Systematic review: faecal microbiota transplantation in the management of inflammatory bowel disease

Authors


  • This uncommissioned systematic review was subject to full peer-review.

Correspondence to:

Dr K. Whelan, King's College London, Diabetes and Nutritional Sciences Division, 150 Stamford Street, London, SE1 9NH, UK.

E-mail: kevin.whelan@kcl.ac.uk

Summary

Background

The intestinal microbiota is involved in the pathogenesis of inflammatory bowel disease (IBD). Faecal microbiota transplantation (FMT) has been used for the management of IBD as well as infectious diarrhoea.

Aim

To undertake a systematic review of FMT in patients with IBD.

Methods

The systematic review followed Cochrane and PRISMA recommendations. Nine electronic databases were searched in addition to hand searching and contacting experts. Inclusion criteria were reports (RCT, nonrandomised trials, case series and case reports) of FMT in patients with IBD.

Results

Of the 5320 articles identified, 17 fulfilled the inclusion criteria, none of which were controlled trials. There were nine case series/case reports of patients receiving FMT for management of their IBD, and eight where FMT was for the treatment of infectious diarrhoea in IBD. These 17 articles reported on 41 patients with IBD (27 UC, 12 Crohn's, 2 unclassified) with a follow-up period of between 2 weeks and 13 years. Where reported, FMT was administered via colonoscopy/enema (26/33) or via enteral tube (7/33). In patients treated for their IBD, the majority experienced a reduction of symptoms (19/25), cessation of IBD medications (13/17) and disease remission (15/24). There was resolution of C. difficile infection in all those treated for such (15/15).

Conclusions

Whilst the available evidence is limited and weak, it suggests that faecal microbiota transplantation has the potential to be an effective and safe treatment for IBD, at least when standard treatments have failed. Well-designed randomised controlled trials are required to investigate these findings.

Introduction

Inflammatory bowel disease (IBD) is a chronic, relapsing and remitting disease, with both ulcerative colitis (UC) and Crohn's disease (CD) causing significant morbidity.[1] The precise aetiology of IBD is unclear, however, its development, progression and phenotype are multifactorial with genetics and environment playing a role.[2] There is increasing evidence supporting a microbial influence in the pathogenesis of IBD resulting from an inappropriate immune response towards components of the commensal microbiota.[2, 3] Although there is inconclusive evidence for a specific pathogen causing IBD,[2] evidence suggests that there is a reduced diversity of luminal microbiota in IBD, with a decrease in Firmicutes such as bifidobacteria, lactobacillus and Faecalibacterium prausnitzii and an increase in mucosal-adherent bacteria.[4]

Firmicutes are major producers of short-chain fatty acids (SCFA's) such as butyrate, which is a substrate with immunoregulatory properties.[4] Indeed, whilst not conclusive, studies have suggested possible benefits of butyrate enemas when used as a supplement to standard drugs, particularly in refractory UC.[5, 6]

The treatment of IBD is rapidly evolving and many conventional and novel drug treatments have proven efficacy, including steroids, aminosalicylates, immunosuppressants and biological therapies.[7] However, some patients become refractory to standard management and some have significant adverse side effects with many patients requiring surgery.[8, 9] Despite medical treatment, a significant number of patients live with mild active symptoms and have a poor quality of life.[7, 10]

Given the role of the gastrointestinal microbiota in driving inflammation in IBD, treatments that manipulate the microbiota have been investigated including the use of probiotics and prebiotics, with variable evidence for their efficacy.[11, 12] An additional alternative treatment for the management of IBD is faecal microbiota transplantation (FMT), which is the transfer of gastrointestinal microbiota from a healthy donor, via infusion of a liquid stool suspension, to restore the intestinal microbiota of a diseased individual.[13-15] Although it was first documented back in 1958,[16] it has recently become prominent as a treatment for refractory and recurrent Clostridium difficile infection. Two systematic reviews have recently been published in relation to FMT for the management of C. difficile. They failed to identify any controlled trials, although a number of case series and case reports indicated FMT as a potentially effective therapy where standard treatments had failed.[13, 17] Interestingly, C.difficile infection is more common in patients with IBD,[18] with one study reporting a higher prevalence among patients with UC (3.7%) and CD (1.1%) compared with the background general population (0.5%).[18, 19] Symptoms may mimic an exacerbation of IBD or precipitate a genuine relapse.

