📘

Definitions

2

Amsel clinical criteria

Three of four bedside findings: thin white-to-yellow homogeneous discharge, more than 20% clue cells on microscopy, vaginal fluid pH above 4.5, and a fishy odour on adding 10% potassium hydroxide. The set is worth reading mechanistically rather than as a checklist - each item reports a different node of this entry: the discharge and clue cells report epithelial adhesion, the pH reports loss of lactic acid, and the whiff test reports biogenic amine production. It remains the entry criterion for contemporary randomised trials.

DIAGNOSTIC_CRITERIA Office diagnosis of bacterial vaginosis in symptomatic and asymptomatic reproductive-aged women.

Inclusion criteria

Thin, white or yellow, homogeneous vaginal discharge
More than 20% clue cells on microscopic examination
Vaginal fluid pH greater than 4.5
Fishy odour on addition of 10% potassium hydroxide (whiff test)

Show evidence (2 references)

PMID:6600371 SUPPORT Human Clinical

"Practical diagnostic criteria for standard clinical use are proposed. Application of such criteria should assist in clinical management of nonspecific vaginitis and in further study of the microbiologic and biochemical correlates and the pathogenesis of this mild but quite prevalent disease."

The originating publication proposing these criteria.

PMID:32402161 SUPPORT Human Clinical

"If a potentially eligible woman met at least three of four Amsel criteria (i.e., thin, white or yellow, homogeneous discharge; >20% clue cells on microscopic examination; vaginal fluid with a pH of >4.5; and release of a fishy odor when 10% potassium hydroxide is added to a vaginal specimen)"

Enumerates all four items and the three-of-four threshold, as operationalised for trial entry.

Nugent score (Gram stain)

A 0-10 weighted count of three morphotypes on Gram stain - lactobacilli, small gram-variable or gram-negative rods (Gardnerella/Bacteroides), and curved gram-variable rods - with 7 or above read as BV. Its design principle is not diagnostic accuracy but reproducibility: morphotypes with poor intercentre agreement (gram-positive cocci) were deliberately dropped and the most reliable one (curved rods) weighted most heavily, raising intercentre correlation from 0.61 to 0.82. That is why it, rather than Amsel, became the research standard, and why the intermediate band 4-6 exists as a graded state rather than a diagnostic failure.

DIAGNOSTIC_CRITERIA Standardised laboratory diagnosis of bacterial vaginosis from a Gram-stained vaginal smear; the reference standard for research and the comparator for molecular assays.

Show evidence (3 references)

PMID:1706728 SUPPORT Human Clinical

"The scoring system (0 to 10) was described as a weighted combination of the following morphotypes: lactobacilli, Gardnerella vaginalis or bacteroides (small gram-variable rods or gram-negative rods), and curved gram-variable rods."

Specifies the three scored morphotypes and the score range.

PMID:1706728 SUPPORT Human Clinical

"For comparison with the Spiegel criteria, a score of 7 or higher was considered indicative of bacterial vaginosis."

Gives the diagnostic threshold.

PMID:1706728 SUPPORT Human Clinical

"The standardized score had improved intercenter reliability (r = 0.82) compared with the Spiegel criteria (r = 0.61)."

Quantifies the reproducibility gain that is this score's reason for existing.

Notes: A contemporary review still describes the Nugent score as the gold-standard diagnostic tool despite its known limitations, which is why both it and the Amsel criteria are curated here rather than only one.

◈

Mechanistic Hypotheses

2

Relapse from a surviving adherent biofilm

biofilm_persistence_relapse CANONICAL

Evidence balance 1 support

Recurrence is re-emergence of the original community from a Gardnerella-dominated adherent biofilm that tolerated the antibiotic course. Marked CANONICAL because the biofilm is directly demonstrated on vaginal biopsy and because it is the account most treatment development has been built around - not because it has been shown to be the operative route in any individual recurrence.

Pathograph links

Epithelial Adhesion and Gardnerella-Dominated Adherent Biofilm -> Biofilm Persistence After Antimicrobial Therapy DIRECT The adherent biofilm is the proposed reservoir from which BV-associated bacteria re-emerge after a course of antibiotics.

Biofilm Persistence After Antimicrobial Therapy -> Recurrent Bacterial Vaginosis INDIRECT UNKNOWN INTERMEDIATES Re-emergence of BV-associated bacteria from the surviving adherent layer.

Show evidence (1 reference)

PMID:16260520 SUPPORT Human Clinical

"A biofilm comprised of confluent G vaginalis with other bacterial groups incorporated in the adherent layer is a prominent feature of bacterial vaginosis."

Establishes that the structure this hypothesis requires exists in vivo.

Reinfection from an untreated sexual partner

sexual_reinfection EMERGING

Evidence balance 1 support

Recurrence is reacquisition of BV-associated bacteria from a regular untreated partner. Marked EMERGING rather than ALTERNATIVE: six earlier partner-treatment trials were negative and guidelines still do not recommend partner treatment, but the 2025 StepUp randomised trial was stopped early for benefit, which is a stronger form of evidence than anything supporting the biofilm route. This status should be revisited when guidelines respond.

Pathograph links

Reinfection from an Untreated Sexual Partner -> Recurrent Bacterial Vaginosis DIRECT Reinoculation with BV-associated bacteria from an untreated partner re-establishes the dysbiotic community.

Show evidence (1 reference)

PMID:40043236 SUPPORT Human Clinical

"The addition of combined oral and topical antimicrobial therapy for male partners to treatment of women for bacterial vaginosis resulted in a lower rate of recurrence of bacterial vaginosis within 12 weeks than standard care."

The randomised result on which this hypothesis's EMERGING status rests.

?

Discussions and Knowledge Gaps

4

What precipitates the shift from a lactobacillus-dominant to a lactobacillus-deficient vaginal community in an individual woman?

KNOWLEDGE GAP OPEN bv_initiating_event_unknown

Attached to

pathophysiology#Loss of Lactobacillus Dominance and Vaginal Acidification

This entry curates a detailed chain downstream of the community shift and nothing upstream of it, because there is nothing to curate: the trigger node has no evidenced incoming edge other than two epidemiological exposures. Risk factors are known (sexual activity and intrauterine device use are the two cited here), but no sequence of events from a normal community to a BV community has been demonstrated. This matters practically, not just theoretically - every intervention in this entry acts after the shift has happened, and none prevents it.

Proposed experiments

Densely sampled prospective cohort through incident BV

bv_incident_transition_cohort

Daily or near-daily self-sampling in a cohort of women with a stable L. crispatus-dominant community, with metagenomics, metabolomics (lactic acid isomers, biogenic amines), pH and behavioural diaries, powered on incident transitions rather than prevalent BV. The measurement that is missing is not another cross-sectional comparison of BV versus not-BV; it is the order of events across the transition.

Decision criterion

Whether a consistent temporal ordering (for example, amine rise preceding lactobacillus loss, or the reverse) is observed across incident transitions.

Show evidence (1 reference)

PMID:35118003 SUPPORT Human Clinical

"The etiology of this dysbiosis remains unknown, but its health consequences are significant, including obstetrical complications, increased risk of sexually transmitted infections and urogenital infections."

States the gap directly, alongside the consequences that make it worth closing.

In a woman whose BV recurs after treatment, can relapse from a surviving biofilm be distinguished from reinfection by an untreated partner?

KNOWLEDGE GAP OPEN bv_recurrence_route_indistinguishable

Attached to

pathophysiology#Recurrent Bacterial Vaginosis pathophysiology#Biofilm Persistence After Antimicrobial Therapy pathophysiology#Reinfection from an Untreated Sexual Partner

Two curated hypothesis groups converge on the recurrence node and there is currently no assay that resolves which operated. The clinical presentation is identical, point-of-care tests cannot separate them, and next-generation sequencing has not identified a signature that distinguishes relapse from reacquisition. This is why the entry retains both routes rather than promoting one: the StepUp trial shows the reinfection route is real at a population level, which is a different claim from it being the route in any given woman.

Proposed experiments

Strain-resolved tracking of BV-associated bacteria across a treated episode

bv_strain_resolved_recurrence_tracking

Deep metagenomic strain typing of the woman's pre-treatment community, her partner's penile and urethral community, and her post-treatment recurrent community. Relapse predicts recovery of the woman's own pre-treatment strains; reinfection predicts strains matching the partner's and not present in her pre-treatment sample.

Decision criterion

Whether recurrent-episode strains are shared with the woman's own pre-treatment community, with the partner's community, or with both.

Show evidence (1 reference)

PMID:34470644 SUPPORT Human Clinical

"Frustratingly for both clinicians and patients, the contribution of reinfection and vaginal relapse cannot be separated, as the clinical presentation of both mechanisms of recurrence is identical."

States the indistinguishability that defines this gap.

How much of BV pathogenesis can any animal model address, given that vaginolysin requires human CD59 and no animal has a lactobacillus-dominant acidic vagina?

HUMAN MODEL MISMATCH OPEN bv_no_animal_model_vaginolysin_human_specific

Attached to

pathophysiology#Vaginolysin-Mediated Epithelial Pore Formation and Signalling pathophysiology#Loss of Lactobacillus Dominance and Vaginal Acidification

The species restriction here is not incidental, it is receptor-level: vaginolysin lyses cells only where human CD59 is present, and transfecting human CD59 into hamster cells is what makes them susceptible. Separately, the healthy state this disease departs from - a lactobacillus-dominant, lactic-acid-acidified vagina - is close to unique to humans, so there is no normal community for a model organism to lose. The consequence is that the murine work curated above can test one downstream arm (T-cell activation) and nothing else, and that essentially all mechanism in this entry rests on human samples and human cell lines. Reviews of BV have named the absence of a suitable animal model as a limiting factor for two decades.

Proposed experiments

Human-relevant model systems for the two blocked arms

bv_humanised_cd59_and_hydrogel_models

For the toxin arm, a human-CD59 transgenic mouse or a human vaginal epithelial organ-chip perfused with defined consortia. For the community arm, a cervicovaginal-mucus hydrogel or organ-chip supporting stable L. crispatus dominance into which BV taxa can be introduced, giving a system in which the transition itself is observable.

Decision criterion

Whether either system reproduces vaginolysin-dependent epithelial responses and sialidase-mediated glycocalyx loss with human-like kinetics.

Show evidence (2 references)

PMID:18390664 SUPPORT Other

"Mechanistic studies of BV and its adverse consequences have been limited by the absence of definitive diagnostic testing and a suitable animal model"

Cited as background rather than as a result: this is the authors' statement about the state of the field in their introduction, not an outcome of their own experiments, so it is tagged OTHER. It names the model gap explicitly, in the same paper that establishes the receptor-level reason for it (quoted separately below from the companion study).

PMID:24082080 SUPPORT In Vitro

"transfection of nonhuman cells (CHO-K1) with hCD59 renders them susceptible to toxin-induced membrane blebbing"

Demonstrates that human CD59 is the sufficient species determinant, which is what makes this a receptor-level rather than a generic model limitation.

If more than half of women meeting the case definition have no symptoms, and the barrier defect is present in asymptomatic disease, what is the right target population for treatment?

OPEN QUESTION OPEN bv_asymptomatic_majority_and_screening

Attached to

pathophysiology#Loss of Cervicovaginal Mucus Adhesive Barrier Function

Two curated findings sit awkwardly together. Asymptomatic BV is the majority of BV, and the HIV-relevant mucus barrier defect is present regardless of symptoms - which argues for screening. But the entry's own obstetric node records that screening-and-treatment has not reliably prevented preterm birth, and treatment does not durably restore an L. crispatus community. This is recorded as an open question rather than a knowledge gap because the missing item is a policy-relevant trial result, not a mechanism.

Show evidence (2 references)

PMID:31971984 SUPPORT In Vitro

"HIV virions had significantly increased mobility in CVM from women with BV compared to CVM from women with Lactobacillus crispatus-dominant microbiota, regardless of whether symptoms were present"

Establishes that the barrier defect does not depend on symptoms, which is what makes the target-population question live.

PMID:34470644 SUPPORT Human Clinical

"Although up to half of BV-affected women do not experience symptoms"

Establishes the size of the asymptomatic population at issue.

⚙

Pathophysiology

15

Loss of Lactobacillus Dominance and Vaginal Acidification

The initiating state change. A healthy premenopausal vagina is dominated by Lactobacillus species whose copious lactic acid holds the lumen below pH 4.5; in BV that community is replaced, lactic acid production falls and pH rises. Two points matter for everything downstream. The species identity is not interchangeable - L. crispatus-dominated communities are the most protective, whereas L. iners, which makes only the L-isomer of lactic acid, sits at an intermediate and unstable position and is the species most associated with transition into and relapse after BV. And the trigger itself is unexplained: what precipitates the shift in an individual woman is the central unanswered question of the field, curated as a knowledge gap below. `GO:0006885 regulation of pH` is deliberately not bound here: the pH claim this node makes is about the chemistry of the vaginal lumen, not about a cellular homeostatic process, and GO annotates the latter. The mechanism that produces the pH change is carried by the lactate term below, and the luminal measurement itself is curated as an Amsel criterion under `definitions`.

lactate biosynthetic process GO:0019249 ↓ DECREASED

vagina UBERON:0000996

Downstream

Polymicrobial Anaerobic Overgrowth Relief of lactic-acid-mediated suppression permits expansion of facultative and strict anaerobes.

Show evidence (4 references)

PMID:35118003 SUPPORT Human Clinical

"BV is the result of a disturbance in the vaginal ecosystem; i.e., a sudden replacement of Lactobacilli by anaerobic bacteria such as Gardnerella vaginalis, Atopobium vaginae, Ureaplasma urealyticum, Mycoplasma hominis, and others."

States the defining community shift that this node encodes.

PMID:35118003 SUPPORT Human Clinical

"Under normal conditions, 70-90% of the vaginal bacterial species in healthy premenopausal women are Lactobacilli"

Establishes the baseline lactobacillus dominance whose loss defines this node.

PMID:37234911 SUPPORT Human Clinical

"Lactobacillus iners is distinct from L. crispatus, L. gasseri, and L. jensenii by its high global prevalence in vaginal microbiomes, relatively small genome, production of only L-lactic acid, and inconsistent associations with genital health outcomes."

Supports the claim that lactobacillus species identity, not merely lactobacillus presence, determines how protective this node's baseline state is.

+ 1 more reference

Polymicrobial Anaerobic Overgrowth

Expansion of a dense, diverse consortium - Gardnerella spp., Prevotella (Hoylesella) spp., Fannyhessea (Atopobium) vaginae, Mobiluncus, Megasphaera, Sneathia and others - to concentrations orders of magnitude above baseline. This node is deliberately named for the community rather than for an organism: no single member reproduces the syndrome on inoculation, and each is recoverable from women without disease. Treating the consortium as the unit is what makes the broad anaerobic coverage of metronidazole and clindamycin the rational therapy, and it is why an organism-specific agent has never been developed.

response to bacterium GO:0009617 ↑ INCREASED

vagina UBERON:0000996

Downstream

Biogenic Amine Production and Amplified Lactobacillus Suppression Anaerobic amino acid decarboxylation generates the biogenic amines that give BV its odour and further suppress the residual lactobacilli.

Epithelial Adhesion and Gardnerella-Dominated Adherent Biofilm Members of the expanded consortium adhere to the epithelium and organise into a structured biofilm.

Bacterial Sialidase and Glycosidase Degradation of the Mucus Barrier The consortium secretes the mucin-degrading enzyme activities that are elevated in BV vaginal fluid.

Anaerobe Ribosomal Translation (Clindamycin and Tetracycline Target) Sustained expansion of the consortium depends on bacterial protein synthesis, which is the molecular target exploited by the lincosamide arm of therapy.

Show evidence (2 references)

PMID:34470644 SUPPORT Human Clinical

"This non-optimal microbiological state involves a reduction in protective lactobacilli, and an increase in bacterial diversity and facultative and strict anaerobes, including Gardnerella spp., Atopobium vaginae, Prevotella spp., and others, referred to as BV-associated bacteria (BVAB)"

Names the consortium taxa and frames BV as a community state rather than a single-organism infection.

PMID:6600371 SUPPORT Human Clinical

"A clinical diagnosis of nonspecific vaginitis, based on simple office procedures, was correlated with both the presence and the concentration of Gardnerella vaginalis (Hemophilus vaginalis) in vaginal discharge, and with characteristic biochemical findings in vaginal discharge."

The founding correlation between organism concentration - not mere presence - and clinical disease, which is the quantitative claim this node makes.

Biogenic Amine Production and Amplified Lactobacillus Suppression

Amino acid decarboxylation by the expanded anaerobes produces putrescine, cadaverine, tyramine and trimethylamine. These are usually treated as a symptom - they are what the whiff test detects and what patients describe as a fishy odour - but the measurements curated here make them a mechanism as well: they inhibit the growth of vaginal lactobacilli and independently reduce their lactic acid output. That closes a positive feedback loop back onto the trigger node, which is why the dysbiotic state is self-reinforcing rather than self-limiting, and it is the reason this node is curated separately from the odour phenotype it produces.

amine biosynthetic process GO:0009309 ↑ INCREASED

Downstream

Loss of Lactobacillus Dominance and Vaginal Acidification Biogenic amines slow lactobacillus growth and cut lactic acid production, feeding back on and entrenching the initiating state change.

Thin homogeneous malodorous vaginal discharge BV-associated amines produce the malodor detected by the whiff test.

Show evidence (3 references)

PMID:33674429 SUPPORT Human Clinical

"Increases in cadaverine, putrescine, and tyramine were associated with greater odds of women transitioning from L. crispatus-dominated vaginal microbiota to microbiota that have a paucity of Lactobacillus spp. and from Nugent scores of 0 to 3 to Nugent scores of 7 to 10, consistent with BV."

The in vivo longitudinal association between rising biogenic amines and transition into BV, which is what licenses treating this node as upstream rather than merely concurrent.

PMID:33674429 SUPPORT In Vitro

"BAs were associated with reduced production of d- and l-lactic acid by vaginal Lactobacillus spp., and this effect was independent of their effect upon Lactobacillus species growth."

The axenic-culture experiment establishing the feedback edge onto lactic acid production, and establishing it as separable from a simple growth effect.

PMID:33674429 SUPPORT In Vitro

"Exposure to putrescine lengthened the lag time and/or slowed the growth of all vaginal Lactobacillus spp. except L. jensenii 62G."

Direct growth inhibition of lactobacilli by a BV-associated amine, with the single exception recorded rather than smoothed over.

Epithelial Adhesion and Gardnerella-Dominated Adherent Biofilm

Gardnerella adheres to the vaginal epithelium and forms a confluent, structured biofilm into which other consortium members are incorporated. This is the one structural feature that FISH on vaginal biopsies found specific to BV rather than merely more abundant in it, and it is the physical object that makes BV a persistent rather than an episodic condition: a sessile, matrix-embedded community tolerates antimicrobial exposure that clears planktonic organisms. Prevotella (Hoylesella) timonensis adheres comparably well, so adhesion is not a Gardnerella monopoly even though biofilm architecture appears to be.

Relapse from a surviving adherent biofilm

vaginal squamous epithelial cell CL:1001578

biofilm formation GO:0042710 ↑ INCREASED cell adhesion involved in biofilm formation GO:0043708 ↑ INCREASED

vagina UBERON:0000996

Downstream

Exfoliation of Bacteria-Coated Epithelial Cells (Clue Cells) Adherent organisms are shed with the epithelial cells they coat, producing the clue cells seen on wet mount.

Vaginolysin-Mediated Epithelial Pore Formation and Signalling Close apposition of Gardnerella to the epithelium delivers its cholesterol-dependent cytolysin to the host cell membrane.

Biofilm Persistence After Antimicrobial Therapy The adherent biofilm is the proposed reservoir from which BV-associated bacteria re-emerge after a course of antibiotics.

Show evidence (3 references)

PMID:16260520 SUPPORT Human Clinical

"A biofilm comprised of confluent G vaginalis with other bacterial groups incorporated in the adherent layer is a prominent feature of bacterial vaginosis."

Describes the composite structure this node asserts - a Gardnerella scaffold with other taxa incorporated.

PMID:16260520 SUPPORT Human Clinical

"Bacterial vaginosis was associated with greater occurrence and higher concentrations of a variety of bacterial groups. However, only Gardnerella vaginalis developed a characteristic adherent biofilm that was specific for bacterial vaginosis."

Draws the distinction this node depends on: many taxa are more abundant, but only one forms a BV-specific adherent structure.

PMID:39162399 SUPPORT In Vitro

"We demonstrate that P. timonensis, but not P. bivia, strongly adheres to vaginal and endocervical cells to a similar level as G. vaginalis but did not elicit a comparable proinflammatory epithelial response."

The study's own adhesion assay establishes that adherence to vaginal and endocervical epithelium is a property of the consortium rather than of Gardnerella alone, and that it is not shared by every anaerobe (P. bivia does not adhere).

Exfoliation of Bacteria-Coated Epithelial Cells (Clue Cells)

Squamous epithelial cells so heavily coated with adherent organisms that their borders are obscured are shed into the vaginal fluid. Immunofluorescence on clue cells identifies the adherent rods as Gardnerella rather than Mobiluncus, Bacteroides or Fusobacterium, which is the observation that ties the microscopic diagnostic sign to a specific adhesion mechanism rather than to generic anaerobic overgrowth. This node is curated as a cellular event, not as a diagnostic test; the test that reads it is curated under `definitions`.

vaginal squamous epithelial cell CL:1001578

vagina UBERON:0000996

Show evidence (2 references)

PMID:2668431 SUPPORT Human Clinical

"Clue cells are epithelial cells covered by adherent gram-negative rods, observed in vaginal smears from women with bacterial vaginosis."

Defines the cellular object this node names.

PMID:2668431 SUPPORT Human Clinical

"Gardnerella vaginalis was most often observed adhering to the surface of clue cells and was detected on the surface of exfoliated vaginal epithelial cells significantly more frequently and in higher numbers than were Mobiluncus, Bacteroides, and Fusobacterium, suggesting that this species of..."

Attributes clue cell formation specifically to Gardnerella adhesion, which is the causal claim on the incoming edge from the biofilm node.

Vaginolysin-Mediated Epithelial Pore Formation and Signalling

Gardnerella secretes vaginolysin, a cholesterol-dependent cytolysin whose receptor is human CD59. Two consequences matter and they are different in kind. At lytic concentrations the toxin kills; at sublytic concentrations - which is the physiological situation - it triggers rapid membrane blebbing, activates epithelial p38 MAPK and induces IL-8. The receptor dependence is the reason BV has no natural animal model: vaginolysin is human-specific because CD59 is, and transfecting human CD59 into non-human cells is what confers susceptibility. That constraint is curated as a human/model mismatch below rather than left implicit.

vaginal squamous epithelial cell CL:1001578

cytolysis GO:0019835 ↑ INCREASED killing of cells of another organism GO:0031640 ↑ INCREASED

Downstream

Epithelial Barrier Disruption and Mucosal Immune Activation Sublytic pore formation drives p38 MAPK signalling and IL-8 release from the vaginal epithelium.

Show evidence (4 references)

PMID:18390664 SUPPORT In Vitro

"We demonstrate that G. vaginalis produces a toxin (vaginolysin [VLY]) that is a member of the cholesterol-dependent cytolysin (CDC) family, most closely related to intermedilysin from Streptococcus intermedius."

