable rates in the vaginal microbiome, include genital mycoplasmas and various strict anaerobes, including species of BVAB1, BVAB2, BVAB3, *Atopobium*, *Leptotrichia*, *Megasphaera*, *Prevotella*, and *Dialister*. However, the assortment of strict anaerobes found in individuals with BV is heterogeneous — different women harbor different combinations.

Several published studies have detected *G. vaginalis* in women who did not meet the clinical criteria for BV. For example, Teixeira et al found *G. vaginalis* in only **17.6% of "healthy" women** without a clinical diagnosis of BV, but notably, there was no documentation that these women had none of the Amsel criteria (the standard clinical diagnostic criteria) or an optimal Nugent score of 0–3 (the standard microscopic scoring system). Similarly, Burton et al reported *G. vaginalis* present in **19% of women without BV** but provided no Nugent score at all.

The authors argue that finding *G. vaginalis* in women without clinical evidence of BV — or possibly with intermediate flora (Nugent scores of 4–6, which meet neither the definition of BV nor normal flora) — likely represents **asymptomatic infection**, just as is seen with all sexually transmitted diseases.

As the authors put it: "If *G. vaginalis* were a part of the normal vaginal flora, one would expect it to be present in all women and in prepubertal girls as well." It is not. The specific host-pathogen interactions that limit the virulence of the colonizing pathogen in these asymptomatic cases are still poorly understood.

## G. vaginalis in the Male Genital Tract

For a sexually transmitted infection to be credible, the pathogen must be found in men as well. Indeed, *G. vaginalis* has been recovered from the urethra and from seminal fluid in several studies. The presence of BV-related microorganisms in the male genital tract suggests a possible reservoir and supports the theory of sexual transmission of BV.

Some researchers had hypothesized that semen itself — because of its alkaline (basic) properties — might alter the acidic pH of the vagina and lead to BV. However, documented sexual transmission of BV without exposure to semen in a heterosexual couple strongly suggests that it is not the semen itself that contributes to the development of BV, but rather the microbes transmitted via sexual activity. This is further supported by the apparent sexual transmission of BV among women who have sex with women (WSW) only.

Using sophisticated sequence variation techniques, Eren et al were able to show that **sexual partners shared the same strains of *G. vaginalis*** — a powerful piece of evidence for sexual transmission.

Where does *G. vaginalis* live in men? Insight comes from a study of the microbiome of adolescent boys. Nelson et al found *G. vaginalis* in **28% of urine samples** but failed to detect it in samples from the coronal sulcus (the groove behind the head of the penis). This makes biological sense: in females, BV is an infection of the **squamous epithelium** (flat, scale-like cells), not the columnar epithelium of the cervix. The male genital tract's only squamous epithelium is the distal urethra.

Holst showed that BV-related organisms, including *G. vaginalis*, seem to transiently colonize male partners of women with BV. As with *Trichomonas vaginalis* in men (another pathogen that infects the squamous epithelium of the vagina), there is probably a high spontaneous resolution rate in men due to the inhospitable environment of the distal urethra.

BV may be more common among WSW than in heterosexual women. Sexual exchange of infected vaginal fluid may be a more efficient mechanism for transmission between WSW than the behaviors that occur during heterosexual sex — which would account for the higher prevalence of BV in this population.

## How G. vaginalis Attaches to the Body's Cells

Once transmitted, how does *G. vaginalis* actually start an infection? Examination of vaginal biopsy specimens demonstrates that BV is a **biofilm community** adherent to the vaginal epithelium — and *G. vaginalis* is the **predominant component** of that biofilm mass.