Faecal microbiota transplantation is also being used as a therapy in IBD with reports of patients with positive outcomes. However, there is currently a lack of cohesive assimilation of the available information on which to inform future robust clinical trials. The aim was to undertake a systematic review of FMT in patients with IBD to provide clinicians with a thorough and clear summary of the available evidence on which to guide current practice and future research.

Methods

Where possible, the systematic review was undertaken in line with the recommendations of the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions.[20] This systematic review adheres to the relevant criteria of the PRISMA statement (Preferred Reporting Items for Systematic reviews and Meta-Analyses).[21] The following methods used in the systematic review, including identification, screening, eligibility and inclusion, were agreed between the authors in advance.

References were identified through electronic database searching, hand searching of conference abstracts and key reference lists and contacting experts in the area. The search strategy was developed by the authors in conjunction with a senior information specialist and was peer reviewed by an independent researcher.

Electronic searching of the following nine electronic databases was undertaken: MEDLINE (from 1948); EMBASE (from 1947); CENTRAL (The Cochrane Library, for all years); CINAHL (for all years); Web of Science (from 1899); BIOSIS (from 1969); SCOPUS (for all years); BIOMED CENTRAL (for all years); NHS Evidence (for all years). The final search date was October 2011.

The search for the term ‘faecal microbiota transplantation’ included a wide range of derivations to ensure as wide a search strategy as possible. Terms were used as keywords or free-text words only, as no relevant medical subject headings were identified in any of the databases. ‘Inflammatory Bowel Disease’ was not included as a search term to ensure that articles with IBD patients within mixed patient groups were captured. A list of the search strategy used is available online (Data S1).

Hand searching of abstracts from the annual conferences of the following organisations was undertaken to obtain conference reports that would not be identifiable through electronic searching: Digestive Disease Week (Gastroenterology) 2001–2012; the British Society of Gastroenterology (Gut) 2001–2012; the European Crohn's and Colitis Organization (Journal of Crohn's and Colitis) 2007–2012; the Crohn's and Colitis Foundation of America (Inflammatory Bowel Diseases) 2003–2010; and the United European Gastroenterology Federation (Gut) 2008–2011. In addition, hand searching of the reference lists of relevant reviews and included studies was undertaken in identify further relevant references. Experts in FMT(n = 7) were contacted to obtain published references not identified during electronic or hand searching.

The research question and inclusion and exclusion criteria were developed using a PICOS structure (Patient, Intervention, Comparators, Outcome, Study design). The inclusion criteria were patients of any age with IBD treated with FMT either for their IBD or for infectious diarrhoea. In studies with mixed patient groups, data were only extracted for those patients with IBD. Published studies or conference abstracts of any language were eligible to be included. Eligibility and data extraction for non-English language articles were undertaken by a native speaker in conjunction with one of the authors. Where two reports related to the same patient group, the most complete report was included. Details of the inclusion and exclusion criteria are described in Data S2.

The references were imported into a bibliographic database to automatically exclude duplicates. Two researchers independently reviewed the title and abstract of each reference to assess its eligibility, excluding those that clearly did not reach the inclusion criteria. Then, for all potentially eligible references, the full article was obtained and the inclusion/exclusion criteria were applied to each. When articles contained insufficient information to assess their eligibility or to extract relevant data, the corresponding author was contacted for further information and this occurred for 12 articles.[22-33] Authors from 7 of the 12 articles were able to provide additional information.

The two researchers independently extracted the data from eligible articles. Data relating to the patient or group, the intervention, the comparator group (where relevant), the outcomes measured (adverse events, mortality, morbidity) and the study design were extracted as detailed (Data S2). Where there was disagreement between eligibility or data extraction, consensus was reached through discussion and where this could not be reached a third researcher arbitrated.

The studies were categorised into reports where FMT was provided for (i) the maintenance or treatment of IBD; or (ii) the prevention or treatment of infectious diarrhoea in patients with IBD.

Results

A total of 5320 nonduplicated articles were identified in the search. The titles and abstracts were reviewed and only 78 were deemed potentially eligible. Following review of the full article, seventeen fulfilled the inclusion criteria, all of which were case series/case reports (4 journal articles, 11 abstracts, 2 letters). Nine of these related to FMT for the maintenance or treatment of IBD[23, 26, 28, 30-32, 34-36] and eight related to the treatment of infectious diarrhoea in IBD[22, 24, 25, 27, 29, 33, 37, 38] (Figure 1). A meta-analysis was not conducted due to the absence of RCTs, the small number of reports and the heterogeneity of treatment protocols and outcome reports.

