Gut microbiome of patients with IBD, whether CD or UC, is characterized by reduced taxonomic diversity and phylumlevel decrease in Firmicutes and an increase in Proteobacteria abundance [2]. Even healthy individuals with IBD-susceptibility genes (high “IBD genetic risk score”) have differences in microbiota from those without these susceptibility genes, such as a decrease in Roseburia genus responsible for acetate to butyrate conversion, that may precede the clinical onset of IBD [3]. It is unclear whether IBD produces dysbiosis or dysbiosis leads to IBD, however, several microbiota-altering exposures such as widespread antibiotic use, exposure to antibiotics in the first year of life, and consumption of a diet rich in fats, polyunsaturated fatty acids, proteins and low in dietary fiber, have been associated with incident IBD [4-8]. Initial studies found an association between IBD and infection with Mycobacterium avium paratuberculosis or adherent-invasive Escherichia coli, however, the data was inconsistent and therapeutic trials to eradicate these infections have not altered disease course [9-12]. Firmicutes phyla seem to play a central role in the anti-inflammatory microbial profile of a healthy population. Key members of this phyla include Faecalibacterium and Roseburia both of which have been found to produce anti-inflammatory metabolites, are depleted in patients with IBD, are repleted by FMT, and repletion correlates with response to therapy [13]. The evidence of this association between IBD and specific microbiota changes (e.g., abundance of Bacteroides and depletion of Firmicutes) has been strengthened by their association with several key aspects: aggressive disease [14], occurrence of complications, postsurgical recurrence, or pouchitis following colectomy for UC [15], and response to therapy with biological agents including antitumor necrosis factor agents [16,17] or anti-integrin agents [18]. However, the information obtained from microbiota analysis is limited in generalizability across regions due to the influence of genetics, diet and environment on the microbiome. The gut microbiome of Chinese and Western (from the Human Microbiome Project) IBD patients have marked differences with different predominant species even within specific IBD types [19].

Stark differences in the microbiome of IBD and non-IBD patients contribute to functional changes in the metabolic milieu, or metabolome. The metabolome is considered to be the route of interaction between the immune cells and microbiome through the production of pro-inflammatory or anti-inflammatory metabolites. Short-chain fatty acids (SCFAs), case in point, including butyrate, propionate, and acetate, are derived from the anaerobic bacteria-mediated breakdown of dietary fiber and modulate host mucosal cells by regulating cell proliferation and immune response [20]. Two G protein-coupled receptors (GPR), GPR41 and GPR43, and regulatory T cells (Treg) cells are activated by SCFAs [21]. SCFAs promote the accumulation of Treg CD4 cells in the colonic mucosa and protect against colitis [22]. Similarly, CARD9 deficient mice are at an increased risk of developing CD due to reduced microbial production of tryptophan, and the risk is attenuated by supplementing with tryptophan metabolizing Lactobacillus species [23]. Metabolomic and metagenomic profiling of the IBD fecal samples indicate enrichment of certain metabolites (bile acids, sphingolipids) and depletion of others (tetrapyrroles, triacylglycerols), indicating increased activity of functions such as those allowing adaptation to oxidative stress in IBD [1]. The microbiome and metabolome profiles could differentiate IBD from non-IBD stool samples with an area under the curve of 0.92. Vich Vila et al. [24] analyzed fecal metabolome and metagenomics in 424 patients with IBD. They found that the fecal metabolites in IBD were characterized by lower levels of saccharolytic fermentation derivatives and increased metabolites from proteolytic fermentation. Higher sphingolipids, ethanolamine, and primary bile acids characterized the gut signature of patients with IBD versus controls and the ratio of sphingolipid-Lurobilin carried high accuracy (area under the receiver operating characteristic curve = 0.85) to differentiate the two. Importantly, their findings supported that gut microbiota composition is the main determinant of fecal metabolite content, compared with patient lifestyle, genetics, or clinical phenotypes, highlighting the centrality and distinct role of gut microbiome. Machine learning algorithms incorporating clinical, microbiome, and metabolome data, have been designed to predict response to anti-integrin therapy in IBD with high accuracy (area under the curve = 0.872) [18].