The initial steps of establishing infection include three critical capabilities:

1.
**Adherence** to host receptor sites on the vaginal cells
1.
**Production of cytotoxic substances** (toxins that damage host cells) specific for human cells
1.
**Biofilm formation** (creating a protective community of bacteria)

In the case of *G. vaginalis*, production of **vaginolysin** — a cholesterol-dependent cytolysin (a toxin that punches holes in cells) — is species-specific for human cells. Vaginolysin encodes a pore-forming toxin that binds to the **CD59** human complement regulatory molecule (a protein that normally protects cells from the immune system's "complement" attack). This cytotoxin assists in the initial adherence of *G. vaginalis* to host epithelial cells.

Researchers have recently compared the virulence factors of *G. vaginalis* against other BV-associated bacteria. A key study by Patterson et al examined adherence, biofilm formation, and cytotoxicity in the lab for *G. vaginalis* strains isolated from women with BV, as well as other BV-associated bacteria including *Atopobium*, *Prevotella*, and *Mobiluncus*. The results were striking: **only *G. vaginalis* demonstrated all 3 virulence factors.** The authors suggested that the other organisms may be relatively avirulent opportunists that colonize after the infection has been initiated by *G. vaginalis*.

Furthermore, recent work by Machado et al found that, in the lab, among BV-associated bacteria, *G. vaginalis* had the **greatest capacity to adhere to epithelial cells** even in the presence of *Lactobacillus crispatus* — one of the most protective species of lactobacilli found in healthy vaginas.

## The Mechanisms of G. vaginalis Infection

Subsequent to initial adherence to the host cell, the invading bacteria multiply and may produce a biofilm community as a means of future survival. Biofilm production is critical to the survival of *G. vaginalis* in the vagina. **Sialidase**, an enzyme produced by some strains of *G. vaginalis*, may enhance the production of biofilm through its **mucinase activity** (breaking down the protective mucus layer of the vagina).

Laboratory studies have shown that biofilms of *G. vaginalis* are **far more tolerant of lactic acid and hydrogen peroxide** (the very substances lactobacilli use to keep the vagina healthy) than are free-floating (planktonic) forms of *G. vaginalis*. This means once the bacteria form a biofilm, they become much harder for the body's natural defenses — and for antibiotics — to eliminate.

Once this critical mass of bacteria is present, **quorum sensing** (or bacterial "cross-talk") occurs. This is a sophisticated communication system that allows bacteria to coordinate their behavior as a community, effectively managing their new territory. Quorum sensing allows the bacteria to act together, turning on genes that help them survive and thrive.

At the same time, it is necessary for the invading pathogens to engage in "chemical warfare" with the native species. In the case of BV, the native species is lactobacilli. Research has shown that:

- Certain species of lactobacilli can produce inhibitory compounds (**bacteriocins**) that are active against *G. vaginalis*

- Conversely, *G. vaginalis* produces bacteriocins that are active against lactobacilli in the lab
-
*G. vaginalis* also demonstrates **antibiosis** (antagonistic interactions) via other, undetermined mechanisms

Back in 1991, Naggy et al published their work on antibiosis among vaginal organisms and concluded: "It is probable that the growth inhibition of lactobacilli caused by certain strains of *G. vaginalis* is one of the first steps in the change in flora characteristic of BV." This statement from over 30 years ago remains a cornerstone of the current model.

Evidence for direct competition between lactobacilli and *G. vaginalis* can be demonstrated from daily vaginal Gram stains of women with intermediate vaginal flora. In these women, lactobacilli and *G. vaginalis* visibly alternate as the dominant organism throughout the menstrual cycle.

Notably, *G. vaginalis* dominates **at the time of menses** (menstrual bleeding), suggesting competition for **iron** as a substrate for the bacteria. It is known that heme (the iron-containing component of blood) favors the growth of proteolytic organisms, such as *G. vaginalis*. The observed fluctuations in the levels of lactobacilli and *G. vaginalis* during the cycle most likely represent competition for resources between these two groups of bacteria.

## The BV Syndrome: Symbiosis in the Vaginal Microbiome

After *G. vaginalis* establishes itself, a cascade of events unfolds. *G. vaginalis*, as a **facultative anaerobe** (able to survive with or without oxygen), can tolerate the relatively high oxidation-reduction (redox) potential of a healthy vaginal microbiome — unlike strict anaerobes.