Figure 1.

Flow chart of studies of faecal microbiota transplantation in IBD.

Patients

The 17 case series/case reports provided data on 41 patients (20 males, 18 females, 3 unknown) ranging in age from 11 to 78 years, of whom 27 were diagnosed with UC, 12 with CD and two with unclassified UC/CD. Nine articles reported 26 patients (18 UC, 6 CD, 2 UC/CD) receiving FMT due to resistance to standard IBD management and recurrence of symptoms when medications were reduced or withdrawn (Table 1). The remaining eight articles reported 15 patients (nine UC, six CD) in whom the primary indication was recurrent or refractory C.difficile infection (Table 2). In all 41 patients, the duration of follow-up ranged between 2 weeks and 13 years.

Table 1. Case series and case reports of faecal microbiota transplant for the management of IBD
ReferencePatientInterventionOutcomes
n IBD/n totalDiseaseIBD MedicationsPatient preparationStool preparation g/ml Volume instilledRoute Number infusionsDonor relationshipIBD MedicationsIBD Symptomsa IBD Diseaseb Duration of Follow Up
  1. CDI, Clostridium difficile infection; IBD, inflammatory bowel disease; UC, ulcerative colitis; CD, Crohn's disease; NR, not reported; NM, not measured.

  2. a

     Resolution = complete cessation of symptoms (no objective measures).

  3. b

     Resolution = absence of active disease confirmed endoscopically and histologically.

  4. c

     Measured by CDAI, Crohn's Disease Activity Index; CDEIS, Crohn's Disease Endoscopic Index of Severity; SES_CD, Simple Endoscopic Score for Crohn's Disease; CRP, C Reactive Protein.

  5. d

     Measured by Mayo Score with a marginal decrease from 11 to 9 and in one patient, a decrease in Mayo endoscopic score of 3 to 2.

  6. e

     Unclear if volume of stool is the amount used to make the infusion and/or the amount administered to patient.

Bennet[35] 1/1UCα-tocopherylquinone ‘sterilise bowel’NRNREnema>1NRAll ceasedResolutionResolution6 months
Borody et al.[25] 2/5UCOlsalazine SulphasalazineNRNRNRNRNRNRAll ceasedResolutionResolution 3 months
  CD

Sulphasalazine

Prednisone

NRNRNRNRNRNRAll ceased ResolutionNR (Relapse after 18 months)1.5 years
Borody et al.[30] 3/6UC

Prednisone

Azathioprine

Mesalazine (Mesalamine)

Vancomycin 500 mg b.d.

Metronidazole 400 mg b.d.

Rifampicin 150 mg b.d.

NR/200ml saline and Psyllium NREnema6Healthy adultsAll ceasedResolutionResolution8–16 months
  UCAs aboveAs aboveAs aboveNREnema6Healthy adultsAll ceasedResolutionResolution8–16 months
  UCAs aboveAs aboveAs aboveNREnema6Healthy adultsAll ceasedResolutionResolution8–16 months
Borody et al.[34] 6/6UC

Prednisone 25 mg/day

Salazopyrin 1 g/day

Vancomycin 500 mg b.d.

Metronidazole 400 mg b.d.

Rifampicin 150 mg b.d.

Polyethylene glycol 3–4 L

200–300 g/200–300 mL salineNREnema6Female partnerAll ceasedResolutionResolution 13 years
UCSalazopyrin 2 g/dayAs aboveAs aboveNREnema6Unrelated maleAll ceasedResolutionResolution12 years
UC

Prednisone 30 g/day

Azathioprine 200 mg/day

Mesalazine 6 g/day Loperamide 12 mg/day

As aboveAs aboveNREnema6BrotherAll ceasedResolutionResolution 4 years
UC

Prednisone 40 mg/day

Olsalazine 3 g/day

Mercaptopurin 75 mg/day

As aboveAs aboveNREnema6Brother in lawAll ceasedResolutionResolution2 years
UC

Prednisone 25 mg/day

Azathioprine 200 mg/day Mercaptopurine 1.5 g/d

As aboveAs aboveNREnema6BrotherAll ceasedResolutionResolution 1 year
UC

Prednisone 25 mg/day

Azathioprine 125 mg/day

Mercaptopurine 2 g/day

As aboveAs aboveNREnema6BrotherAll ceasedResolutionResolution1 year
Grehan et al.[23] 1/10CD

Mesalazine

Azathioprine

Vancomycin 500 mg b.d.

