Oral bacterium Fusobacterium nucleatum and endometriosis: what the evidence really shows

A bacterium best known for living between our teeth has become one of the most talked-about suspects in endometriosis research. Fusobacterium nucleatum, a Gram-negative anaerobe that dominates periodontal pockets and has already been linked to colorectal cancer and adverse pregnancy outcomes, has been proposed as a potential driver of a disease that affects roughly 190 million women worldwide. The idea is seductive: if an oral microbe can travel to the uterus and ignite the chronic inflammation and fibrosis that define endometriotic lesions, it might open new routes to diagnosis and even treatment. But a new systematic review published in Frontiers in Cellular and Infection Microbiology finds that the evidence, while biologically plausible, is far too thin and contradictory to support such bold claims.

The review, led by Maria E. Ramos-Nino, Aniruddha Doshi, and Prakash V. A. K. Ramdass of St. George’s University in Grenada, followed PRISMA 2020 guidelines and searched PubMed/MEDLINE, Embase, Web of Science, and Google Scholar from database inception through August 31, 2026. The team screened 35 records, whittled them down through full-text assessment, and ultimately included just four human studies encompassing 384 participants at the study level, along with the animal component of one translational study. Risk of bias was evaluated with an adapted ROBINS-I framework for human observational studies and the SYRCLE tool for animal work, and certainty was graded using the GRADE framework. Because the studies differed so dramatically in anatomical compartment, detection method, and reported outcomes, the authors judged meta-analysis inappropriate and synthesized the findings qualitatively.

The heart of the controversy lies in two tissue-based studies that simply do not agree. The first, a 2023 translational study by Muraoka and colleagues in Japan, reported a striking signal: Fusobacterium was detected in 64.3 percent of endometrial samples from women with endometriosis compared with just 7.1 percent of controls in a 42-versus-42 tissue subset, and 52.4 percent of endometriotic lesions were also positive. The same group went further, showing in a mouse model that F. nucleatum exposure increased endometriotic lesion weight and promoted fibrosis associated with TGF-β signaling, effects that antibiotic treatment attenuated. That combination of human detection and experimental mechanism made headlines and fueled the oral microbe-host axis hypothesis.

But the second tissue study tells a very different story. Graciano-España and colleagues, working with the University of Pennsylvania Endomarker cohort, analyzed eutopic endometrium from 55 women with endometriosis and 37 controls using genus- and species-specific quantitative PCR. They found no significant difference in Fusobacterium species abundance between cases and controls, and no significant difference in F. nucleatum specifically. Subgroup analyses by disease severity were also null. The review notes that menstrual status was not uniform across groups, with a higher proportion of endometriosis cases menstruating at biopsy, but these differences do not by themselves explain the discordant results. In short, the strongest human tissue association has not been replicated.

The remaining two studies provide only indirect evidence because they never assessed endometrial or endometriotic tissue. Hicks and colleagues in Australia studied 64 participants and found oral Fusobacterium enrichment among patients with moderate or severe endometriosis compared with minimal or mild disease, while vaginal and stool samples showed no analogous signal. That anatomical specificity hints at a relationship between oral microbial composition and disease severity, but as the review emphasizes, it does not establish oral-to-reproductive-tract translocation or causality. Jimenez and colleagues in the United States examined vaginal and rectal microbiomes in women with chronic pelvic pain with and without endometriosis and found that F. animalis, historically classified as F. nucleatum subsp. animalis, was depleted in vaginal samples from the endometriosis group. No endometrial tissue was assessed, so this finding cannot be described as tissue-specific evidence for F. nucleatum.

Methodological heterogeneity is likely a major contributor to the divergent findings. The two tissue studies differed in population characteristics, disease phenotype, sampling procedures, menstrual status, DNA extraction approaches, and molecular targets. FISH provides spatial localization but its taxonomic specificity depends on probe design and hybridization conditions. Quantitative PCR offers sensitive targeted detection but hinges on primer specificity, extraction efficiency, and target selection. 16S rRNA sequencing profiles broader communities but may lack species-level resolution depending on the sequenced region and analytical pipeline. Anatomical sampling varied just as widely, from eutopic endometrium and endometriotic lesions to oral, vaginal, rectal, and stool samples, and evidence at one site cannot be assumed to reflect bacterial presence within lesions.