Identifies the toxin and its family, the molecular claim of this node.

PMID:18390664 SUPPORT In Vitro

"In addition to causing erythrocyte lysis, VLY activates the conserved epithelial p38 mitogen-activated protein kinase pathway and induces interleukin-8 production by human epithelial cells."

Establishes the sublytic signalling output that carries the outgoing edge to epithelial activation, distinct from frank lysis.

PMID:24082080 SUPPORT In Vitro

"Binding of VLY to its human-specific receptor (hCD59) is required for bleb formation, as antibody inhibition of either toxin or hCD59 abrogates this response, and transfection of nonhuman cells (CHO-K1) with hCD59 renders them susceptible to toxin-induced membrane blebbing."

The gain- and loss-of-function pair establishing hCD59 dependence, and with it the species restriction recorded in the human/model-mismatch discussion.

+ 1 more reference

Bacterial Sialidase and Glycosidase Degradation of the Mucus Barrier

The consortium secretes sialidases, fucosidases, beta-galactosidase and beta-N-acetylhexosaminidase that strip terminal sugars from cervicovaginal mucins and from the epithelial glycocalyx. This is the arm of BV pathogenesis with the clearest line to hard clinical outcomes, and it has been misattributed for two decades: biochemical work focused on Gardnerella, but vaginal Prevotella species are now shown to be a major and globally conserved source of vaginal sialidase, with P. (Hoylesella) timonensis carrying both sialidases and an unusual complement of fucosidases. Notably, the same 1999 study that found sialidase, beta-galactosidase and beta-N-acetylhexosaminidase elevated in BV found no increase in O-glycanase, proteinase, sulphatase or whole-mucinase activity - the degradation is glycosidic and selective, not general proteolysis, and that distinction is preserved here rather than flattened into "mucinase activity".

exo-alpha-sialidase activity GO:0004308 ↑ INCREASED

mucus UBERON:0000912

Downstream

Loss of Cervicovaginal Mucus Adhesive Barrier Function Removal of terminal sialic acid and fucose from mucin glycans degrades the adhesive trapping function of the mucus gel.

Ascending Infection and Adverse Pregnancy Outcomes Enzymatic breach of the cervicovaginal mucosal barrier is the proposed route by which BV organisms reach the uterus; detectable vaginal sialidase in preterm labour identifies women at higher risk of early preterm birth.

Show evidence (6 references)

PMID:10454178 SUPPORT Human Clinical

"Significant increases in activity were detected in BV samples for sialidase using a mucin (BSM P<0.005) and serum type glycoprotein (AGP P<0.005) substrates, beta-galactosidase (P<0.001), and beta-N-acetylhexosaminidase (P<0.01)."

Quantifies the specific glycosidase activities elevated in BV vaginal fluid.

PMID:10454178 SUPPORT Human Clinical

"No significant increases in BV patients were detected in O-glycanase, proteinase, arylesterase, sulphatase or whole mucinase activities."

The negative half of the same study, retained because it is what makes the degradation glycosidic and selective rather than general. It supports this node by exclusion: selectivity is part of what the node asserts, so a null result for proteinase, sulphatase and whole-mucinase activity is evidence for the glycosidic mechanism rather than a qualification of it.

PMID:10454178 SUPPORT Human Clinical

"These results support the hypothesis that certain BV-associated enzymes may detrimentally affect the mucosal barrier, permitting bacteria access to the uterus."

States the barrier-breach consequence carried by this node's outgoing edges.

+ 3 more references

Loss of Cervicovaginal Mucus Adhesive Barrier Function

Cervicovaginal mucus normally traps virions and particles. In BV that trapping fails: fluorescently labelled HIV virions move significantly faster through mucus from women with BV than through mucus from L. crispatus-dominant women, whether or not the woman has symptoms. The mechanistically important finding is *how* it fails - electron microscopy and nanoparticle work localise the defect to reduced adhesive interactions, with the physical pore structure of the gel intact. The barrier is not torn open; it stops being sticky. The same loss of trapping is seen with L. iners-dominant mucus, which is why treating BV to a Nugent-normal endpoint does not necessarily restore the barrier.

mucus UBERON:0000912 uterine cervix UBERON:0000002

Downstream

Increased Susceptibility to HIV and Other Genital Tract Infections Loss of mucus trapping allows virions to reach the epithelium.

Show evidence (3 references)

PMID:31971984 SUPPORT In Vitro

"we found that HIV virions had significantly increased mobility in CVM from women with BV compared to CVM from women with Lactobacillus crispatus-dominant microbiota, regardless of whether symptoms were present"

The measurement this node is built on, including the point that the defect is present in asymptomatic BV.

PMID:31971984 SUPPORT In Vitro

"We confirmed using nanoparticles and scanning electron microscopy that the impaired barrier function was due to reduced adhesive barrier properties without an obvious degradation of the physical CVM pore structure."

Localises the defect to adhesion rather than to structural breakdown - the distinction this node insists on.

PMID:31971984 SUPPORT In Vitro

"We further confirmed a similar increase in HIV mobility in CVM from women with Lactobacillus iners-dominant microbiota, the species most associated with transitions to BV and that persists after antibiotic treatment for BV."

Supports the caveat that a post-treatment L. iners community does not restore the barrier, which is the therapeutic rationale for L. crispatus repletion.

Epithelial Barrier Disruption and Mucosal Immune Activation

Despite the "-osis" in its name, BV is not immunologically silent. Vaginal IL-1alpha and soluble E-cadherin - a biomarker of epithelial barrier disruption - are elevated, and a lactobacillus-deficient community is accompanied by increased numbers of activated mucosal CD4+ T cells. The causal reading is supported by intervention: repleting L. crispatus with a live biotherapeutic lowered both IL-1alpha and soluble E-cadherin, so the inflammatory state tracks the community rather than merely co-occurring with it. This node is what converts a microbiological state into an HIV-relevant one, because activated mucosal CD4+ T cells are the target cell.

CD4-positive T cell CL:0000492 vaginal squamous epithelial cell CL:1001578

inflammatory response GO:0006954 ↑ INCREASED T cell activation GO:0042110 ↑ INCREASED

Animal models

Murine cervicovaginal colonisation with BV-associated anaerobes Mouse partially recapitulates Reproduces the target-cell recruitment arm of this node - high-risk bacteria increased activated genital CD4+ T cells - but not the node as a whole.

Downstream

Increased Susceptibility to HIV and Other Genital Tract Infections Recruitment and activation of mucosal CD4+ T cells supplies the target cells for HIV infection.

Show evidence (3 references)

PMID:28087240 SUPPORT Human Clinical

"we found that individuals with diverse genital bacterial communities dominated by anaerobes other than Gardnerella were at over 4-fold higher risk of acquiring HIV and had increased numbers of activated mucosal CD4+ T cells compared to those with Lactobacillus crispatus-dominant communities"

Couples the activated-CD4 phenotype to prospective HIV acquisition in the same cohort, which is the pairing this node depends on.

PMID:35659905 SUPPORT Human Clinical

"Bacterial vaginosis might increase HIV risk by eliciting genital inflammation and epithelial barrier disruption, whereas vaginal Lactobacillus crispatus is associated with immune quiescence and HIV protection."

States the barrier-disruption and inflammation mechanism this node encodes.

PMID:35659905 SUPPORT Human Clinical

"The primary outcomes were vaginal levels of IL-1α and soluble E-cadherin at 24 weeks"

Names the two immunological endpoints - vaginal IL-1alpha and soluble E-cadherin - that this node reports, as prespecified primary outcomes of a randomised trial rather than post-hoc observations.

Increased Susceptibility to HIV and Other Genital Tract Infections

The consequence that dominates BV's public-health importance. Meta-analysis of incidence studies puts the relative risk of HIV acquisition at 1.6, and a prospective South African cohort found over four-fold higher risk for the most diverse lactobacillus-deficient communities. Two distinct mechanisms feed this node and are curated separately upstream: loss of mucus trapping (a physical barrier failure) and recruitment of activated mucosal CD4+ T cells (target-cell supply). The evidence is not uniform across every downstream infection, and the PID arm in particular is a documented disagreement between two prospective cohorts rather than a settled result. Both are retained below. The Longitudinal Study of Vaginal Flora (N = 2956) found both Nugent-BV (aHR 1.53) and symptomatic Amsel-BV (aHR 2.15) associated with incident PID at the subsequent visit; the two intervals overlap and the study does not test the difference between them, so the figures do not establish a gradient by exposure definition. A community-based cohort found no association between a Gardnerella-dominated microbiome and subsequent PID. The two differ in ascertainment on both sides - Nugent and Amsel criteria against microbiome sequencing for the exposure, and clinic-based tenderness criteria against community follow-up for the outcome - and the negative study reports few PID cases, so this node records the association as contested rather than resolving it in either direction.

response to bacterium GO:0009617 ↕ DYSREGULATED

Show evidence (5 references)

PMID:18614873 SUPPORT Human Clinical

"Bacterial vaginosis was associated with an increased risk of HIV acquisition in HIV-incidence studies (relative risk = 1.6, 95% confidence interval: 1.2, 2.1)."

The pooled incidence-study estimate quantifying this node.

PMID:31971984 SUPPORT Other

"Women with BV are at 60% increased risk for HIV acquisition and are 3-times more likely to transmit HIV to an uninfected partner."

Adds the transmission direction, which the acquisition meta-analysis does not cover. Cited as background rather than as a result: this is the authors' framing sentence in their introduction, not an observation from their own mucus-barrier experiments, and it is tagged OTHER for that reason. The transmission half in particular carries no primary source in this paper, so the node rests on it only for direction, not magnitude.

PMID:34396403 SUPPORT Human Clinical

"BV was associated with incident PID in a large prospective cohort, controlling for behavioral factors and sexually transmitted infections (STIs)."

The positive half of the PID disagreement, and the authors' own statement of it. From the larger of the two prospective cohorts (N = 2956, quarterly follow-up for 12 months, BV measured at the visit preceding the PID diagnosis), it supports extending this node beyond HIV to ascending genital infection. The adjustment for concurrent and untreated chlamydia is what makes it more than a confounded association.

+ 2 more references

Ascending Infection and Adverse Pregnancy Outcomes

In pregnancy the same barrier failure permits organisms to reach the choriodecidual space, and BV detected early in pregnancy is associated with late miscarriage and preterm delivery independent of prior preterm birth. The enzyme measurement makes this more than an association: among women already in preterm labour with BV or intermediate flora, those with detectable vaginal sialidase had a higher rate of early preterm birth, which links the outcome to the specific molecular activity curated upstream rather than to the Nugent score alone. Screening-and-treatment trials have nonetheless been inconsistent, so this node records a mechanism and a risk association, not a demonstrated preventable fraction.

uterine cervix UBERON:0000002

Downstream

Premature birth BV-associated ascending infection is linked to preterm delivery risk.

Show evidence (4 references)

PMID:8124116 SUPPORT Human Clinical

"A further logistic analysis of data from women recruited before 16 weeks' gestation showed that preterm deliveries or late miscarriages occurred more often in women with bacterial vaginosis (12/77; 5.5; 2.3 to 13.3; P < 0.001)."

Prospective cohort estimate for early-pregnancy BV and late miscarriage or preterm delivery.

PMID:8124116 SUPPORT Human Clinical

"Late miscarriage and preterm delivery are associated with the presence of bacterial vaginosis in early pregnancy. This is independent of recognised risk factors such as previous preterm delivery."

Establishes independence from the dominant confounder, prior preterm delivery.

PMID:12388968 SUPPORT Human Clinical

"Women in preterm labor with bacterial vaginosis or intermediate flora and detectable sialidase are at increased risk of early preterm birth."

Ties the obstetric outcome to sialidase specifically, supporting the incoming edge from the glycosidase node rather than from BV status alone.

+ 1 more reference

Biofilm Persistence After Antimicrobial Therapy

The first of two competing accounts of why BV comes back. Metronidazole and clindamycin cure 70-85% of episodes within a month, yet more than half of women relapse within six months. On this account the adherent Gardnerella biofilm survives the antibiotic course and the BV-associated bacteria re-emerge from it - a relapse of the same community rather than a new acquisition. Note what this node does not claim: no assay currently distinguishes a surviving biofilm from a reacquired one in an individual woman, so this is an inference from the biofilm's demonstrated existence and known antimicrobial tolerance, not a measured event.

Relapse from a surviving adherent biofilm

biofilm formation GO:0042710 ↑ INCREASED

Downstream

Recurrent Bacterial Vaginosis Re-emergence of BV-associated bacteria from the surviving adherent layer.

Show evidence (2 references)

PMID:34470644 SUPPORT Human Clinical

"The low rate of sustained cure highlights our limited understanding of the pathogenesis of BV recurrence, which has been attributed to possible persistence and re-emergence of BV-associated bacteria (BVAB) or a BV-associated biofilm following antimicrobials and/or reinfection occurring from..."

States the persistence account and, in the same sentence, the competing reinfection account - which is why both are curated as hypothesis groups rather than one being asserted.

PMID:34470644 SUPPORT Human Clinical

"These regimens have similar efficacy and cure ~70–85% of women with BV within 1 month"

Quantifies short-term cure, the denominator against which the relapse rate is the surprising figure.

Reinfection from an Untreated Sexual Partner

The second account. BV-associated bacteria are exchanged between partners during sex, and on this reading a woman cured by antibiotics is reinoculated by an untreated regular partner. Six partner-treatment trials in the 1980s and 1990s were negative and the hypothesis fell out of favour; guidelines do not recommend partner treatment. The 2025 StepUp randomised trial reopened it decisively - it was stopped early by its data and safety monitoring board because treating the woman alone was inferior to treating both partners, with 12-week recurrence 35% versus 63%. This node is therefore curated as a real causal route, not as a historical hypothesis, while the biofilm route is retained alongside it.

Reinfection from an untreated sexual partner

Downstream

Recurrent Bacterial Vaginosis Reinoculation with BV-associated bacteria from an untreated partner re-establishes the dysbiotic community.

Show evidence (3 references)

PMID:40043236 SUPPORT Human Clinical

"The trial was stopped by the data and safety monitoring board after 150 couples had completed the 12-week follow-up period because treatment of the woman only was inferior to treatment of both the woman and her male partner."

The randomised evidence that an untreated partner is a causal contributor to recurrence - the strongest support for this node.

PMID:40043236 SUPPORT Human Clinical

"Evidence of sexual exchange of bacterial vaginosis-associated organisms between partners suggests that male-partner treatment may increase the likelihood of cure."

States the exchange mechanism this node encodes.

PMID:34470644 SUPPORT Human Clinical

"There is a robust body of evidence to support the exchange of bacteria between partners during sexual activity, and while the hypothesis that women treated for BV are subsequently reinfected with BVAB following sex with an untreated sexual partner is not new, failure of past partner treatment..."

Records both the microbiological support for exchange and the trial history that had discredited it, written before the StepUp result.

Recurrent Bacterial Vaginosis

Re-establishment of the dysbiotic community after apparently successful treatment - the outcome that defines the clinical problem, since short-term cure is not the difficulty. Both upstream routes converge here and neither can be excluded in an individual woman, because they present identically. That indistinguishability is not a curation shortcut; it is the field's stated position and is recorded as a knowledge gap below. The node is not a terminal state: a recurrent episode is diagnosed by the same Amsel and Nugent criteria as an incident one, so it re-enters the graph at the trigger node and the entry is a cycle rather than a chain. That is the structural reason short-term cure and sustained cure come apart.

Downstream

Loss of Lactobacillus Dominance and Vaginal Acidification A recurrent episode is the re-established dysbiotic community state, not a separate endpoint - which is why an intervention that only clears anaerobes leaves the cycle intact, and why the one curated intervention acting on this node (LACTIN-V, which repopulates the niche) is also the only one shown to lower recurrence.

Show evidence (2 references)

PMID:35118003 SUPPORT Human Clinical

"Standard antibiotic therapy often fails, with an estimated relapse rate of 50% at six months follow-up"

Quantifies the recurrence burden this node represents.

PMID:32402161 SUPPORT Human Clinical

"After treatment with an antibiotic agent, 20 to 75% of women have recurrent bacterial vaginosis within 3 months."

An independent recurrence estimate over a shorter window, showing the range across populations and definitions.

Anaerobe Ribosomal Translation (Clindamycin and Tetracycline Target)

Protein synthesis in the BV consortium is the molecular target of clindamycin, the lincosamide alternative to metronidazole and the agent applied topically to the penile skin of male partners in the StepUp regimen. It is curated as its own node so that the clindamycin arm of therapy attaches to a mechanism rather than to the disease as a whole - metronidazole, whose nitroimidazole mechanism is reductive DNA damage in anaerobes, does not act here and is deliberately not attached to this node.

translation GO:0006412 ↓ DECREASED

Show evidence (2 references)

PMID:35118003 SUPPORT Human Clinical

"Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended."

Establishes clindamycin as a recommended agent, which is what makes its ribosomal target a therapeutic vulnerability in this disease.

PMID:40043236 SUPPORT Human Clinical

"the male partner received oral and topical antimicrobial treatment (metronidazole 400-mg tablets and 2% clindamycin cream applied to penile skin, both twice daily for 7 days)"

Documents the topical clindamycin component of the partner-treatment regimen that acts on this target.

⬡

Pathograph

Use the checkboxes to hide or show graph categories. Hover nodes for evidence and cross-linked metadata.

●

Phenotypes

2

Genitourinary 1

Thin homogeneous malodorous vaginal discharge FREQUENT Abnormal vaginal discharge HP:0034269

Show evidence (3 references)

PMID:32402161 SUPPORT Human Clinical

"thin, white or yellow, homogeneous discharge"

The discharge character as specified in the Amsel criteria used for trial entry.

PMID:34470644 SUPPORT Human Clinical

"it is the symptoms themselves, including malodour and vaginal discharge, that cause significant distress to women and impact on their quality of life and relationships"

Names malodour and discharge as the symptoms of the disease and their impact.

PMID:6600371 SUPPORT Human Clinical

"we diagnosed nonspecific vaginitis in up to 25 percent of our study population; asymptomatic disease was recognized in more than 50 percent of those with nonspecific vaginitis"

Basis for the FREQUENT rather than VERY_FREQUENT band: more than half of women meeting the case definition had no symptoms at all, so symptomatic discharge is present in a minority-to-half of cases.

Prenatal and Birth 1

Premature birth OCCASIONAL HP:0001622

Show evidence (1 reference)

PMID:8124116 SUPPORT Human Clinical

"Multiple logistic analysis showed that there was an increased incidence of preterm delivery in women with a previous preterm delivery (9/24; odds ratio 25; 95% confidence interval 9 to 70; P < 0.001) and bacterial vaginosis (9/115; 2.8; 1.1 to 7.4; P = 0.04)."

Gives the absolute count behind the OCCASIONAL band: 9 of 115 pregnant women with BV delivered preterm, which is within the 5-29% range for that frequency term.

💊

Medical Actions

7

Metronidazole

Action: Antibiotic Therapy NCIT:C15620

Agent: metronidazole CHEBI:6909

Platform: Small molecule

First-line therapy, oral or as 0.75% intravaginal gel. A nitroimidazole prodrug reductively activated inside anaerobes, so its selectivity is for the metabolic environment of the consortium rather than for any BV organism specifically - which is exactly the property a polymicrobial dysbiosis needs. It cures most episodes and prevents few recurrences, and this entry attaches it to the overgrowth node only, not to the biofilm node, because clearing planktonic anaerobes is what it is shown to do.

Mechanism Target:

INHIBITS Polymicrobial Anaerobic Overgrowth — Broad anaerobic coverage suppresses the expanded consortium, which is what produces the 70-85% one-month cure rate.

Show evidence (1 reference)

PMID:34470644 SUPPORT Human Clinical

"These regimens have similar efficacy and cure ~70–85% of women with BV within 1 month"

Quantifies the effect of first-line therapy on the microbiological state, which is the claim this edge makes.

Show evidence (2 references)

PMID:35118003 SUPPORT Human Clinical

"Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended."

Establishes metronidazole as recommended therapy.

PMID:32402161 SUPPORT Human Clinical

"Potentially eligible women completed a standard 5-day course of vaginal 0.75% metronidazole within 30 days before the screening visit."

Documents the standard intravaginal regimen and dose.

Secnidazole

Action: Antibiotic Therapy NCIT:C15620

Agent: secnidazole CHEBI:140628

Platform: Small molecule

A 5-nitroimidazole given as a single 2 g oral dose of granules, FDA-approved for BV in 2017. Mechanistically it is metronidazole's class-mate - the same reductive activation inside anaerobes - so it attaches to the same overgrowth node and not to the biofilm node. What it changes is exposure kinetics rather than target: a roughly 17-hour half-life against metronidazole's 8 hours is what makes one dose sufficient, which matters for a disease whose first-line regimens run 5 to 7 days and whose recurrence is partly a compliance problem. Superiority was shown against placebo rather than against metronidazole, so this entry does not claim it is more effective than first-line therapy.

Mechanism Target:

INHIBITS Polymicrobial Anaerobic Overgrowth — A single dose suppresses the expanded anaerobic consortium sufficiently to normalise discharge, the whiff test and clue-cell burden in about half of treated women.

Show evidence (1 reference)

PMID:28867602 SUPPORT Human Clinical

"Single-dose secnidazole 2 g was superior to placebo for the primary and all secondary efficacy measures in the modified intent-to-treat population, with clinical outcome responder rates of 53.3% (57/107) vs 19.3% (11/57; P < .001)."

Quantifies the effect of a single secnidazole dose on the Amsel-defined microbiological state, which is the claim this edge makes.

Show evidence (1 reference)

PMID:28867602 SUPPORT Human Clinical

"A novel single oral dose granule formulation of secnidazole 2 g, a 5-nitroimidazole with a longer half-life (∼17 hours) than metronidazole (∼8 hours), is being developed to treat bacterial vaginosis."

States the pharmacokinetic difference from metronidazole that is the whole rationale for the single-dose regimen, and confirms the shared nitroimidazole class.

Tinidazole

Action: Antibiotic Therapy NCIT:C15620

Agent: tinidazole CHEBI:63627

Platform: Small molecule

A second-generation 5-nitroimidazole, FDA-approved for BV and a CDC-recommended alternative regimen, given orally as 1 g daily for 5 days or 2 g daily for 2 days. Like secnidazole it shares metronidazole's reductive-activation mechanism and is attached to the overgrowth node only. It is curated here partly because its pivotal trial used the strict five-criterion FDA cure definition, which is why its reported cure rates (27-37%) look far worse than the 70-85% quoted for first-line regimens elsewhere in this entry - the difference is the endpoint, not the drug, and that discrepancy is itself worth recording in a disease whose literature mixes Amsel, Nugent and molecular definitions.

Mechanism Target:

INHIBITS Polymicrobial Anaerobic Overgrowth — Both licensed oral regimens suppress the anaerobic consortium enough to clear all five FDA cure criteria in a significant minority of women, against a 5% placebo rate.

Show evidence (1 reference)

PMID:17666604 SUPPORT Human Clinical

"Superior efficacy was demonstrated by tinidazole for the 1 g once daily for 5 days regimen (36.8% cured, P<.001, number needed to treat 3.2) and for the 2 g once daily for 2 days regimen (27.4% cured, P<.001, number needed to treat 4.5), when compared with placebo (5.1% cured) in the primary..."

