Fecal Microbiota Transplant (FMT)

Gut · §61

For someone stuck in a cycle of recurrent C. difficile — a gut infection that keeps coming back after antibiotics — a transplant of stool from a healthy screened donor stops the diarrhoea within days and keeps it stopped in roughly 85–95% of cases. That's a higher single-treatment cure rate than any antibiotic ever managed, and it works by an almost embarrassingly simple idea: put back the gut bacteria the antibiotics killed. For other conditions where stool transplant gets tried — ulcerative colitis, IBS, autism, obesity — the evidence is still too thin to call it standard medicine.

Respond · As-needed Evidence Strong Chapter Gut

Stuck in a C. diff recurrence cycle? Ask your doctor about fecal transplant after the second episode, push for it after the third. It's a single procedure or a three-day pill course, it's covered by most insurance for that use, and the odds you walk out with a normal gut again sit in the high 80s. For anything else, slow down. The headlines about microbiome miracle cures are mostly small trials with mixed results, and a handful of people have died when donor screening missed dangerous bacteria.

What broad-spectrum antibiotics do to your gut is take a wrecking ball through it. Most of the bacteria living in a healthy colon aren't C. difficile's rivals — they're its police. They produce specific bile-acid byproducts that keep C. diff spores from waking up, they outcompete it for food, and they cover the colon wall so it has nowhere to anchor. Strip that community with vancomycin and you remove the police force. C. diff, whose spores survive antibiotics easily, walks right back in. That's the recurrence cycle: every new antibiotic course resolves the immediate infection and re-enables the next one.

Fecal transplant interrupts the loop. Stool from a healthy donor carries the full microbial community — hundreds of species, including the specific ones that produce the protective bile acids and crowd C. diff out. Within days of receiving it, the recipient's colon goes from "no police" to "normal police force," and the infection has nowhere to grow. The transplanted community is a mosaic of donor strains and whatever survived in the recipient — engraftment is partial, but enough Ianiro 2022.

How sure are we, and about what

For recurrent C. difficile, this isn't fringe medicine anymore. The trial that broke it open was so lopsided the safety board stopped it early.

Replications followed and have been consistent. A Boston double-blind trial showed 91% cure with donor stool against 63% with the patient's own stool as control Kelly 2016. A Canadian trial showed pill capsules work as well as colonoscopy — 96% cure in both arms Kao 2017. A Danish head-to-head pitted transplant against the two best antibiotics and the transplant won both matchups Hvas 2019. The largest real-world dataset — over 5,000 procedures across 681 US hospitals — reported 81% cure outside trial conditions.

In late 2022 and early 2023 the FDA approved the first two standardised products. Rebyota is a single rectal enema of pre-screened donor stool Khanna 2022; Vowst is a three-day oral course of purified spores Feuerstadt 2022. Both are indicated for preventing further recurrence after standard antibiotics; both clear the bar against placebo in their pivotal trials, though with somewhat lower cure rates than conventional colonoscopic transplant.

For other conditions the picture is murkier. Intensive multidonor stool given for ulcerative colitis roughly doubles 8-week remission over placebo across four randomised trials Paramsothy 2017 Costello 2019, but durability past three months is poorly studied. IBS trials are split — stool delivered by colonoscope helps, oral capsules have sometimes made symptoms worse El-Salhy 2020. Autism trials are small, mostly single-centre, mostly Chinese Kang 2019. Obesity trials show transient improvements in insulin sensitivity but no weight loss Vrieze 2012 Allegretti 2020. A small hepatic encephalopathy trial in cirrhotic patients showed real cognitive benefit Bajaj 2017. The 2024 American Gastroenterological Association guideline says yes for recurrent C. diff, not yet for anything else AGA 2024.

How it actually happens

Four roads to the same destination. Which one you take depends on what your hospital offers, what your insurance covers, and what your gastroenterologist recommends.

Conventional transplant via colonoscopy. Highest single-treatment cure rate, mid-90s percent. Finish your antibiotic course, wait one to three days, do a bowel prep the night before, and the gastroenterologist delivers about 250 mL of donor stool suspension to your colon during a standard colonoscopy. Same-day discharge.
Conventional transplant via capsules. Cure rate matches colonoscopy in the head-to-head trial. Frozen or freeze-dried stool packed into a few dozen acid-resistant capsules, swallowed across one or two days. No scope, no sedation.
Rebyota. One pre-packaged enema, given once in clinic, 24–72 hours after your last antibiotic dose. No bowel prep.
Vowst. Four capsules daily for three days in a row, preceded by a one-day magnesium-citrate bowel prep, starting two to four days after antibiotics end.

Resume a normal diet the same day. The one rule that matters afterward: avoid further non-essential antibiotics for at least eight weeks. Antibiotics will undo the transplant.