The fungal and viral components of the microbiota are important but have been neglected due to difficulty in sequencing techniques. Fungal mycobiome is usually sequenced using the “universal fungal barcode sequence” of the ribosomal RNA called internal transcribed spacer [25]. Ascomycota and Basidiomycota are the most prevalent taxa in the mucosa of healthy subjects. The abundance of the fungus Candida tropicalis is higher in patients with familial CD compared to non-affected relatives, and it positively correlated with levels of anti-Saccharomyces cerevisiae antibodies that are purported to be against Candida antigens [26]. Further, the abundance of C. tropicalis correlated with that of Serratia marcescens and E. coli, suggesting that these organisms interact in the gut and a biofilm formed from these 3 organisms was thicker than those of any single- and double-species biofilms along with morphological changes in hyphae formation when the 3 organisms coexisted. A probiotic combination of Saccharomyces, 3 bacterial species, and amylase was effective in preventing and treating this biofilm formation [27]. Another report from Japan found similar alpha diversity in patients with IBD and controls, but increased abundance of Candida in CD compared to healthy controls and UC [28]. Two studies linked different fungal genera (Debaryomyces and Malassezia) with mucosal inflammation in CD and demonstrated their pro-inflammatory potential in animal models [29,30]. Evidently mycobiome holds more relevance in CD than UC.

Virome signature interpretation is even more difficult–longitudinal analysis has shown that the virome is less stable and shows interpersonal variability in viral taxa, frequent fluctuation of taxa even within an individual, and vast fluctuations in the abundance of temporal viruses [31]. An in-depth analysis of free virions enriched intestinal preparations showed significant enhancement of Caudovirales bacteriophages in both CD and UC [32]. The mucosal virome is distinct from the fecal virome and houses certain phages that escape detection in stool samples [33]. The alpha and beta diversity of viromes is unable to differentiate IBD from controls, unlike the bacteriome.

The fecal microbiome is the usual source of microbiome information, being easily accessible and noninvasive, and allowing longitudinal sampling. However, there appear to be additional microbiome strata that line the mucosal lining, as well as the crypts. Recent advancements in single-cell and spatial transcriptomics have enabled us to characterize these niches [34]. The mucosa-associated microbiome and crypt-associated microbiomes are “deeper” and thus closer to the host epithelial lining and immune cells. They have been found to have distinct microbial profiles from the fecal microbiomes. The mucosa-associated microbiome is more “abnormal” and differentiates better from controls than the fecal microbiome, and has a uniform increase in Proteobacteria and reduction in Firmicutes regardless of IBD type and severity [35,36]. The mucin of the epithelial lining in UC has a lower degree of sulphation and correspondingly lower abundance of sulfated-mucin metabolizing species Desulfovibrio [37]. Early studies found that only a few crypts in IBD patients are colonized by bacteria unlike acute self-limiting colitis, however, these used fluorescent in situ hybridization techniques that have been since surpassed by laser-microdissection of tissues and 16S rRNA gene sequencing [38]. The composition of crypt-associated microbiota (CAM) in health, as well as disease, is dominated by aerobic members, which can be attributed to higher oxygen tension in the crypts [39]. Microbiome composition analysis showed mucosa-associated microbiome being dominated by members of phyla Firmicutes (45%), followed by Bacteroidetes (26%), Proteobacteria (16%), and Actinobacteria (6%), while CAM comprised predominantly of aerobic members of Actinobacteria (54%) and Proteobacteria (38%). Dysbiosis has been reported in salivary samples in IBD, and some strains of Streptococcus are shared in saliva and fecal samples from patients with IBD [40]. Atarashi et al. [41] elegantly demonstrated the potential pathological role of ectopic oral microbial species (Klebsiella) when they colonized the gut in gnotobiotic mice, and drive intestinal inflammation via the T_H1 cell pathway. They hypothesized that the oral cavity may serve as a reservoir for potential intestinal pathobionts.

There is a stringent screening of the donors for viral pathogens such as hepatitis A, B, and C viruses, human immunodeficiency virus, Ebstein-Barr virus, and cytomegalovirus, severe acute respiratory syndrome coronavirus 2, bacterial pathogens such as Clostridioides difficile, drug-resistant bacteria particularly Shigatoxin-producing E. coli and enteropathogenic E. coli using antimicrobial resistance genes, and parasitic infestations [48,49]. Appropriate screening is often time and resource intensive with around 90% of prospective donors failing screening, half of them due to lack of follow-up [50]. A complete list may be accessed through several consensus guidelines [51,52].

