Endometrial Microbiome Dysbiosis and IVF Outcomes: From Association to Clinical Utility: A Narrative Literature Review

Background: The female reproductive tract, once assumed to be sterile above the cervix, is now recognised to harbour a distinct, low-biomass microbial community. Molecular studies over the past decade have proposed that the composition of this endometrial microbiota particularly the relative dominance of Lactobacillus species is associated with embryo implantation and pregnancy outcomes following in vitro fertilisation (IVF). However, the field remains divided between researchers who view endometrial dysbiosis as a clinically actionable biomarker and others who question whether a genuine endometrial microbiome exists at all, given the confounding effect of low-biomass contamination.

Objective: This review synthesises the current evidence on endometrial microbiome dysbiosis and IVF outcomes, tracing the field from its foundational association studies to the emerging, but still contested, question of clinical utility.

Methods: A narrative synthesis of peer-reviewed literature (2016–2026) was performed, drawing on studies indexed in PubMed, Scopus and related repositories that examined endometrial or genital-tract microbiota composition, chronic endometritis, and reproductive outcomes in women undergoing IVF or intracytoplasmic sperm injection (ICSI).

Results: A Lactobacillus-dominated endometrial microbiota (generally defined as >90% relative abundance) is repeatedly associated with higher implantation, clinical pregnancy and live birth rates compared with a non-Lactobacillus-dominated, dysbiotic profile. Dysbiosis correlates with local pro-inflammatory cytokine activity and frequently co-occurs with, but is not identical to, histologically defined chronic endometritis. Commercial next-generation-sequencing panels (e.g. EMMA/ALICE) and small prospective cohorts report improved pregnancy rates after antibiotic and/or probiotic correction of dysbiosis, yet larger and better-controlled studies have failed to reproduce a benefit, and methodological concerns, sampling contamination, absence of validated normative thresholds, and heterogeneous outcome definitions, temper confidence in causality.

Conclusion: Endometrial dysbiosis is a reproducible association with adverse IVF outcomes, but current evidence does not yet establish it as a validated, actionable clinical biomarker. Routine endometrial microbiome testing and empirical treatment are not yet justified outside of research settings or, arguably, recurrent implantation failure (RIF) work-ups; adequately powered randomised controlled trials (RCTs) with standardised sequencing methodology and contamination controls are required before the transition from association to clinical utility can be considered complete.

Infertility affects a substantial proportion of reproductive-age couples worldwide, and assisted reproductive technologies, particularly IVF, remain the mainstay of treatment for many aetiologies (Cox et al., 2022). Despite continuous refinement of ovarian stimulation protocols, laboratory culture systems and embryo selection tools, live birth rates per embryo transfer have plateaued at well below 50% in most programmes, focusing renewed attention on the receiving environment of the embryo, the endometrium, rather than the embryo itself. For much of the twentieth century, the uterine cavity was considered sterile in health, with bacterial presence equated with infection. This paradigm has been challenged by culture-independent, sequencing-based studies demonstrating that the endometrium, like the gut and vagina, hosts a distinct, low-biomass microbial community that differs in composition from the vaginal microbiota (Moreno et al., 2016). This community appears to be dominated in fertile, reproductively healthy women by Lactobacillus species, mirroring the protective, lactic-acid-producing ecology long described in the vagina. Deviation from this Lactobacillus-dominated state — termed endometrial dysbiosis — has been proposed as a novel, modifiable factor in implantation failure, recurrent pregnancy loss and IVF failure.

