Microbiome treatments have occupied an unusual position in medicine for the better part of two decades. The underlying biology has been credible and increasingly well characterised, while the products reaching patients have largely been dietary supplements sold without any requirement to demonstrate clinical benefit. That gap is now closing at the regulated end of the market.
Since late 2022, two faecal microbiota products have been licensed by the United States Food and Drug Administration as biological products, both indicated for the prevention of recurrent Clostridioides difficile infection in adults following antibacterial treatment. These are prescription microbiome therapies in the full regulatory sense: manufactured under defined controls, tested in randomised placebo-controlled trials, and approved for a specific indication in a specific population.
The distinction between these products and the probiotics on a pharmacy shelf is not a matter of degree. It is a difference in category, evidence standard and legal status, and it is one that patients rarely appreciate when they raise the subject in clinic. Understanding where the gut microbiome has produced licensed therapeutics, where it has produced disappointing trial results, and where the remaining uncertainty sits is becoming a practical requirement across gastroenterology, infectious diseases, oncology and primary care. The commercial narrative around microbiome science has consistently run ahead of the clinical evidence, and separating the two is now part of routine practice.
This article looks at what has been licensed, what the current evidence supports, and what remains speculative. If microbiome questions are coming up in your consultations, the comments section at the end is open.
This article covers:
What Two Decades of Microbiome Science Established
Sequencing-based characterisation shifted the field from culture-dependent description to community-level analysis, and produced a reasonably consistent picture. Microbial communities differ substantially between individuals, vary by body site, and shift with diet, age, geography, and drug exposure. Antimicrobial exposure in particular produces measurable and sometimes prolonged reductions in diversity.
Associations between altered community composition and disease have been reported across inflammatory bowel disease, metabolic disease, colorectal cancer, hepatic disease and several neuropsychiatric conditions. The volume of association data is considerable. The causal data is much thinner, and this remains the central interpretive problem in the field. Disease states alter diet, transit time, inflammation, and medication exposure, all of which alter the microbiota independently. Distinguishing driver from passenger has proved difficult outside a small number of conditions.
Recurrent Clostridioides difficile infection is the clearest exception, and the reason it became the first therapeutic target. Here the mechanism is coherent rather than correlative: antibiotic exposure disrupts colonisation resistance, permitting germination and toxin production, and restoring a diverse microbial community re establishes that resistance. The causal chain is short, the intervention is testable, and the clinical endpoint is unambiguous.
That specificity matters when interpreting the wider literature. The success of microbiome treatments in recurrent infection does not generalise to conditions where dysbiosis is a marker rather than a mechanism, and much commercial activity has not observed that distinction.
Why Probiotics and Prescription Microbiome Therapies Are Different Categories
The regulatory separation is sharper than most patients realise. Probiotic products are generally marketed as foods or dietary supplements*. In the United States, that framework permits structure and function claims but not disease claims, and requires no premarket demonstration of efficacy. Strain identity, viability at the point of use and dose consistency are not independently verified as standard.
Licensed microbiome therapeutics sit under the biological products framework, requiring characterised composition, controlled manufacture, donor screening where the source material is human derived, and efficacy demonstrated in controlled trials before approval.
The consequences of the softer framework are not theoretical. In 2023 the FDA issued a warning to healthcare providers* regarding the use of probiotic products in hospitalised preterm infants, following the death of an infant from sepsis in which genomic sequencing matched the causative organism to a strain in the administered product. No probiotic product is approved as a drug or biological product for use in infants.
Guideline positions reflect this unevenness. The American Gastroenterological Association reviewed probiotics across gastrointestinal indications* and issued strain specific recommendations for a narrow set of situations, including pouchitis and prevention of Clostridioides difficile infection in patients receiving antibiotics. For acute infectious gastroenteritis in children, the recommendation was against use, and for irritable bowel syndrome, the panel found insufficient evidence to recommend either way.
The practical implication is that generic probiotic advice is difficult to defend. Effects appear strain specific and indication specific, and a product supported for one purpose carries no transferable evidence for another.