Where this can go wrong

The risk that matters is donor pathogen transmission. Modern screened stool — whether from a stool bank, a hospital lab, or an FDA-approved product — is tested for HIV, hepatitis, parasites, drug-resistant bacteria, and a long list of stool pathogens. But the screen is not perfect.

The two FDA-approved products are licensed for adults 18 and over only; children, severely immunocompromised patients, and people with the fulminant form of C. diff (toxic megacolon, ICU-level illness) go through individual physician orders or expanded-access protocols. Pregnancy is excluded from all major trials; case reports describe successful use under careful consent for severe recurrent C. diff in pregnancy. Active gut bleeding, suspected toxic megacolon, and recent abdominal surgery are situational reasons to wait. Do not try this at home with an unscreened household donor — the screen is what makes the procedure safe.

What this isn't

Four things people get wrong. First, stool transplant is for recurrent C. diff, not first-episode. If this is your first run, antibiotics — vancomycin or fidaxomicin — remain the right answer. Transplant comes onto the table after the second recurrence and definitely belongs there by the third Kelly 2021. Second, it's not a microbiome cure-all. Many of the clinics — most outside US jurisdiction — selling stool transplant for IBS, autism, autoimmune disease, depression, or obesity are running well ahead of the evidence. Some of those uses have honest early signals; none has the trial weight that C. diff does. Third, probiotic supplements aren't a small version of this. The lactobacillus and saccharomyces strains in capsules at the pharmacy don't reconstitute the diverse anaerobic communities that produce the protective bile acids and crowd out C. diff — they're a different intervention entirely. Fourth, a "do-it-yourself" transplant with a household donor is dangerous in a way the procedure isn't. The 2019 death came from a screened donor whose screen missed one organism; unscreened donor stool can carry HIV, hepatitis B and C, parasites, drug-resistant bacteria, and a long tail of pathogens you would not knowingly accept.

What else works

For a first episode of C. diff, oral fidaxomicin or vancomycin clears the infection in 80–90% of patients — antibiotics are still the right call. For a first recurrence, fidaxomicin slightly outperforms vancomycin on preventing the next one. Bezlotoxumab — a one-time IV antibody against the C. diff toxin — added on to standard antibiotics cuts 12-week recurrence by roughly ten percentage points. By the second or third recurrence, fecal transplant or one of the standardised products is the highest-cure option Hvas 2019 Johnson 2021.

Among the transplant options themselves, no head-to-head trial yet compares Rebyota, Vowst, and conventional stool. Indirect comparison favours conventional colonoscopic transplant on single-treatment efficacy (mid-90s percent versus roughly 70%), but the standardised products are easier to access, regulated, and free from the logistics of sourcing screened stool. A reasonable rule of thumb: if your hospital has an established conventional transplant programme, it usually wins on cure rate; if not, the FDA-approved products are the practical default.

Cost and access in the US

Three price points. Conventional transplant material from a stool bank runs roughly $500–$1,700 plus the colonoscopy or office-visit cost (another $1,000–$3,000). Rebyota lists at roughly $9,000 per enema. Vowst lists at roughly $19,680 per three-day capsule course. Most US private insurers and Medicare Part B now cover the FDA-approved products for confirmed recurrent C. diff; conventional transplant coverage is patchier and often needs prior authorisation.

One quirk of access worth knowing: the nonprofit stool bank OpenBiome — which supplied tens of thousands of preparations to over 1,300 US hospitals between 2013 and 2022 — was forced into near-shutdown by an FDA stool-bank rule and now operates as a foundation. Conventional stool is still available through hospital-sourced programmes and physician-initiated requests, particularly for patients who can't access the approved products or have failed them. For children and severely immunocompromised patients, who are excluded from the approved products' labels, individual hospital programmes are usually the only option.

What life is like without it

For someone with recurrent C. diff, life shrinks to bathroom geography. Six to ten loose stools a day. Cramps that wake you. Weight comes off. The job becomes hard to hold. The social life is the first thing to go — you stop accepting dinner invitations because you can't promise you'll get through them, and after a while people stop asking. Each recurrence carries a 40–60% chance of yet another, and for older or already-sick patients each acute episode carries an 8–25% chance of dying within thirty days — not from the infection elegantly, but from the cascade of dehydration, electrolyte loss, hospitalisation, and the things that go wrong in hospitals Johnson 2021.

The pre-transplant path was months of progressively longer vancomycin tapers, the rebuilding-after-each-relapse rhythm, with no clear exit. People disappear from their lives. The neighbour stops asking how you're doing because the answer is always the same. The grandkids' visits get put off again. The version of you that planned trips and showed up for things becomes the version negotiating with a body that won't stop hurting.