Placebo-controlled effect size for both licensed regimens against the consortium, which is the claim this edge makes.

Show evidence (2 references)

PMID:17666604 SUPPORT Human Clinical

"Both tinidazole regimens studied provided effective treatment for bacterial vaginosis."

Establishes tinidazole as effective therapy for BV in a multicentre placebo-controlled trial.

PMID:17666604 SUPPORT Human Clinical

"Using more traditional criteria for cure, efficacy was greater."

Records the endpoint dependence explicitly: the same trial reports a higher cure rate under the conventional definition, so the 27-37% figures are not comparable with the 70-85% quoted from Amsel-based series elsewhere in this entry.

Dequalinium Chloride

Action: antiseptic therapy Ontology label: Pharmacotherapy NCIT:C15986

Agent: dequalinium chloride CHEBI:31466

Platform: Small molecule

A broad-spectrum quaternary-ammonium antiseptic given as a 10 mg vaginal tablet for six days, widely used in Europe and non-inferior to oral metronidazole in a phase 4 double-dummy trial. It is curated as a distinct arm rather than as another antibiotic because it is not one: a membrane-active antiseptic exerts no selection pressure of the kind that drives nitroimidazole resistance, which is the stated reason the trial was run. It attaches to the same overgrowth node, since the evidence is a clinical cure rate and not a demonstration of action on the biofilm.

Mechanism Target:

INHIBITS Polymicrobial Anaerobic Overgrowth — Six days of intravaginal dequalinium chloride clears the Amsel-defined state as often as a seven-day oral metronidazole course.

Show evidence (1 reference)

PMID:38696172 SUPPORT Human Clinical

"The clinical cure rates at visit 1 were 64 of 69 (92.8%) for dequalinium chloride vs 69 of 74 (93.2%) for metronidazole in the intention-to-treat population"

Head-to-head cure rates against the first-line comparator already curated here, supporting an equivalent effect on the consortium.

Show evidence (2 references)

PMID:38696172 SUPPORT Human Clinical

"This randomized clinical trial showed that dequalinium chloride was not inferior to metronidazole for the treatment of BV."

Establishes the non-inferiority conclusion that justifies curating a non-antibiotic alternative alongside the two first-line antimicrobials.

PMID:38696172 SUPPORT Human Clinical

"Due to the increase in antibiotic resistance, effective nonantibiotic treatments for BV are needed."

States the rationale for a non-antibiotic arm, which is why this treatment is curated separately rather than folded into the nitroimidazole group.

Clindamycin

Action: Antibiotic Therapy NCIT:C15620

Agent: clindamycin CHEBI:3745

Platform: Small molecule

Lincosamide alternative to metronidazole, given as 2% intravaginal cream, and the topical component applied to penile skin in the StepUp partner-treatment regimen. Unlike metronidazole it has a defined molecular target in this entry - the bacterial ribosome - which is why it carries a module-conforming target node.

Mechanism Target:

INHIBITS Anaerobe Ribosomal Translation (Clindamycin and Tetracycline Target) — Clindamycin binds the 50S ribosomal subunit and blocks translation in the BV-associated anaerobes.

Show evidence (1 reference)

PMID:40043236 SUPPORT Human Clinical

"the male partner received oral and topical antimicrobial treatment (metronidazole 400-mg tablets and 2% clindamycin cream applied to penile skin, both twice daily for 7 days)"

Documents the clindamycin exposure whose target this edge names.

INHIBITS Polymicrobial Anaerobic Overgrowth — Broad anaerobic coverage equivalent to metronidazole in short-term cure.

Show evidence (1 reference)

PMID:34470644 SUPPORT Human Clinical

"These regimens have similar efficacy and cure ~70–85% of women with BV within 1 month"

Establishes equivalence of the two first-line regimens against the consortium.

Show evidence (1 reference)

PMID:35118003 SUPPORT Human Clinical

"Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended."

Establishes clindamycin as recommended therapy.

LACTIN-V (Lactobacillus crispatus CTV-05 live biotherapeutic)

Action: vaginal live biotherapeutic (Lactobacillus crispatus) administration Ontology label: Pharmacotherapy NCIT:C15986

Platform: Other

A vaginally applied live biotherapeutic containing a naturally occurring human vaginal strain of L. crispatus, given after a course of metronidazole. It is the only intervention curated here that acts on the trigger node rather than on the consortium: it does not kill anaerobes, it repopulates the niche. Recurrence at 12 weeks fell from 45% to 30%, and a randomised immunology substudy showed lower vaginal IL-1alpha and soluble E-cadherin, so the effect reaches the epithelial-activation node as well as the community-state node.

Mechanism Target:

RESTORES Loss of Lactobacillus Dominance and Vaginal Acidification — Exogenous L. crispatus colonises the vagina and re-establishes the protective lactobacillus-dominant community state that defines this node's healthy baseline.

Show evidence (1 reference)

PMID:32402161 SUPPORT Human Clinical

"At the 12-week visit, L. crispatus CTV-05 was detected in 79% of participants in the Lactin-V group."

Demonstrates that the intervention achieves the colonisation this edge asserts, not merely the clinical endpoint.

INHIBITS Epithelial Barrier Disruption and Mucosal Immune Activation — Repletion with L. crispatus lowered vaginal IL-1alpha and soluble E-cadherin, the two prespecified markers of inflammation and barrier disruption at this node.

Show evidence (1 reference)

PMID:35659905 SUPPORT Human Clinical

"The primary outcomes were vaginal levels of IL-1α and soluble E-cadherin at 24 weeks"

Names the endpoints; the trial reported both significantly lower in the LACTIN-V arm, which is the basis for the INHIBITS direction on this edge.

Show evidence (2 references)

PMID:32402161 SUPPORT Human Clinical

"The use of Lactin-V after treatment with vaginal metronidazole resulted in a significantly lower incidence of recurrence of bacterial vaginosis than placebo at 12 weeks."

The primary efficacy result.

PMID:32402161 SUPPORT Human Clinical

"recurrence of bacterial vaginosis by week 12 occurred in 46 participants (30%) in the Lactin-V group and in 34 participants (45%) in the placebo group"

The absolute recurrence rates behind that result.

Concurrent male-partner antimicrobial treatment

Action: Antibiotic Therapy NCIT:C15620

Agent: metronidazole CHEBI:6909 clindamycin CHEBI:3745

Platform: Small molecule

Treating the woman's regular male partner with 7 days of oral metronidazole plus 2% clindamycin cream applied to penile skin, alongside her own first-line therapy. This is the only intervention here aimed at the reinfection route rather than at the woman's own vaginal community, and it is a reversal: six trials in the 1980s-90s were negative and guidelines do not recommend it, but the 2025 StepUp trial was stopped early because withholding it was inferior. Curated as a treatment because the trial result is randomised and positive, with the guideline lag recorded in the notes.

Mechanism Target:

INHIBITS Reinfection from an Untreated Sexual Partner — Eradicating BV-associated bacteria carried by the partner removes the reinoculating source, cutting 12-week recurrence from 63% to 35%.

Show evidence (1 reference)

PMID:40043236 SUPPORT Human Clinical

"recurrence occurred in 24 of 69 women (35%) in the partner-treatment group (recurrence rate, 1.6 per person-year; 95% confidence interval [CI], 1.1 to 2.4) and in 43 of 68 women (63%) in the control group"

The randomised effect size on which this edge rests.

Show evidence (1 reference)

PMID:40043236 SUPPORT Human Clinical

"Adverse events in treated men included nausea, headache, and metallic taste."

Records the harms borne by a person who is not the patient, which is the specific ethical feature of this intervention.

🌍

Environmental Factors

2

Intrauterine device use and non-barrier contraception

No `exposure_term` is bound. ECTO was searched for an intrauterine-device or contraception exposure term and none was found, and binding to a broader device-exposure concept would assert more than the source supports.

In the founding case-control series, clinical BV was correlated with current use of non-barrier contraceptive methods and particularly with an intrauterine device. The mechanism is not established - a foreign body in the endometrial cavity plausibly supports biofilm, but that is inference, not evidence - so this is curated as a predisposing exposure onto the trigger node rather than as a causal trigger.

Show evidence (1 reference)

PMID:6600371 SUPPORT Human Clinical

"current use of nonbarrier contraceptive methods, and, particularly, use of an intrauterine device"

Establishes that this exposure is a real correlate of the disease at all, which is the claim the entry itself makes; the mechanism link above carries the separate claim about which node it acts on. The correlation is from the study that established the modern case definition, so it is a founding observation rather than a later replication.

Mechanism Target:

PREDISPOSES Loss of Lactobacillus Dominance and Vaginal Acidification — Associated with the dysbiotic community shift; the intervening steps are not identified.

Show evidence (1 reference)

PMID:6600371 SUPPORT Human Clinical

"Nonspecific vaginitis was also correlated with a history of sexual activity, a history of previous trichomoniasis, current use of nonbarrier contraceptive methods, and, particularly, use of an intrauterine device."

The epidemiological association this exposure records.

Sexual activity with a regular untreated male partner

No `exposure_term` is bound: ECTO was searched for a sexual-activity or sexual-intercourse exposure term and returned nothing suitable.

Sexual activity has been associated with BV since the original case definition, and BV-associated bacteria are demonstrably exchanged between partners. This entry separates the two claims the association can carry: exposure to a partner as a predisposing factor, curated here, and reinfection from a specific untreated partner as a mechanism of recurrence, curated as its own pathophysiology node with randomised support.

Show evidence (2 references)

PMID:40043236 SUPPORT Human Clinical

"Evidence of sexual exchange of bacterial vaginosis-associated organisms between partners suggests that male-partner treatment may increase the likelihood of cure."

Establishes the exposure itself - an ongoing regular male sexual partner - as relevant to the disease, which is the entry-level claim. The randomised result that the exposure is modifiable is carried by the partner-treatment entry rather than asserted here.

PMID:6600371 SUPPORT INDIRECT Human Clinical

"Nonspecific vaginitis was also correlated with a history of sexual activity"

The founding epidemiological correlation, INDIRECT for the same reason as on the mechanism link: a history of sexual activity is a weaker exposure than an ongoing regular untreated partner, so the entry-level claim follows by an inference step.

Mechanism Target:

PREDISPOSES Reinfection from an Untreated Sexual Partner — Continued sexual contact with an untreated regular partner is the exposure route by which BV-associated bacteria are reacquired after treatment.

Show evidence (2 references)

PMID:34470644 SUPPORT Human Clinical

"There is a robust body of evidence to support the exchange of bacteria between partners during sexual activity"

Establishes partner-to-partner bacterial exchange as the exposure mechanism.

PMID:6600371 SUPPORT INDIRECT Human Clinical

"Nonspecific vaginitis was also correlated with a history of sexual activity"

The original epidemiological correlation. INDIRECT because a history of sexual activity is not the same claim as reinfection from a specific current partner - the link follows from the quote only by an inference step.

📊

Prevalence

2

Reproductive-aged women, worldwide

Point Prevalence 33000.0 per 100,000 >1 in 1,000

Approximately one third of reproductive-aged women, as stated in the StepUp trial report. This is a prevalence of the microbiological state (Nugent/Amsel-defined), not of symptomatic disease - roughly half of affected women report no symptoms - so it should not be read as a burden of clinical illness.

Show evidence (1 reference)

PMID:40043236 SUPPORT Human Clinical

"Bacterial vaginosis affects one third of reproductive-aged women, and recurrence is common."

Source for the one-in-three figure used as the worldwide point prevalence.

Women of reproductive age, United States

Point Prevalence 32500.0 per 100,000 (15000.0–50000.0) >1 in 1,000

The 15-50% range quoted by the LACTIN-V trial. The width of the range is itself informative: it reflects real variation by population and by diagnostic method (Amsel versus Nugent versus molecular assay) rather than measurement noise.

Show evidence (1 reference)

PMID:32402161 SUPPORT Human Clinical

"Bacterial vaginosis affects 15 to 50% of women of reproductive age, and recurrence is common after treatment with an antibiotic agent."

Source for the reproductive-age prevalence range recorded here.

🦠

Infectious Agent

2

Gardnerella vaginalis

A gram-variable, facultatively anaerobic member of the Bifidobacteriales. It is present in essentially all cases of BV and is the organism that forms the adherent biofilm, but it is also carried by a minority of asymptomatic women, so its presence is necessary rather than sufficient for the syndrome.

Gardnerella vaginalis NCBITaxon:2702

Show evidence (2 references)

PMID:16260520 SUPPORT Human Clinical

"only Gardnerella vaginalis developed a characteristic adherent biofilm that was specific for bacterial vaginosis"

Identifies G. vaginalis as the biofilm-forming organism of the BV consortium.

PMID:18390664 SUPPORT Other

"is present in essentially all cases of BV but can also be detected in a minority of asymptomatic women"

Cited as background rather than as a result: this is the authors' review of the clinical-epidemiological literature in their introduction, not an observation from their own in-vitro cytolysin experiments, and it is tagged OTHER for that reason. It supports listing G. vaginalis as an agent while recording that carriage alone does not constitute disease - the reason this entry treats the consortium, not a single organism, as the pathogenic unit.

Prevotella timonensis

An anaerobe of the BV consortium, reclassified as Hoylesella timonensis. It adheres to vaginal and endocervical epithelium as strongly as Gardnerella but provokes less of a proinflammatory epithelial response, and it carries an unusually large complement of mucus-degrading fucosidases and sialidases. It is curated here because the barrier-degrading arm of this entry is not a Gardnerella monopoly. The `term:` binds the current NCBI taxonomy name, Hoylesella timonensis; the literature cited here, including the 2024 papers, still writes Prevotella timonensis, which is retained as the preferred term.

Prevotella timonensis NCBITaxon:386414

Show evidence (1 reference)

PMID:39162399 SUPPORT In Vitro

"We demonstrate that P. timonensis, but not P. bivia, strongly adheres to vaginal and endocervical cells to a similar level as G. vaginalis but did not elicit a comparable proinflammatory epithelial response."

Establishes P. timonensis as an adherent member of the consortium with a distinct host-response profile from Gardnerella.

↔️

Transmission

3

Endogenous overgrowth of resident vaginal anaerobes

The organisms of the BV consortium are already vaginal residents; what changes is their abundance relative to the lactobacilli that normally dominate. There is no exogenous agent to acquire, which is the reason BV has never been classified as a sexually transmitted infection despite its epidemiology, and it is why this record describes a route within the host first.

Show evidence (2 references)

PMID:35118003 SUPPORT REVIEW SYNTHESIS Human Clinical

"BV is characterized by a change in the vaginal flora composition, with a dramatic depletion of Lactobacilli due to a significant overgrowth of obligate or facultative anaerobes previously a minority in the vagina"

States that the organisms were already present as a vaginal minority and that the disease is their overgrowth, which is the endogenous route recorded here.

PMID:35118003 SUPPORT REVIEW SYNTHESIS Human Clinical

"Although the absence of a known causal agent makes it difficult to characterize BV as a sexually transmitted infection (STI)"

Records the reviewers' judgement that BV is not classifiable as an STI, which is why sexual exposure is curated below as an associated route rather than as acquisition of a pathogen.

Sexual exchange of consortium organisms between partners

Sexual activity is nevertheless a route: the consortium organisms are carried by male partners, and treating the partner as well as the woman roughly halves recurrence in a randomised trial stopped early for benefit. That result is the strongest evidence that something is exchanged sexually, whatever the classification of the syndrome.

Show evidence (2 references)

PMID:40043236 SUPPORT BACKGROUND Human Clinical

"Evidence of sexual exchange of bacterial vaginosis-associated organisms between partners suggests that male-partner treatment may increase the likelihood of cure."

States the sexual-exchange premise. Quoted from the trial's background paragraph, which is where the prior evidence for the route is summarised.

PMID:40043236 SUPPORT Human Clinical

"The addition of combined oral and topical antimicrobial therapy for male partners to treatment of women for bacterial vaginosis resulted in a lower rate of recurrence of bacterial vaginosis within 12 weeks than standard care."

The trial's own result: eliminating partner carriage reduces recurrence, which is direct interventional evidence that the partner is a source.

Exposure to semen and other alkalinising sexual exposures

A second, non-microbial sexual route. Contact with highly alkaline semen raises vaginal pH and removes the acid conditions that keep the lactobacilli dominant, so intercourse can precipitate the shift without transferring an organism at all. BV prevalence rises with lifetime partner number and is lowest, though not zero, in women who have never been sexually active.

Show evidence (2 references)

PMID:35118003 SUPPORT REVIEW SYNTHESIS Human Clinical

"it is strongly associated with sexual activities and has some characteristics of a sexually transmitted disease not by microorganism transfer, but by mechanical or chemical interaction such as contact with highly alkaline semen"

Names the mechanical and chemical sexual route, explicitly distinguishing it from transfer of an organism.

PMID:35118003 SUPPORT REVIEW SYNTHESIS Human Clinical

"It has been evaluated at 18.8% for non-sexually active women, 22.4% for women with one lifetime partner and 43.4% and 58% for women having 2-3 lifetime sex partners and those having"

Quantifies the dose-response with lifetime partner number while showing that BV also occurs in women who have never been sexually active, which is the shape a route that contributes without being necessary should have.

🔬

Clinical Trials

3

NCT02766023 PHASE_II COMPLETED

LACTIN-V phase 2b: randomised, double-blind, placebo-controlled trial of vaginal L. crispatus CTV-05 given after a 5-day course of metronidazole gel, with recurrence at 12 weeks as the primary outcome.

Target Phenotypes: Thin homogeneous malodorous vaginal discharge HP:0034269

Show evidence (1 reference)

clinicaltrials:NCT02766023 SUPPORT Other

"To assess the safety of LACTIN-V over 24 weeks by comparing the incidence of AEs between individuals randomized to LACTIN-V or placebo."

Registry record establishing the trial's design and duration.

NCT00229216 PHASE_III COMPLETED

The registration trial for oral tinidazole in bacterial vaginosis: multicentre, double-blind, double-dummy, placebo-controlled, comparing 1 g for 5 days and 2 g for 2 days against placebo under the strict five-criterion FDA cure definition (PMID:17666604).

Target Phenotypes: Thin homogeneous malodorous vaginal discharge HP:0034269

Show evidence (1 reference)

clinicaltrials:NCT00229216 SUPPORT Other

"The purpose of this study is to confirm the safety and efficacy of oral tinidazole for the treatment of bacterial vaginosis."

Registry record establishing the trial's identity and objective.

ACTRN12619000196145 PHASE_III COMPLETED

StepUp: open-label randomised controlled trial of treating male partners of women being treated for bacterial vaginosis, to reduce recurrence. Registered on ANZCTR and therefore keyed on its WHO ICTRP identifier rather than an NCT number; reported in 2025 (PMID:40043236) after being stopped early by its data and safety monitoring board.

Show evidence (2 references)

ICTRP:ACTRN12619000196145 SUPPORT Other

"Scientific title: Treating male partners of women with bacterial vaginosis (BV) to reduce recurrence: randomised controlled trial"

WHO ICTRP registration record establishing the trial's identity and objective.

ICTRP:ACTRN12619000196145 SUPPORT Other

"Male partners randomised to the Intervention Group will receive oral MTZ 400mg tablets twice daily and topical 2% clindamycin cream to be applied to the glans penis and upper shaft (under the foreskin if uncircumcised) twice daily for 7 days."

The registered intervention, matching the regimen curated under the partner treatment entry.

🐁

Animal Models

1

Murine cervicovaginal colonisation with BV-associated anaerobes

Mice colonised with high-risk cervicovaginal taxa identified in the South African cohort. Used to test whether the association between a lactobacillus-deficient community and activated mucosal CD4+ T cells is causal rather than confounded, since that direction cannot be established observationally in women.

Species

Mouse

Genotype

Wild type

Publication

PMID:28087240

Pathograph links

Epithelial Barrier Disruption and Mucosal Immune Activation partially recapitulates fidelity: low Reproduces the target-cell recruitment arm of this node - high-risk bacteria increased activated genital CD4+ T cells - but not the node as a whole.

Limitations: The mouse has no lactobacillus-dominant vaginal community and no acidic vaginal lumen, so the community state whose loss defines this disease does not exist to be perturbed. The model can show that particular taxa activate mucosal T cells; it cannot model dysbiosis. It also cannot address the vaginolysin arm at all, because that toxin is restricted to human CD59.

Readouts

Activated genital CD4+ T cell number

Direction: INCREASED

Interpretation: Colonisation with high-risk taxa raises the number of activated mucosal CD4+ T cells, the HIV target-cell population this node supplies.

Show evidence (1 reference)

PMID:28087240 SUPPORT Model Organism

"high-risk bacteria increased numbers of activated genital CD4+ T cells in a murine model"

Reports the measurement and its direction in the mouse.

Show evidence (1 reference)

PMID:28087240 SUPPORT Model Organism

"We identified specific bacterial taxa linked with reduced (L. crispatus) or elevated (Prevotella, Sneathia, and other anaerobes) inflammation and HIV infection and found that high-risk bacteria increased numbers of activated genital CD4+ T cells in a murine model."

Supports treating this model as informative for the immune-activation node, and names the taxa tested.