The choice between frozen versus fresh feces being used for FMT is an area of ongoing debate. Frozen feces can be stored at –80°C for up to 2 years, allowing the setup of “stool banks” at several FMT centers and ready access to clinicians for use. This allows testing of each sample prior to administration and it has been recommended for this safety factor [52]. On the other hand, freezing of feces is known to impact microbiome composition. It leads to a 4-fold reduction in living bacterial cells as detected by flow cytometry (70% to 15%), similar alpha diversity, and definite differences in beta diversity on principal coordinates analysis from fresh fecal samples [53]. Prolonged frozen storage leads to an increase in Firmicutes to Bacteroidetes ratio, as some Gram-negative species are depleted during storage and some hypothesize that trials in which frozen FMT was found to be effective also depend on this phenomena of selective depletion of administered Bacteroidetes. A network metaanalysis of 8 studies in patients with recurrent C. difficile infection (CDI) found similar efficacy in fresh versus frozen FMT (93% vs. 88%, P= 0.18) and concluded that the trend towards improved efficacy with fresh FMT was offset by safety, accessibility, and practicality of frozen FMT [54]. Similar findings have been observed in UC from meta-analyses of 22 studies where the efficacy of fresh (34.4%) versus frozen (46.8%) was similar [55]. However, an understanding of the biological effect of freezing on microbiome and its effect on the sustenance of response remains to be seen.

Washed FMT is a technique developed in Chinese FMT centers using an automated washing machine GenFMTer (FMT Medical, Nanjing, China). It includes additional steps of filtration, centrifugation and resuspension three times in an attempt to retain intact bacteria and remove fecal particles, parasite eggs and fungi. Chinese databases found a reduction in adverse events due to FMT from 38% to 12% upon switching from conventional FMT preparation practices [56]. Intraperitoneal injection of washed FMT into mice after increasing number of centrifugation cycles showed a reduction in viruses, and pro-inflammatory metabolites (leukotrienes, prostaglandin G2, corticosterone), and prevented death of these experimental mice [56]. It has been shown to not impair efficacy in either CD or UC, however, has not been validated in other countries [57]. Another technique fecal filtrate transplantation instead removes intact bacteria and only transfers bacterial debris, metabolites, and oligonucleotides [58]. It has been utilized successfully in CDI and is also being evaluated in UC in a clinical trial [59].

Anaerobic preparation of FMT involves processing frozen FMT under anaerobic conditions, at the steps of homogenization, freeze-thaw cycles, and the time between defecation till transplant preparation, with the benefit of preservation of obligate anaerobes from the donor. Costello et al. [60] found that all microbes associated with treatment response of FMT in UC were anaerobes (mostly obligate anaerobes). Obligate aerobes (such as Faecalibacterium prausnitzii and Bacteroides fragilis) tend to decrease with mixing during preparation but anaerobic preparation mitigates this decrease, while facultative anaerobes (such as Enterobacteriaceae, Streptococcus) proliferate during mixing [61]. Commercially available systems (such as anaerobic chamber Bactron IV Work Station, USA) use a nitrogen (85%) + hydrogen (10%) + carbon dioxide (5%) atmosphere to maintain anaerobic conditions. Chu et al. [62] used propidium monoazide-based sequencing to differentiate viable and dead bacterial cells and assess the effect of various preparation methods. They found that oxygen exposure degraded fecal microbial composition (particularly abundance of F. prausnitzii), while freeze-thaw cycles and lag time between donor defecation and transplant preparation had smaller effects. Another group found drastically lower viability (19%) in aerobic preparation under ambient air, compared to anaerobic preparation (50%) [63]. A shorter duration between defecation by donor and administration of FMT has been found in those who respond to FMT versus nonresponders (median 2.1 hours vs. 2.6 hours, P= 0.002) with a drastic reduction in response rate at time gap > 3 hours [64]. Consensus guidelines suggest the use of fresh stools within 6 hours, and frozen processed stools used within 6 hours of thawing, in an effort to preserve anaerobic species [51]. Thus, anaerobic preparation is superior in composition of beneficial bacteria and should be practiced consistently, but its direct impact on clinical outcomes remains to be established.

Pretreatment with antibiotics improves engraftment of donor FMT. Weekly sequencing of fecal microbiota during an 8-week course of FMT demonstrated that higher alpha diversity and donor engraftment strongly correlate with clinical response [65]. Additionally, “post-antibiotic pre-FMT” time point alpha diversity, i.e., residual diversity after antibiotic therapy, correlates well with both engraftment and subsequent clinical response. This is in contrast to the lack of independent efficacy of antibiotics or probiotics in typical non-fistulizing CD and UC [66]. Initial studies in mice compared pretreatment with antibiotics, colonic lavage or no pretreatment prior to FMT and found superiority of the antibiotic regimen. They demonstrated that this treatment allows colonization by the xenomicrobiota (from the donor human feces), by eliminating antibiotic-sensitive bacteria and generating niches for colonization [67]. Ishikawa et al. [68] reported using triple antibiotics (amoxicillin, metronidazole, and fosfomycin) prior to FMT, with the intent to eradicate harmful Bacteroidetes species. A systematic review of 28 articles (including 6 randomized trials), in which 11% of the patients were pre-administered antibiotics, found an improvement in response rate (82% vs. 58%) and remission rates (66% vs. 31%) with antibiotic pretreatment that was consistent with an increase in alpha-diversity and enrichment with SCFA producing anaerobes [44]. Higher baseline microbiota diversity is also linked with greater response rates, suggesting that the dysbiosis can potentially exceed the capability of modulation with FMT [69].