What changes when it works

Days, not weeks. The diarrhoea typically stops within 48–72 hours of the transplant. The cramping fades over the next few days. Appetite comes back by the end of the first week. By two weeks the energy floor starts to lift and you sleep through the night again.

By a month, most patients are back at work and the people around them stop asking if they're okay. By three months the recurrence risk has fallen from "almost certain" to "unlikely" — about 85–95% of single-transplant recipients stay cured at the 8-week endpoint and remain symptom-free at the one- and two-year marks Kelly 2016. For the ~10–15% whose C. diff does come back, a second transplant salvages the majority Kelly 2021. The honest framing: this isn't a marginal improvement on antibiotics. It's the difference between a chronic illness and a finished one.

A 70-year detour

The first description in modern Western medicine is from 1958: Ben Eiseman and a team of Denver surgeons treated four critically ill patients with fulminant pseudomembranous colitis — before anyone knew C. difficile was the cause — by giving them stool enemas. All four recovered Eiseman 1958. The technique then sat almost unused for half a century, dismissed as too unpleasant to take seriously. Australian gastroenterologist Thomas Borody began using it for refractory bowel disease in the 1980s. The arrival of a hypervirulent C. difficile strain in the 2000s, paired with widespread broad-spectrum antibiotic use, drove a recurrence epidemic that conventional medicine couldn't keep up with. Van Nood's 2013 trial broke the dam van Nood 2013. The first two FDA-approved standardised products arrived a decade later Khanna 2022 Feuerstadt 2022.

Adjacent topics

If this entry sent you down the microbiome rabbit hole, a few places worth your time: the antibiotic-stewardship case for using narrower-spectrum drugs first (the upstream lever on C. diff incidence); fidaxomicin as the first-recurrence drug of choice; how everyday diet shapes the gut community day to day; and — for the curious — whether minimally-defined bacterial cocktails can replace whole-stool transplant for indications beyond C. diff.

Related in the handbook

Alternatives

— FMT and these newer FDA-approved live biotherapeutics are the same idea; the products are the regulated version.

— FMT works by replacing the gut bacteria antibiotics wiped out — the extreme end of microbiome recovery.
— Recurrent C. diff is a leading cause of antibiotic-driven chronic diarrhea - when the cycle won't break, a stool transplant usually ends it.
— FMT is the industrial-strength version of the reseed-the-gut idea that probiotics gesture at.

Substance + claimed effects

Fecal microbiota transplant (FMT) is the introduction of a screened donor's stool — processed into liquid suspension, frozen slurry, or oral capsules — into a recipient's gastrointestinal tract via colonoscopy, enema, nasoduodenal tube, or oral capsule, with the goal of reconstituting a depleted or pathologically altered gut microbial community. The single overwhelmingly supported indication is the prevention of further episodes of recurrent Clostridioides difficile infection (rCDI) in adults who have failed standard antibiotic therapy; pivotal randomised trials and pragmatic registries place cure rates between roughly 70% and 96% depending on route, dose, and donor preparation van Nood 2013 Kelly 2016 Kao 2017. Beyond rCDI, FMT has been investigated for ulcerative colitis (moderate signal), Crohn's disease (mixed/weak), irritable bowel syndrome (heterogeneous, route-dependent), hepatic encephalopathy (small but promising trials), metabolic syndrome and obesity (transient metabolic effects, no durable weight loss), autism spectrum disorder (small early-phase trials), and graft-versus-host disease decolonisation of multidrug-resistant organisms (early). This entry covers the substance and its full spectrum of meaningful consequences — primarily the cure of rCDI, with secondary coverage of the wider investigational landscape and the credibility of those claims.

Evidence by addressing question

mechanism

Recurrent CDI is fundamentally a disease of microbial community collapse. Broad-spectrum antibiotics destroy the commensal anaerobes (especially Bacteroidetes and Clostridia clusters IV and XIVa) that normally suppress C. difficile through several mechanisms: competition for nutrients (succinate, sialic acid, primary bile acids), production of secondary bile acids (deoxycholate, lithocholate) by 7α-dehydroxylating Clostridia that inhibit C. difficile germination and vegetative growth, and saturation of mucosal binding sites van Nood 2013. Transplanted donor microbiota re-establish this ecological resistance — termed "colonisation resistance" — within days to weeks. Strain-tracking metagenomic studies show that recipient communities post-FMT are typically a mosaic of donor and pre-existing recipient strains, with engraftment patterns predictable from baseline community state and donor–recipient compatibility Ianiro 2022. For non-CDI indications the mechanism story is fuzzier: in ulcerative colitis it is presumed to involve dampening of pro-inflammatory community signals and restoration of short-chain fatty acid production; in hepatic encephalopathy, reduction of ammoniagenic and urease-positive taxa; in metabolic syndrome, transfer of bile-acid-modifying and SCFA-producing communities that improve insulin signalling Vrieze 2012. None of these non-CDI mechanisms is as cleanly demonstrated as the colonisation-resistance story.