The clinical appeal of this hypothesis is considerable: unlike embryo aneuploidy or age-related oocyte decline, a dysbiotic endometrial microbiome is, in theory, correctable with antibiotics or probiotics, offering a tractable target for improving IVF success. This has driven the rapid commercialisation of endometrial microbiome sequencing panels, most notably the Endometrial Microbiome Metagenomic Analysis (EMMA) and Analysis of Infectious Chronic Endometritis (ALICE) tests, which are now marketed directly to IVF patients, particularly those with recurrent implantation failure (RIF) (Moreno et al., 2022). However, the endometrial microbiome field is also one of the more contested areas of reproductive microbiology. Because the endometrial bacterial load is extremely low, close to or indistinguishable from background reagent and extraction-kit contamination (the 'kitome'), some investigators using stringent contamination controls and transabdominal sampling have questioned whether a genuine resident endometrial microbiota exists at all, and have suggested that earlier Lactobacillus-dominance findings may reflect vaginal or cervical contamination during transcervical sampling (Winters et al., 2019). This tension between a clinically attractive, commercially available biomarker and a still-unsettled basic science question — motivates the present review.

This review aims to: (i) summarise the evidence linking endometrial microbiome composition to IVF outcomes; (ii) examine the relationship between endometrial dysbiosis and the overlapping but distinct entity of chronic endometritis; (iii) appraise the evidence for microbiota-targeted interventions (antibiotics, probiotics) in improving reproductive outcomes; and (iv) critically evaluate whether the field has moved convincingly from association to genuine clinical utility.

This is a narrative, rather than systematic, literature review, reflecting the aim of providing a critical, integrative synthesis across a methodologically heterogeneous and rapidly evolving field rather than a quantitative pooled estimate of effect.

Search strategy

A structured literature search was conducted across PubMed/MEDLINE, Scopus and related full-text repositories for articles published between 2016 (the year of the first major endometrial microbiota–IVF outcome study) and 2026. Search terms combined the concepts of 'endometrial microbiome' OR 'endometrial microbiota' OR 'uterine microbiome' with 'dysbiosis', 'in vitro fertilisation' OR 'IVF' OR 'assisted reproduction', 'implantation', 'chronic endometritis', and 'pregnancy outcome' OR 'live birth'. Reference lists of identified systematic reviews and meta-analyses were hand-searched for additional primary studies.

Eligibility criteria

Original research articles (prospective and retrospective cohorts, case-control studies, pilot studies and randomised trials), together with systematic reviews, meta-analyses and narrative reviews, were eligible if they reported endometrial or genital-tract microbiota composition (by culture, quantitative PCR or next-generation sequencing) in relation to IVF/ICSI outcomes, chronic endometritis, or microbiota-targeted therapeutic interventions in infertile women. Studies restricted purely to vaginal microbiota without an endometrial or intrauterine component were included selectively where they provided relevant mechanistic or ascending-infection context.

Data synthesis

Given the marked heterogeneity in sampling technique (transcervical catheter, endometrial biopsy, hysterectomy specimen), sequencing methodology (16S rRNA gene sequencing, shotgun metagenomics, culturomics, quantitative PCR), dysbiosis definitions (most commonly <90>

A total of 14 studies and reviews forming the core evidence base of this review are summarised in Table 1, spanning foundational cohort studies, diagnostic comparison studies, systematic reviews/meta-analyses, and mechanistic/interventional work published between 2011 and 2025. Detailed narrative synthesis by theme follows in Sections 3.1-3.6.

Table 1: Characteristics of included studies

Establishing an endometrial microbiota distinct from the vagina

The foundational study in this field, by Moreno et al. (2016), used pyrosequencing of endometrial fluid obtained before endometrial receptivity biopsy in 35 women undergoing IVF, and demonstrated a bacterial community profile in the endometrium that differed from paired vaginal aspirates, while remaining relatively stable across the pre-receptive and receptive phases of the cycle. Clustering of samples according to Lactobacillus, Bifidobacterium, Gardnerella and Streptococcus abundance allowed classification into Lactobacillus-dominated (LD, ≥90% Lactobacillus) and non-Lactobacillus-dominated (NLD, <90>

This work was substantially extended in a larger, multicentre, prospective observational study of 342 infertile patients (Moreno et al., 2022), which again used 16S rRNA gene sequencing of endometrial fluid and biopsy samples collected before embryo transfer. A dysbiotic profile enriched for Atopobium, Bifidobacterium, Chryseobacterium, Gardnerella, Haemophilus, Klebsiella, Neisseria, Staphylococcus and Streptococcus was associated with unsuccessful reproductive outcomes, and the authors concluded that pre-transfer endometrial microbiota composition constitutes a potentially useful biomarker for predicting reproductive outcome.