Where Microbiome Treatments Are Being Applied Now
Recurrent Clostridioides difficile infection remains the only indication with licensed products. Two are approved in the United States: a rectally administered faecal microbiota preparation (fecal microbiota, live-jslm), licensed in November 2022*, and an orally administered purified bacterial spore preparation (faecal microbiota spores, live-brpk), licensed in April 2023*. Both are indicated for prevention of recurrence in adults following antibacterial treatment, not for treatment of an active first episode. In the pivotal phase 3 trial supporting the rectal product, treatment success was reported in 70.6 per cent of recipients compared with 57.5 per cent of those receiving placebo.
Oncology is the most closely watched area beyond infection. The composition of the gut microbiota has been associated with response to immune checkpoint inhibition, and early phase trials have tested whether modulating it can alter that response. A multicentre phase 1 trial* combined healthy donor faecal microbiota transplantation with nivolumab or pembrolizumab in twenty previously untreated patients with advanced melanoma. The objective response rate was 65 per cent, including four complete responses. No grade 3 adverse events were attributed to the transplantation itself, although a quarter of patients experienced grade 3 immune related events from the combination.
These are small, uncontrolled, early phase results in a selected population, and they should be read as hypothesis generating. Randomised data will determine whether the signal holds.
Inflammatory bowel disease has produced more cautionary results. Defined bacterial consortia developed for this indication advanced into phase 2 and were discontinued* after failing to meet primary endpoints, a reminder that mechanistic plausibility and early signal do not reliably predict controlled trial performance.
Evidence, Regulation and the Adoption Problem
Several obstacles are shaping how quickly microbiome based therapies reach practice. Trial design is difficult. Engraftment varies considerably between recipients, and durable colonisation appears to correlate with benefit, which means the delivered dose is not the biologically active dose. Placebo control is complicated when the intervention involves stool derived material. Endpoints in chronic inflammatory and metabolic disease are less clean than recurrence of a defined infection.
Manufacture and supply create their own constraints. Donor derived products require extensive screening and carry an irreducible transmission risk that must be managed through both screening and downstream processing. Defined consortia grown from characterised strains reduce batch variability but require the correct organisms to have been identified in the first place.
Adoption is also affected by logistics. A rectally administered product may fall outside the usual practice of the specialties most likely to prescribe it, and reimbursement pathways for a biological product delivered in an outpatient setting are not always established at launch.
Patient expectation is a separate problem. Direct to consumer microbiome testing and supplementation are widely marketed, generally without the evidence base implied by the marketing. Patients frequently arrive having spent substantially on interventions that no guideline supports, and the clinical conversation often begins by distinguishing a licensed therapy with a specific indication from a supplement sold on association data.
What the Next Phase May Look Like
Three directions are visible in current development.
Defined consortia are the most immediate. Products built from a fixed number of characterised, individually cultured strains avoid donor dependency and offer reproducible composition. One eight strain oral consortium for recurrent Clostridioides difficile infection reported a substantial reduction in recurrence risk against placebo in phase 2 and has progressed to phase 3. If defined consortia can match donor derived products, the supply and screening constraints largely disappear.
Engineered strains represent a different approach, using bacteria as delivery vehicles rather than as ecosystem replacements. Candidates designed to degrade specific metabolites have entered early clinical testing in inherited metabolic disease, where the therapeutic goal is a defined enzymatic function rather than restoration of a community.
Microbiota as a modifier of other therapies may prove more clinically significant than microbiota as a therapy in itself. Antimicrobial exposure around the initiation of immune checkpoint inhibition has been associated with poorer outcomes, and stewardship in that window is available now, without waiting for licensed products.
The realistic expectation is incremental: two licensed products in one indication, several instructive failures, and methodological problems that are being addressed rather than solved.
The useful distinction is between a licensed microbiome therapy with an approved indication and a supplement sold on association data. For recurrent Clostridioides difficile infection there are now prescribable options with trial evidence behind them. Elsewhere, including oncology, the data remain early, and the honest position is that the mechanism is plausible and benefit unproven.
Are microbiome questions coming up in your consultations, and what has been most useful in answering them? Share your experience in the comments section below.
Patient expectation is a separate problem. Direct to consumer microbiome testing and supplementation are widely marketed, generally without the evidence base implied by the marketing. Patients frequently arrive having spent substantially on interventions that no guideline supports, and the clinical conversation often begins by distinguishing a licensed therapy with a specific indication from a supplement sold on association data.