evidence

Recurrent CDI — settled. The landmark trial is van Nood et al. (NEJM 2013): 43 patients with rCDI randomised to a short vancomycin course + bowel lavage + nasoduodenal donor stool infusion vs vancomycin alone vs vancomycin + lavage. The trial was stopped early for efficacy: 81% cure with a single FMT (94% after a second), vs 31% with vancomycin alone and 23% with vancomycin + lavage (P<0.001) van Nood 2013. Subsequent RCTs replicated this: Cammarota (90% colonoscopic FMT vs 26% vancomycin), Kelly et al. 2016 (91% donor FMT vs 63% autologous control, double-blinded) Kelly 2016, Hvas et al. 2019 (FMT 71% vs fidaxomicin 33% vs vancomycin 19% at 8 weeks in a head-to-head Danish RCT) Hvas 2019. Youngster et al. JAMA 2014 showed 70–90% cure with oral frozen capsules from unrelated donors Youngster 2014. Kao et al. JAMA 2017 demonstrated noninferiority of capsules to colonoscopy: 96.2% cure in both arms at 12 weeks (n=116) Kao 2017. OpenBiome's real-world registry of 5,344 patients across 681 centres reported 81% cure. Pivotal trials of the two FDA-approved standardised products: Khanna et al. PUNCH CD3 of RBX2660/Rebyota delivered as a single rectal enema, 70.6% success at 8 weeks vs 57.5% placebo (Bayesian primary analysis) Khanna 2022; Feuerstadt et al. NEJM 2022 of SER-109/Vowst (purified Firmicutes spores, oral capsules ×4 days), 12% recurrence vs 40% placebo through 8 weeks (relative risk 0.32) Feuerstadt 2022.

Ulcerative colitis — moderate, route- and intensity-dependent. Four placebo-controlled RCTs (Moayyedi 2015, Rossen 2015, Paramsothy 2017, Costello 2019) collectively show that intensive multidonor FMT roughly doubles 8-week remission rates over placebo — Paramsothy's Lancet trial reported 27% steroid-free clinical and endoscopic remission vs 8% placebo with 5 enemas/week for 8 weeks Paramsothy 2017; Costello reported 32% vs 9% with anaerobically-prepared multidonor FMT delivered colonoscopically + enema Costello 2019. Pooled meta-analyses across 13–14 RCTs (~580–600 patients) confirm RR ≈1.7 for clinical remission, with effect strongest for oral and multidonor preparations. Durability beyond 8–12 weeks remains poorly characterised; the AGA does not recommend FMT for UC outside clinical trials AGA 2024.

IBS — weak and contradictory. Meta-analyses of 7–12 RCTs are split: El-Salhy's single-donor "super-donor" colonoscopic trial showed dose-responsive symptom improvement and durability to 3 years El-Salhy 2020, but oral capsule trials (Halkjær 2018, Aroniadis 2019) showed FMT inferior to placebo. Pooled effects do not reach statistical significance; GRADE quality is very low.

Hepatic encephalopathy — small but consistent. Bajaj et al. (Hepatology 2017) randomised 20 cirrhotic patients with recurrent HE to FMT enema from a rationally-selected donor vs standard care; FMT recipients had significantly fewer HE episodes and hospitalisations over a year, with improved cognition Bajaj 2017. Capsule follow-up replicated the signal. Sample sizes remain <100 across all HE trials.

Obesity / metabolic syndrome — transient metabolic effects, no weight change. Vrieze et al. (Gastroenterology 2012) showed that lean-donor microbiota improved peripheral insulin sensitivity at 6 weeks in 18 men with metabolic syndrome, with effect waning by 18 weeks Vrieze 2012. Allegretti's obesity capsule trial (CGH 2020) showed engraftment but no weight or BMI change at 12 weeks Allegretti 2020. The Finnish bariatric-FMT RCT showed no weight-loss advantage.

Autism — small early trials, larger ones underway. Kang et al.'s open-label Microbiota Transfer Therapy in 18 children with ASD showed 47% reduction in Childhood Autism Rating Scale scores sustained at 2-year follow-up Kang 2019. Wan et al. (2024) reported a randomised double-blind placebo-controlled trial in 103 ASD children showing improvement in Vineland social-domain scores. Meta-analysis of five trials reports weighted mean difference of −14.96 on the Autism Behaviour Checklist favouring FMT. All trials small, several open-label, mostly single-centre.