Metronidazole 400 mg b.d. Polyethylene glycol 3–4 L

NR/250 mL saline200–400 mL

Colonoscopy

Enema

1

9

Unrelated maleNMNMNM6 months
Borody et al.[30] 3/3UC/CD

Corticosteroids

Biologic therapies

NRNRNR

Colonoscopy

Enema

1

33

Clinic donorNRResolutionResolution1 year
  UC/CD

Corticosteroids

Immunosuppressants

Biologic therapies

Antibiotics

NRNRNREnema60+FamilyNRResolutionResolution1.5 years
  UC

5-ASA

Immunosuppressants

Antibiotics

Probiotics

Acetarsol

NRNRNREnema (self-administer)70PartnerAll ceasedResolutionResolution4 years
Borody et al.[28] 1/1UC

Corticosteroids

Immunosuppressants

NRNRNRNRNRNRNRReductionNRNR
Vermeire et al.[32] 4/4CD

Corticosteroids

Immunosuppressants

Biologic therapies

Polyethylene glycol200 g/NR200 gNasojejunal3Healthy donorNo change or ceased due to not workingNo resolutionNo resolution c 8 weeks
  CDAs abovePolyethylene glycol200 g/NR200 gNasojejunal3Healthy donorNo change or ceased due to not workingNo resolutionNo resolution c 8 weeks
  CDAs abovePolyethylene glycol200 g/NR200 gNasojejunal3Healthy donorNo change or ceased due to not workingNo resolutionNo resolution c 8 weeks
  CDAs abovePolyethylene glycol200g/NR200 gNasojejunal3Healthy donorNo change or ceased due to not workingNo resolutionNo resolution c 8 wks
Angelberger et al.[36] 5/5UC

NR

Immunosuppressants stopped prior to FMT

Antibiotics

Probiotics

Bowel lavage

6–25 g/NRNR e

Nasojejunal

Enema

1

2

Healthy adultNRNRDeterioration12 weeks
  UCNRImmunosuppressants stopped prior to FMT

Antibiotics

Probiotics

Bowel lavage

6–25 g/NRNR e

Nasojejunal

Enema

1

2

Healthy adultNRNRDeterioration12 weeks
  UC

NR

Immunosuppressants stopped prior to FMT

Antibiotics

Probiotics

Bowel lavage

6–25 g/NRNR e

Nasojejunal

Enema

1

2

Healthy adultNR

NR

Improved wellbeing

No resolution d 12 weeks
  UC

NR

Immunosuppressants stopped prior to FMT

Antibiotics

Probiotics

Bowel lavage

6–25 g/NRNR e

Nasojejunal

Enema

1

2

Healthy adultNR

NR

Improved wellbeing

No resolution

d

12 weeks
  UC

NR

Immunosuppressants stopped prior to FMT

Antibiotics

Bowel lavage

6–25 g/NRNR e

Nasojejunal

Enema

1

2

Healthy adultNR

NR

Improved wellbeing

No resolution d 12 weeks
Table 2. Case series and case reports of faecal microbiota transplant for the management of CDI in patients with IBD
ReferencePatientInterventionOutcomes
n IBD/n totalDisease Patient preparationStool preparation g/ml salineVolume instilledRoute Number of infusionsDonor Relationship C. difficile toxin Symptomsb IBD Medicationsa Duration of follow-up
  1. CDI, Clostridium difficile infection; IBD, inflammatory bowel disease; UC, ulcerative colitis; CD, Crohn's disease; NR, not reported; NM, not measured; FMT, faecal microbiota transplant.

  2. a

     Responded = Previously refractory to IBD medications or unable to use secondary to CDI and responding post FMT.

  3. b

     Symptoms = One or more of diarrhoea, abdominal pain, mucus, flatulence, urgency, rectal bleeding.

  4. c

     Unclear if volume of stool is the amount used to make the infusion and/or the amount administered to patient.