Bowel preparation (with agents like polyethylene glycol) is commonly used prior to FMT, regardless of whether FMT is administered via colonoscopy, nasoduodenal tube or gastroscope [51,70]. It aids cecal intubation, removal of vegetative cells, and allows engraftment of transplanted microbes [71]. The lavage impact does not alter the gut microbiota independently over long periods but may have short-term reduction in the total bacterial abundance of fecal microbiota as well as mucosa-associated microbiota [72-74]. No controlled studies have assessed its clinical impact on FMT in IBD but data from recurrent CDI suggest that poor bowel preparation is associated with failure of FMT [75].

Typically fecal matter is converted to a slurry and administered via either one of several routes: during colonoscopy, during gastroscopy, via nasogastric or nasojejunal tube, as enemas, or as oral ingestion of capsules. FMT administered via nasogastric tube was as effective as colonoscopy in treating recurrent CDI in a pilot randomized controlled trial (RCT) [76]. Similarly, administration through a gastroscope or colonoscope did not affect efficacy or adverse events in a trial of CD [77]. Meta-analyses have suggested higher response rates with colonoscopic instillation instead of upper gastrointestinal administration (31% in 137 patients vs. 8% in 42 patients) [78]. The only negative RCT on FMT in UC utilized the nasoduodenal route, suggesting that the lower gastrointestinal route (colonoscopic or rectal enema) would be preferred in UC [79]. The efficacy of oral capsules in both induction and maintenance of remission in mild to moderate UC was demonstrated in the LOTUS trial [80]. Oral capsules as FMT have thus far been limited to recurrent CDI but not IBD [81]. A large trial and subsequently non-randomized study proved non-inferiority of capsulized frozen FMT and colonoscopic administration in preventing recurrent CDI with roughly half the adverse event rate [82].

The dose of donor feces administered per session and cumulatively is probably one of the most heterogeneous aspects of the practice of FMT. As little as 30 g, and as much as 600 g, have been utilized in clinical trials of IBD. Zhao et al. [55] identified 275 g as the median dose used in the 29 studies they included in their meta-analysis. Studies using higher doses of donor feces ( > 275 g) were associated with higher response rates (52% vs. 30%) than those using lower doses. Similarly, subgroup analysis by Wei et al. [83] found benefit of FMT in induction of remission of UC only when the cumulative dose was higher than 300 g (risk ratio [RR], 1.85; 95% confidence interval [CI], 1.22–2.83) but not in lower doses. To achieve these doses, several sessions may be required, and there may be differences in optimum doses for induction versus maintenance of remission in IBD. A higher load (60 g vs. 30 g) of infused donor feces led to a greater reduction of irritable bowel syndrome (IBS) symptoms (89% vs. 76%) following FMT [84]. There are fallacies with using fecal weight as a “currency” and there is a need for developing ubiquitous and validated measures to assess the dose of active bacteria being administered.

Repeated sessions allow repeated exposures to the donor microbiome and allow an opportunity for donor microbial engraftment, which is correlated closely to clinical respons [44,85]. This effect is seen with repeat sessions independent of the disease and mimics dose escalation of conventional therapy. Response rates for repeat FMT in recurrent CDI were superior (91% vs. 84%) with a consequently lower number needed to treat (1.5 vs. 2.9) in a large meta-analysis [86]. In a small study of patients with IBS, those who did not improve with FMT initially had an excellent response following a repeat session with double dose of transplant feces [87].