{ }

Source YAML

click to show

name: Bacterial Vaginosis
creation_date: "2026-08-20T05:30:00Z"
category: Infectious Disease
disease_term:
  preferred_term: bacterial vaginosis
  term:
    id: MONDO:0005316
    label: bacterial vaginosis
synonyms:
- BV
- nonspecific vaginitis
- vaginal dysbiosis
- anaerobic vaginosis
parents:
- Female reproductive system disorder
- Bacterial infectious disease
description: >
  Bacterial vaginosis is a polymicrobial dysbiosis in which the lactobacillus-dominated
  vaginal community is replaced by a dense consortium of facultative and strict anaerobes
  - Gardnerella, Prevotella (Hoylesella), Fannyhessea/Atopobium, Mobiluncus and others -
  with loss of lactic acid production and a rise in vaginal pH. Three features make it
  mechanistically distinctive. First, no single organism satisfies Koch's postulates: the
  disease is a community state, and the same taxa are recoverable from women without
  disease, so the pathogenic unit is the consortium and its adherent biofilm rather than a
  pathogen. Second, the damage that matters is done to a barrier rather than to a tissue -
  bacterial sialidases and glycosidases strip the cervicovaginal mucus of its protective
  glycans, and the resulting loss of mucus adhesive function, not inflammation, is what
  raises the risk of HIV acquisition and of ascending infection in pregnancy. Third, the
  clinical problem is recurrence, not cure: first-line antimicrobials clear most episodes
  and more than half of women relapse within six months, and the two candidate
  explanations - survival of the adherent biofilm and reinfection from an untreated sexual
  partner - are clinically indistinguishable in an individual woman. A 2025 randomised
  trial of male-partner treatment that was stopped early for benefit is the strongest
  evidence yet that the second route is real.
references:
- reference: PMID:35118003
  title: "Bacterial Vaginosis: What Do We Currently Know?"
  findings:
  - statement: >
      Contemporary review of the vaginal microbiome, the community-state-type framework,
      the Amsel/Nugent diagnostic standards and first-line antimicrobial therapy.
  - statement: >
      States plainly that the etiology of the dysbiosis is unknown and that relapse after
      standard therapy approaches 50% at six months.
- reference: PMID:34470644
  title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
  findings:
  - statement: >
      Sets out the two competing accounts of recurrence - biofilm/BVAB persistence versus
      reinfection from an untreated partner - and the evidence for sexual exchange of
      BV-associated bacteria.
  - statement: >
      Notes that the two routes cannot be separated clinically because their presentation
      is identical.
- reference: PMID:40043236
  title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
  findings:
  - statement: >
      Open-label randomised trial (StepUp) stopped early by its DSMB because treating the
      woman alone was inferior to treating both partners.
- reference: PMID:16260520
  title: "Adherent biofilms in bacterial vaginosis."
  findings:
  - statement: >
      Vaginal biopsy FISH study identifying a confluent Gardnerella-dominated adherent
      biofilm as the structural feature specific to bacterial vaginosis.
- reference: PMID:31971984
  title: "The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis."
  findings:
  - statement: >
      HIV virion mobility is increased in cervicovaginal mucus from women with BV, and the
      defect is in the adhesive rather than the structural properties of the mucus.

infectious_agent:
- name: Gardnerella vaginalis
  description: >-
    A gram-variable, facultatively anaerobic member of the Bifidobacteriales. It is
    present in essentially all cases of BV and is the organism that forms the adherent
    biofilm, but it is also carried by a minority of asymptomatic women, so its presence
    is necessary rather than sufficient for the syndrome.
  infectious_agent_term:
    preferred_term: Gardnerella vaginalis
    term:
      id: NCBITaxon:2702
      label: Gardnerella vaginalis
  evidence:
  - reference: PMID:16260520
    reference_title: "Adherent biofilms in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "only Gardnerella vaginalis developed a characteristic adherent biofilm that was specific for bacterial vaginosis"
    explanation: >-
      Identifies G. vaginalis as the biofilm-forming organism of the BV consortium.
  - reference: PMID:18390664
    reference_title: "Functional and phylogenetic characterization of Vaginolysin, the human-specific cytolysin from Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "is present in essentially all cases of BV but can also be detected in a minority of asymptomatic women"
    explanation: >-
      Cited as background rather than as a result: this is the authors' review of the
      clinical-epidemiological literature in their introduction, not an observation from
      their own in-vitro cytolysin experiments, and it is tagged OTHER for that reason. It
      supports listing G. vaginalis as an agent while recording that carriage alone does
      not constitute disease - the reason this entry treats the consortium, not a single
      organism, as the pathogenic unit.
- name: Prevotella timonensis
  description: >-
    An anaerobe of the BV consortium, reclassified as Hoylesella timonensis. It adheres to
    vaginal and endocervical epithelium as strongly as Gardnerella but provokes less of a
    proinflammatory epithelial response, and it carries an unusually large complement of
    mucus-degrading fucosidases and sialidases. It is curated here because the
    barrier-degrading arm of this entry is not a Gardnerella monopoly. The `term:` binds
    the current NCBI taxonomy name, Hoylesella timonensis; the literature cited here,
    including the 2024 papers, still writes Prevotella timonensis, which is retained as
    the preferred term.
  infectious_agent_term:
    preferred_term: Prevotella timonensis
    term:
      id: NCBITaxon:386414
      label: Hoylesella timonensis
  evidence:
  - reference: PMID:39162399
    reference_title: "Prevotella timonensis degrades the vaginal epithelial glycocalyx through high fucosidase and sialidase activities."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We demonstrate that P. timonensis, but not P. bivia, strongly adheres to vaginal and endocervical cells to a similar level as G. vaginalis but did not elicit a comparable proinflammatory epithelial response."
    explanation: >-
      Establishes P. timonensis as an adherent member of the consortium with a distinct
      host-response profile from Gardnerella.

transmission:
- name: Endogenous overgrowth of resident vaginal anaerobes
  description: >-
    The organisms of the BV consortium are already vaginal residents; what changes is
    their abundance relative to the lactobacilli that normally dominate. There is no
    exogenous agent to acquire, which is the reason BV has never been classified as a
    sexually transmitted infection despite its epidemiology, and it is why this record
    describes a route within the host first.
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "BV is characterized by a change in the vaginal flora composition, with a dramatic depletion of Lactobacilli due to a significant overgrowth of obligate or facultative anaerobes previously a minority in the vagina"
    explanation: >-
      States that the organisms were already present as a vaginal minority and that the
      disease is their overgrowth, which is the endogenous route recorded here.
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "Although the absence of a known causal agent makes it difficult to characterize BV as a sexually transmitted infection (STI)"
    explanation: >-
      Records the reviewers' judgement that BV is not classifiable as an STI, which is why
      sexual exposure is curated below as an associated route rather than as acquisition
      of a pathogen.
- name: Sexual exchange of consortium organisms between partners
  description: >-
    Sexual activity is nevertheless a route: the consortium organisms are carried by male
    partners, and treating the partner as well as the woman roughly halves recurrence in
    a randomised trial stopped early for benefit. That result is the strongest evidence
    that something is exchanged sexually, whatever the classification of the syndrome.
  evidence:
  - reference: PMID:40043236
    reference_title: Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: BACKGROUND
    snippet: "Evidence of sexual exchange of bacterial vaginosis-associated organisms between partners suggests that male-partner treatment may increase the likelihood of cure."
    explanation: >-
      States the sexual-exchange premise. Quoted from the trial's background paragraph,
      which is where the prior evidence for the route is summarised.
  - reference: PMID:40043236
    reference_title: Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The addition of combined oral and topical antimicrobial therapy for male partners to treatment of women for bacterial vaginosis resulted in a lower rate of recurrence of bacterial vaginosis within 12 weeks than standard care."
    explanation: >-
      The trial's own result: eliminating partner carriage reduces recurrence, which is
      direct interventional evidence that the partner is a source.
- name: Exposure to semen and other alkalinising sexual exposures
  description: >-
    A second, non-microbial sexual route. Contact with highly alkaline semen raises
    vaginal pH and removes the acid conditions that keep the lactobacilli dominant, so
    intercourse can precipitate the shift without transferring an organism at all. BV
    prevalence rises with lifetime partner number and is lowest, though not zero, in
    women who have never been sexually active.
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "it is strongly associated with sexual activities and has some characteristics of a sexually transmitted disease not by microorganism transfer, but by mechanical or chemical interaction such as contact with highly alkaline semen"
    explanation: >-
      Names the mechanical and chemical sexual route, explicitly distinguishing it from
      transfer of an organism.
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    quote_role: REVIEW_SYNTHESIS
    snippet: "It has been evaluated at 18.8% for non-sexually active women, 22.4% for women with one lifetime partner and 43.4% and 58% for women having 2-3 lifetime sex partners and those having"
    explanation: >-
      Quantifies the dose-response with lifetime partner number while showing that BV also
      occurs in women who have never been sexually active, which is the shape a route that
      contributes without being necessary should have.

prevalence:
- population: Reproductive-aged women, worldwide
  measure_type: POINT_PREVALENCE
  prevalence_class: ABOVE_1_IN_1000
  rate_per_100000: 33000.0
  notes: >
    Approximately one third of reproductive-aged women, as stated in the StepUp trial
    report. This is a prevalence of the microbiological state (Nugent/Amsel-defined), not
    of symptomatic disease - roughly half of affected women report no symptoms - so it
    should not be read as a burden of clinical illness.
  evidence:
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Bacterial vaginosis affects one third of reproductive-aged women, and recurrence is common."
    explanation: >-
      Source for the one-in-three figure used as the worldwide point prevalence.
- population: Women of reproductive age, United States
  measure_type: POINT_PREVALENCE
  prevalence_class: ABOVE_1_IN_1000
  rate_per_100000: 32500.0
  rate_low: 15000.0
  rate_high: 50000.0
  notes: >
    The 15-50% range quoted by the LACTIN-V trial. The width of the range is itself
    informative: it reflects real variation by population and by diagnostic method
    (Amsel versus Nugent versus molecular assay) rather than measurement noise.
  evidence:
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Bacterial vaginosis affects 15 to 50% of women of reproductive age, and recurrence is common after treatment with an antibiotic agent."
    explanation: >-
      Source for the reproductive-age prevalence range recorded here.

pathophysiology:
- name: Loss of Lactobacillus Dominance and Vaginal Acidification
  biological_scale: TISSUE
  role: trigger
  description: >
    The initiating state change. A healthy premenopausal vagina is dominated by
    Lactobacillus species whose copious lactic acid holds the lumen below pH 4.5; in BV
    that community is replaced, lactic acid production falls and pH rises. Two points
    matter for everything downstream. The species identity is not interchangeable -
    L. crispatus-dominated communities are the most protective, whereas L. iners, which
    makes only the L-isomer of lactic acid, sits at an intermediate and unstable position
    and is the species most associated with transition into and relapse after BV. And the
    trigger itself is unexplained: what precipitates the shift in an individual woman is
    the central unanswered question of the field, curated as a knowledge gap below.
    `GO:0006885 regulation of pH` is deliberately not bound here: the pH claim this node
    makes is about the chemistry of the vaginal lumen, not about a cellular homeostatic
    process, and GO annotates the latter. The mechanism that produces the pH change is
    carried by the lactate term below, and the luminal measurement itself is curated as an
    Amsel criterion under `definitions`.
  biological_processes:
  - preferred_term: lactate biosynthetic process
    term:
      id: GO:0019249
      label: lactate biosynthetic process
    modifier: DECREASED
  locations:
  - preferred_term: vagina
    term:
      id: UBERON:0000996
      label: vagina
  downstream:
  - target: Polymicrobial Anaerobic Overgrowth
    description: >-
      Relief of lactic-acid-mediated suppression permits expansion of facultative and
      strict anaerobes.
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "BV is the result of a disturbance in the vaginal ecosystem; i.e., a sudden replacement of Lactobacilli by anaerobic bacteria such as Gardnerella vaginalis, Atopobium vaginae, Ureaplasma urealyticum, Mycoplasma hominis, and others."
    explanation: >-
      States the defining community shift that this node encodes.
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Under normal conditions, 70-90% of the vaginal bacterial species in healthy premenopausal women are Lactobacilli"
    explanation: >-
      Establishes the baseline lactobacillus dominance whose loss defines this node.
  - reference: PMID:37234911
    reference_title: "Lactobacillus iners and genital health: molecular clues to an enigmatic vaginal species."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Lactobacillus iners is distinct from L. crispatus, L. gasseri, and L. jensenii by its high global prevalence in vaginal microbiomes, relatively small genome, production of only L-lactic acid, and inconsistent associations with genital health outcomes."
    explanation: >-
      Supports the claim that lactobacillus species identity, not merely lactobacillus
      presence, determines how protective this node's baseline state is.
  - reference: PMID:37234911
    reference_title: "Lactobacillus iners and genital health: molecular clues to an enigmatic vaginal species."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Vaginal lactobacilli are recognized as important drivers of genital health including protection against bacterial vaginosis and sexually transmitted infections."
    explanation: >-
      Establishes the protective role whose loss initiates the chain.

- name: Polymicrobial Anaerobic Overgrowth
  biological_scale: TISSUE
  description: >
    Expansion of a dense, diverse consortium - Gardnerella spp., Prevotella (Hoylesella)
    spp., Fannyhessea (Atopobium) vaginae, Mobiluncus, Megasphaera, Sneathia and others -
    to concentrations orders of magnitude above baseline. This node is deliberately named
    for the community rather than for an organism: no single member reproduces the
    syndrome on inoculation, and each is recoverable from women without disease. Treating
    the consortium as the unit is what makes the broad anaerobic coverage of metronidazole
    and clindamycin the rational therapy, and it is why an organism-specific agent has
    never been developed.
  biological_processes:
  - preferred_term: response to bacterium
    term:
      id: GO:0009617
      label: response to bacterium
    modifier: INCREASED
  locations:
  - preferred_term: vagina
    term:
      id: UBERON:0000996
      label: vagina
  downstream:
  - target: Biogenic Amine Production and Amplified Lactobacillus Suppression
    description: >-
      Anaerobic amino acid decarboxylation generates the biogenic amines that give BV its
      odour and further suppress the residual lactobacilli.
    causal_link_type: DIRECT
  - target: Epithelial Adhesion and Gardnerella-Dominated Adherent Biofilm
    description: >-
      Members of the expanded consortium adhere to the epithelium and organise into a
      structured biofilm.
    causal_link_type: DIRECT
  - target: Bacterial Sialidase and Glycosidase Degradation of the Mucus Barrier
    description: >-
      The consortium secretes the mucin-degrading enzyme activities that are elevated in
      BV vaginal fluid.
    causal_link_type: DIRECT
  - target: Anaerobe Ribosomal Translation (Clindamycin and Tetracycline Target)
    description: >-
      Sustained expansion of the consortium depends on bacterial protein synthesis, which
      is the molecular target exploited by the lincosamide arm of therapy.
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "This non-optimal microbiological state involves a reduction in protective lactobacilli, and an increase in bacterial diversity and facultative and strict anaerobes, including Gardnerella spp., Atopobium vaginae, Prevotella spp., and others, referred to as BV-associated bacteria (BVAB)"
    explanation: >-
      Names the consortium taxa and frames BV as a community state rather than a
      single-organism infection.
  - reference: PMID:6600371
    reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A clinical diagnosis of nonspecific vaginitis, based on simple office procedures, was correlated with both the presence and the concentration of Gardnerella vaginalis (Hemophilus vaginalis) in vaginal discharge, and with characteristic biochemical findings in vaginal discharge."
    explanation: >-
      The founding correlation between organism concentration - not mere presence - and
      clinical disease, which is the quantitative claim this node makes.

- name: Biogenic Amine Production and Amplified Lactobacillus Suppression
  biological_scale: MOLECULAR
  description: >
    Amino acid decarboxylation by the expanded anaerobes produces putrescine, cadaverine,
    tyramine and trimethylamine. These are usually treated as a symptom - they are what
    the whiff test detects and what patients describe as a fishy odour - but the
    measurements curated here make them a mechanism as well: they inhibit the growth of
    vaginal lactobacilli and independently reduce their lactic acid output. That closes a
    positive feedback loop back onto the trigger node, which is why the dysbiotic state is
    self-reinforcing rather than self-limiting, and it is the reason this node is curated
    separately from the odour phenotype it produces.
  biological_processes:
  - preferred_term: amine biosynthetic process
    term:
      id: GO:0009309
      label: amine biosynthetic process
    modifier: INCREASED
  chemical_entities:
  - preferred_term: putrescine
    term:
      id: CHEBI:17148
      label: putrescine
    modifier: INCREASED
  - preferred_term: cadaverine
    term:
      id: CHEBI:18127
      label: cadaverine
    modifier: INCREASED
  - preferred_term: trimethylamine
    term:
      id: CHEBI:18139
      label: trimethylamine
    modifier: INCREASED
  downstream:
  - target: Loss of Lactobacillus Dominance and Vaginal Acidification
    description: >-
      Biogenic amines slow lactobacillus growth and cut lactic acid production, feeding
      back on and entrenching the initiating state change.
    causal_link_type: DIRECT
  - target: Thin homogeneous malodorous vaginal discharge
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    description: BV-associated amines produce the malodor detected by the whiff test.
  evidence:
  - reference: PMID:33674429
    reference_title: "Biogenic Amines Increase the Odds of Bacterial Vaginosis and Affect the Growth of and Lactic Acid Production by Vaginal Lactobacillus spp."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Increases in cadaverine, putrescine, and tyramine were associated with greater odds of women transitioning from L. crispatus-dominated vaginal microbiota to microbiota that have a paucity of Lactobacillus spp. and from Nugent scores of 0 to 3 to Nugent scores of 7 to 10, consistent with BV."
    explanation: >-
      The in vivo longitudinal association between rising biogenic amines and transition
      into BV, which is what licenses treating this node as upstream rather than merely
      concurrent.
  - reference: PMID:33674429
    reference_title: "Biogenic Amines Increase the Odds of Bacterial Vaginosis and Affect the Growth of and Lactic Acid Production by Vaginal Lactobacillus spp."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "BAs were associated with reduced production of d- and l-lactic acid by vaginal Lactobacillus spp., and this effect was independent of their effect upon Lactobacillus species growth."
    explanation: >-
      The axenic-culture experiment establishing the feedback edge onto lactic acid
      production, and establishing it as separable from a simple growth effect.
  - reference: PMID:33674429
    reference_title: "Biogenic Amines Increase the Odds of Bacterial Vaginosis and Affect the Growth of and Lactic Acid Production by Vaginal Lactobacillus spp."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Exposure to putrescine lengthened the lag time and/or slowed the growth of all vaginal Lactobacillus spp. except L. jensenii 62G."
    explanation: >-
      Direct growth inhibition of lactobacilli by a BV-associated amine, with the single
      exception recorded rather than smoothed over.

- name: Epithelial Adhesion and Gardnerella-Dominated Adherent Biofilm
  biological_scale: TISSUE
  description: >
    Gardnerella adheres to the vaginal epithelium and forms a confluent, structured
    biofilm into which other consortium members are incorporated. This is the one
    structural feature that FISH on vaginal biopsies found specific to BV rather than
    merely more abundant in it, and it is the physical object that makes BV a persistent
    rather than an episodic condition: a sessile, matrix-embedded community tolerates
    antimicrobial exposure that clears planktonic organisms. Prevotella (Hoylesella)
    timonensis adheres comparably well, so adhesion is not a Gardnerella monopoly even
    though biofilm architecture appears to be.
  cell_types:
  - preferred_term: vaginal squamous epithelial cell
    term:
      id: CL:1001578
      label: vagina squamous cell
  biological_processes:
  - preferred_term: biofilm formation
    term:
      id: GO:0042710
      label: biofilm formation
    modifier: INCREASED
  - preferred_term: cell adhesion involved in biofilm formation
    term:
      id: GO:0043708
      label: cell adhesion involved in biofilm formation
    modifier: INCREASED
  locations:
  - preferred_term: vagina
    term:
      id: UBERON:0000996
      label: vagina
  downstream:
  - target: Exfoliation of Bacteria-Coated Epithelial Cells (Clue Cells)
    description: >-
      Adherent organisms are shed with the epithelial cells they coat, producing the clue
      cells seen on wet mount.
    causal_link_type: DIRECT
  - target: Vaginolysin-Mediated Epithelial Pore Formation and Signalling
    description: >-
      Close apposition of Gardnerella to the epithelium delivers its cholesterol-dependent
      cytolysin to the host cell membrane.
    causal_link_type: DIRECT
  - target: Biofilm Persistence After Antimicrobial Therapy
    description: >-
      The adherent biofilm is the proposed reservoir from which BV-associated bacteria
      re-emerge after a course of antibiotics.
    causal_link_type: DIRECT
    hypothesis_groups:
    - biofilm_persistence_relapse
  evidence:
  - reference: PMID:16260520
    reference_title: "Adherent biofilms in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A biofilm comprised of confluent G vaginalis with other bacterial groups incorporated in the adherent layer is a prominent feature of bacterial vaginosis."
    explanation: >-
      Describes the composite structure this node asserts - a Gardnerella scaffold with
      other taxa incorporated.
  - reference: PMID:16260520
    reference_title: "Adherent biofilms in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Bacterial vaginosis was associated with greater occurrence and higher concentrations of a variety of bacterial groups. However, only Gardnerella vaginalis developed a characteristic adherent biofilm that was specific for bacterial vaginosis."
    explanation: >-
      Draws the distinction this node depends on: many taxa are more abundant, but only
      one forms a BV-specific adherent structure.
  - reference: PMID:39162399
    reference_title: "Prevotella timonensis degrades the vaginal epithelial glycocalyx through high fucosidase and sialidase activities."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We demonstrate that P. timonensis, but not P. bivia, strongly adheres to vaginal and endocervical cells to a similar level as G. vaginalis but did not elicit a comparable proinflammatory epithelial response."
    explanation: >-
      The study's own adhesion assay establishes that adherence to vaginal and endocervical
      epithelium is a property of the consortium rather than of Gardnerella alone, and that
      it is not shared by every anaerobe (P. bivia does not adhere).

- name: Exfoliation of Bacteria-Coated Epithelial Cells (Clue Cells)
  biological_scale: CELLULAR
  description: >
    Squamous epithelial cells so heavily coated with adherent organisms that their borders
    are obscured are shed into the vaginal fluid. Immunofluorescence on clue cells
    identifies the adherent rods as Gardnerella rather than Mobiluncus, Bacteroides or
    Fusobacterium, which is the observation that ties the microscopic diagnostic sign to a
    specific adhesion mechanism rather than to generic anaerobic overgrowth. This node is
    curated as a cellular event, not as a diagnostic test; the test that reads it is
    curated under `definitions`.
  cell_types:
  - preferred_term: vaginal squamous epithelial cell
    term:
      id: CL:1001578
      label: vagina squamous cell
  locations:
  - preferred_term: vagina
    term:
      id: UBERON:0000996
      label: vagina
  evidence:
  - reference: PMID:2668431
    reference_title: "Clue cells in bacterial vaginosis: immunofluorescent identification of the adherent gram-negative bacteria as Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Clue cells are epithelial cells covered by adherent gram-negative rods, observed in vaginal smears from women with bacterial vaginosis."
    explanation: >-
      Defines the cellular object this node names.
  - reference: PMID:2668431
    reference_title: "Clue cells in bacterial vaginosis: immunofluorescent identification of the adherent gram-negative bacteria as Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Gardnerella vaginalis was most often observed adhering to the surface of clue cells and was detected on the surface of exfoliated vaginal epithelial cells significantly more frequently and in higher numbers than were Mobiluncus, Bacteroides, and Fusobacterium, suggesting that this species of gram-negative bacteria is responsible for clue cell formation."
    explanation: >-
      Attributes clue cell formation specifically to Gardnerella adhesion, which is the
      causal claim on the incoming edge from the biofilm node.