Aas et al.[22] 2/18 CD

Vancomycin 250 mg 8 hrs

Omeprazole 20 mg

30 g/50–70 mL25 mLNasogastric1Family members or clinic staffResolutionResolution of diarrhoeaNM90 days
  UC

Vancomycin 250 mg 8 hrs

Omeprazole 20 mg

30 g/50–70 mL25 mLNasogastric1As aboveNRNRNM90 days
Wettstein et al.[25] 1/16UC

Vancomycin 500 mg b.d. or

Metronidazole 400 mg b.d.

or Rifampicin 150 mg b.d.

Polyethylene glycol 3L

200–300 g/200–300 mL (Saline and psyllium)NR

Colonoscopy

Enema

1

5–24

Relatives or unrelated

NR for IBD patient

Resolution in 15/16

NR for IBD patient specifically

Resolution in 8/16

Reduction in 7/16

NM4–6  weeks
Borody et al.[24] 6/6UCVancomycin 500 mg b.d. Polyethylene glycol 3–4 L 200–300g/200–300 mLNR

Colonoscopy

Enema

1≥5 Relatives or friendResolution Resolution, reduction5/6 responded8  weeks–3 months
  UCVancomycin 500 mg b.d. Polyethylene glycol 3–4 L 200–300 g/200–300 mLNR

Colonoscopy

Enema

1≥5 Relatives or friendResolution Resolution, reduction5/6 responded8  weeks–3 months
  UCVancomycin 500 mg b.d. Polyethylene glycol 3–4 L 200–300 g/200–300 mLNR

Colonoscopy

Enema

1≥5 Relatives or friendResolution Resolution, reduction5/6 responded8  weeks–3 months
  UCVancomycin 500 mg b.d. Polyethylene glycol 3–4 L 200–300 g/200–300 mLNR

Colonoscopy

Enema

1≥5 Relatives or friendResolution Resolution, reduction5/6 responded8  weeks–3 months
  CDVancomycin 500 mg b.d. Polyethylene glycol 3–4 L 200–300 g/200–300 mLNR

Colonoscopy

Enema

1≥5 Relatives or friendResolution Resolution, reduction5/6 responded8  weeks–3 months
  CDVancomycin 500 mg b.d. Polyethylene glycol 3–4 L 200–300 g/200–300 mLNR

Colonoscopy

Enema

1≥5 Relatives or friendResolution Resolution, reduction5/6 responded8  weeks–3 months
Mellow et al.[27] 1/13CD NRNR/NR300–500  mLColonoscopy1NR

NR for IBD patient

Resolution in 12/13

NR for IBD patient specifically

Resolution of diarrhoea 12/13

NM

Died at 5 months

(Not FMT related)

You et al.[29] 1/1CDNR75 g/200 mLNRNasogastric1HusbandResolutionReductionResponded13 days
Watson et al.[37] 1/1UCNRNRNRColonoscopy1WifeNo relapse of CDIDeveloped diarrhoea and abdominal crampCommenced mesalamine due to UC flare33+  days
Neelakanta et al.[38] 2/2CDNR250 g/NR c NRColonoscopy1MotherResolutionReductioǹAble to commence Adalimumab>1 year
̀̀̀NRNR/NRNRColonoscopy1MotherResolutionReductionAble to commence steroids5 months
Zainah et al.[33] 1/1UCPolyethylene glycol300 mL/NR c NRColonoscopy1WifeNo CDI recurrenceResolutionNR8 months

There was incomplete reporting of IBD-specific data such as disease duration, location and activity, with some data available predominantly from the nine articles reporting FMT as the primary treatment of IBD. Where reported (n = 23), the duration of disease ranged from 18 months to >20 years. Active disease was confirmed via endoscopy only (n = 21) or via endoscopy and histology (n = 15), and only two articles[32, 36] reported objective measures of disease activity, such as the CD Endoscopic Index of Severity, Simple Endoscopic Score for CD, CD Activity Index and the Mayo Index. One patient treated with FMT for C. difficile infection showed no endoscopic evidence of active disease.[37] There was variable disease distribution of rectum (n = 1), rectum, sigmoid & descending colon (n = 2), plus transverse colon (n = 2), plus ascending colon (n = 3) in UC and terminal ileum (n = 1), ileocolonic (n = 4) and isolated colitis (n = 1) in CD patients. Where reported, all patients with active IBD were on concurrent IBD medication at the time of FMT, and their duration ranged between 5 and 48 months.