Crohn's disease — too thin to weigh. Small open-label series and two underpowered RCTs; no clear signal. AGA explicitly does not recommend AGA 2024.

protocol

Three pathways for rCDI in 2026: (1) FDA-approved Rebyota (fecal microbiota, live-jslm) — single rectal enema (~150 mL) delivered 24–72 hours after completing standard antibiotic course, no bowel prep required Khanna 2022. (2) FDA-approved Vowst (fecal microbiota spores, live-brpk) — four oral capsules daily for three consecutive days, starting 2–4 days after antibiotic completion, preceded by magnesium citrate bowel prep Feuerstadt 2022. (3) Conventional FMT via colonoscopy (preferred — highest single-treatment efficacy ~94%), oral capsules (multiple capsules of frozen/lyophilised stool from a stool bank such as OpenBiome — comparable efficacy, lower procedural risk), or enema/nasoduodenal (less common). Standard sequence: complete vancomycin or fidaxomicin course; wash out 24–72 hours; deliver FMT; resume normal diet immediately. Repeat FMT for failures has 80–90% cumulative cure Kelly 2021. Donor stool from screened banks is the practical default; directed donors (typically a household contact) carry additional individual screening burden.

contraindications

Severe immunosuppression (active chemotherapy, post–solid-organ transplant on induction immunosuppression, CAR-T, advanced HIV) is the principal warning, on the basis of the 2019 sentinel event in which two immunocompromised adults developed bacteraemia with ESBL-producing E. coli traced to a single unscreened donor; one died DeFilipp 2019. The FDA mandated MDRO donor screening as a result FDA 2019. Subsequent alerts addressed STEC and EPEC transmissions from a US stool bank. Other relative contraindications: ileus or toxic megacolon (procedural risk), active GI bleeding, pregnancy (precautionary), severe food allergies matching donor diet, recent abdominal surgery. The two FDA-approved products are licensed only for immunocompetent adults ≥18; paediatric, severely immunocompromised, and fulminant-CDI patients require conventional FMT under physician judgement or expanded-access INDs AGA 2024.

misconceptions

Three widespread errors: (a) "FMT cures C. difficile in general" — incorrect; FMT is for recurrent or refractory CDI after antibiotic failure, not first-episode CDI (where antibiotics remain first-line; an LFS trial showed FMT was comparable to vancomycin for primary CDI but is not yet guideline-endorsed for that indication). (b) "FMT cures IBS / autism / depression" — premature; signals exist but are small, route-dependent, and unreplicated at scale. (c) "Do it yourself with a household donor" — strongly discouraged outside physician oversight: the death case in DeFilipp 2019 came from a screened donor whose screening missed MDROs, and unscreened household FMTs carry transmission risk for HIV, hepatitis B/C, syphilis, norovirus, EBV/CMV, and parasites DeFilipp 2019. (d) "Probiotics are equivalent to FMT" — incorrect; probiotic species (Lactobacillus, Saccharomyces) do not reconstitute the diverse anaerobic communities that mediate colonisation resistance.

practicalities

Cost spread is wide. Conventional stool-bank FMT runs roughly $500–$1,700 for the inoculum plus the procedural cost (colonoscopy ~$1,000–$3,000); Rebyota lists at ~$9,000 per single-dose enema; Vowst at ~$17,500–$19,680 per 12-capsule course. Most US private insurers and Medicare Part B now cover Rebyota under HCPCS J1440 paired with the FMT procedure code 0780T; Vowst coverage is improving but uneven; conventional FMT coverage is patchier and often requires prior authorisation. OpenBiome — the dominant US stool bank that supplied ~68,000 preparations to 1,300+ hospitals since 2013 — was forced into near-shutdown by FDA 2022 enforcement and now operates as a foundation; physicians can still source under enforcement discretion for confirmed rCDI patients who have failed standard therapy or the approved products. Pediatric and immunocompromised patients increasingly rely on individual hospital-sourced FMT or expanded-access INDs.

history

The use of human faeces as therapy is ancient — 4th-century Chinese physician Ge Hong described "yellow soup" oral suspensions for severe diarrhoea, and 17th-century European veterinary medicine practised "transfaunation" for ruminants. Modern Western medical use begins with Eiseman et al. (Surgery, 1958), who treated four patients with fulminant pseudomembranous colitis — before C. difficile was identified as the cause — using fecal enemas, with rapid recovery Eiseman 1958. The field remained niche until Thomas Borody's Australian work in the 1980s–2000s and the explosion of CDI cases driven by the hypervirulent NAP1/BI/027 strain in the 2000s. Van Nood's 2013 NEJM RCT — the field's first — was the inflection point van Nood 2013. The FDA classified FMT as an investigational biologic in 2012, then issued enforcement discretion for rCDI in 2013. Standardised products arrived in 2022–23: Rebyota (Nov 2022) and Vowst (April 2023) Khanna 2022 Feuerstadt 2022. The 2019 ESBL E. coli death reshaped donor screening permanently DeFilipp 2019.