- name: Vaginolysin-Mediated Epithelial Pore Formation and Signalling
  biological_scale: MOLECULAR
  description: >
    Gardnerella secretes vaginolysin, a cholesterol-dependent cytolysin whose receptor is
    human CD59. Two consequences matter and they are different in kind. At lytic
    concentrations the toxin kills; at sublytic concentrations - which is the physiological
    situation - it triggers rapid membrane blebbing, activates epithelial p38 MAPK and
    induces IL-8. The receptor dependence is the reason BV has no natural animal model:
    vaginolysin is human-specific because CD59 is, and transfecting human CD59 into
    non-human cells is what confers susceptibility. That constraint is curated as a
    human/model mismatch below rather than left implicit.
  biological_processes:
  - preferred_term: cytolysis
    term:
      id: GO:0019835
      label: cytolysis
    modifier: INCREASED
  - preferred_term: killing of cells of another organism
    term:
      id: GO:0031640
      label: killing of cells of another organism
    modifier: INCREASED
  cell_types:
  - preferred_term: vaginal squamous epithelial cell
    term:
      id: CL:1001578
      label: vagina squamous cell
  downstream:
  - target: Epithelial Barrier Disruption and Mucosal Immune Activation
    description: >-
      Sublytic pore formation drives p38 MAPK signalling and IL-8 release from the
      vaginal epithelium.
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:18390664
    reference_title: "Functional and phylogenetic characterization of Vaginolysin, the human-specific cytolysin from Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We demonstrate that G. vaginalis produces a toxin (vaginolysin [VLY]) that is a member of the cholesterol-dependent cytolysin (CDC) family, most closely related to intermedilysin from Streptococcus intermedius."
    explanation: >-
      Identifies the toxin and its family, the molecular claim of this node.
  - reference: PMID:18390664
    reference_title: "Functional and phylogenetic characterization of Vaginolysin, the human-specific cytolysin from Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "In addition to causing erythrocyte lysis, VLY activates the conserved epithelial p38 mitogen-activated protein kinase pathway and induces interleukin-8 production by human epithelial cells."
    explanation: >-
      Establishes the sublytic signalling output that carries the outgoing edge to
      epithelial activation, distinct from frank lysis.
  - reference: PMID:24082080
    reference_title: "Vaginolysin drives epithelial ultrastructural responses to Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Binding of VLY to its human-specific receptor (hCD59) is required for bleb formation, as antibody inhibition of either toxin or hCD59 abrogates this response, and transfection of nonhuman cells (CHO-K1) with hCD59 renders them susceptible to toxin-induced membrane blebbing."
    explanation: >-
      The gain- and loss-of-function pair establishing hCD59 dependence, and with it the
      species restriction recorded in the human/model-mismatch discussion.
  - reference: PMID:24082080
    reference_title: "Vaginolysin drives epithelial ultrastructural responses to Gardnerella vaginalis."
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: IN_VITRO
    snippet: "VLY-induced epithelial cell membrane blebbing in the vaginal mucosa may play a role in the pathogenesis of BV"
    explanation: >-
      The authors' own hedged statement of relevance to disease. INDIRECT because the
      blebbing was demonstrated in cultured epithelium and its in vivo contribution to BV
      pathogenesis is proposed rather than shown, so the node's claim follows from the
      quote only by an inference step from the model system to the mucosa.

- name: Bacterial Sialidase and Glycosidase Degradation of the Mucus Barrier
  biological_scale: MOLECULAR
  description: >
    The consortium secretes sialidases, fucosidases, beta-galactosidase and
    beta-N-acetylhexosaminidase that strip terminal sugars from cervicovaginal mucins and
    from the epithelial glycocalyx. This is the arm of BV pathogenesis with the clearest
    line to hard clinical outcomes, and it has been misattributed for two decades:
    biochemical work focused on Gardnerella, but vaginal Prevotella species are now shown
    to be a major and globally conserved source of vaginal sialidase, with
    P. (Hoylesella) timonensis carrying both sialidases and an unusual complement of
    fucosidases. Notably, the same 1999 study that found sialidase, beta-galactosidase and
    beta-N-acetylhexosaminidase elevated in BV found no increase in O-glycanase,
    proteinase, sulphatase or whole-mucinase activity - the degradation is glycosidic and
    selective, not general proteolysis, and that distinction is preserved here rather than
    flattened into "mucinase activity".
  molecular_functions:
  - preferred_term: exo-alpha-sialidase activity
    term:
      id: GO:0004308
      label: exo-alpha-sialidase activity
    modifier: INCREASED
  chemical_entities:
  - preferred_term: sialic acid
    term:
      id: CHEBI:26667
      label: sialic acid
    modifier: DECREASED
  locations:
  - preferred_term: mucus
    term:
      id: UBERON:0000912
      label: mucus
  downstream:
  - target: Loss of Cervicovaginal Mucus Adhesive Barrier Function
    description: >-
      Removal of terminal sialic acid and fucose from mucin glycans degrades the adhesive
      trapping function of the mucus gel.
    causal_link_type: DIRECT
  - target: Ascending Infection and Adverse Pregnancy Outcomes
    description: >-
      Enzymatic breach of the cervicovaginal mucosal barrier is the proposed route by
      which BV organisms reach the uterus; detectable vaginal sialidase in preterm labour
      identifies women at higher risk of early preterm birth.
    causal_link_type: INDIRECT_KNOWN_INTERMEDIATES
  evidence:
  - reference: PMID:10454178
    reference_title: "Mucinase and sialidase activity of the vaginal microflora: implications for the pathogenesis of preterm labour."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Significant increases in activity were detected in BV samples for sialidase using a mucin (BSM P<0.005) and serum type glycoprotein (AGP P<0.005) substrates, beta-galactosidase (P<0.001), and beta-N-acetylhexosaminidase (P<0.01)."
    explanation: >-
      Quantifies the specific glycosidase activities elevated in BV vaginal fluid.
  - reference: PMID:10454178
    reference_title: "Mucinase and sialidase activity of the vaginal microflora: implications for the pathogenesis of preterm labour."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "No significant increases in BV patients were detected in O-glycanase, proteinase, arylesterase, sulphatase or whole mucinase activities."
    explanation: >-
      The negative half of the same study, retained because it is what makes the
      degradation glycosidic and selective rather than general. It supports this node by
      exclusion: selectivity is part of what the node asserts, so a null result for
      proteinase, sulphatase and whole-mucinase activity is evidence for the glycosidic
      mechanism rather than a qualification of it.
  - reference: PMID:10454178
    reference_title: "Mucinase and sialidase activity of the vaginal microflora: implications for the pathogenesis of preterm labour."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "These results support the hypothesis that certain BV-associated enzymes may detrimentally affect the mucosal barrier, permitting bacteria access to the uterus."
    explanation: >-
      States the barrier-breach consequence carried by this node's outgoing edges.
  - reference: PMID:39186657
    reference_title: "Prevotella are major contributors of sialidases in the human vaginal microbiome."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Here, we show that vaginal Prevotella species produce sialidases that possess variable activity toward mucin substrates."
    explanation: >-
      Establishes Prevotella as a sialidase source, correcting the Gardnerella-only
      attribution this node warns about.
  - reference: PMID:39186657
    reference_title: "Prevotella are major contributors of sialidases in the human vaginal microbiome."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "Elevated bacterial sialidase activity in the female genital tract is strongly associated with poor health outcomes including preterm birth and bacterial vaginosis (BV). These negative effects may arise from sialidase-mediated degradation of the protective mucus layer in the cervicovaginal environment."
    explanation: >-
      States the sialidase-to-outcome link and the proposed mucus-degradation mechanism
      that organises this node.
  - reference: PMID:39162399
    reference_title: "Prevotella timonensis degrades the vaginal epithelial glycocalyx through high fucosidase and sialidase activities."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "The P. timonensis genome uniquely encodes a large set of mucus-degrading enzymes, including four putative fucosidases and two putative sialidases, PtNanH1 and PtNanH2."
    explanation: >-
      Identifies the specific enzymes behind the Prevotella contribution to this node.

- name: Loss of Cervicovaginal Mucus Adhesive Barrier Function
  biological_scale: TISSUE
  description: >
    Cervicovaginal mucus normally traps virions and particles. In BV that trapping fails:
    fluorescently labelled HIV virions move significantly faster through mucus from women
    with BV than through mucus from L. crispatus-dominant women, whether or not the woman
    has symptoms. The mechanistically important finding is *how* it fails - electron
    microscopy and nanoparticle work localise the defect to reduced adhesive interactions,
    with the physical pore structure of the gel intact. The barrier is not torn open; it
    stops being sticky. The same loss of trapping is seen with L. iners-dominant mucus,
    which is why treating BV to a Nugent-normal endpoint does not necessarily restore the
    barrier.
  locations:
  - preferred_term: mucus
    term:
      id: UBERON:0000912
      label: mucus
  - preferred_term: uterine cervix
    term:
      id: UBERON:0000002
      label: uterine cervix
  downstream:
  - target: Increased Susceptibility to HIV and Other Genital Tract Infections
    description: >-
      Loss of mucus trapping allows virions to reach the epithelium.
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:31971984
    reference_title: "The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "we found that HIV virions had significantly increased mobility in CVM from women with BV compared to CVM from women with Lactobacillus crispatus-dominant microbiota, regardless of whether symptoms were present"
    explanation: >-
      The measurement this node is built on, including the point that the defect is
      present in asymptomatic BV.
  - reference: PMID:31971984
    reference_title: "The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We confirmed using nanoparticles and scanning electron microscopy that the impaired barrier function was due to reduced adhesive barrier properties without an obvious degradation of the physical CVM pore structure."
    explanation: >-
      Localises the defect to adhesion rather than to structural breakdown - the
      distinction this node insists on.
  - reference: PMID:31971984
    reference_title: "The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "We further confirmed a similar increase in HIV mobility in CVM from women with Lactobacillus iners-dominant microbiota, the species most associated with transitions to BV and that persists after antibiotic treatment for BV."
    explanation: >-
      Supports the caveat that a post-treatment L. iners community does not restore the
      barrier, which is the therapeutic rationale for L. crispatus repletion.

- name: Epithelial Barrier Disruption and Mucosal Immune Activation
  biological_scale: TISSUE
  description: >
    Despite the "-osis" in its name, BV is not immunologically silent. Vaginal IL-1alpha
    and soluble E-cadherin - a biomarker of epithelial barrier disruption - are elevated,
    and a lactobacillus-deficient community is accompanied by increased numbers of
    activated mucosal CD4+ T cells. The causal reading is supported by intervention:
    repleting L. crispatus with a live biotherapeutic lowered both IL-1alpha and soluble
    E-cadherin, so the inflammatory state tracks the community rather than merely
    co-occurring with it. This node is what converts a microbiological state into an
    HIV-relevant one, because activated mucosal CD4+ T cells are the target cell.
  cell_types:
  - preferred_term: CD4-positive T cell
    term:
      id: CL:0000492
      label: CD4-positive helper T cell
  - preferred_term: vaginal squamous epithelial cell
    term:
      id: CL:1001578
      label: vagina squamous cell
  biological_processes:
  - preferred_term: inflammatory response
    term:
      id: GO:0006954
      label: inflammatory response
    modifier: INCREASED
  - preferred_term: T cell activation
    term:
      id: GO:0042110
      label: T cell activation
    modifier: INCREASED
  downstream:
  - target: Increased Susceptibility to HIV and Other Genital Tract Infections
    description: >-
      Recruitment and activation of mucosal CD4+ T cells supplies the target cells for
      HIV infection.
    causal_link_type: DIRECT
  evidence:
  - reference: PMID:28087240
    reference_title: "Lactobacillus-Deficient Cervicovaginal Bacterial Communities Are Associated with Increased HIV Acquisition in Young South African Women."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "we found that individuals with diverse genital bacterial communities dominated by anaerobes other than Gardnerella were at over 4-fold higher risk of acquiring HIV and had increased numbers of activated mucosal CD4+ T cells compared to those with Lactobacillus crispatus-dominant communities"
    explanation: >-
      Couples the activated-CD4 phenotype to prospective HIV acquisition in the same
      cohort, which is the pairing this node depends on.
  - reference: PMID:35659905
    reference_title: "Sustained effect of LACTIN-V (Lactobacillus crispatus CTV-05) on genital immunology following standard bacterial vaginosis treatment: results from a randomised, placebo-controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Bacterial vaginosis might increase HIV risk by eliciting genital inflammation and epithelial barrier disruption, whereas vaginal Lactobacillus crispatus is associated with immune quiescence and HIV protection."
    explanation: >-
      States the barrier-disruption and inflammation mechanism this node encodes.
  - reference: PMID:35659905
    reference_title: "Sustained effect of LACTIN-V (Lactobacillus crispatus CTV-05) on genital immunology following standard bacterial vaginosis treatment: results from a randomised, placebo-controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The primary outcomes were vaginal levels of IL-1α and soluble E-cadherin at 24 weeks"
    explanation: >-
      Names the two immunological endpoints - vaginal IL-1alpha and soluble E-cadherin -
      that this node reports, as prespecified primary outcomes of a randomised trial
      rather than post-hoc observations.

- name: Increased Susceptibility to HIV and Other Genital Tract Infections
  biological_scale: ORGANISM
  description: >
    The consequence that dominates BV's public-health importance. Meta-analysis of
    incidence studies puts the relative risk of HIV acquisition at 1.6, and a prospective
    South African cohort found over four-fold higher risk for the most diverse
    lactobacillus-deficient communities. Two distinct mechanisms feed this node and are
    curated separately upstream: loss of mucus trapping (a physical barrier failure) and
    recruitment of activated mucosal CD4+ T cells (target-cell supply). The evidence is
    not uniform across every downstream infection, and the PID arm in particular is a
    documented disagreement between two prospective cohorts rather than a settled result.
    Both are retained below. The Longitudinal Study of Vaginal Flora (N = 2956) found
    both Nugent-BV (aHR 1.53) and symptomatic Amsel-BV (aHR 2.15) associated with incident
    PID at the subsequent visit; the two intervals overlap and the study does not test the
    difference between them, so the figures do not establish a gradient by exposure
    definition. A community-based cohort found no association between a
    Gardnerella-dominated microbiome and subsequent PID. The two differ in ascertainment
    on both sides - Nugent and Amsel criteria against microbiome sequencing for the
    exposure, and clinic-based tenderness criteria against community follow-up for the
    outcome - and the negative study reports few PID cases, so this node records the
    association as contested rather than resolving it in either direction.
  biological_processes:
  - preferred_term: response to bacterium
    term:
      id: GO:0009617
      label: response to bacterium
    modifier: DYSREGULATED
  evidence:
  - reference: PMID:18614873
    reference_title: "Bacterial vaginosis and HIV acquisition: a meta-analysis of published studies."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Bacterial vaginosis was associated with an increased risk of HIV acquisition in HIV-incidence studies (relative risk = 1.6, 95% confidence interval: 1.2, 2.1)."
    explanation: >-
      The pooled incidence-study estimate quantifying this node.
  - reference: PMID:31971984
    reference_title: "The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Women with BV are at 60% increased risk for HIV acquisition and are 3-times more likely to transmit HIV to an uninfected partner."
    explanation: >-
      Adds the transmission direction, which the acquisition meta-analysis does not cover.
      Cited as background rather than as a result: this is the authors' framing sentence in
      their introduction, not an observation from their own mucus-barrier experiments, and
      it is tagged OTHER for that reason. The transmission half in particular carries no
      primary source in this paper, so the node rests on it only for direction, not
      magnitude.
  - reference: PMID:34396403
    reference_title: "Bacterial Vaginosis and Behavioral Factors Associated With Incident Pelvic Inflammatory Disease in the Longitudinal Study of Vaginal Flora."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "BV was associated with incident PID in a large prospective cohort, controlling for behavioral factors and sexually transmitted infections (STIs)."
    explanation: >-
      The positive half of the PID disagreement, and the authors' own statement of it. From
      the larger of the two prospective cohorts (N = 2956, quarterly follow-up for 12
      months, BV measured at the visit preceding the PID diagnosis), it supports extending
      this node beyond HIV to ascending genital infection. The adjustment for concurrent
      and untreated chlamydia is what makes it more than a confounded association.
  - reference: PMID:34396403
    reference_title: "Bacterial Vaginosis and Behavioral Factors Associated With Incident Pelvic Inflammatory Disease in the Longitudinal Study of Vaginal Flora."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "symptomatic Amsel-BV (aHR, 2.15 [95% CI, 1.23-3.75]), and vaginal douching (aHR, 1.47 [95% CI, 1.03-2.09]) were associated with incident PID."
    explanation: >-
      Quantifies the same association where the exposure is defined by symptoms and Amsel
      criteria rather than by Gram stain. The point estimate is higher for symptomatic
      Amsel-BV (aHR 2.15) than for Nugent-BV (aHR 1.53, 95% CI 1.05-2.21) in the same
      model, but the two confidence intervals overlap across most of their range and the
      paper does not test the difference, so these figures do not establish a gradient by
      ascertainment. What the item does add is that the association survives a second
      exposure definition, which is why the microbiome-sequencing study below can reach a
      different answer while defining the exposure a third way.
  - reference: PMID:35086915
    reference_title: "Vaginal microbiota in ethnically diverse young women who did or did not develop pelvic inflammatory disease: community-based prospective study."
    supports: REFUTE
    evidence_source: HUMAN_CLINICAL
    snippet: "There was no association between a more diverse, G. vaginalis dominated microbiome and subsequent PID, although increased Shannon diversity was associated with black ethnicity (p=0.002) and bacterial vaginosis (diagnosed by Gram stain p<0.0001)."
    explanation: >-
      A prospective negative result against the specific claim that a BV-type microbiome
      predicts pelvic inflammatory disease. Retained as REFUTE because generalising this
      node from HIV to every ascending genital infection is exactly the inference it
      contradicts; the authors note the small number of PID cases. It is not superseded by
      the positive cohort above - the exposure is defined differently (Gardnerella-dominated
      microbiome by sequencing, not Nugent or Amsel criteria), so the two are answering
      adjacent questions and the disagreement is curated rather than adjudicated.

- name: Ascending Infection and Adverse Pregnancy Outcomes
  biological_scale: ORGANISM
  description: >
    In pregnancy the same barrier failure permits organisms to reach the choriodecidual
    space, and BV detected early in pregnancy is associated with late miscarriage and
    preterm delivery independent of prior preterm birth. The enzyme measurement makes this
    more than an association: among women already in preterm labour with BV or
    intermediate flora, those with detectable vaginal sialidase had a higher rate of early
    preterm birth, which links the outcome to the specific molecular activity curated
    upstream rather than to the Nugent score alone. Screening-and-treatment trials have
    nonetheless been inconsistent, so this node records a mechanism and a risk
    association, not a demonstrated preventable fraction.
  locations:
  - preferred_term: uterine cervix
    term:
      id: UBERON:0000002
      label: uterine cervix
  evidence:
  - reference: PMID:8124116
    reference_title: "Abnormal bacterial colonisation of the genital tract and subsequent preterm delivery and late miscarriage."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A further logistic analysis of data from women recruited before 16 weeks' gestation showed that preterm deliveries or late miscarriages occurred more often in women with bacterial vaginosis (12/77; 5.5; 2.3 to 13.3; P < 0.001)."
    explanation: >-
      Prospective cohort estimate for early-pregnancy BV and late miscarriage or preterm
      delivery.
  - reference: PMID:8124116
    reference_title: "Abnormal bacterial colonisation of the genital tract and subsequent preterm delivery and late miscarriage."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Late miscarriage and preterm delivery are associated with the presence of bacterial vaginosis in early pregnancy. This is independent of recognised risk factors such as previous preterm delivery."
    explanation: >-
      Establishes independence from the dominant confounder, prior preterm delivery.
  - reference: PMID:12388968
    reference_title: "Vaginal hydrolytic enzymes, immunoglobulin A against Gardnerella vaginalis toxin, and risk of early preterm birth among women in preterm labor with bacterial vaginosis or intermediate flora."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Women in preterm labor with bacterial vaginosis or intermediate flora and detectable sialidase are at increased risk of early preterm birth."
    explanation: >-
      Ties the obstetric outcome to sialidase specifically, supporting the incoming edge
      from the glycosidase node rather than from BV status alone.
  - reference: PMID:12388968
    reference_title: "Vaginal hydrolytic enzymes, immunoglobulin A against Gardnerella vaginalis toxin, and risk of early preterm birth among women in preterm labor with bacterial vaginosis or intermediate flora."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Prolidase and anti-Gvh IgA did not predict early preterm birth."
    explanation: >-
      The discriminating negative from the same study: of the three markers tested only
      sialidase predicted the outcome, which is why this entry gives sialidase its own
      node and does not curate a prolidase or anti-vaginolysin-IgA arm.
  downstream:
  - target: Premature birth
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    description: BV-associated ascending infection is linked to preterm delivery risk.

- name: Biofilm Persistence After Antimicrobial Therapy
  biological_scale: TISSUE
  role: resistance_mechanism
  description: >
    The first of two competing accounts of why BV comes back. Metronidazole and
    clindamycin cure 70-85% of episodes within a month, yet more than half of women
    relapse within six months. On this account the adherent Gardnerella biofilm survives
    the antibiotic course and the BV-associated bacteria re-emerge from it - a relapse of
    the same community rather than a new acquisition. Note what this node does not claim:
    no assay currently distinguishes a surviving biofilm from a reacquired one in an
    individual woman, so this is an inference from the biofilm's demonstrated existence
    and known antimicrobial tolerance, not a measured event.
  biological_processes:
  - preferred_term: biofilm formation
    term:
      id: GO:0042710
      label: biofilm formation
    modifier: INCREASED
  downstream:
  - target: Recurrent Bacterial Vaginosis
    description: >-
      Re-emergence of BV-associated bacteria from the surviving adherent layer.
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    hypothesis_groups:
    - biofilm_persistence_relapse
  evidence:
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The low rate of sustained cure highlights our limited understanding of the pathogenesis of BV recurrence, which has been attributed to possible persistence and re-emergence of BV-associated bacteria (BVAB) or a BV-associated biofilm following antimicrobials and/or reinfection occurring from sexual partners."
    explanation: >-
      States the persistence account and, in the same sentence, the competing reinfection
      account - which is why both are curated as hypothesis groups rather than one being
      asserted.
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "These regimens have similar efficacy and cure ~70–85% of women with BV within 1 month"
    explanation: >-
      Quantifies short-term cure, the denominator against which the relapse rate is the
      surprising figure.

- name: Reinfection from an Untreated Sexual Partner
  biological_scale: ORGANISM
  description: >
    The second account. BV-associated bacteria are exchanged between partners during sex,
    and on this reading a woman cured by antibiotics is reinoculated by an untreated
    regular partner. Six partner-treatment trials in the 1980s and 1990s were negative and
    the hypothesis fell out of favour; guidelines do not recommend partner treatment. The
    2025 StepUp randomised trial reopened it decisively - it was stopped early by its data
    and safety monitoring board because treating the woman alone was inferior to treating
    both partners, with 12-week recurrence 35% versus 63%. This node is therefore curated
    as a real causal route, not as a historical hypothesis, while the biofilm route is
    retained alongside it.
  downstream:
  - target: Recurrent Bacterial Vaginosis
    description: >-
      Reinoculation with BV-associated bacteria from an untreated partner re-establishes
      the dysbiotic community.
    causal_link_type: DIRECT
    hypothesis_groups:
    - sexual_reinfection
  evidence:
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The trial was stopped by the data and safety monitoring board after 150 couples had completed the 12-week follow-up period because treatment of the woman only was inferior to treatment of both the woman and her male partner."
    explanation: >-
      The randomised evidence that an untreated partner is a causal contributor to
      recurrence - the strongest support for this node.
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Evidence of sexual exchange of bacterial vaginosis-associated organisms between partners suggests that male-partner treatment may increase the likelihood of cure."
    explanation: >-
      States the exchange mechanism this node encodes.
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "There is a robust body of evidence to support the exchange of bacteria between partners during sexual activity, and while the hypothesis that women treated for BV are subsequently reinfected with BVAB following sex with an untreated sexual partner is not new, failure of past partner treatment trials has eroded confidence in this concept."
    explanation: >-
      Records both the microbiological support for exchange and the trial history that
      had discredited it, written before the StepUp result.