Faecal microbiota transplantation

No details were available regarding any aspects of the FMT protocol for three articles[26, 28, 35] and details in the remaining articles were incomplete and variable. Donors were healthy adults who had no antibiotics in the preceding 6–8 weeks (n = 16) or preceding 6 months (n = 10) and underwent viral screening for human immunodeficiency virus, hepatitis, cytomegalovirus, Epstein–Barr virus and C. difficile toxin as well as screening for parasites, ova and bacterial pathogens (n = 34). The relationship of the donor to the patient varied and comprised healthy relatives, partners, friends or unrelated donors.

Incomplete data were reported for the preparation and administration of FMT with no article reporting complete data enabling protocol replication. Where reported, faecal samples were collected within 10 min (n = 9) or within 6 h (n = 5) before FMT and the volume was usually 200–300 g of stool homogenised in 200–300 ml of normal saline.

Twenty-six patients were administered FMT via the lower gastrointestinal tract (GIT) and seven via the upper GIT (via enteral tube). In one article (n = 5), a combination of administration via the upper and lower GIT was reported.[36] The majority received follow-up enemas for between 2 days to 59 weeks, with only eight patients receiving a single infusion, all of whom were receiving treatment for C. difficile infection. Following training, one patient self-administered all infusions using a home infusion protocol.

Management of IBD

There were 26 patients receiving FMT for treatment of IBD (Table 1). In the 17 patients in whom it was reported, 13/17 ceased IBD medications within 6 weeks. In the 16 patients in whom IBD symptoms were measured and reported prior to FMT, all experienced symptom reduction or resolution within 4 months and 15 patients experienced complete resolution within a year. Two articles reported improved ‘general well-being’ in 3/5 UC patients and ‘no clinical benefit’ in 4/4 CD patients, but both reported short-term adverse symptoms during and following FMT administration. Therefore, improvement in symptoms was experienced by 19/25 (76%). However, disease activity were only reported in 24 patients, with 15 (63%) having no evidence of active disease 3–36 months after FMT. Two recent articles using objective measures of disease activity reported no endoscopic benefit (n = 4 CD), marginal improvement (n = 3 UC) or deterioration in UC (n = 2).[32, 36] Where ‘disease remission’ following FMT was reported (n = 15), there was no disease recurrence in long-term follow-up ranging from 3 to 6 months (n = 3/15) and 1–13 years (n = 12/15), however, one author noted in correspondence that one patient with CD relapsed after 18 months, though it was suspected they had undiagnosed C. difficile infection.[26]

Management of C. difficile in patients with IBD

There were 15 patients receiving FMT for treatment of C. difficile infections in patients with IBD (Table 2), however, outcome data could only be extracted for 12/15 patients.[22, 24, 29, 33, 37, 38] In all patients, there was resolution of C. difficile as measured by negative stool sample enterotoxin and in 11/12 patients, there was a marked reduction or complete resolution of diarrhoea. Two articles reported that FMT resulted in improved response to IBD medications in 6/7 patients. The death of a patient with CD from superior mesenteric vein thrombosis 5 months post FMT was reported; however, there was no indication that FMT contributed to the death.[27]

The impact of FMT on luminal microbiota were measured in only two articles (n = 5 CD). The microbiota in one patient consisted predominantly of bacteria derived from the healthy donor.[23] In contrast, the luminal microbiota from four patients did not cluster with their donors on several phylogenetic levels following FMT, and whilst there were changes in microbial composition at 2–4 weeks, by 8 weeks these had returned to the host baseline levels.[32]

Adverse events were reported in three articles.[32, 36, 37] In two articles where FMT was administered for treatment of IBD, one reported 3/4 CD patients with high fever and abdominal tenderness for 2 days following FMT.[32] The other reported 5/5 UC patients with high fever and diarrhoea and 1/5 with vomiting during FMT and 2/5 with a deterioration of UC in the 12 weeks following FMT.[36] One UC patient with inactive disease, treated with FMT for C. difficiIe infection, subsequently developed diarrhoea, abdominal cramping and active disease requiring medication.[37]

There was no formal assessment of patient attitude and perception of FMT, nor cost analysis of FMT in any of the articles.