stakes

For the rCDI patient stuck in a recurrence cycle: each episode of recurrent CDI carries roughly a 40–60% probability of further recurrence on standard antibiotic re-treatment, and chronic recurrent disease is associated with malnutrition, dehydration, loss of work, social isolation around bathroom proximity, and — in older or comorbid patients — 30-day mortality of 8–25% per episode. Annual US rCDI burden is ~75,000–175,000 cases. Without microbiota restoration, the disease often resolves only after months of progressively longer vancomycin tapers, with substantial cumulative cost and morbidity. For the much broader research landscape (UC, IBS, metabolic, autism), the stake is opportunity cost — patients spending out-of-pocket for unregulated FMT at clinics abroad chasing unproven signals.

payoff

For rCDI: the dominant clinical reality is resolution of diarrhoea within 1–3 days post-FMT, with most patients reporting normalisation of stool form, return of appetite, and resumption of normal activity within a week. Eight-week sustained cure rates of 80–95% with single FMT (and ~90% cumulative with a repeat FMT in initial failures) Kelly 2021. For successfully treated patients, the recurrence curve flattens: the typical post-FMT rCDI patient returns to baseline microbiome diversity within months and remains symptom-free at 1- and 2-year follow-up at high rates Kelly 2016. Felt-experience scale: from multiple-month bathroom-bound illness to "the diarrhoea stops within 48 hours and stays stopped." For non-CDI indications the payoff narrative remains speculative.

alternatives

For rCDI: standard-of-care antibiotics (fidaxomicin preferred over vancomycin for recurrence prevention per IDSA 2021), vancomycin tapered/pulsed regimens, bezlotoxumab (a monoclonal antibody against C. difficile toxin B; reduces recurrence ~10 percentage points), and the two FDA-approved live biotherapeutics (Rebyota, Vowst). Head-to-head Hvas 2019 showed FMT superior to fidaxomicin and vancomycin in recurrent disease Hvas 2019. Among microbiota-based options, no head-to-head RCT exists comparing Rebyota vs Vowst vs conventional colonoscopic FMT; indirect comparisons favour conventional FMT in single-treatment efficacy but the standardised products avoid sourcing burden and have regulatory backing.

failure-modes

Common reasons FMT fails or recurrence occurs: (1) continued non-CDI antibiotic exposure after FMT, which re-disrupts the transplanted community; (2) ongoing PPI use, which is associated with both initial CDI and post-FMT recurrence; (3) older age and haemodialysis, identified as independent predictors of failure in registry data; (4) inadequate dose or fragmented engraftment (most relevant for capsule routes with insufficient stool mass); (5) misdiagnosis — toxin-negative diarrhoea attributed to CDI that is actually post-infectious IBS, which FMT does not reliably treat Kelly 2021. Repeat FMT salvages the majority of single-FMT failures.

out-of-scope

This entry does not cover: gut microbiome composition as a longevity/metabolism lever in general (separate entry candidate); probiotics; prebiotic / dietary fibre interventions; antibiotic stewardship for primary CDI prevention; bezlotoxumab as a stand-alone CDI relapse-prevention strategy.

The credibility range

Optimist case

For rCDI, the optimist case is overwhelming and uncontested: FMT is among the highest-magnitude single interventions in modern infectious disease, with cure rates that exceed those of any conventional antibiotic regimen for recurrent disease, sustained at 1–2 years, with a clean and well-understood mechanism (colonisation resistance via secondary bile acids and ecological niche occupation) and a favourable safety record across >10,000 documented procedures when proper donor screening is in place. Beyond rCDI, the optimist points to (a) consistent UC remission signals across multiple RCTs with intensified protocols and multidonor preparations Paramsothy 2017 Costello 2019, (b) the El-Salhy IBS data showing dose-responsive 3-year durability with single-donor super-donor FMT El-Salhy 2020, (c) the Kang autism trial's sustained 2-year improvements Kang 2019, (d) Bajaj's hepatic encephalopathy reduction Bajaj 2017, and (e) the broader argument that the gut microbiome is a real, evolutionarily ancient organ whose dysfunction plausibly contributes to a wider spectrum of chronic disease than the literature has yet formally documented.