Similar to facultative anaerobes involved in the initiation of oral diseases (like gum disease), *G. vaginalis* begins the process of creating a **lower redox potential** — a chemical environment that is then suitable for the overgrowth of strict anaerobes, which are normally present in very low numbers.

But *G. vaginalis* doesn't just change the environment; it actively feeds the anaerobes. Here's how the symbiotic relationship works:

-
*G. vaginalis* is a **proteolytic bacterium** (protein-digesting) that produces **amino acids** through its metabolism
-
*Prevotella bivia*, a strict anaerobe, uses these amino acids as its fuel source
- As *P. bivia* consumes the amino acids, it produces **ammonia** as a byproduct
-
*G. vaginalis* then uses that ammonia for its own growth

This creates a two-way feeding loop between *G. vaginalis* and the anaerobes. Moreover, this symbiotic relationship with the production of ammonia would cause a shift to a more **alkaline pH** (less acidic), which is inhospitable to lactobacilli.

In vitro work has confirmed this: the addition of anaerobes to a *G. vaginalis* biofilm **enhances the growth of *G. vaginalis***. This synergistic relationship is responsible for the microbiological findings that define BV, as well as the typical clinical vaginal signs:

-
**"Clue cells"** — sloughing of vaginal epithelial cells coated with bacteria, visible under a microscope
-
**Amine odor** — the classic fishy smell resulting from metabolic by-products of the increased numbers of BV-associated anaerobes

## Genetic Diversity: Are There "Good" and "Bad" Strains?

Not all *G. vaginalis* may be equally dangerous. Genomic sequencing has recently shown differences in virulence factors among strains of *G. vaginalis*.

Harwich et al examined virulence factors for a strain of *G. vaginalis* from a woman with BV and another from a woman without BV. They found **impaired adherence in the non-BV isolate** and suggested that there may be both commensal (harmless) and pathogenic (disease-causing) strains of *G. vaginalis*.

However, the authors of this review note an important caveat: as in other studies, there is no mention of the Amsel criteria or Gram stain characteristics of the woman without BV. It's possible she had intermediate flora or asymptomatic infection. Furthermore, this type of work needs to be replicated with multiple isolates before firm conclusions can be drawn.

A recent comparative genomic analysis of **17 clinical isolates** of *G. vaginalis* suggested that the species can be subdivided into **4 distinct clades** (genetic groups) — or perhaps that there may even be multiple separate species of *G. vaginalis*. Ahmed et al found that the degree of diversity among the strains was **exceptionally high for a single species** — remarkable genetic variety that may explain why some women develop symptomatic BV while others carry the organism without symptoms.

## Koch's Postulates: Proof That G. vaginalis Causes BV

One of the most powerful arguments in favor of *G. vaginalis* as the cause of BV comes from **Koch's postulates** — the gold standard criteria for proving a causal relationship between a microbe and a disease. Developed in 1884, these four conditions must be met:

1. The microorganism must be found in abundance in all organisms suffering from the disease and not present in those without it
1. The microorganism must be isolated from a diseased organism and grown in pure culture
1. The cultured microorganism should cause disease when introduced into a healthy organism
1. The microorganism must be reisolated from the inoculated, diseased experimental host and identified as being identical to the original specific causative agent

In the case of *G. vaginalis* as the causative agent for BV, the authors argue these postulates have been fulfilled:

-
*G. vaginalis* is found in **nearly 100% of cases of BV**

- It is isolated as the **predominant bacteria** from women with BV, along with small amounts of indigenous vaginal flora
- Most importantly, **Criswell et al caused clinical BV in healthy women** who had negative cultures for *G. vaginalis* and no clinical signs of BV, by vaginal inoculation with a pure culture of *G. vaginalis* in its logarithmic growth phase
- They then **reisolated *G. vaginalis*** from these women after the initiation of BV

To address the objection that *G. vaginalis* has been isolated from women not meeting the Amsel criteria for BV, the authors point out that this does not mean those women are without disease. In many instances, the causative organism is present in the host but not causing overt disease — just as with asymptomatic carriage of STIs and gastrointestinal pathogens such as *Salmonella typhi* (the cause of typhoid fever).