Smaller cohorts have produced broadly concordant findings. In an exploratory study of 141 women whose first IVF/ICSI cycle had failed, women who achieved a live birth within 12 months had a significantly higher relative abundance of Lactobacillus crispatus specifically, rather than Lactobacillus species in general, and a smaller proportion had L. crispatus abundance below 10%, suggesting that species-level, and not only genus-level, resolution may be clinically informative. A culturomics-based analysis of the endometrial microbiota in 93 women undergoing single embryo transfer similarly linked a dysbiotic culturable profile to lower ongoing pregnancy rates, indicating that the association is not an artefact confined to sequencing-based methods alone.

Vaginal dysbiosis and its relationship to endometrial and IVF outcomes

Because ascending colonisation from the vagina is proposed as one route by which the endometrial cavity acquires its microbial community, vaginal dysbiosis has been studied extensively as a proxy, or contributing cause, of poor IVF outcomes. A systematic PRISMA review and meta-analysis of 17 studies and 3,543 patients found a vaginal dysbiosis prevalence of 18%, and demonstrated that vaginal dysbiosis was a significant risk factor for early pregnancy loss in IVF (relative risk 1.71, 95% CI 1.29–2.27); a pre-specified sub-analysis restricted to molecular diagnostic methods additionally showed a significantly reduced clinical pregnancy rate in women with vaginal dysbiosis compared with those with a normal vaginal microbiota (relative risk 0.55, 95% CI 0.32–0.93) (Skafte-Holm et al., 2021). A more recent systematic review and meta-analysis of the vaginal microbiome and assisted reproduction similarly concluded that women with a favourable, Lactobacillus-dominated vaginal microbiome have better reproductive outcomes than those with an unfavourable profile, proposing that vaginal and endometrial dysbiosis share a common mechanistic pathway through pro-inflammatory cytokine up-regulation, local immune overactivation, and consequently impaired endometrial receptivity.

A comparative analysis of the genital tract microbiota of IVF patients and oocyte donors similarly reported that genital tract dysbiosis correlates with poorer reproductive outcomes, and cited earlier work demonstrating an increased risk of a Gardnerella-dominated biofilm attached to the endometrium in women with bacterial vaginosis, consistent with an ascending-infection model linking vaginal dysbiosis to endometrial pathology.

Mechanistic evidence: dysbiosis, inflammation and endometrial receptivity

Beyond simple association, several studies have sought a mechanistic bridge between microbial composition and impaired implantation. Cela et al. (2022) compared endometrial tissue cytokine profiles in women with RIF stratified into eubiosis (≥90% endometrial lactobacilli) and dysbiosis (<90>

Chronic endometritis: an overlapping but distinct entity

Chronic endometritis (CE) a persistent, often subclinical inflammation of the endometrium classically diagnosed by the presence of CD138-positive stromal plasma cells on immunohistochemistry, sometimes supported by hysteroscopic findings of mucosal oedema, focal hyperaemia or micropolyps — is frequently discussed alongside endometrial dysbiosis, and the two concepts are often conflated in clinical practice, but the available evidence indicates they are related, overlapping, yet distinct phenomena. CE prevalence estimates in RIF populations range widely, from around 30% up to 56–67

Taken together, the literature supports a consistent, moderate-to-large association between a non-Lactobacillus-dominated (dysbiotic) endometrial or vaginal microbial profile and adverse IVF outcomes, spanning implantation, clinical pregnancy, ongoing pregnancy and live birth rates, and this association is biologically plausible given accompanying evidence of local pro-inflammatory cytokine dysregulation in dysbiotic endometria. This represents genuine progress from the pre-2016 assumption of endometrial sterility, and has legitimately reframed the endometrium as an active site of host–microbe interaction relevant to reproductive success.