Discussion

Approaches to modify the gastrointestinal microbiota, such as probiotics and prebiotics, have been used in the treatment and maintenance of IBD, with variable efficacy,[1] and have also been used successfully in the management of C. difficile infection.[39] Here, we describe the current knowledge in relation to patients with IBD who received FMT either for treatment of active disease or for C. difficile infection. The evidence is limited and weak, with no controlled trials (randomised, nonrandomised) and only 17 case reports/series, of which only four were full text reports. However, although limited, the majority of evidence suggests that FMT may be an effective and largely safe alternative treatment for those with IBD that warrants further investigation in well-designed randomised controlled trials.

The use of FMT for the management of active IBD resulted in a reduction or complete resolution of symptoms in 76% patients, cessation of all IBD medications in 76% patients and ‘prolonged remission’ of active disease in 63% of patients where they were measured and reported. In patients treated with FMT for C. difficile infection, all had complete resolution of their C. difficile infection and the majority (92%) had a marked reduction or resolution in diarrhoea. Most patients (86%) who were previously refractory to, or limited to subtherapeutic doses of, IBD medications subsequently responded to them.

The majority of articles reported no relapses nor any reported adverse events directly related to the FMT. However, three recent articles reported relapsing UC following FMT and short-term adverse symptoms during and following administration of FMT. All future studies should screen and report adverse events.

In the majority of cases, FMT was administered solely via the lower GIT (79%), with the initial infusion either via enema or colonoscope, followed by multiple enema infusions. This differs to the evidence of FMT in the treatment of infectious diarrhoea in general patients, where approximately 69% administered solely via the lower GIT with the use of a single infusion in 53% of patients.[13] It has been suggested that multiple and recurrent infusions are required to achieve prolonged remission particularly in severely ill or complicated patients.[14, 30]

The exact mechanism of action for FMT in the treatment of active IBD and of C. difficile infection is unknown, however, by replacing the disrupted microbiota with diverse bacterial populations it is thought to restore the composition and function of the intestinal microbiota in diseased patients.[13, 40] Two case series have demonstrated that the impact of FMT on the microbiota of the recipient is durable (for between 14 days and up to 24 weeks) and becomes similar to that of the healthy donor's microbiota, however, in only one of these studies did any of the patients have IBD.[23, 40] This concurs with studies in patients with antibiotic-associated diarrhoea, which indicate that FMT results in microbiota and SCFA profiles similar to healthy subjects.[41, 42] However, a recent report has shown no durable change in luminal microbiota in patients with IBD after 8 weeks.[32]

Restoring the microbiota of patients may be efficacious in both the treatment of active IBD and in C. difficile infection as both are associated with gastrointestinal dysbiosis. For example, reduced diversity of luminal microbiota exists in IBD, including lower F. Prausnitzii.[4] Similarly patients with recurrent C. difficile infection also have decreased species diversity.[43] Importantly, the luminal microbiota function to inhibit enteropathogenic growth through colonisation inhibition and the release of end products of microbial fermentation (e.g. SCFA),[4] which might play a role in treating active IBD or C. difficile infection. Furthermore, the proportions of bifidobacteria and F. prausnitzii in the lumen are known to have important associations with dendritic cell phenotype.[44] However, it is the mucosal microbiota that are likely to be most important in mucosal immune regulation in the context of IBD, as demonstrated by high mucosal F. prausnitzii and its association with longer remission rates.[45] However, the studies that have measured the impact of FMT on the host microbiota have investigated luminal rather than mucosal effects. Research on the impact of FMT on mucosal microbiota and immune regulation is warranted.

Alternative approaches to modifying the gastrointestinal microbiota in IBD have been trialled, including the use of probiotics and prebiotics; however, a recent Cochrane review of the use of probiotics in the maintenance of remission in UC, found insufficient evidence from which to make conclusions.[46] There is also little evidence to support the use of probiotics in the treatment or maintenance of CD[11] or the use of prebiotics in either UC[47] or CD.[12]

It has been suggested that FMT introduces bacteria already adapted to the GIT and applying a regime to reduce colonisation resistance such as antibiotics and bowel lavage is likely to result in a more persistent change in the resident microbiota than with the use of probiotics, although studies that directly compare the two have not been undertaken.[3, 23] In contrast, data from animal studies have shown that pre-treatment with antibiotics prior to FMT interferes with the establishment of an exogenous microbial community.[48]

The greatest body of evidence for the efficacy of FMT is in the treatment of recurrent C. difficile infection. A recent systematic review of FMT for the treatment of recurrent C. difficile infection and pseudomembranous colitis in general patients showed a 92% resolution in 317 patients across 27 case series and case reports.[13] Another systematic review showed 83% of 124 patients with C. difficile-associated disease with improved symptoms.[17] However, this is the first ever systematic review to investigate the role of FMT in the context of IBD.