## What This Means for Patients and Treatment

This conceptual model has profound implications for how BV should be understood, prevented, and treated. If BV is truly initiated by the sexual transmission of *G. vaginalis*, then:

-
**Prevention strategies** should focus on preventing transmission of this specific pathogen, much like we prevent transmission of other STIs
-
**Treatment strategies** should be specifically designed to eliminate *G. vaginalis* and its biofilm, rather than just addressing the overgrown anaerobes that cause symptoms
-
**Partner treatment** may need to be reconsidered — as with any STI, treating only the index patient may lead to reinfection from an untreated partner
-
**Biofilm-targeted therapies** are needed, since *G. vaginalis* biofilms are far more tolerant of lactobacilli's natural defenses (lactic acid and hydrogen peroxide) than free-floating bacteria — a finding that likely extends to antibiotic tolerance as well

The authors conclude: "BV is an important public health issue, yet its pathogenesis remains controversial. Its epidemiology strongly favors the hypothesis that it is a sexually transmitted infection. A significant body of data supports the pathogenicity of *G. vaginalis*. Studies have confirmed that it is the dominant component of the BV biofilm, strongly suggesting that other bacteria in the biofilm are opportunistic secondary intruders."

The presence of virulence factors in *G. vaginalis*, shown to be present over 30 years ago, has been recently confirmed by several studies. The authors recommend that "prevention and treatment strategies specifically for *G. vaginalis* should be tested to determine their efficacy in decreasing rates of BV and its complications."

## Study Limitations

As a review article proposing a conceptual model, this paper has inherent limitations that patients should understand:

- The model is based on synthesis of existing research rather than new experimental data
- Some studies cited did not rigorously define "normal" vaginal flora, making it difficult to interpret findings of *G. vaginalis* in asymptomatic women
- The genetic diversity studies suggesting "commensal" versus "pathogenic" strains of *G. vaginalis* have not been replicated with multiple isolates
- The specific host-pathogen interactions that limit virulence in asymptomatic carriers remain poorly understood
- While Koch's postulates have been fulfilled for *G. vaginalis* alone, the contribution of the complex polymicrobial community in established BV is not fully explained by this model alone
- The authors did not disclose any conflicts of interest, and the paper underwent standard peer review

## Recommendations for Patients

Based on this research, here's what patients should know and consider:

1.
**BV is not a "normal" condition.** It is an infection, and the leading evidence suggests it is sexually transmitted — not caused by poor hygiene, as was once believed.
1.
**Recurrent BV may be related to reinfection.** If you have a regular partner, it may be worth discussing BV with your healthcare provider as a potentially sexually transmitted condition that might merit partner evaluation.
1.
**Women who have sex with women are at increased risk.** BV is more common in WSW, likely because exchanging vaginal fluid is an efficient transmission route. This has implications for safer-sex practices between female partners.
1.
**Current treatments may be incomplete.** Because *G. vaginalis* forms a protective biofilm that standard antibiotics may not fully penetrate, recurrence is common. If your BV keeps coming back, ask your doctor about biofilm-disrupting strategies or extended treatment regimens.
1.
**The health implications go beyond discomfort.** BV is linked to preterm birth and increased risk of HIV and other STIs. Treating and preventing BV is not just about symptom relief — it's about protecting your overall reproductive health.
1.
**Ask your provider about the evidence.** Not all healthcare providers are aware of the latest research on *G. vaginalis* and biofilms. Bringing printed information from reputable sources can help facilitate a more informed conversation about your treatment options.