Skeptic case

The skeptic concedes the rCDI evidence entirely and pushes back hard on everything else. (a) Non-CDI RCTs are small, single-centre, often unblinded, with substantial placebo responses and inconsistent endpoints; effect sizes shrink with rigour and larger samples. (b) Engraftment is heterogeneous and unpredictable; "FMT works" reduces to "sometimes some donor strains take, sometimes they don't, and we cannot prospectively match" Ianiro 2022. (c) The death from ESBL E. coli transmission and the STEC/EPEC outbreaks demonstrate that even well-intentioned screening misses pathogens, and as we extend FMT to immunocompromised or metabolically-targeted use the risk-benefit calculation tightens DeFilipp 2019. (d) For IBS, autism, and obesity, the most reasonable read of pooled data is that there is no robust benefit and that positive trials are heterogeneity-driven. (e) Commercial pressure (Seres, Ferring, hundreds of "stem cell tourism"–style FMT clinics) is now substantial and will likely produce overreach.

Author's call

For recurrent CDI: settled. Highest possible evidence rating; FMT or one of the two approved live biotherapeutics is the standard of care after the second recurrence and should be considered after the first. For ulcerative colitis: real but modest signal, not yet ready for routine practice — AGA 2024 explicitly does not recommend, but a thoughtful patient with refractory UC discussing intensive multidonor FMT with a specialist in a clinical-trial context is not pursuing nothing. For everything else (IBS, autism, obesity, metabolic syndrome, hepatic encephalopathy outside cirrhosis trials): evidence is too thin to recommend; reasonable to pursue inside formal trials. The article lands on rCDI as the primary frame, with secondary acknowledgement of the wider research landscape and explicit calibration of how unsettled it remains.

Stakeholder + incentive map

Commercial: Ferring Pharmaceuticals (Rebyota — ~$9,000/dose), Seres Therapeutics with Nestlé Health Science (Vowst — ~$19,680/course). Both have substantial post-approval marketing budgets. Pre-approval, dozens of "FMT clinics" (some legitimate, many predatory) operated abroad targeting IBS, autism, autoimmune patients.
Nonprofit / public: OpenBiome — the major US stool bank, near-shut-down by 2022 FDA enforcement, now operating as a foundation. Academic medical centres (MGH, Mayo, Brown, Calgary, Amsterdam) run in-house programmes.
Professional / guidelines: AGA (2024 guideline endorses for rCDI, not for IBD/IBS) AGA 2024; ACG (2021, FMT for ≥2 recurrences) Kelly 2021; IDSA/SHEA (2021, FMT after ≥3 episodes as expert opinion) Johnson 2021; ESCMID European guidance broadly aligned.
Regulatory: FDA — investigational status since 2012, enforcement discretion for rCDI since 2013, donor-screening mandate since 2019, stool-bank restriction since 2022, two approved products since 2022–23 FDA 2019.
Patient communities: Strong online communities (Peggy Lillis Foundation, C diff Foundation) advocating for broader access; pediatric and immunocompromised patient families particularly active on access issues. Autism FMT advocacy is loud, contested, and commercially exploited.
Skeptic / counter: Antibiotic-stewardship community (focused on primary CDI prevention rather than FMT salvage); microbiome researchers cautious about extrapolation beyond CDI; ethicists raising consent issues for paediatric and ASD use.

Population variability

Age: Adults are the studied population. Paediatric data are smaller but consistent with adult cure rates; the FDA-approved products are licensed ≥18 only. Older adults (>75, dialysis-dependent) have somewhat lower cure rates and higher post-FMT recurrence in registries.
Immune status: Mildly-to-moderately immunocompromised adults can safely receive FMT with appropriate donor screening per AGA 2024 AGA 2024; severely immunocompromised (active induction post-transplant, CAR-T, advanced HIV) carry elevated transmission risk and require individualised review.
Pregnancy: Excluded from all major trials; case reports describe successful use under careful informed consent for severe rCDI in pregnancy.
Baseline microbiome state: Patients with prolonged or repeated antibiotic exposure have more depleted communities and may benefit more from FMT; engraftment is higher in more-depleted recipients Ianiro 2022.
Disease context: Fulminant CDI (toxic megacolon, ileus, ICU-level disease) is a special case — small case series support FMT as rescue, but evidence base is thin and the procedural setting is intensive.
Underlying IBD: rCDI patients with concurrent IBD have somewhat lower cure rates and higher post-FMT IBD flare rates (~15%); requires careful coordination with IBD specialist.

Knowledge gaps

Donor selection. Why do some donor stools produce 95% cure and others 70%? The "super-donor" phenomenon is real in retrospective data but no prospective screening criterion reliably predicts it.
Optimal preparation. Fresh vs frozen vs lyophilised, aerobic vs anaerobic processing, single vs pooled donor — answered for rCDI by pragmatic equivalence; unanswered for non-CDI indications where the effect size is small enough that processing details may matter.
Long-term outcomes. 10+ year follow-up data on FMT recipients are minimal. Whether transferring a donor microbiome carries long-term metabolic, immune, or oncologic consequences is not yet answerable.
Non-CDI mechanism. What specifically does FMT do in UC, IBS, HE, autism? Without mechanistic clarity, trial design remains shotgun.
Defined consortia. Whether minimal-defined bacterial consortia (e.g., Vowst's purified Firmicutes spores, or VE303) can replace whole-stool FMT across the indication spectrum. Vowst is the proof-of-concept that it works for CDI; replicating that simplification for UC or IBS is an active research frontier.
Standardised vs conventional comparison. No RCT yet comparing Rebyota, Vowst, and conventional FMT head-to-head; indirect comparison favours conventional in single-treatment efficacy but is confounded by patient selection.