A meta-analysis by Mocanu et al. [44] found higher pooled response rates (70% vs. 53%) and remission rates (43% vs. 30%) to FMT in those given repeated doses, as compared to a single dose. Regimens applied to patients with IBD have wide variability, and the following have been used in various clinical studies: daily enemas [88], alternate day enemas or nasojejunal or colonoscopy [89], initial colonoscopy followed by twice weekly enemas [60], weekly colonoscopies [90,91], 2 weekly colonoscopies [69], and some have utilized regimens similar to those for biological agents with an induction and maintenance phase [92]. Ren et al. [85] evaluated “low intensity” FMT sessions for two sessions (0 and 2 months) with a single super donor and found remission rates of 21% and 55% after the first and second FMT sessions. No comparison is available between the different regimens due to small number of available studies, and differences in FMT composition and patient populations receiving the therapy.

Importantly, there is no single optimum way to perform FMT and each institute should establish standard operating procedures to minimize heterogeneity in technique. Centers catering to patients with IBD and establishing FMT in their practice should set up stool banks to provide a steady supply of frozen stool for FMT.

An easy workaround for enriching the microbial diversity of donor fecal samples is to pool multiple donors. Pooling of fecal samples proposes to provide a wider spectrum of microbes such that “deficiencies” of the dysbiotic recipient microbiome can be resolved. Mathematical modeling based on the RCT by Moayyedi et al. [90] suggested that pooling of 2 or 3 donors can increase remission rates in UC [93]. Multi-donor or pooled FMT have been compared with single donor FMT in a meta-analysis of 14 studies that indicate the superiority of the former (RR, 2.31; P= 0.04) in terms of treatment response in IBD [94]. Another meta-analysis that did not detect any effect of single or pooled donors attributed this to the possibility of “super donors” in those receiving from a single donor. Importantly, head-to-head comparisons are lacking since institutes typically use a uniform policy for sourcing donor samples, and such trials are expected in the future.

The donor fecal microbiome is important in determining response to therapy. Focus, thus far, during the selection of donors has been on safety–particularly preventing transmission of pathogenic or virulent organisms or genes. Recurrent CDI required the replacement of healthy gut microbiome to replace niches for C. difficile and thus trials were not focused on selecting donors for their effect on efficacy. However, in contrast to CDI, there is emerging evidence linking properties of donor microbiome with clinical outcomes in IBD. In an early trial reporting the efficacy of FMT in inducing remission in UC, remission was attained in 9 patients (24%) in the FMT arm, 7 of whom received FMT from a single donor [90]. This donor, characterized by enrichment for the family Lachnospiraceae and the genera Ruminococcus, showed much higher remission rates (39% vs. 10%). Further, there was a clear association between achieving a microbiome similar to the donor at week 6 and achieving clinical remission after FMT. This suggested the phenomena of “super donors”: donors whose fecal microbiota characteristics greatly improve clinical response to FMT [95,96]. A smaller donor-recipient age difference (< 10 years) was associated with improved maintenance rates in a trial of FMT following antibiotics in UC [97]. A prime parameter of the donor fecal microbiota responsible for this appears to be high alpha diversity and absolute richness. A single-arm study utilized a single donor with high microbial diversity and fecal butyrate concentration, considered to represent healthy gut microbiome, for FMT to 10 pediatric patients with recurrent CDI and found complete response in all patients at 10 weeks, as well as a sustained increase in phylogenic diversity of recipient microbiome over this period [98]. Similarly, in UC patients undergoing FMT, higher alpha diversity was associated with better recipient outcomes in 2 trials [69,43]. Longitudinal analysis of 2 donors from the LOTUS trial identified long-term microbiome stability and donor species evenness as novel factors associated with FMT success [99]. There have been several individual microorganisms that have been proposed to influence response, but these findings have been limited to individual trials and few could be replicated in other studies, suggesting a need for broader assessment than individual organisms [95]. A metagenomic analysis of 14 public databases of FMT trials in CDI and IBD identified that donor fecal microbiome with donor/p phenotype, indicating Prevotella as the dominant taxa, were associated with a greater response to FMT versus donor/b phenotype (65.1% vs. 40.1%), rich in Bacteriodetes [100]. Clinical response to these donor phenotypes correlated well with an increase in recipient alpha diversity. However, Olesen and Gerardin [101] reevaluated the statistical evidence related to super donor phenomenon in studies on FMT in IBD and suggested uncertainty on the magnitude of donor effect and the need for larger studies specifically designed to assess donor phenomena. When possible, donors should be unrelated to the recipient.