Searches of clinical trial databases indicate at least five trials of FMT in progress, three of which are in patients with IBD. One is investigating the microbiological, immunological and clinical sequelae of FMT in the treatment of chronic pouchitis. Another is a RCT of FMT for the induction of endoscopic and clinical remission of active UC and there is also a pilot study of the effect of FMT on clinical symptoms of paediatric patients with UC or colonic CD. In addition, two trials are in the treatment of infectious diarrhoea, comparing FMT with vancomycin. These trials will provide much needed high-grade evidence for the efficacy of FMT in IBD and in C. difficile infection.

Barriers to the adoption of FMT as a therapeutic tool include the absence of high-quality evidence for its efficacy, patient and clinician acceptance and potential safety concerns. It has been suggested that perhaps it is clinicians rather than patients who are resistant to the use of FMT as a viable treatment option.[49] In a survey of 73 physicians, 25 (34%) indicated they would be unwilling to perform the procedure with the most common concerns being patient acceptance and tolerability (71%) followed by safety (60%) and efficacy (57%).[50] In contrast, in a recent qualitative study exploring readiness for FMT in patients with UC, the majority of patients wished the procedure was already available and most viewed it as being safer than current available therapies, especially steroids or biological therapies.[51] In a quantitative extension of this study, the majority of UC patients were interested or willing to consider FMT despite reporting disease control with IBD medications.[52] The most important concerns were around donor selection and screening as a means of eliminating the potential risk of transmission of infectious agents. This risk can be reduced with the implementation of thorough screening protocols. Interestingly, this systematic review highlights that close relatives are frequently selected as stool donors. However, recent evidence indicates that the relatives of patients with IBD themselves have altered microbiota,[53] and that this dysbiosis may itself be partly responsible for their elevated risk of IBD.[54] Investigation of the ideal stool donor in FMT in the context of IBD is warranted.

Limitations

Reporting on the efficacy of FMT for the treatment of IBD is limited by the absence of well-designed randomised controlled trials, with evidence only from a small number of case series and case reports. These are subject to publication bias and patient selection bias, whilst adverse-event reporting is notoriously limited even in controlled trials.[55] In addition, the inclusion of all publication types resulted in the inclusion of eleven articles available in abstract form only, which have been shown to be biased by selective reporting.[20, 21]

There were significant gaps in the measuring and reporting of information from articles that were not always resolved by contacting authors for further information. For example, there was often absent or incomplete data for the FMT procedure, including preparation and administration of the faecal infusion and patient preparation. Practice guidelines for the administration of FMT in C. difficile infection have recently been published.[56] In the absence of adequate evidence for FMT specifically in IBD, these guidelines can be used together with the data from case reports presented here, to guide the development of future research and practice.

Important IBD variables were frequently unreported or were not clearly defined. For example, none reported standardised criteria for disease phenotype (e.g. Montreal classification) and only two articles reported indices of disease activity (e.g. CD Activity Index, Mayo Index).[8, 57, 58] The use of uniform and consistent classification of disease is important to facilitate comparisons between studies. Furthermore, it is essential in the development of future research that investigates in which disease (UC, CD), disease phenotype or disease activity FMT is most efficacious, and whether it should be used in the early management of IBD or only as a strategy when all other treatments have failed. Also outcome variables such as symptom resolution were not well defined, making comparison between studies difficult.

Conclusions

Despite the relative paucity of information specifically in IBD, there is limited, weak evidence that FMT has the potential to be an effective and safe treatment for IBD, at least when standard treatments have failed or are unacceptable to the patient. Well-designed randomised, controlled trials are necessary to confirm the positive findings from case reports, to evaluate safety and to develop optimal protocols for the use of FMT in IBD, prior to this becoming a standard part of clinical therapy.

Acknowledgements

Declaration of personal interests: KW has served as a speaker for DDW, DDF and Danone, and has received research funding from Crohn's and Colitis UK, British Dietetic Association, Healthcare Quality Improvement Partnership and Californian Dried Plum Board. KW is an employee of King's College London. Declaration of funding interests: The systematic review was internally funded.

Ancillary