Research and evidence

The study that established the >90% Lactobacillus-dominant threshold; in IVF patients, non-dominant profiles showed implantation at 23.1% versus 60.7%, and live birth at 6.7% versus 58.8%.

Sample: n=35 participants. Am J Obstet Gynecol, 2016.

Multicentre prospective study across 13 ART centres on four continents (n=342), in which a dysbiotic signature featuring Atopobium, Gardnerella, Klebsiella, Streptococcus and others was independently associated with failed live birth.

Sample: n=342 participants. Microbiome, 2022.

Across 73 RIF patients, endometrial dysbiosis prevalence was 53.4%, hysteroscopic chronic endometritis 56.2%, CD138-positive chronic endometritis 49.3%; the three tests identified largely non-overlapping populations; post-treatment clinical pregnancy reached 88.9% in dysbiosis-positive patients (OR 6.29, p=0.031).

Sample: n=73 participants. Sci Rep, 2025.

A published walk-through of a test-identify-treat-retest pathway. All nine treated non-Lactobacillus-dominant patients converted to Lactobacillus dominance after antibiotic plus probiotic therapy; five of the nine (55.6%) subsequently achieved pregnancy. The evidential strength is in establishing the workflow rather than in the size of the cohort.

Sample: n=92 participants. Reprod Med Biol, 2019.

In women with repeated unexplained implantation failure, live birth reached 60.8% in those whose chronic endometritis resolved after antibiotic therapy, versus 13.3% where it persisted.

Sample size not stated. Hum Reprod, 2015.

The paper that introduced the molecular approach to diagnosing chronic endometritis; compared histology, cultures, hysteroscopy, and molecular microbiology head-to-head in the same cohort and found each test identifies a different subset of patients. Hiratsuka 2025 reported a similar pattern seven years later.

Sample: n=113 participants. Am J Obstet Gynecol, 2018.

Paired samples across six reproductive-tract sites (lower vagina, posterior fornix, cervical mucus, endometrium, fallopian tubes and peritoneal fluid) showed a continuous bacterial transition rather than discrete compartments, establishing the upper reproductive tract as a distinct microbial niche.

Sample size not stated. Nat Commun, 2017.

One of the earliest deep-sequencing studies to characterise a reproducible endometrial microbial signal, using a transcervical device designed to minimise cervicovaginal contamination. The paper established what was measurable with careful methodology.

Sample size not stated. PeerJ, 2016.

A review of why low-biomass sequencing is difficult: it sets out the contamination risks in uterine microbiome studies and the methodological standards labs should meet.

Sample size not stated. Biol Reprod, 2019.

Prospective IVF cohort showing a meaningful drop in clinical pregnancy when vaginal microbiota was abnormal at treatment, defined by the presence of Gardnerella vaginalis and/or Atopobium vaginae.

Sample size not stated. Hum Reprod, 2016.

Prospective study using IS-pro molecular typing; an unfavourable microbiome profile was identified in around 17% of patients pre-treatment and associated with significantly reduced implantation.

Sample size not stated. Hum Reprod, 2019.

Fusobacterium was detected in the endometrium of 64% of endometriosis patients versus <10% of controls; in mouse models, antibiotic eradication significantly reduced lesion formation. Pairs human observational data with mechanistic rescue experiments, suggesting a candidate causal link between a bacterial species and endometriosis.

Sample size not stated. Sci Transl Med, 2023.

Prospective 16S rRNA study of 80 RIF patients (40 with chronic endometritis, 40 without) showing CE patients carry a distinct, more diverse endometrial microbiota with reduced Lactobacillus and enrichment of Proteobacteria; the bacterial signature tracked with adverse reproductive outcomes and altered immune-related metabolic pathways.

Sample: n=80 participants. Ann Clin Microbiol Antimicrob, 2024.

A study of 1,000 people found no difference in gut microbiome composition between individuals with and without endometriosis. An important negative result: it argues against a simple gut-microbiome explanation for endometriosis and is a counterweight to smaller positive cohorts.

Sample: n=1000 participants. BMC Med, 2024.

Non-Lactobacillus-dominant endometrial microbiota was not unique to repeated implantation failure: it was present in 44.4% of RIF patients versus 52.7% of controls. A significant caveat on using dominance alone to explain implantation failure.

Sample size not stated. Arch Gynecol Obstet, 2022.

A systematic review of evidence linking gut, vaginal, cervical and endometrial microbiome alterations to endometriosis.

Sample size not stated. BJOG, 2020.

A review of how the microbiome may act: uterine NK cells, cytokine profiles and immune tolerance.

Sample size not stated. Hum Reprod Update, 2018.

Paired vaginal and endometrial samples collected at hysterectomy showed bacterial colonisation of the upper genital tract in 95% of women (n=58), with Prevotella detected in around one-third of upper-tract samples, establishing the upper genital tract as routinely non-sterile.

Sample size not stated. Am J Obstet Gynecol, 2015.