Scope vs brief. The brief listed two areas: cure rates for recurrent C. difficile, and the limits of evidence for other indications. The article covers both. Recurrent C. diff gets the centre of gravity (dek, evidence, protocol, contraindications, payoff, stakes) because it's the one settled indication and the dominant clinical reality. Non-CDI indications (UC, IBS, autism, obesity, hepatic encephalopathy) are handled in the evidence section and the misconceptions section, calibrated to "honest signals, not standard care." This matches the brief's framing precisely; no narrowing.

Hard scoring calls.

longevity = 2: meaningful only for the older, comorbid rCDI subpopulation (8–25% per-episode 30-day mortality). For the median catalogue reader the longevity effect is small. Held at 2 to reflect real-but-narrow effect.
health_short_term = 5: justified by the magnitude of the rCDI cure (months-of-illness to symptom-free in days), even though the population is narrow. The dimension scores the substance's effect when it applies, not the population fraction.
cost_burden = 3: awkward call. Standardised products are five-figure list price (Vowst $19,680, Rebyota $9,000) but largely insurance-covered for rCDI; conventional FMT is <$3,000 all-in. Net out-of-pocket for most US patients with the approved indication is modest. Scored 3 (substantial) as a midpoint between the sticker-price reality and the typical insured experience.
controversy = 3: rCDI consensus is settled (0–1), but the broader landscape — non-CDI overreach, the 2022 stool-bank enforcement that nearly killed OpenBiome, paediatric/ASD ethics, conventional vs standardised product access — keeps the entry's overall controversy nontrivial.
action = respond, cadence = as-needed: this is a trigger-based therapy keyed to a specific clinical event (recurrent CDI), not a habit or screening. "Respond" + "as-needed" fits better than "decide" because once the trigger is met the choice to do FMT is largely settled by guidelines.

Contraindication tokens. Closed vocabulary doesn't include "immunocompromised", which is the real headline contraindication. pregnancy is the unambiguous one. autoimmune is included as the closest proxy for the immunosuppression population that needs extra screening — imperfect (FMT is studied as a treatment for some autoimmune conditions), but flagging this audience for caution matters more than the false-positive cost. The article's contraindications section gives the nuance.

Separate-entry candidates.

Recurrent C. difficile infection (the condition itself) — currently referenced only in the context of FMT; deserves its own entry covering antibiotic choice (fidaxomicin vs vancomycin), bezlotoxumab, isolation, and recurrence-prevention strategy across the full pathway.
Gut microbiome composition as a general health lever — diet, fibre, antibiotic-sparing, etc. Currently linked only by implication; warrants its own entry.
Antibiotic stewardship — the upstream lever on C. diff incidence and broader resistance. Flagged in out-of-scope.
Bezlotoxumab — could stand alone for the rCDI-prone patient considering one-time IV prophylaxis with the next antibiotic course.
Vowst / Rebyota individually — probably do not warrant separate entries; they belong nested under this one as treatment options.

Future links to wire up. recurrent-c-difficile, fidaxomicin, bezlotoxumab, gut-microbiome, antibiotic-stewardship, ulcerative-colitis, colonoscopy. Related field already names placeholders for the most-relevant.

Exclusions from the article body.

Detailed donor-screening protocols — too technical, irrelevant to the reader's decision; the article just notes that screening exists and the 2019 sentinel event reshaped it.
Detailed metagenomic engraftment data — referenced via Ianiro 2022 in mechanism but not unpacked; the reader doesn't need strain-tracking numbers to understand why FMT works.
Veterinary and ancient-history detail (Ge Hong, transfaunation) — cut for length; the 1958 Eiseman starting point is enough modern-medicine anchoring.
Non-US regulatory landscape (EMA, NICE) — the article is US-centric on cost/access; explicitly flagged as "in the US" in the practicalities section.

Rating difficulty worth flagging for the reviewer: the evidence dimension is unambiguously 5 for rCDI but would be 1–2 for any of the secondary indications. Scored 5 holistically because rCDI is the dominant clinical reality, and the article calibrates honestly elsewhere. If a future reviewer thinks this overweights, the call to revisit is whether to split into "FMT for recurrent C. diff" + a separate "FMT for other indications" — currently judged not worth the split.