- name: Recurrent Bacterial Vaginosis
  biological_scale: ORGANISM
  description: >
    Re-establishment of the dysbiotic community after apparently successful treatment -
    the outcome that defines the clinical problem, since short-term cure is not the
    difficulty. Both upstream routes converge here and neither can be excluded in an
    individual woman, because they present identically. That indistinguishability is not a
    curation shortcut; it is the field's stated position and is recorded as a knowledge
    gap below. The node is not a terminal state: a recurrent episode is diagnosed by the
    same Amsel and Nugent criteria as an incident one, so it re-enters the graph at the
    trigger node and the entry is a cycle rather than a chain. That is the structural
    reason short-term cure and sustained cure come apart.
  downstream:
  - target: Loss of Lactobacillus Dominance and Vaginal Acidification
    description: >-
      A recurrent episode is the re-established dysbiotic community state, not a separate
      endpoint - which is why an intervention that only clears anaerobes leaves the cycle
      intact, and why the one curated intervention acting on this node (LACTIN-V, which
      repopulates the niche) is also the only one shown to lower recurrence.
    causal_link_type: DIRECT
    evidence:
    - reference: PMID:32402161
      reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "The primary efficacy outcome was the percentage of participants who had recurrent bacterial vaginosis (defined by the presence of at least three Amsel criteria and a Nugent score of 4 to 10) at any follow-up visit up to and including the week 12 visit."
      explanation: >-
        Recurrence is operationalised as re-meeting the same diagnostic criteria as an
        incident case, which is what makes this edge a return to the trigger state rather
        than a new outcome.
    - reference: PMID:34470644
      reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Factors including persistence of a BV-associated biofilm, failure to recolonise the vagina with lactobacilli, reinfection from an untreated partner, and host genetic and/or immune factors may all play a role in recurrence"
      explanation: >-
        Names failure to restore lactobacillus dominance - the content of the target node -
        among the drivers of recurrence, closing the loop this edge asserts.
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Standard antibiotic therapy often fails, with an estimated relapse rate of 50% at six months follow-up"
    explanation: >-
      Quantifies the recurrence burden this node represents.
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "After treatment with an antibiotic agent, 20 to 75% of women have recurrent bacterial vaginosis within 3 months."
    explanation: >-
      An independent recurrence estimate over a shorter window, showing the range across
      populations and definitions.

- name: Anaerobe Ribosomal Translation (Clindamycin and Tetracycline Target)
  biological_scale: MOLECULAR
  role: therapeutic_vulnerability
  conforms_to: "bacterial_protein_synthesis_inhibition#Bacterial mRNA Translation by the Ribosome"
  description: >
    Protein synthesis in the BV consortium is the molecular target of clindamycin, the
    lincosamide alternative to metronidazole and the agent applied topically to the penile
    skin of male partners in the StepUp regimen. It is curated as its own node so that the
    clindamycin arm of therapy attaches to a mechanism rather than to the disease as a
    whole - metronidazole, whose nitroimidazole mechanism is reductive DNA damage in
    anaerobes, does not act here and is deliberately not attached to this node.
  biological_processes:
  - preferred_term: translation
    term:
      id: GO:0006412
      label: translation
    modifier: DECREASED
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended."
    explanation: >-
      Establishes clindamycin as a recommended agent, which is what makes its ribosomal
      target a therapeutic vulnerability in this disease.
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "the male partner received oral and topical antimicrobial treatment (metronidazole 400-mg tablets and 2% clindamycin cream applied to penile skin, both twice daily for 7 days)"
    explanation: >-
      Documents the topical clindamycin component of the partner-treatment regimen that
      acts on this target.

mechanistic_hypotheses:
- hypothesis_group_id: biofilm_persistence_relapse
  hypothesis_label: Relapse from a surviving adherent biofilm
  status: CANONICAL
  description: >
    Recurrence is re-emergence of the original community from a Gardnerella-dominated
    adherent biofilm that tolerated the antibiotic course. Marked CANONICAL because the
    biofilm is directly demonstrated on vaginal biopsy and because it is the account most
    treatment development has been built around - not because it has been shown to be the
    operative route in any individual recurrence.
  evidence:
  - reference: PMID:16260520
    reference_title: "Adherent biofilms in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A biofilm comprised of confluent G vaginalis with other bacterial groups incorporated in the adherent layer is a prominent feature of bacterial vaginosis."
    explanation: >-
      Establishes that the structure this hypothesis requires exists in vivo.
- hypothesis_group_id: sexual_reinfection
  hypothesis_label: Reinfection from an untreated sexual partner
  status: EMERGING
  description: >
    Recurrence is reacquisition of BV-associated bacteria from a regular untreated
    partner. Marked EMERGING rather than ALTERNATIVE: six earlier partner-treatment trials
    were negative and guidelines still do not recommend partner treatment, but the 2025
    StepUp randomised trial was stopped early for benefit, which is a stronger form of
    evidence than anything supporting the biofilm route. This status should be revisited
    when guidelines respond.
  evidence:
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The addition of combined oral and topical antimicrobial therapy for male partners to treatment of women for bacterial vaginosis resulted in a lower rate of recurrence of bacterial vaginosis within 12 weeks than standard care."
    explanation: >-
      The randomised result on which this hypothesis's EMERGING status rests.

phenotypes:
- name: Thin homogeneous malodorous vaginal discharge
  category: Clinical
  description: >
    The presenting complaint in symptomatic disease: a thin, white-to-yellow, homogeneous
    discharge with a fishy odour that is accentuated by alkali. HPO's `Abnormal vaginal
    discharge` covers anomalous amount, odour and consistency in one term, so the
    malodour is captured here rather than as a separate unbound phenotype.
  phenotype_term:
    preferred_term: Thin homogeneous malodorous vaginal discharge
    term:
      id: HP:0034269
      label: Abnormal vaginal discharge
  frequency: FREQUENT
  evidence:
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "thin, white or yellow, homogeneous discharge"
    explanation: >-
      The discharge character as specified in the Amsel criteria used for trial entry.
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "it is the symptoms themselves, including malodour and vaginal discharge, that cause significant distress to women and impact on their quality of life and relationships"
    explanation: >-
      Names malodour and discharge as the symptoms of the disease and their impact.
  - reference: PMID:6600371
    reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "we diagnosed nonspecific vaginitis in up to 25 percent of our study population; asymptomatic disease was recognized in more than 50 percent of those with nonspecific vaginitis"
    explanation: >-
      Basis for the FREQUENT rather than VERY_FREQUENT band: more than half of women
      meeting the case definition had no symptoms at all, so symptomatic discharge is
      present in a minority-to-half of cases.
- name: Premature birth
  category: Clinical
  description: >
    Delivery before 37 weeks, and late miscarriage between 16 and 24 weeks, occur more
    often in women with BV detected early in pregnancy. Recorded as a phenotype of the
    disease in pregnancy, not of the disease in general.
  phenotype_term:
    preferred_term: Premature birth
    term:
      id: HP:0001622
      label: Premature birth
  frequency: OCCASIONAL
  evidence:
  - reference: PMID:8124116
    reference_title: "Abnormal bacterial colonisation of the genital tract and subsequent preterm delivery and late miscarriage."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Multiple logistic analysis showed that there was an increased incidence of preterm delivery in women with a previous preterm delivery (9/24; odds ratio 25; 95% confidence interval 9 to 70; P < 0.001) and bacterial vaginosis (9/115; 2.8; 1.1 to 7.4; P = 0.04)."
    explanation: >-
      Gives the absolute count behind the OCCASIONAL band: 9 of 115 pregnant women with
      BV delivered preterm, which is within the 5-29% range for that frequency term.

definitions:
- name: Amsel clinical criteria
  definition_type: DIAGNOSTIC_CRITERIA
  derivation_basis: ESTABLISHED_CRITERIA
  scope: >-
    Office diagnosis of bacterial vaginosis in symptomatic and asymptomatic
    reproductive-aged women.
  description: >
    Three of four bedside findings: thin white-to-yellow homogeneous discharge, more than
    20% clue cells on microscopy, vaginal fluid pH above 4.5, and a fishy odour on adding
    10% potassium hydroxide. The set is worth reading mechanistically rather than as a
    checklist - each item reports a different node of this entry: the discharge and clue
    cells report epithelial adhesion, the pH reports loss of lactic acid, and the whiff
    test reports biogenic amine production. It remains the entry criterion for
    contemporary randomised trials.
  attaches_to:
  - "pathophysiology#Loss of Lactobacillus Dominance and Vaginal Acidification"
  - "pathophysiology#Exfoliation of Bacteria-Coated Epithelial Cells (Clue Cells)"
  - "pathophysiology#Biogenic Amine Production and Amplified Lactobacillus Suppression"
  inclusion_criteria:
  - preferred_term: Thin, white or yellow, homogeneous vaginal discharge
  - preferred_term: More than 20% clue cells on microscopic examination
  - preferred_term: Vaginal fluid pH greater than 4.5
  - preferred_term: Fishy odour on addition of 10% potassium hydroxide (whiff test)
  evidence:
  - reference: PMID:6600371
    reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Practical diagnostic criteria for standard clinical use are proposed. Application of such criteria should assist in clinical management of nonspecific vaginitis and in further study of the microbiologic and biochemical correlates and the pathogenesis of this mild but quite prevalent disease."
    explanation: >-
      The originating publication proposing these criteria.
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "If a potentially eligible woman met at least three of four Amsel criteria (i.e., thin, white or yellow, homogeneous discharge; >20% clue cells on microscopic examination; vaginal fluid with a pH of >4.5; and release of a fishy odor when 10% potassium hydroxide is added to a vaginal specimen)"
    explanation: >-
      Enumerates all four items and the three-of-four threshold, as operationalised for
      trial entry.
- name: Nugent score (Gram stain)
  definition_type: DIAGNOSTIC_CRITERIA
  derivation_basis: ESTABLISHED_CRITERIA
  scope: >-
    Standardised laboratory diagnosis of bacterial vaginosis from a Gram-stained vaginal
    smear; the reference standard for research and the comparator for molecular assays.
  description: >
    A 0-10 weighted count of three morphotypes on Gram stain - lactobacilli, small
    gram-variable or gram-negative rods (Gardnerella/Bacteroides), and curved gram-variable
    rods - with 7 or above read as BV. Its design principle is not diagnostic accuracy but
    reproducibility: morphotypes with poor intercentre agreement (gram-positive cocci) were
    deliberately dropped and the most reliable one (curved rods) weighted most heavily,
    raising intercentre correlation from 0.61 to 0.82. That is why it, rather than Amsel,
    became the research standard, and why the intermediate band 4-6 exists as a graded
    state rather than a diagnostic failure.
  attaches_to:
  - "pathophysiology#Loss of Lactobacillus Dominance and Vaginal Acidification"
  - "pathophysiology#Polymicrobial Anaerobic Overgrowth"
  evidence:
  - reference: PMID:1706728
    reference_title: "Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The scoring system (0 to 10) was described as a weighted combination of the following morphotypes: lactobacilli, Gardnerella vaginalis or bacteroides (small gram-variable rods or gram-negative rods), and curved gram-variable rods."
    explanation: >-
      Specifies the three scored morphotypes and the score range.
  - reference: PMID:1706728
    reference_title: "Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "For comparison with the Spiegel criteria, a score of 7 or higher was considered indicative of bacterial vaginosis."
    explanation: >-
      Gives the diagnostic threshold.
  - reference: PMID:1706728
    reference_title: "Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The standardized score had improved intercenter reliability (r = 0.82) compared with the Spiegel criteria (r = 0.61)."
    explanation: >-
      Quantifies the reproducibility gain that is this score's reason for existing.
  notes: >-
    A contemporary review still describes the Nugent score as the gold-standard diagnostic
    tool despite its known limitations, which is why both it and the Amsel criteria are
    curated here rather than only one.

treatments:
- name: Metronidazole
  description: >
    First-line therapy, oral or as 0.75% intravaginal gel. A nitroimidazole prodrug
    reductively activated inside anaerobes, so its selectivity is for the metabolic
    environment of the consortium rather than for any BV organism specifically - which is
    exactly the property a polymicrobial dysbiosis needs. It cures most episodes and
    prevents few recurrences, and this entry attaches it to the overgrowth node only, not
    to the biofilm node, because clearing planktonic anaerobes is what it is shown to do.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Antibiotic Therapy
    term:
      id: NCIT:C15620
      label: Antibiotic Therapy
    therapeutic_agent:
    - preferred_term: metronidazole
      term:
        id: CHEBI:6909
        label: metronidazole
  target_mechanisms:
  - target: Polymicrobial Anaerobic Overgrowth
    treatment_effect: INHIBITS
    description: >-
      Broad anaerobic coverage suppresses the expanded consortium, which is what produces
      the 70-85% one-month cure rate.
    evidence:
    - reference: PMID:34470644
      reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "These regimens have similar efficacy and cure ~70–85% of women with BV within 1 month"
      explanation: >-
        Quantifies the effect of first-line therapy on the microbiological state, which is
        the claim this edge makes.
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended."
    explanation: >-
      Establishes metronidazole as recommended therapy.
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Potentially eligible women completed a standard 5-day course of vaginal 0.75% metronidazole within 30 days before the screening visit."
    explanation: >-
      Documents the standard intravaginal regimen and dose.

- name: Secnidazole
  description: >
    A 5-nitroimidazole given as a single 2 g oral dose of granules, FDA-approved for BV in
    2017. Mechanistically it is metronidazole's class-mate - the same reductive activation
    inside anaerobes - so it attaches to the same overgrowth node and not to the biofilm
    node. What it changes is exposure kinetics rather than target: a roughly 17-hour
    half-life against metronidazole's 8 hours is what makes one dose sufficient, which
    matters for a disease whose first-line regimens run 5 to 7 days and whose recurrence is
    partly a compliance problem. Superiority was shown against placebo rather than against
    metronidazole, so this entry does not claim it is more effective than first-line
    therapy.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Antibiotic Therapy
    term:
      id: NCIT:C15620
      label: Antibiotic Therapy
    therapeutic_agent:
    - preferred_term: secnidazole
      term:
        id: CHEBI:140628
        label: secnidazole
  target_mechanisms:
  - target: Polymicrobial Anaerobic Overgrowth
    treatment_effect: INHIBITS
    description: >-
      A single dose suppresses the expanded anaerobic consortium sufficiently to normalise
      discharge, the whiff test and clue-cell burden in about half of treated women.
    evidence:
    - reference: PMID:28867602
      reference_title: "A phase-3, double-blind, placebo-controlled study of the effectiveness and safety of single oral doses of secnidazole 2 g for the treatment of women with bacterial vaginosis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Single-dose secnidazole 2 g was superior to placebo for the primary and all secondary efficacy measures in the modified intent-to-treat population, with clinical outcome responder rates of 53.3% (57/107) vs 19.3% (11/57; P < .001)."
      explanation: >-
        Quantifies the effect of a single secnidazole dose on the Amsel-defined
        microbiological state, which is the claim this edge makes.
  evidence:
  - reference: PMID:28867602
    reference_title: "A phase-3, double-blind, placebo-controlled study of the effectiveness and safety of single oral doses of secnidazole 2 g for the treatment of women with bacterial vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A novel single oral dose granule formulation of secnidazole 2 g, a 5-nitroimidazole with a longer half-life (∼17 hours) than metronidazole (∼8 hours), is being developed to treat bacterial vaginosis."
    explanation: >-
      States the pharmacokinetic difference from metronidazole that is the whole rationale
      for the single-dose regimen, and confirms the shared nitroimidazole class.

- name: Tinidazole
  description: >
    A second-generation 5-nitroimidazole, FDA-approved for BV and a CDC-recommended
    alternative regimen, given orally as 1 g daily for 5 days or 2 g daily for 2 days. Like
    secnidazole it shares metronidazole's reductive-activation mechanism and is attached to
    the overgrowth node only. It is curated here partly because its pivotal trial used the
    strict five-criterion FDA cure definition, which is why its reported cure rates (27-37%)
    look far worse than the 70-85% quoted for first-line regimens elsewhere in this entry -
    the difference is the endpoint, not the drug, and that discrepancy is itself worth
    recording in a disease whose literature mixes Amsel, Nugent and molecular definitions.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Antibiotic Therapy
    term:
      id: NCIT:C15620
      label: Antibiotic Therapy
    therapeutic_agent:
    - preferred_term: tinidazole
      term:
        id: CHEBI:63627
        label: tinidazole
  target_mechanisms:
  - target: Polymicrobial Anaerobic Overgrowth
    treatment_effect: INHIBITS
    description: >-
      Both licensed oral regimens suppress the anaerobic consortium enough to clear all five
      FDA cure criteria in a significant minority of women, against a 5% placebo rate.
    evidence:
    - reference: PMID:17666604
      reference_title: "Effectiveness of two tinidazole regimens in treatment of bacterial vaginosis: a randomized controlled trial."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Superior efficacy was demonstrated by tinidazole for the 1 g once daily for 5 days regimen (36.8% cured, P<.001, number needed to treat 3.2) and for the 2 g once daily for 2 days regimen (27.4% cured, P<.001, number needed to treat 4.5), when compared with placebo (5.1% cured) in the primary endpoint analysis."
      explanation: >-
        Placebo-controlled effect size for both licensed regimens against the consortium,
        which is the claim this edge makes.
  evidence:
  - reference: PMID:17666604
    reference_title: "Effectiveness of two tinidazole regimens in treatment of bacterial vaginosis: a randomized controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Both tinidazole regimens studied provided effective treatment for bacterial vaginosis."
    explanation: >-
      Establishes tinidazole as effective therapy for BV in a multicentre placebo-controlled
      trial.
  - reference: PMID:17666604
    reference_title: "Effectiveness of two tinidazole regimens in treatment of bacterial vaginosis: a randomized controlled trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Using more traditional criteria for cure, efficacy was greater."
    explanation: >-
      Records the endpoint dependence explicitly: the same trial reports a higher cure rate
      under the conventional definition, so the 27-37% figures are not comparable with the
      70-85% quoted from Amsel-based series elsewhere in this entry.

- name: Dequalinium Chloride
  description: >
    A broad-spectrum quaternary-ammonium antiseptic given as a 10 mg vaginal tablet for six
    days, widely used in Europe and non-inferior to oral metronidazole in a phase 4
    double-dummy trial. It is curated as a distinct arm rather than as another antibiotic
    because it is not one: a membrane-active antiseptic exerts no selection pressure of the
    kind that drives nitroimidazole resistance, which is the stated reason the trial was
    run. It attaches to the same overgrowth node, since the evidence is a clinical cure
    rate and not a demonstration of action on the biofilm.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: antiseptic therapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: dequalinium chloride
      term:
        id: CHEBI:31466
        label: dequalinium chloride
  target_mechanisms:
  - target: Polymicrobial Anaerobic Overgrowth
    treatment_effect: INHIBITS
    description: >-
      Six days of intravaginal dequalinium chloride clears the Amsel-defined state as often
      as a seven-day oral metronidazole course.
    evidence:
    - reference: PMID:38696172
      reference_title: "Efficacy of Dequalinium Chloride vs Metronidazole for the Treatment of Bacterial Vaginosis: A Randomized Clinical Trial."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "The clinical cure rates at visit 1 were 64 of 69 (92.8%) for dequalinium chloride vs 69 of 74 (93.2%) for metronidazole in the intention-to-treat population"
      explanation: >-
        Head-to-head cure rates against the first-line comparator already curated here,
        supporting an equivalent effect on the consortium.
  evidence:
  - reference: PMID:38696172
    reference_title: "Efficacy of Dequalinium Chloride vs Metronidazole for the Treatment of Bacterial Vaginosis: A Randomized Clinical Trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "This randomized clinical trial showed that dequalinium chloride was not inferior to metronidazole for the treatment of BV."
    explanation: >-
      Establishes the non-inferiority conclusion that justifies curating a non-antibiotic
      alternative alongside the two first-line antimicrobials.
  - reference: PMID:38696172
    reference_title: "Efficacy of Dequalinium Chloride vs Metronidazole for the Treatment of Bacterial Vaginosis: A Randomized Clinical Trial."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Due to the increase in antibiotic resistance, effective nonantibiotic treatments for BV are needed."
    explanation: >-
      States the rationale for a non-antibiotic arm, which is why this treatment is curated
      separately rather than folded into the nitroimidazole group.

- name: Clindamycin
  description: >
    Lincosamide alternative to metronidazole, given as 2% intravaginal cream, and the
    topical component applied to penile skin in the StepUp partner-treatment regimen.
    Unlike metronidazole it has a defined molecular target in this entry - the bacterial
    ribosome - which is why it carries a module-conforming target node.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Antibiotic Therapy
    term:
      id: NCIT:C15620
      label: Antibiotic Therapy
    therapeutic_agent:
    - preferred_term: clindamycin
      term:
        id: CHEBI:3745
        label: clindamycin
  target_mechanisms:
  - target: Anaerobe Ribosomal Translation (Clindamycin and Tetracycline Target)
    treatment_effect: INHIBITS
    description: >-
      Clindamycin binds the 50S ribosomal subunit and blocks translation in the
      BV-associated anaerobes.
    evidence:
    - reference: PMID:40043236
      reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "the male partner received oral and topical antimicrobial treatment (metronidazole 400-mg tablets and 2% clindamycin cream applied to penile skin, both twice daily for 7 days)"
      explanation: >-
        Documents the clindamycin exposure whose target this edge names.
  - target: Polymicrobial Anaerobic Overgrowth
    treatment_effect: INHIBITS
    description: >-
      Broad anaerobic coverage equivalent to metronidazole in short-term cure.
    evidence:
    - reference: PMID:34470644
      reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "These regimens have similar efficacy and cure ~70–85% of women with BV within 1 month"
      explanation: >-
        Establishes equivalence of the two first-line regimens against the consortium.
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Prescription of antibiotics such as metronidazole, clindamycin, etc. is recommended."
    explanation: >-
      Establishes clindamycin as recommended therapy.

- name: LACTIN-V (Lactobacillus crispatus CTV-05 live biotherapeutic)
  description: >
    A vaginally applied live biotherapeutic containing a naturally occurring human vaginal
    strain of L. crispatus, given after a course of metronidazole. It is the only
    intervention curated here that acts on the trigger node rather than on the consortium:
    it does not kill anaerobes, it repopulates the niche. Recurrence at 12 weeks fell from
    45% to 30%, and a randomised immunology substudy showed lower vaginal IL-1alpha and
    soluble E-cadherin, so the effect reaches the epithelial-activation node as well as
    the community-state node.
  therapeutic_modality: OTHER
  treatment_term:
    preferred_term: vaginal live biotherapeutic (Lactobacillus crispatus) administration
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
  target_mechanisms:
  - target: Loss of Lactobacillus Dominance and Vaginal Acidification
    treatment_effect: RESTORES
    description: >-
      Exogenous L. crispatus colonises the vagina and re-establishes the protective
      lactobacillus-dominant community state that defines this node's healthy baseline.
    evidence:
    - reference: PMID:32402161
      reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "At the 12-week visit, L. crispatus CTV-05 was detected in 79% of participants in the Lactin-V group."
      explanation: >-
        Demonstrates that the intervention achieves the colonisation this edge asserts,
        not merely the clinical endpoint.
  - target: Epithelial Barrier Disruption and Mucosal Immune Activation
    treatment_effect: INHIBITS
    description: >-
      Repletion with L. crispatus lowered vaginal IL-1alpha and soluble E-cadherin, the
      two prespecified markers of inflammation and barrier disruption at this node.
    evidence:
    - reference: PMID:35659905
      reference_title: "Sustained effect of LACTIN-V (Lactobacillus crispatus CTV-05) on genital immunology following standard bacterial vaginosis treatment: results from a randomised, placebo-controlled trial."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "The primary outcomes were vaginal levels of IL-1α and soluble E-cadherin at 24 weeks"
      explanation: >-
        Names the endpoints; the trial reported both significantly lower in the LACTIN-V
        arm, which is the basis for the INHIBITS direction on this edge.
  evidence:
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The use of Lactin-V after treatment with vaginal metronidazole resulted in a significantly lower incidence of recurrence of bacterial vaginosis than placebo at 12 weeks."
    explanation: >-
      The primary efficacy result.
  - reference: PMID:32402161
    reference_title: "Randomized Trial of Lactin-V to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "recurrence of bacterial vaginosis by week 12 occurred in 46 participants (30%) in the Lactin-V group and in 34 participants (45%) in the placebo group"
    explanation: >-
      The absolute recurrence rates behind that result.
  notes: >-
    `treatment_term` is bound to the generic NCIT Pharmacotherapy action because NCIT has
    no clinical-action term for administration of a live biotherapeutic product; the more
    specific intent is carried by `preferred_term`. No CHEBI or NCIT agent term exists for
    L. crispatus CTV-05, so `therapeutic_agent` is deliberately omitted rather than bound
    to something broader.