Ultimately, this review makes a compelling case that BV is not a vague "imbalance" of vaginal bacteria, but a specific infectious disease caused by a single initiating pathogen — *Gardnerella vaginalis* — with all the public health implications that come with that designation.

## Frequently Asked Questions

### What causes bacterial vaginosis (BV)?

BV is triggered by sexual transmission of a single bacterium, Gardnerella vaginalis. It attaches to vaginal cells, creates a protective biofilm, and outcompetes healthy lactobacilli. Other anaerobic bacteria then overgrow, causing the symptoms of BV. This is a specific infection, not just a vague imbalance.

### Is BV sexually transmitted?

Yes, strong evidence indicates BV is sexually transmitted. The bacteria G. vaginalis is found in the male genital tract, sexual partners often share the same strain, and BV occurs in women who have sex with women. Even transmission without semen exposure has been documented.

### Why does my BV keep coming back after treatment?

G. vaginalis forms a biofilm that is much more tolerant of natural defenses like lactic acid and hydrogen peroxide, and likely also more tolerant of antibiotics. Recurrent BV may also result from reinfection by an untreated partner. Ask your doctor about biofilm-disrupting strategies or extended treatment regimens.

### Is Gardnerella vaginalis part of a healthy vaginal microbiome?

No. G. vaginalis is not found in healthy prepubertal girls, and it is present in nearly 100% of women with clinically diagnosed BV. It can sometimes be found in women without BV, but that likely represents asymptomatic infection, just like other sexually transmitted infections.

### Are women who have sex with women at higher risk for BV?

Yes, BV appears more common among women who have sex with women than in heterosexual women. This is likely because exchanging vaginal fluid is a more efficient way to transmit G. vaginalis. Safer-sex practices between female partners may help reduce this risk.

### Should my partner be treated if I have BV?

Because BV is sexually transmitted, treating only you might lead to reinfection from an untreated partner. The article suggests that partner treatment may need to be reconsidered. You may want to discuss BV with your healthcare provider as a potentially sexually transmitted condition that might merit partner evaluation.

### When should I seek a second opinion for recurring bacterial vaginosis?

If your bacterial vaginosis keeps returning after treatment, a second opinion is worth considering. Recurrence is common because Gardnerella vaginalis creates a biofilm that standard antibiotics may not fully eradicate, and the infection can be sexually transmitted, meaning an untreated partner may cause reinfection. A specialist can review your history and discuss whether biofilm-disrupting therapy, extended regimens, or partner evaluation is appropriate for your situation. Diagnostic Detectives Network provides independent expert second opinions.

## Source Information

**Original Article Title:** Gardnerella Bacterial Vaginosis

**DOI:** [10.1093/infdis/jiu089](https://doi.org/10.1093/infdis/jiu089)

**Authors:** Jane R. Schwebke, MD (Department of Medicine, University of Alabama at Birmingham); Christina A. Muzny, MD (Department of Medicine, University of Alabama at Birmingham); and William E. Josey, MD (Department of Gynecology and Obstetrics, Emory University, Atlanta, Georgia)

**Journal:** The Journal of Infectious Diseases 2014;210:338–43

**Publication Details:** Received December 16, 2013; accepted February 3, 2014; electronically published February 7, 2014. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. DOI: 10.1093/infdis/jiu089

**Note:** This patient-friendly article is based on peer-reviewed research. All authors reported no conflicts of interest. Dr. Charles Rivers assisted with graphics for the original manuscript.


---
Publisher: Diagnostic Detectives Network (https://diagnosticdetectives.com) — independent multi-expert medical second opinions, worldwide, private-pay.
Author byline: Anton Titov, MD, PhD. Contact: https://diagnosticdetectives.com/pages/contact
Canonical page: https://diagnosticdetectives.com/products/understanding-bacterial-vaginosis-how-one-bacterium-gardnerella-vaginalis-starts-the-infection