As important as selecting appropriate donors is identifying those who will benefit from FMT, vis-à-vis several options for pharmacotherapy. A prospective cohort treating patients with moderate to severe UC analyzed clinical predictors of response to FMT: younger patients, left-sided colitis disease extent (E2 disease, vs. either proctitis or pancolitis), and endoscopic Mayo score of 2 were associated with remission, while severe disease (Mayo Clinic score ≥ 2), Mayo endoscopic score 3, and pancolitis (E3 extent) were associated with failure of therapy [92]. A systematic review of 25 studies analyzed response predictors to FMT in IBD [102]. Donor fecal microbiome SCFA production, higher alpha diversity, and greater abundance of Bacteroidetes and Clostridia clusters were associated with clinical response. Recipient clinical parameters of young age, less severe disease, and shorter duration of disease, and microbiome characterized by higher fecal species richness, greater abundance of Candida, and similarity to donor profile were also significant predictors.

Current technology and computational setups allow a detailed analysis of the gut microbiome. However, it is not ready for clinical application because of the involved cost and turn-around time. Retrospective analysis of donor-recipient matching provides insights into the deficiencies of the diseased microbiome that are being filled by the donor microbiome. He et al. [100] developed a machine learning model for selecting donors based on recipients (enterotype-donors selection, EDS) and validated this model on a population of 42 patients with IBD. They found 93% clinical response in the patient-donor pairs who could be matched with the model and 37% in those who were not selected by the model. An analytical-hierarchy based donor-recipient matching model was proposed by Zhang et al. [103] based on sequencing data from 350 patients with UC receiving FMT. Important indicators for this model included alphadiversity (28% of weight) and presence of beneficial taxa and metabolic pathways (24% and 22% of weight), and this model showed favorable classification ( > 70%) for FMT effectiveness in 2 prior trials. A matching strategy based on ex-vivo response assessment of patient’s rectal immune cells to candidate donor specimens has been explored in a case report [47].

Engraftment of donor fecal species into the gut of the recipient, and correction of dysbiosis, is the basis of FMT therapy. However, the degree to which new taxa are added and existing taxa are displaced from the microbiome pool is variable. Patients who responded well to FMT and achieved remission had greater diversity both at baseline as well as post-FMT, compared to those who did not respond [43]. Patients in remission after FMT had enrichment of Eubacterium and Roseburia post-FMT and increased levels of SCFA biosynthesis and secondary bile acids, versus enrichment of Fusobacterium, Sutterella, and Escherichia species and increased levels of heme and lipopolysaccharide biosynthesis in those who did not respond. A recent study conducted by our group explored the effect of FMT on the composition of fecal and mucosal microbiota and metabolome in patients with UC [104]. An 8-week pooled multi-donor FMT in our patient cohort mediated enhancement of alpha diversity, with increased abundances of beneficial gut microbial taxa such as Eubacterium, Alloprevotella, Blautia in the fecal microbiota and Alistipes, Eubacterium in the mucosal microbiota. These beneficial taxa have previously been reported to be active SCFAs and indole-based AhR ligand producers and bile acid transformers. These shifts in abundances were coupled with a reduction in abundances of pathobionts such as Pseudomonas, Streptococcus, Enterococcus, Klebsiella, and Staphylococcus. As seen here, there is semblance in the bacterial species and metabolite changes post-FMT across 2 different cohorts.

Several thousands of patients with UC have undergone FMT as part of clinical trials as well as patient care. Large meta-analyses have shown the efficacy of FMT in UC, and clinical trials for this indication are listed in Table 1 [60,79,80,90,91,105-113]. El Hage Chehade et al. [114] published their meta-analyses limited to double-blind placebo-controlled trials for induction of remission in UC and found a good composite clinical and endoscopic response with odds ratio (OR) of 4.1 (95% CI, 2.2–7.7). All included trials included patients with mild-moderate UC based on Mayo score between 3 to 10, and the longest duration of follow-up was 3 months. Subgroup analyses showed no significant influence on the efficacy of several parameters that are discussed above: fresh versus frozen FMT, single versus pooled donor, pretreatment with antibiotics, bowel lavage, concomitant antibiotic therapy, and route of FMT delivery. Zhao et al. [55] included both cohort studies (n = 30) and randomized clinical trials (n = 7) in their meta-analyses and found complete remission response rates of 35.9% in uncontrolled studies and OR of 3.39 (95% CI, 2.2–5.2) in controlled cohort studies and RCTs. Serious adverse event rates were low but significant (5.9%). Cochrane Society analyzed 12 RCTs published till 2022 and found a RR of induction of remission in UC to be 1.79 (95% CI, 1.13–2.84) but was not clear on its role in the maintenance of remission of active UC after induction (2 trials: RR, 2.97; 95% CI, 0.26–34.42) [115]. A recent network meta-analysis compared available pharmacotherapy and FMT for induction of remission in patients with UC [116]. Four of 19 included articles were trials for FMT, and they found a pooled OR of 2.8 (95% CI, 1.4–5.8) for remission compared to placebo–statistically similar efficacy as other biological agents. Overall efficacy ranking from highest to lowest was infliximab, tofacitinib, ustekinumab, FMT, and least vedolizumab and adalimumab. As of this review, consensus statements recommend providing FMT as part of clinical trials as there is insufficient evidence for routine clinical use [52].