However, several considerations temper any claim that the field has definitively progressed from association to validated clinical utility. First, diagnostic heterogeneity remains substantial: sampling route (transcervical catheter versus biopsy versus hysterectomy), sequencing platform, bioinformatic pipeline, and the threshold used to define dysbiosis (most commonly, but not universally, <90>

Fourth, evidence for the pivotal claim that correcting dysbiosis (via antibiotics and/or Lactobacillus probiotics) improves reproductive outcomes remains inconsistent. While several small-to-moderate prospective cohorts report improved pregnancy rates after EMMA/ALICE-guided treatment, at least one comparative study found no benefit over untested, standard-care controls, and systematic reviews of microbiota-targeted therapy for CE-associated RIF describe considerable study heterogeneity and low-to-moderate methodological quality, without a consistent RCT evidence base. This pattern — encouraging observational and single-arm cohort data, without confirmatory adequately powered RCT evidence — is a recognised precursor to over-adoption of unproven interventions in reproductive medicine, and mirrors historical experience with other 'add-on' IVF interventions that were widely adopted on the strength of association data before rigorous trials clarified, or refuted, their benefit.

From a clinical-utility standpoint, three criteria are generally required before a biomarker-guided intervention can be considered validated for routine practice: analytical validity (the test reliably and reproducibly measures what it claims to measure), clinical validity (the test result reliably predicts the clinical outcome of interest), and clinical utility (acting on the test result, typically through treatment, demonstrably improves that outcome, ideally in a randomised comparison against no testing or sham treatment). The reviewed literature provides reasonably consistent evidence for clinical validity — dysbiotic profiles do correlate with worse IVF outcomes across multiple independent cohorts — but weaker and more contested evidence for analytical validity, given the low-biomass contamination problem, and still-immature evidence for clinical utility, given the paucity of adequately powered RCTs directly comparing microbiome-guided treatment against standard care.

Practically, this has implications for how clinicians should currently counsel patients. Endometrial microbiome testing may reasonably be considered, with appropriate counselling about the limitations of current evidence, in the specific context of unexplained RIF after exclusion of other established causes (embryo aneuploidy, uterine structural anomalies, thrombophilia), where it may be integrated alongside, rather than as a replacement for, established chronic endometritis assessment. It should not, on current evidence, be recommended as a routine, first-line investigation for all IVF patients, nor should a positive dysbiosis result alone be assumed automatically to justify empirical antibiotic or probiotic therapy outside a shared decision-making conversation acknowledging the uncertain magnitude of benefit.

Future research priorities that would meaningfully advance the field from association to clinical utility include: standardisation of sampling technique and contamination controls across studies; consensus, externally validated dysbiosis thresholds derived from large, diverse reference populations; adequately powered, blinded RCTs of microbiome-guided antibiotic/probiotic intervention against standard care, with live birth rate as the primary outcome; and mechanistic studies clarifying whether the endometrial microbial signal reflects a genuinely resident, functionally active community or a secondary marker of local immune and inflammatory status.

The past decade has transformed the endometrium from a presumed-sterile organ into a putative site of microbial ecology with plausible relevance to implantation and IVF success. A dysbiotic, non-Lactobacillus-dominated endometrial or vaginal microbial profile is consistently associated with reduced implantation, pregnancy and live birth rates across independent cohorts, and this association is supported by mechanistic evidence of local pro-inflammatory dysregulation. However, unresolved methodological concerns regarding low-biomass sampling contamination, inconsistent diagnostic thresholds, imperfect concordance with chronic endometritis, and — most critically — inconsistent evidence that correcting dysbiosis actually improves reproductive outcomes, mean that the field has not yet completed its transition from association to validated clinical utility. Endometrial microbiome testing and empirical treatment may have a role in carefully selected patients with unexplained recurrent implantation failure, but should currently be regarded as an evolving adjunct rather than a validated, routine component of IVF care, pending the results of larger, methodologically rigorous, contamination-controlled randomised trials.