- name: Concurrent male-partner antimicrobial treatment
  description: >
    Treating the woman's regular male partner with 7 days of oral metronidazole plus 2%
    clindamycin cream applied to penile skin, alongside her own first-line therapy. This
    is the only intervention here aimed at the reinfection route rather than at the
    woman's own vaginal community, and it is a reversal: six trials in the 1980s-90s were
    negative and guidelines do not recommend it, but the 2025 StepUp trial was stopped
    early because withholding it was inferior. Curated as a treatment because the trial
    result is randomised and positive, with the guideline lag recorded in the notes.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Antibiotic Therapy
    term:
      id: NCIT:C15620
      label: Antibiotic Therapy
    therapeutic_agent:
    - preferred_term: metronidazole
      term:
        id: CHEBI:6909
        label: metronidazole
    - preferred_term: clindamycin
      term:
        id: CHEBI:3745
        label: clindamycin
  target_mechanisms:
  - target: Reinfection from an Untreated Sexual Partner
    treatment_effect: INHIBITS
    description: >-
      Eradicating BV-associated bacteria carried by the partner removes the reinoculating
      source, cutting 12-week recurrence from 63% to 35%.
    evidence:
    - reference: PMID:40043236
      reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "recurrence occurred in 24 of 69 women (35%) in the partner-treatment group (recurrence rate, 1.6 per person-year; 95% confidence interval [CI], 1.1 to 2.4) and in 43 of 68 women (63%) in the control group"
      explanation: >-
        The randomised effect size on which this edge rests.
  evidence:
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Adverse events in treated men included nausea, headache, and metallic taste."
    explanation: >-
      Records the harms borne by a person who is not the patient, which is the specific
      ethical feature of this intervention.
  notes: >-
    Partner treatment was not recommended in guidelines at the time of the source review
    (PMID:34470644), which predates the StepUp result. The applicability of the trial is
    also bounded by its enrolment criterion - women in a monogamous relationship with a
    regular male partner - and should not be extended to other partnership contexts
    without evidence.

clinical_trials:
- name: NCT02766023
  phase: PHASE_II
  status: COMPLETED
  description: >
    LACTIN-V phase 2b: randomised, double-blind, placebo-controlled trial of vaginal
    L. crispatus CTV-05 given after a 5-day course of metronidazole gel, with recurrence
    at 12 weeks as the primary outcome.
  target_phenotypes:
  - preferred_term: Thin homogeneous malodorous vaginal discharge
    term:
      id: HP:0034269
      label: Abnormal vaginal discharge
  evidence:
  - reference: clinicaltrials:NCT02766023
    reference_title: "Phase II-b Randomized Double-Blind Placebo-Controlled Trial of Lactobacillus Crispatus CTV-05 (LACTIN-V) to Prevent the Recurrence of Bacterial Vaginosis"
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "To assess the safety of LACTIN-V over 24 weeks by comparing the incidence of AEs between individuals randomized to LACTIN-V or placebo."
    explanation: >-
      Registry record establishing the trial's design and duration.
- name: NCT00229216
  phase: PHASE_III
  status: COMPLETED
  description: >
    The registration trial for oral tinidazole in bacterial vaginosis: multicentre,
    double-blind, double-dummy, placebo-controlled, comparing 1 g for 5 days and 2 g for
    2 days against placebo under the strict five-criterion FDA cure definition
    (PMID:17666604).
  target_phenotypes:
  - preferred_term: Thin homogeneous malodorous vaginal discharge
    term:
      id: HP:0034269
      label: Abnormal vaginal discharge
  evidence:
  - reference: clinicaltrials:NCT00229216
    reference_title: "A Phase III Randomized, Multi-center, Double-blind, Double-dummy, Placebo-controlled Treatment Trial of Bacterial Vaginosis With Tinidazole Oral Tablets."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "The purpose of this study is to confirm the safety and efficacy of oral tinidazole for the treatment of bacterial vaginosis."
    explanation: >-
      Registry record establishing the trial's identity and objective.
- name: ACTRN12619000196145
  phase: PHASE_III
  status: COMPLETED
  description: >
    StepUp: open-label randomised controlled trial of treating male partners of women
    being treated for bacterial vaginosis, to reduce recurrence. Registered on ANZCTR and
    therefore keyed on its WHO ICTRP identifier rather than an NCT number; reported in
    2025 (PMID:40043236) after being stopped early by its data and safety monitoring
    board.
  evidence:
  - reference: ICTRP:ACTRN12619000196145
    reference_title: "Treating male partners of women being treated for bacterial vaginosis (BV): randomised controlled trial"
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Scientific title: Treating male partners of women with bacterial vaginosis (BV) to reduce recurrence: randomised controlled trial"
    explanation: >-
      WHO ICTRP registration record establishing the trial's identity and objective.
  - reference: ICTRP:ACTRN12619000196145
    reference_title: "Treating male partners of women being treated for bacterial vaginosis (BV): randomised controlled trial"
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Male partners randomised to the Intervention Group will receive oral MTZ 400mg tablets twice daily and topical 2% clindamycin cream to be applied to the glans penis and upper shaft (under the foreskin if uncircumcised) twice daily for 7 days."
    explanation: >-
      The registered intervention, matching the regimen curated under the partner
      treatment entry.

environmental:
- name: Intrauterine device use and non-barrier contraception
  description: >
    In the founding case-control series, clinical BV was correlated with current use of
    non-barrier contraceptive methods and particularly with an intrauterine device. The
    mechanism is not established - a foreign body in the endometrial cavity plausibly
    supports biofilm, but that is inference, not evidence - so this is curated as a
    predisposing exposure onto the trigger node rather than as a causal trigger.
  influences_mechanisms:
  - target: Loss of Lactobacillus Dominance and Vaginal Acidification
    environmental_effect: PREDISPOSES
    causal_link_type: UNKNOWN
    description: >-
      Associated with the dysbiotic community shift; the intervening steps are not
      identified.
    evidence:
    - reference: PMID:6600371
      reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "Nonspecific vaginitis was also correlated with a history of sexual activity, a history of previous trichomoniasis, current use of nonbarrier contraceptive methods, and, particularly, use of an intrauterine device."
      explanation: >-
        The epidemiological association this exposure records.
  evidence:
  - reference: PMID:6600371
    reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "current use of nonbarrier contraceptive methods, and, particularly, use of an intrauterine device"
    explanation: >-
      Establishes that this exposure is a real correlate of the disease at all, which is
      the claim the entry itself makes; the mechanism link above carries the separate
      claim about which node it acts on. The correlation is from the study that
      established the modern case definition, so it is a founding observation rather than
      a later replication.
  notes: >-
    No `exposure_term` is bound. ECTO was searched for an intrauterine-device or
    contraception exposure term and none was found, and binding to a broader
    device-exposure concept would assert more than the source supports.
- name: Sexual activity with a regular untreated male partner
  description: >
    Sexual activity has been associated with BV since the original case definition, and
    BV-associated bacteria are demonstrably exchanged between partners. This entry
    separates the two claims the association can carry: exposure to a partner as a
    predisposing factor, curated here, and reinfection from a specific untreated partner
    as a mechanism of recurrence, curated as its own pathophysiology node with randomised
    support.
  influences_mechanisms:
  - target: Reinfection from an Untreated Sexual Partner
    environmental_effect: PREDISPOSES
    causal_link_type: DIRECT
    description: >-
      Continued sexual contact with an untreated regular partner is the exposure route by
      which BV-associated bacteria are reacquired after treatment.
    evidence:
    - reference: PMID:34470644
      reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "There is a robust body of evidence to support the exchange of bacteria between partners during sexual activity"
      explanation: >-
        Establishes partner-to-partner bacterial exchange as the exposure mechanism.
    - reference: PMID:6600371
      reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
      supports: SUPPORT
      directness: INDIRECT
      evidence_source: HUMAN_CLINICAL
      snippet: "Nonspecific vaginitis was also correlated with a history of sexual activity"
      explanation: >-
        The original epidemiological correlation. INDIRECT because a history of sexual
        activity is not the same claim as reinfection from a specific current partner -
        the link follows from the quote only by an inference step.
  evidence:
  - reference: PMID:40043236
    reference_title: "Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Evidence of sexual exchange of bacterial vaginosis-associated organisms between partners suggests that male-partner treatment may increase the likelihood of cure."
    explanation: >-
      Establishes the exposure itself - an ongoing regular male sexual partner - as
      relevant to the disease, which is the entry-level claim. The randomised result that
      the exposure is modifiable is carried by the partner-treatment entry rather than
      asserted here.
  - reference: PMID:6600371
    reference_title: "Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations."
    supports: SUPPORT
    directness: INDIRECT
    evidence_source: HUMAN_CLINICAL
    snippet: "Nonspecific vaginitis was also correlated with a history of sexual activity"
    explanation: >-
      The founding epidemiological correlation, INDIRECT for the same reason as on the
      mechanism link: a history of sexual activity is a weaker exposure than an ongoing
      regular untreated partner, so the entry-level claim follows by an inference step.
  notes: >-
    No `exposure_term` is bound: ECTO was searched for a sexual-activity or
    sexual-intercourse exposure term and returned nothing suitable.

animal_models:
- name: Murine cervicovaginal colonisation with BV-associated anaerobes
  species: Mouse
  genotype: Wild type
  description: >
    Mice colonised with high-risk cervicovaginal taxa identified in the South African
    cohort. Used to test whether the association between a lactobacillus-deficient
    community and activated mucosal CD4+ T cells is causal rather than confounded, since
    that direction cannot be established observationally in women.
  publication: PMID:28087240
  modeled_mechanisms:
  - target: Epithelial Barrier Disruption and Mucosal Immune Activation
    relationship: PARTIALLY_RECAPITULATES
    fidelity: LOW
    description: >-
      Reproduces the target-cell recruitment arm of this node - high-risk bacteria
      increased activated genital CD4+ T cells - but not the node as a whole.
    limitations: >-
      The mouse has no lactobacillus-dominant vaginal community and no acidic vaginal
      lumen, so the community state whose loss defines this disease does not exist to be
      perturbed. The model can show that particular taxa activate mucosal T cells; it
      cannot model dysbiosis. It also cannot address the vaginolysin arm at all, because
      that toxin is restricted to human CD59.
    readouts:
    - name: Activated genital CD4+ T cell number
      target: Epithelial Barrier Disruption and Mucosal Immune Activation
      direction: INCREASED
      interpretation: >-
        Colonisation with high-risk taxa raises the number of activated mucosal CD4+ T
        cells, the HIV target-cell population this node supplies.
      evidence:
      - reference: PMID:28087240
        reference_title: "Lactobacillus-Deficient Cervicovaginal Bacterial Communities Are Associated with Increased HIV Acquisition in Young South African Women."
        supports: SUPPORT
        evidence_source: MODEL_ORGANISM
        snippet: "high-risk bacteria increased numbers of activated genital CD4+ T cells in a murine model"
        explanation: >-
          Reports the measurement and its direction in the mouse.
    evidence:
    - reference: PMID:28087240
      reference_title: "Lactobacillus-Deficient Cervicovaginal Bacterial Communities Are Associated with Increased HIV Acquisition in Young South African Women."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "We identified specific bacterial taxa linked with reduced (L. crispatus) or elevated (Prevotella, Sneathia, and other anaerobes) inflammation and HIV infection and found that high-risk bacteria increased numbers of activated genital CD4+ T cells in a murine model."
      explanation: >-
        Supports treating this model as informative for the immune-activation node, and
        names the taxa tested.

discussions:
- discussion_id: bv_initiating_event_unknown
  prompt: >-
    What precipitates the shift from a lactobacillus-dominant to a
    lactobacillus-deficient vaginal community in an individual woman?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - "pathophysiology#Loss of Lactobacillus Dominance and Vaginal Acidification"
  rationale: >
    This entry curates a detailed chain downstream of the community shift and nothing
    upstream of it, because there is nothing to curate: the trigger node has no evidenced
    incoming edge other than two epidemiological exposures. Risk factors are known
    (sexual activity and intrauterine device use are the two cited here), but no
    sequence of events from a normal community to a BV community has been
    demonstrated. This matters practically, not just
    theoretically - every intervention in this entry acts after the shift has happened,
    and none prevents it.
  proposed_experiments:
  - experiment_id: bv_incident_transition_cohort
    name: Densely sampled prospective cohort through incident BV
    description: >
      Daily or near-daily self-sampling in a cohort of women with a stable
      L. crispatus-dominant community, with metagenomics, metabolomics (lactic acid
      isomers, biogenic amines), pH and behavioural diaries, powered on incident
      transitions rather than prevalent BV. The measurement that is missing is not another
      cross-sectional comparison of BV versus not-BV; it is the order of events across the
      transition.
    decision_criterion: >
      Whether a consistent temporal ordering (for example, amine rise preceding
      lactobacillus loss, or the reverse) is observed across incident transitions.
  evidence:
  - reference: PMID:35118003
    reference_title: "Bacterial Vaginosis: What Do We Currently Know?"
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The etiology of this dysbiosis remains unknown, but its health consequences are significant, including obstetrical complications, increased risk of sexually transmitted infections and urogenital infections."
    explanation: >-
      States the gap directly, alongside the consequences that make it worth closing.

- discussion_id: bv_recurrence_route_indistinguishable
  prompt: >-
    In a woman whose BV recurs after treatment, can relapse from a surviving biofilm be
    distinguished from reinfection by an untreated partner?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - "pathophysiology#Recurrent Bacterial Vaginosis"
  - "pathophysiology#Biofilm Persistence After Antimicrobial Therapy"
  - "pathophysiology#Reinfection from an Untreated Sexual Partner"
  rationale: >
    Two curated hypothesis groups converge on the recurrence node and there is currently
    no assay that resolves which operated. The clinical presentation is identical,
    point-of-care tests cannot separate them, and next-generation sequencing has not
    identified a signature that distinguishes relapse from reacquisition. This is why the
    entry retains both routes rather than promoting one: the StepUp trial shows the
    reinfection route is real at a population level, which is a different claim from it
    being the route in any given woman.
  proposed_experiments:
  - experiment_id: bv_strain_resolved_recurrence_tracking
    name: Strain-resolved tracking of BV-associated bacteria across a treated episode
    description: >
      Deep metagenomic strain typing of the woman's pre-treatment community, her partner's
      penile and urethral community, and her post-treatment recurrent community. Relapse
      predicts recovery of the woman's own pre-treatment strains; reinfection predicts
      strains matching the partner's and not present in her pre-treatment sample.
    decision_criterion: >
      Whether recurrent-episode strains are shared with the woman's own pre-treatment
      community, with the partner's community, or with both.
  evidence:
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Frustratingly for both clinicians and patients, the contribution of reinfection and vaginal relapse cannot be separated, as the clinical presentation of both mechanisms of recurrence is identical."
    explanation: >-
      States the indistinguishability that defines this gap.

- discussion_id: bv_no_animal_model_vaginolysin_human_specific
  prompt: >-
    How much of BV pathogenesis can any animal model address, given that vaginolysin
    requires human CD59 and no animal has a lactobacillus-dominant acidic vagina?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - "pathophysiology#Vaginolysin-Mediated Epithelial Pore Formation and Signalling"
  - "pathophysiology#Loss of Lactobacillus Dominance and Vaginal Acidification"
  rationale: >
    The species restriction here is not incidental, it is receptor-level: vaginolysin lyses
    cells only where human CD59 is present, and transfecting human CD59 into hamster cells
    is what makes them susceptible. Separately, the healthy state this disease departs from
    - a lactobacillus-dominant, lactic-acid-acidified vagina - is close to unique to humans,
    so there is no normal community for a model organism to lose. The consequence is that
    the murine work curated above can test one downstream arm (T-cell activation) and
    nothing else, and that essentially all mechanism in this entry rests on human samples
    and human cell lines. Reviews of BV have named the absence of a suitable animal model
    as a limiting factor for two decades.
  proposed_experiments:
  - experiment_id: bv_humanised_cd59_and_hydrogel_models
    name: Human-relevant model systems for the two blocked arms
    description: >
      For the toxin arm, a human-CD59 transgenic mouse or a human vaginal epithelial
      organ-chip perfused with defined consortia. For the community arm, a
      cervicovaginal-mucus hydrogel or organ-chip supporting stable L. crispatus dominance
      into which BV taxa can be introduced, giving a system in which the transition itself
      is observable.
    decision_criterion: >
      Whether either system reproduces vaginolysin-dependent epithelial responses and
      sialidase-mediated glycocalyx loss with human-like kinetics.
  evidence:
  - reference: PMID:18390664
    reference_title: "Functional and phylogenetic characterization of Vaginolysin, the human-specific cytolysin from Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: OTHER
    snippet: "Mechanistic studies of BV and its adverse consequences have been limited by the absence of definitive diagnostic testing and a suitable animal model"
    explanation: >-
      Cited as background rather than as a result: this is the authors' statement about the
      state of the field in their introduction, not an outcome of their own experiments, so
      it is tagged OTHER. It names the model gap explicitly, in the same paper that
      establishes the receptor-level reason for it (quoted separately below from the
      companion study).
  - reference: PMID:24082080
    reference_title: "Vaginolysin drives epithelial ultrastructural responses to Gardnerella vaginalis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "transfection of nonhuman cells (CHO-K1) with hCD59 renders them susceptible to toxin-induced membrane blebbing"
    explanation: >-
      Demonstrates that human CD59 is the sufficient species determinant, which is what
      makes this a receptor-level rather than a generic model limitation.

- discussion_id: bv_asymptomatic_majority_and_screening
  prompt: >-
    If more than half of women meeting the case definition have no symptoms, and the
    barrier defect is present in asymptomatic disease, what is the right target population
    for treatment?
  kind: OPEN_QUESTION
  status: OPEN
  attaches_to:
  - "pathophysiology#Loss of Cervicovaginal Mucus Adhesive Barrier Function"
  rationale: >
    Two curated findings sit awkwardly together. Asymptomatic BV is the majority of BV,
    and the HIV-relevant mucus barrier defect is present regardless of symptoms - which
    argues for screening. But the entry's own obstetric node records that
    screening-and-treatment has not reliably prevented preterm birth, and treatment does
    not durably restore an L. crispatus community. This is recorded as an open question
    rather than a knowledge gap because the missing item is a policy-relevant trial result,
    not a mechanism.
  evidence:
  - reference: PMID:31971984
    reference_title: "The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "HIV virions had significantly increased mobility in CVM from women with BV compared to CVM from women with Lactobacillus crispatus-dominant microbiota, regardless of whether symptoms were present"
    explanation: >-
      Establishes that the barrier defect does not depend on symptoms, which is what makes
      the target-population question live.
  - reference: PMID:34470644
    reference_title: "Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Although up to half of BV-affected women do not experience symptoms"
    explanation: >-
      Establishes the size of the asymptomatic population at issue.

notes: >
  Scope. This entry models bacterial vaginosis as a community state, so no single
  organism is curated as "the" cause and no `genetic:` section is present - BV is not a
  Mendelian or susceptibility-gene disease and none of the cited work supports a host
  genetic arm. The 2026-08-25 completeness sweep did surface candidate-gene association
  studies (IL-1B, TNFA, TLR variants), but they are small, single-population and
  unreplicated, and no GWAS exists; the sweep's own recommendation was to leave the section
  absent rather than curate them, and that recommendation is followed here.

  What is deliberately absent. There is no aerobic vaginitis, desquamative inflammatory
  vaginitis, candidiasis or trichomoniasis content here; those are separate entities with
  different communities and different treatments. Metronidazole is not attached to the
  ribosomal-translation node because its nitroimidazole mechanism is unrelated to it.
  No `conforms_to` is claimed against `intracellular_pathogen_persistence`: the BV
  consortium is extracellular and biofilm-associated, and the cell-penetrance constraint
  that module encodes does not apply.

  Provenance. The entry was built directly from primary literature identified by PubMed
  search, and every snippet was checked against the cached reference before commit. A
  deep-research report was produced afterwards rather than before
  (`research/Bacterial_Vaginosis-deep-research-claude.md`, 2026-08-25) as a
  content-completeness sweep in response to PR review; it found the licensed-treatment gap
  that this revision closes, and its remaining recommendations - a host-MMP barrier arm,
  the preterm-birth association-versus-intervention controversy, the virulent-strain
  conceptual model of pathogenesis, and a molecular diagnostic definition - are recorded
  there as follow-ups and are deliberately not folded into this revision.

📚

References & Deep Research

References

5

PMID:35118003

Bacterial Vaginosis: What Do We Currently Know?

2 findings

Contemporary review of the vaginal microbiome, the community-state-type framework, the Amsel/Nugent diagnostic standards and first-line antimicrobial therapy.

States plainly that the etiology of the dysbiosis is unknown and that relapse after standard therapy approaches 50% at six months.

PMID:34470644

Bacterial vaginosis: drivers of recurrence and challenges and opportunities in partner treatment.

2 findings

Sets out the two competing accounts of recurrence - biofilm/BVAB persistence versus reinfection from an untreated partner - and the evidence for sexual exchange of BV-associated bacteria.

Notes that the two routes cannot be separated clinically because their presentation is identical.

PMID:40043236

Male-Partner Treatment to Prevent Recurrence of Bacterial Vaginosis.

1 finding

Open-label randomised trial (StepUp) stopped early by its DSMB because treating the woman alone was inferior to treating both partners.

PMID:16260520

Adherent biofilms in bacterial vaginosis.

1 finding

Vaginal biopsy FISH study identifying a confluent Gardnerella-dominated adherent biofilm as the structural feature specific to bacterial vaginosis.

PMID:31971984

The cervicovaginal mucus barrier to HIV-1 is diminished in bacterial vaginosis.

1 finding

HIV virion mobility is increased in cervicovaginal mucus from women with BV, and the defect is in the adhesive rather than the structural properties of the mucus.

Deep Research

1

Deep research results are used as seeds for research; they do not undergo the same validation as the main records and may contain errors. How we use deep research.