Contamination is an additional concern that the review treats seriously. Endometrial and endometriotic tissues are low-biomass specimens highly susceptible to reagent, environmental, and procedural contamination, and microbial signals generated this way can be difficult to distinguish from genuine tissue-associated bacteria. Inconsistent reporting of negative and procedural controls reduces confidence in some findings. Geographic and population differences may also contribute, since microbiome composition varies with diet, oral health, medication exposure, age, and disease stage. The emerging concept of an endobiota, which challenges the traditional view of the upper reproductive tract as sterile, makes rigorous contamination control all the more essential before any pathobiont can be credibly implicated.

The quality assessment is sobering. All four human observational studies were judged at serious overall risk of bias, driven primarily by potential confounding: adjustment for factors that could influence the microbiome, including antibiotic exposure, hormonal treatment, menstrual-cycle status, oral health, and comorbid gynecologic conditions, was incomplete or inconsistent across studies. Classification of endometriosis status was predominantly at low risk because surgical or histopathological confirmation was reported, but outcome measurement was rated moderate across studies because of differences in detection methods and potential contamination or taxonomic misclassification. For the animal experiments, the SYRCLE assessment found substantial uncertainty across most domains, with sequence generation, allocation concealment, random housing, and blinding all rated unclear because the relevant procedures were insufficiently reported. The GRADE assessment rated certainty of evidence as very low for every human and mechanistic outcome.

What does this mean for patients hoping that antibiotics might treat endometriosis? The review is unambiguous: the current evidence does not support antibiotic treatment based on presumed F. nucleatum involvement. Although antibiotic treatment reduced lesion burden and fibrosis in the experimental mouse model, these findings remain preclinical and have not been validated in randomized human trials. The authors also caution that evidence supporting antibiotic treatment of chronic endometritis, a distinct inflammatory condition, should not be extrapolated to endometriosis, whose bacterial etiology has not been established. Unnecessary antimicrobial exposure may promote resistance and disrupt commensal microbial communities, so antibiotics targeting F. nucleatum should not be recommended outside appropriately designed clinical trials.

Looking ahead, the review lays out a roadmap for testing the oral microbe-host axis properly. Future studies should use standardized, contamination-aware sampling protocols that define anatomical location, menstrual-cycle phase, recent antibiotic exposure, and oral-health status, with appropriate negative and environmental controls. Combining targeted methods such as qPCR or FISH with sequencing-based approaches in the same specimens could help separate methodological discrepancies from genuine biological variation. Critically, prospective studies should collect paired oral samples, preferably saliva and subgingival plaque, together with eutopic endometrium and endometriotic lesions, and apply strain-level genomic analyses to determine whether genetically related F. nucleatum strains occur in the oral cavity and reproductive tissues of the same individuals. Longitudinal designs are needed to establish whether bacterial presence precedes lesion development or occurs secondary to the altered inflammatory environment. Until such evidence exists, the review concludes, F. nucleatum should be regarded as a hypothesis-generating lead rather than an established cause or therapeutic target in endometriosis.

Subject of Research: The role of the oral bacterium Fusobacterium nucleatum in endometriosis

Article Title: The emerging oral microbe-host axis in endometriosis: a systematic review of Fusobacterium nucleatum

Article References: Ramos-Nino, M. E., Doshi, A., & Ramdass, P. V. A. K. (2026). The emerging oral microbe-host axis in endometriosis: a systematic review of Fusobacterium nucleatum. Frontiers in Cellular and Infection Microbiology, 16, Article 1994934. https://doi.org/10.3389/fcimb.2026.1994934

Image Credits: AI Generated

DOI: 10.3389/fcimb.2026.1994934

Keywords: endometriosis, Fusobacterium nucleatum, oral microbiome, systematic review, microbiome, fibrosis, TGF-β, endometrial tissue, microbial translocation, risk of bias, antibiotics, reproductive tract microbiome