Long-term follow-up of trials in FMT show that clinical response is not sustained after induction, indicating that additional microbiome manipulation is required to maintain the response [111,117]. Role of FMT in the maintenance of remission in UC has been explored in a randomized sham-controlled trial by Sood et al. that included patients with UC in remission and found higher rates of 48-week endoscopic (58% vs. 27%, P= 0.03) and histological (45% vs. 16%, P= 0.03) remission in those treated with 8 weekly FMT compared to placebo, although it did not meet primary outcome of clinical remission (87% vs. 66%, P= 0.11) [112]. Lahtinen et al. [113] randomized 48 patients with UC to a single session of FMT versus autologous transplant via colonoscopy (control) and found no difference in remission rates at 12 months (54% vs. 41%, P= 0.66), although quality of life scores were worse in FMT arm. This lack of efficacy is likely due to providing a single session of FMT versus multiple sessions.

Twenty percent of patients with moderate to severe UC remain steroid-dependent on follow-up, and FMT has been evaluated for the treatment of this subgroup in an attempt to avoid steroids and related adverse effects. Cui et al. [118] treated 15 patients of steroid-dependent UC with step-up FMT and 57% (n = 8) achieved clinical response and were able to discontinue steroids, and these patients were likely to have a change in their microbiota composition towards the donor. A real-world prospective cohort analysis of 49 patients with steroid-dependent UC were treated with 4-weekly multisession FMT through colonoscopy and 75% patients achieved clinical response at 24 weeks and could withdraw steroids [119]. A small clinical trial utilized FMT for induction, maintenance and rescue therapy in 27 patients with steroid-dependent UC, 13 of whom remained in remission at 24 weeks after induction with 3 fortnightly sessions of FMT [120]. Ten of 13 patients eventually relapsed at a median duration of 15 months, and 40% of these relapses were managed with rescue FMT alone. While these results are promising, no patient who was on steroids at the time of recruitment (surrogate for steroid-dependent disease) to the RCT by Paramsothy et al. [111]. achieved clinical remission with FMT. Thus, this population of steroid-dependent UC patients is positioned to be important stakeholders for future trials of FMT in UC.

Diet has been an important modulator of gut microbiome and hence it would be an attractive strategy to modify diet along with FMT [121]. We performed an open-label RCT in which we randomized 66 patients with mild-moderate UC to a combination of FMT and anti-inflammatory diet (AID), compared to optimized standard medical therapy [91]. There were high response rates in the FMT-AID arm with over two-thirds of patients having a clinical response at 8 weeks (OR, 3.5; 95% CI, 1.3–9.6). Further, the response to induction in the FMT-AID arm was maintained with AID alone in 25% of patients (vs. 0% in SMT, P= 0.007). The AID consisted of a combination of dietary constituents that have been found to ameliorate gut inflammation in pre-clinical models and epidemiological studies–mainly, avoidance of gluten-based grains, dairy products, processed meats, refined sugars, and increasing amount of fruits and vegetables in diet [122]. Another small trial was stopped due to futility after the recruitment of 62 patients when UC exclusion diet (UCED) was found to be superior to FMT with or without UCED based on higher clinical remission and mucosal healing [106]. This suggests the need for further head-to-head comparison of UCED versus FMT and as adjuncts.

There have been emerging reports on the use of FMT in patients with acute severe UC (ASUC). A patient who developed ASUC on a background medical therapy comprising of infliximab, azathioprine and mesalamine, and did not respond to steroids by day 3 was provided FMT and had clinical as well as endoscopic remission on follow-up [123]. Another case report depicted failure of FMT as a salvage therapy in a 3-year-old patient with ASUC who eventually required colectomy [124]. Microbiome analysis confirmed the lack of alteration in gut microbiome in this patient despite FMT, possibly explaining the lack of response.