Evaluations and curation notes (3)

Record notes

Scope. This entry models bacterial vaginosis as a community state, so no single organism is curated as "the" cause and no `genetic:` section is present - BV is not a Mendelian or susceptibility-gene disease and none of the cited work supports a host genetic arm. The 2026-08-25 completeness sweep did surface candidate-gene association studies (IL-1B, TNFA, TLR variants), but they are small, single-population and unreplicated, and no GWAS exists; the sweep's own recommendation was to leave the section absent rather than curate them, and that recommendation is followed here. What is deliberately absent. There is no aerobic vaginitis, desquamative inflammatory vaginitis, candidiasis or trichomoniasis content here; those are separate entities with different communities and different treatments. Metronidazole is not attached to the ribosomal-translation node because its nitroimidazole mechanism is unrelated to it. No `conforms_to` is claimed against `intracellular_pathogen_persistence`: the BV consortium is extracellular and biofilm-associated, and the cell-penetrance constraint that module encodes does not apply. Provenance. The entry was built directly from primary literature identified by PubMed search, and every snippet was checked against the cached reference before commit. A deep-research report was produced afterwards rather than before (`research/Bacterial_Vaginosis-deep-research-claude.md`, 2026-08-25) as a content-completeness sweep in response to PR review; it found the licensed-treatment gap that this revision closes, and its remaining recommendations - a host-MMP barrier arm, the preterm-birth association-versus-intervention controversy, the virulent-strain conceptual model of pathogenesis, and a molecular diagnostic definition - are recorded there as follow-ups and are deliberately not folded into this revision.

Review response on PR #9073: licensed treatments, evidence retagging, DR sweep · 2026-08-25T12:53:13Z · View source

Response to the 2026-08-20 ai4c-reviewer CHANGES_REQUESTED review on PR #9073, plus a merge of origin/main to clear a conflict the branch had accumulated. Branch refresh. The conflict was confined to two derived cache files, cache/chebi/terms.csv and cache/enums/chemicalentityterm_b3afbb412a9e.csv. Resolved by taking main's rows wholesale and then re-deriving this entry's own rows with `just validate-terms`, rather than hand-placing any row. The resulting cache diff is additive only (CHEBI:18127 cadaverine restored; CHEBI:140628 and CHEBI:31466 added). Blocking finding 1 - no deep-research artifact. Produced research/Bacterial_Vaginosis-deep-research-claude.md, a PubMed E-utilities completeness sweep over eight dimensions (treatments, adverse outcomes, pathogenesis models, host genetics, immune/barrier mechanism, diagnostics, sialidase provenance, datasets). Every PMID, title, journal and publication type in it was read back from the NCBI record rather than recalled. `just preflight-dr` returns SKIP because MONDO:0005316 has no RO:0004003 causal gene, so the gene-identity check cannot discriminate; `just qc-deep-research --only Bacterial_Vaginosis` reports 0 missing and 0 unresolved references. The sweep is recorded honestly as having been run *after* curation rather than before, in the entry's provenance notes. Blocking finding 2 - missing licensed treatments. Added three: - Secnidazole (CHEBI:140628), single 2 g oral dose, FDA-approved 2017. PMID:28867602, phase 3 double-blind placebo-controlled, 189 women at 21 US centres. - Tinidazole (CHEBI:63627), FDA-approved and a CDC alternative regimen. PMID:17666604, multicentre placebo-controlled trial, plus its registration record NCT00229216 added to clinical_trials. - Dequalinium chloride (CHEBI:31466), a non-antibiotic antiseptic. PMID:38696172, phase 4 double-dummy non-inferiority trial against oral metronidazole. All three attach to Polymicrobial Anaerobic Overgrowth and none to the ribosomal translation node, matching the existing treatment of metronidazole: secnidazole and tinidazole share metronidazole's nitroimidazole reductive-activation mechanism, and the dequalinium evidence is a clinical cure rate rather than a demonstrated action on the biofilm. The tinidazole entry deliberately curates both halves of its endpoint-dependent result - the 36.8%/27.4% cure rates under the strict five-criterion FDA definition and the paper's own statement that efficacy was greater under traditional criteria - so that those figures are not read against the 70-85% quoted from Amsel-based series elsewhere in the same entry. Blocking finding 3 - background prose quoted as findings. Three items fixed: - PMID:39162399 on the biofilm/adhesion node re-quoted from the abstract's opening background sentence to the study's own adhesion result, which supports the node directly and is the sentence the reviewer identified. - Two PMID:18390664 items (Gardnerella carriage in asymptomatic women; the absence of a suitable animal model) retagged from IN_VITRO to OTHER, with each explanation now stating that the sentence is the authors' introduction rather than an outcome of their cytolysin experiments. Both are genuinely in that paper's introduction, citing its own references, and were verified against the cached full text. Non-blocking findings also taken: - Finding 5. GO:0006885 regulation of pH removed from the trigger node. The claim is about the chemistry of the vaginal lumen, not a cellular homeostatic process, and GO annotates the latter. The reason is recorded in the node description rather than left as a silent deletion; the mechanism is still carried by GO:0019249 lactate biosynthetic process and the luminal measurement by the Amsel pH definition. - Finding 6. A DIRECT edge added from Recurrent Bacterial Vaginosis back to Loss of Lactobacillus Dominance and Vaginal Acidification, closing the cycle the entry argues for. Evidenced by the LACTIN-V recurrence definition (recurrence is operationalised as re-meeting the same Amsel and Nugent criteria as an incident case) and by PMID:34470644 naming failure to recolonise with lactobacilli among the drivers of recurrence. Finding 4 - three compound node names - was not taken. The reviewer marked it non-blocking and noted that each is a tight mechanistic pair rather than a chain; splitting them would fragment shared evidence. Finding 7 - phenotype coverage - was left for a follow-up. The sweep surfaced the meta-analytic preterm-birth literature and a live controversy in it (a robust association whose causal reading is undermined by the failure of the corresponding intervention), which is a discussion rather than a frequency band and is recorded in the report as a follow-up rather than curated here. Validation run in the worktree: `just validate-disorders` passes schema, terms and references with 99/99 snippets verified against cached references (88/88 before this revision); `just validate-history` clean on this record.

Create: Bacterial Vaginosis (MONDO:0005316) · 2026-08-20T05:13:08Z · View source

New disorder entry created for the obstetrics/gynaecology coverage gap tracked in issue #7837. Modelled as a community-state disease rather than a single-organism infection: no infectious agent is curated as sufficient, and the Gardnerella entry carries a PARTIAL evidence item recording that carriage occurs in asymptomatic women. Fifteen pathophysiology nodes run from loss of Lactobacillus dominance through the adherent Gardnerella biofilm, vaginolysin-mediated epithelial signalling and bacterial sialidase degradation of the cervicovaginal mucus barrier, to HIV susceptibility, adverse pregnancy outcomes, and recurrence. Load-bearing curation decisions: - Recurrence is curated as two competing mechanistic hypotheses converging on one node: biofilm_persistence_relapse (CANONICAL, because the biofilm is directly demonstrated on biopsy) and sexual_reinfection (EMERGING, on the strength of the 2025 StepUp randomised trial stopped early for benefit, PMID:40043236). Neither is asserted as operative in an individual woman; the indistinguishability is curated as a KNOWLEDGE_GAP. - Biogenic amine production is a pathophysiology node with a feedback edge onto the trigger node, not merely the odour phenotype, because the cited work shows amines slow lactobacillus growth and independently cut lactic acid output. - The mucus-barrier node records that the defect is in adhesive rather than structural properties (PMID:31971984), and is deliberately not described as the barrier being torn open. - Negative and constraining results are retained rather than dropped: no increase in O-glycanase, proteinase or whole-mucinase activity in BV (PMID:10454178); prolidase and anti-Gvh IgA did not predict preterm birth (PMID:12388968); and a prospective REFUTE item against the BV-to-PID inference (PMID:35086915). - conforms_to is claimed only at bacterial_protein_synthesis_inhibition#Bacterial mRNA Translation by the Ribosome, for the clindamycin arm. Metronidazole is explicitly not attached to that node, and intracellular_pathogen_persistence is explicitly not claimed. - A HUMAN_MODEL_MISMATCH discussion records the receptor-level reason BV has no animal model (vaginolysin requires human CD59) alongside the absence of a lactobacillus-dominant vagina in model species; the one murine model is curated as PARTIALLY_RECAPITULATES with fidelity LOW. - The StepUp trial has no NCT number and is keyed on ICTRP:ACTRN12619000196145 (ANZCTR) fetched with just ictrp-fetch. - Two environmental exposures are curated with influences_mechanisms but no exposure_term; ECTO was searched for sexual-activity and contraceptive-device exposure concepts and nothing suitable was found, which is recorded in each entry notes field rather than left implicit. Validation: just validate (schema, terms and references) passes, reporting 88 of 88 snippets verified against cached references. just check-duplicate-keys, just check-title-snippets and just check-term-cache-integrity are all clean. Cache diffs are purely additive: 10 new rows across 8 files, all written by the validator, with no reordering churn. No deep-research provider report was used; all references were found by PubMed search and fetched with just fetch-reference.

Claude ▸

Disease Pathophysiology Research — Bacterial Vaginosis (MONDO:0005316)

claude-opus-5 24 citations 2026-08-25T13:00:00Z

Question

Disease Pathophysiology Research — Bacterial Vaginosis (MONDO:0005316)

Focus: content-completeness sweep for the dismech entry created in PR #9073. The entry was curated directly from primary literature without a deep-research provider report, and the PR review (2026-08-20) blocked on the absence of one, specifically because a completeness sweep is the kind of pass that catches missing licensed therapies — which it did.

Method: PubMed E-utilities searches (esearch + esummary + efetch) across eight dimensions, with every PMID, title, journal, year and publication type read back from the NCBI record rather than recalled. Numeric results quoted below are from the fetched abstracts.

Standing caveat. Everything here is a lead. No statement below may be copied into a YAML snippet:. Any PMID used for curation must first be fetched with just fetch-reference and the exact quote verified with just count-verified-snippets.

1. Treatment landscape — the confirmed gap

The entry as first written curated metronidazole, clindamycin, LACTIN-V and partner treatment. The sweep found three further agents with randomised data, two of them FDA-approved. All three have been added to the entry in this revision.

Agent	Status	Key trial	PMID
Secnidazole 2 g single oral dose	FDA-approved 2017	Phase 3, placebo-controlled, 189 women, 21 US centres	28867602
Tinidazole (1 g × 5 d; 2 g × 2 d)	FDA-approved; CDC alternative regimen	Phase 3, placebo-controlled, 235 women, 10 US centres (NCT00229216)	17666604
Dequalinium chloride 10 mg × 6 d	Widely used in Europe; non-antibiotic antiseptic	Phase 4 non-inferiority vs oral metronidazole, 147 women	38696172

Notes that shaped how these were curated:

Secnidazole and tinidazole are nitroimidazole class-mates of metronidazole, so they attach to the same Polymicrobial Anaerobic Overgrowth node and not to the ribosome node that carries clindamycin. Neither was tested against metronidazole for superiority; both were placebo-controlled, so no comparative efficacy claim is warranted.
The tinidazole trial's cure rates (36.8% and 27.4% vs 5.1% placebo) look far worse than the 70–85% quoted elsewhere in the entry. This is an endpoint artefact, not a drug difference — Livengood used the strict five-criterion FDA definition, and the same paper states efficacy was greater under traditional criteria. Curating the number without the caveat would have produced a false comparison inside one entry. Both halves are now curated.
Dequalinium is mechanistically distinct (membrane-active antiseptic, no resistance-selection pressure of the nitroimidazole kind), which is the trial's stated rationale.

Still uncurated treatment leads

Astodrimer sodium 1% gel (dendrimer, approved in Europe/Australia, not US). Mini-review PMID:35246717 reports BV cure rates of 50–57% and a recurrence-prevention result against placebo. A dendrimer that acts by physically blocking adhesion would attach to the biofilm/adhesion node rather than the overgrowth node — i.e. it would be the only agent in the entry targeting the arm that metronidazole is explicitly curated as not reaching. Worth a follow-up pass; the primary trials, not the mini-review, should be the cited source.
Vaginal microbiome transplantation (VMT). PMID:31591599 (Nat Med 2019, NCT02236429) is an exploratory case series of five women with intractable BV; four achieved long-term remission with reconstitution of a Lactobacillus-dominated microbiome, some requiring repeat transplant and one a donor change. This is a five-patient uncontrolled series and should be curated, if at all, as an EMERGING hypothesis-linked intervention with the sample size in the description — not as an established treatment. A 2025 engraftment preprint exists (PMID:40909844) and is explicitly a preprint.
Boric acid and TOL-463 returned no BV-specific randomised evidence in this sweep. Do not curate from the searches above.

2. Adverse outcomes — where the entry is thin, and one genuine controversy

The entry curates Premature birth as OCCASIONAL from a single cohort (9/115). The sweep found the meta-analytic literature, and it contains a live dispute that is more interesting than the point estimate:

PMID:36251068 (2023, 20 articles, 290,397 observations): OR 1.79 (1.32–2.43), RR 1.44 (1.19–1.73). Concludes BV is "undoubtedly associated" with preterm birth and argues for routine screening.
PMID:39442804 (2025, 28 studies, 50,466 patients): OR 1.60 (1.36–1.89), I²=67%. Concludes the association is weaker than previously documented, and explicitly offers that as an explanation for why treating BV in pregnancy has not reduced preterm birth.
PMID:36651636 (2023): individual-participant-data meta-analysis of antibiotic treatment of BV to prevent preterm delivery.

This is a candidate mechanistic_hypotheses / discussions item rather than a frequency band: a robust association whose causal interpretation is undermined by the failure of the corresponding intervention. The entry's existing Ascending Infection and Adverse Pregnancy Outcomes node is the attachment point, and the treatment-failure argument is exactly the sort of constraint this entry already handles well elsewhere (see its retained REFUTE items).

Other outcome leads with real effect sizes:

PID: PMID:34396403 (Longitudinal Study of Vaginal Flora, N=2956, prospective) — Nugent-BV aHR 1.53 (1.05–2.21), symptomatic Amsel-BV aHR 2.15 (1.23–3.75), vaginal douching aHR 1.47 (1.03–2.09). Note this cuts against the entry's retained REFUTE item (PMID:35086915, no association between a Gardnerella-dominated microbiome and subsequent PID). The two are not necessarily inconsistent — one measures a taxon, the other measures the syndrome — but curating both would materially improve the PID arm, and the distinction between them is itself the finding.
HIV acquisition: PMID:18614873 is already curated. Additional: PMID:22745608 (female-to-male transmission, prospective African couples), PMID:21358808 (IPD meta-analysis of intravaginal practices).
Infertility: PMID:23543384 (Hum Reprod 2013 meta-analysis) — BV more prevalent in infertility than antenatal controls (OR 3.32, 1.53–7.20), and more prevalent in tubal infertility specifically (OR 2.77, 1.62–4.75), but not associated with decreased conception rates (OR 1.03, 0.79–1.33) though associated with preclinical pregnancy loss. The negative conception result is the more useful curation target, because it constrains the mechanism.

3. Mechanism — the sexual-transmission hypothesis has a named model

The entry curates recurrence as two hypotheses (biofilm_persistence_relapse CANONICAL, sexual_reinfection EMERGING). The sweep found that the transmission account is not only a recurrence hypothesis but a primary-pathogenesis model with an explicit published statement:

PMID:31369673 — Muzny et al., "An Updated Conceptual Model on the Pathogenesis of Bacterial Vaginosis" (J Infect Dis 2019). States that BV epidemiology supports sexual transmission and that the central debate is primary pathogen versus sexually transmitted polymicrobial consortium. Its update responds to the objection that G. vaginalis is found in virginal women, by proposing that the genus contains 13 species and that healthy women carry non-pathogenic ones while virulent strains cause BV. Also names Prevotella bivia and Atopobium vaginae as model components.
PMID:24511102 — the earlier (2014) version of the same conceptual model.
PMID:30001418 — phenotypic characterisation of G. vaginalis subgroups suggesting they differ in virulence potential.

Why this matters for the entry as written. The entry currently records that P. bivia does not adhere (from PMID:39162399) and treats the consortium as the pathogenic unit. The Muzny model is a genuinely competing framing — a virulent-strain model, where identity within Gardnerella is doing the work that the entry attributes to the community. That is a candidate ALTERNATIVE mechanistic_hypotheses group, and it also bears on the entry's HUMAN_MODEL_MISMATCH, since a strain-resolved model changes what an animal or organ-chip system would have to reproduce.

4. Host genetics — an absent section

The entry has no genetic: section. There is real, if modest, literature:

PMID:17314118 — IL-1B −511 CC (OR 1.5, 1.03–2.14) and +3954 TT (OR 2.8, 1.37–5.88) associated with BV in non-pregnant Italian women; the −511/+3954 T-C haplotype protective (OR 0.7).
PMID:17123692 — TNFA −308G>A and the TNF-α response to altered vaginal flora.
PMID:23021866, PMID:19200604 — TLR variants and BV / Gardnerella–Atopobium carriage.
PMID:32723796 — host genetic factors associated with vaginal microbiome composition (Kenyan women).

Recommendation: curate with restraint or not at all. These are small single-population candidate-gene association studies of the kind that replicate poorly, and no GWAS surfaced. If curated, relationship_type: SUSCEPTIBILITY with the population named in the description, and the absence of replication stated. A KNOWLEDGE_GAP recording that BV has no GWAS-scale host-genetic evidence may be worth more than the individual associations.

5. Immune / barrier mechanism — one strong addition

PMID:32094253 (Infect Immun 2020) — host matrix metalloproteinases secreted in response to BV-associated bacteria disrupt endocervical epithelial polarisation and increase HIV-1 transmigration through the epithelium; an MMP inhibitor reduces the effect, and cervicovaginal fluid with higher BV-associated MMP concentrations increases transmigration.

This is a host-derived barrier-damage mechanism running in parallel to the bacterial glycosidase mechanism the entry already curates (sialidase/fucosidase stripping the glycocalyx). The entry's barrier arm is currently entirely bacterial-enzyme-driven; this would add the host-protease arm and would sit naturally upstream of the HIV-susceptibility node. Strongest single mechanistic addition found by this sweep.

Supporting: PMID:24403560 (species-specific effects of vaginal taxa on innate immunity and barrier properties), PMID:32515473 (Prevotella species modulate barrier function in a 3D endometrial epithelial model — relevant to the entry's HUMAN_MODEL_MISMATCH, since it is a human 3D model rather than an animal one), PMID:18403235 (Atopobium vaginae innate immune response in vitro).

6. Diagnostics — molecular assays

The entry curates Amsel and Nugent as definitions. Molecular assays are absent: PMID:22535982 (validated semiquantitative multitarget PCR), PMID:20814710 (qPCR vs Gram stain accuracy), PMID:38061216 (real-time PCR for intermediate Nugent scores). A molecular definition would be a genuine addition, because the entry already argues that BV's prevalence range is partly an artefact of diagnostic method — a third method with its own operating characteristics is the evidence for that argument.

7. Sialidase source — already correctly curated

PMID:39186657 (PNAS 2024, "Prevotella are major contributors of sialidases in the human vaginal microbiome") is already cited in the entry. The sweep confirms this is the current state of the art on sialidase provenance and that the entry's choice to treat sialidase as a consortium property rather than a Gardnerella property is the correct reading.

8. Datasets

No datasets: block exists on the entry. This sweep did not run a repository search; just discover-datasets Bacterial_Vaginosis and just verify-datasets are the right tools and were not run here. Flagged as an open item rather than answered.

Summary of what this sweep changed

Acted on in this revision: section 1 (three treatments added), plus the review's other two blocking findings.

Recommended follow-ups, in descending value:

Host-MMP barrier-disruption arm (PMID:32094253) — section 5.
Preterm-birth association-vs-intervention controversy as a discussion — section 2.
Muzny virulent-strain conceptual model as an ALTERNATIVE hypothesis — section 3.
PID prospective cohort alongside the existing REFUTE item — section 2.
Astodrimer sodium as a biofilm/adhesion-targeting agent — section 1.
Molecular diagnostic definition — section 6.
Host genetics, with restraint, or as a knowledge gap — section 4.
Dataset discovery — section 8.

These are deliberately not all done in this PR. The PR is a review-response on an entry that is already approved-in-substance; folding eight new arms into it would make the diff unreviewable and would re-open questions the reviewer has already settled. They belong in a follow-up pass.

References surfaced by this sweep

All PMIDs below were returned by PubMed E-utilities during this sweep and their titles read back from the NCBI record.

PMID	Section	Title (abbreviated)
28867602	1	Phase-3 placebo-controlled single-dose secnidazole 2 g for BV
17666604	1	Effectiveness of two tinidazole regimens in treatment of BV
38696172	1	Dequalinium chloride vs metronidazole for BV: RCT
35246717	1	Astodrimer sodium and bacterial vaginosis: a mini review
31591599	1	Vaginal microbiome transplantation in intractable BV
40909844	1	Donation strain engraftment, VMT (preprint)
36251068	2	Effect of BV on preterm birth: a meta-analysis
39442804	2	Reassessing the association between BV and preterm birth
36651636	2	Antibiotic treatment of BV to prevent preterm delivery: IPD MA
34396403	2	BV and behavioral factors in incident PID (LSVF)
22745608	2	BV and female-to-male HIV-1 transmission
21358808	2	Intravaginal practices, BV and HIV: IPD meta-analysis
23543384	2	Risks associated with BV in infertility patients: meta-analysis
31369673	3	An updated conceptual model on the pathogenesis of BV
24511102	3	Role of G. vaginalis in BV pathogenesis: a conceptual model
30001418	3	Phenotypic characterization of G. vaginalis subgroups
17314118	4	IL-1beta and IL-1ra polymorphisms and BV
17123692	4	TNFA-308G>A and the TNF-alpha response to altered vaginal flora
23021866	4	TLR gene variants and BV among HIV-1 infected adolescents
32723796	4	Host genetic factors and vaginal microbiome composition
32094253	5	MMPs disrupt endocervical epithelium, increasing HIV transmigration
24403560	5	Vaginal bacteria alter innate immunity and barrier properties
32515473	5	Prevotella modulate barrier function in 3D endometrial model
22535982	6	Validated semiquantitative multitarget PCR for BV diagnosis

Bacterial Vaginosis

Ask OpenScientist

OpenScientist Report

Definitions

Mechanistic Hypotheses

Discussions and Knowledge Gaps

Pathophysiology

Pathograph

Phenotypes

Medical Actions

Environmental Factors

Prevalence

Infectious Agent

Transmission

Clinical Trials

Animal Models

Source YAML

References & Deep Research

References

Deep Research

Question

Disease Pathophysiology Research — Bacterial Vaginosis (MONDO:0005316)

1. Treatment landscape — the confirmed gap

Still uncurated treatment leads

2. Adverse outcomes — where the entry is thin, and one genuine controversy

3. Mechanism — the sexual-transmission hypothesis has a named model

4. Host genetics — an absent section

5. Immune / barrier mechanism — one strong addition

6. Diagnostics — molecular assays

7. Sialidase source — already correctly curated

8. Datasets

Summary of what this sweep changed

References surfaced by this sweep