A systematic review analyzed 228 patients from 11 cohort studies and 1 randomized trial that evaluated the efficacy of FMT in the induction of remission in CD [125]. They found 57% of patients achieved clinical remission within 2 to 4 weeks after FMT and significantly reduced Crohn’s Disease Activity Index (CDAI) scores within 4 to 8 weeks after FMT. Three of the 12 included studies restricted recruitment to patients with L2 (colonic) disease, possibly due to positive experience in UC. Subgroup analyses found that pre-FMT treatment with antibiotics led to improved remission rates. Further, after FMT there was an increase in microbiome alpha diversity (Shannon index) and a shift towards donor microbiome profile. Yang et al. [77] only included patients with active (clinical disease activity, CDAI scores > 150) colonic disease in their clinical trial. They administered FMT either via gastroscopy or colonoscopy in half patients each. Baseline CDAI scores (mean) were 275 and 290 in the gastroscopy and colonoscopy groups, respectively, with an overall clinical remission rate (CDAI < 150) in 66% of patients which was similar in both groups. Importantly, no clinical trial in CD till date, nor any cohort, has compared FMT in CD with placebo [115]. Thus there is an urgent need for placebo-controlled clinical trials of FMT in CD, and results of recently completed trials are awaited (Clinicaltrials.gov NCT 03078803) [126].

A pilot RCT included 17 patients with colonic or ileocolonic CD and after attaining remission of flare with steroids, were randomized to either FMT or placebo for maintenance of remission [127]. Remission rates at 10 weeks were 87% and 44%, and at 24 weeks were 50% and 33% in the FMT and placebo arms respectively, but the study was underpowered for this comparison (RR for remission at 24 weeks 1.21; 95% CI, 0.36–4.14). Larger controlled studies are needed to ascertain the potential of FMT for CD, including both colonic and ileal disease.

FMT has been accepted widely following revelations of its efficacy, safety, and cost-efficacy in the treatment of recurrent CDI. Authorities regulate FMT differently across countries, as depicted in Table 2 [146-151]. There remains confusion as to whether FMT constitutes a product of human tissue, or a collection of microbes alone, and is unfit for the definition of drug due to the impossibility of similarity between samples. An important milestone for regulatory approval was when the designation of FMT was downgraded from human tissue to medicinal product in the United Kingdom in 2015–which allowed greater ease of access to this therapy for a large number of patients and clinical trials across the nation [152]. FMT remains experimental in the context of use in IBD.

The FDA issued guidance statements in 2012 detailing the chemistry and manufacturing of live microbial products (LMPs), and approved the first fecal microbiota product, REBYOTA, in 2022 for preventing recurrent CDI [149,153]. LMPs include preparations containing live bacteria or viruses for therapeutic or preventative purposes, apart from vaccines. REBYOTA is a biotherapeutic product manufactured from qualified donors, with samples collected at the manufacturing site, in line with good manufacturing practices with a broad spectrum of microbes, representative of the healthy gut. It is administered rectally as 150 mL suspension within 3 days of completing antibiotic therapy and is well tolerated [154]. Success of SER-287 (spore-based microbiome therapeutic) in a phase 1b trial was followed by negative results in the phase II trial [155,156]. Further trials of preformulated LMP preparations in IBD are underway [157].

Colonoscopic transendoscopic tube insertion involves the trans-anal placement of a tube to the cecum, where it is affixed using titanium clips, for the administration of FMT [158]. The placement procedure was successful in all cases, required approximately 15 minutes, and the tube retention period was acceptable to patients in the initial observation study. The tubes remained in place for 2 weeks on average, before falling off, and during this period the patients could be administered FMT or mesalazine through this without significant discomfort.

A recent study looked at a machine learning model to predict the best donors and also the optimal engraftment rates [159]. This was based on an integrated shotgun metagenomic systematic meta-analysis of new and publicly available stool microbiomes collected from 226 triads of donors, pre-FMT recipients, and post-FMT recipients across 8 different disease types from 24 studies. This study employed new strain-resolved metagenomic approaches to elucidate the degree of FMT engraftment and associated clinical outcomes. It was seen that recipients with higher donor strain engraftment were more likely to experience clinical success after FMT. Increased engraftment was seen in individuals receiving FMT from multiple routes (e.g., both via capsules and colonoscopy during the same treatment) as well as in antibiotic-treated recipients with infectious diseases compared with antibiotic-naïve patients with noncommunicable diseases. Bacteroidetes and Actinobacteria species (including Bifidobacteria) displayed higher engraftment than Firmicutes except for 6 under-characterized Firmicutes species.

Much remains to be explored regarding FMT as a routinely available treatment option for IBD. Some gaps in our knowledge that impede this goal are listed in Table 3.