Efficacy and Safety of Fecal Microbiota Transplantation For The Treatment of Diseases Other Than Infection: A Systematic Review and Meta-Analysis
Efficacy and Safety of Fecal Microbiota Transplantation For The Treatment of Diseases Other Than Infection: A Systematic Review and Meta-Analysis
Efficacy and Safety of Fecal Microbiota Transplantation For The Treatment of Diseases Other Than Infection: A Systematic Review and Meta-Analysis
REVIEW
CONTACT Jessica Emily Green [email protected] Food & Mood Centre, School of Medicine, Barwon Health, Deakin University, IMPACT,
the Institute for Mental and Physical Health and Clinical Translation, Geelong, Australia.
*WM and FJ are joint senior authors.
Supplemental data for this article can be accessed on the publisher’s website.
© 2020 The Author(s). Published with license by Taylor & Francis Group, LLC.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.
e1854640-2 J. E. GREEN ET AL.
have potential as a treatment for the many health review, FMT was defined as any process by
conditions linked with the intestinal microbiome.8 which a fecal microbiota suspension was trans
FMT is already widely practiced as a highly effective ferred from the gastrointestinal tract of
treatment for recurrent Clostridium difficile infec a healthy individual into another person with
tion (CDI).9,10 the aim of treating a health condition.35
A wealth of new research is investigating whether ● Comparator: Studies were included if they uti
FMT may be used to treat other health conditions lized a control group.
linked to the intestinal microbiome,11,12 including ● Outcomes: When reporting on efficacy, this review
gastrointestinal,13–17 autoimmune,18,19 metabolic,20,21 used primary outcome measures as described by
and neuropsychiatric22–24 conditions. There is also each study. When the primary outcome did not
promising preclinical evidence supporting the use of relate to efficacy, the secondary outcomes relating
FMT in conditions other than CDI, including Major to clinical efficacy were noted, but results were only
Depressive Disorder,25,26 schizophrenia,27 and cardi considered significant when the primary outcome
ometabolic syndrome.28 measure related to clinical efficacy and was statisti
While reviews of FMT for specific indications cally significant vs the control intervention. Adverse
such as IBS29–31 and IBD exist,32–34 to date there events (AE) were reported as presented by the
have been no comprehensive reviews evaluating included study.
and synthesizing the entire body of data for both
the efficacy and safety of FMT for all conditions In March 2020, searches were carried out using
other than CDI. This systematic review and meta- MEDLINE, EMBASE, Cochrane Central Register of
analysis addresses the question of whether FMT is Controlled Trials, Health Technology Assessment
safe and effective at treating health conditions other Database, Allied and Complementary Medicine
than CDI in humans. (AMED) and PsycINFO. Reference sections of pre
viously published randomized trials, systematic reviews,
and meta-analyses on this and related topics were also
Methods searched.
Protocol and registration Forty-two iterations of the term “FMT” were
identified and used as search terms:
The Preferred Reporting Items for Systematic FMT or fecal microbiota transplant* or fecal
Reviews and Meta-Analyses (PRISMA) guidelines microbiota transplant* or microbiota transfer or
were adhered to as a methodological template for microbiome transfer or microbiota transplant*
this review. The protocol for this Systematic Review OR microbiome transplant* or microbial trans
was prospectively registered with PROSPERO plant* or microbial transfer or fecal transplant*
(CRD42018104243). OR fecal transplant* or feces transplant* OR
feces transplant* or stool transplant* or stool
transfer or fecal flora transplant* OR fecal flora
Search strategy and eligibility criteria
transplant* or microflora transplant* OR fecal
The PICO approach (population, intervention, flora transfer or fecal flora transfer OR fecal
comparator, outcomes) was used to guide the bacteriotherapy OR fecal bacteriotherapy OR
search strategy for this review. The PICO criteria feces bacteriotherapy OR feces bacteriotherapy
used are outlined below: OR rectal bacteriotherapy OR fecal flora bacter
iotherapy OR donor fecal OR donor stool OR
● Population: Humans participants of any age donor feces OR donor fecal or donor feces fecal
with any acute or chronic health condition transfer OR fecal transfer OR fecal reconstitu
other than CDI. Studies were included only if tion OR fecal reconstitution OR flora reconstitu
participants were followed up for at least two tion OR microbiome reconstitution OR feces
weeks post-FMT. reconstitution or feces reconstitution. The fol
● Intervention: All possible variations of human lowing modifiers were applied: studies relating
FMT were included. For the purposes of this to humans and published in English.
GUT MICROBES e1854640-3
The following study types were included: rando This review also assessed whether successful
mized controlled trials (RCTs), non-randomized- “engraftment” occurred of the donor microbiome
controlled studies, and observational studies with in the recipient. For the purposes of this review, the
a comparator arm. In the case of observational term “engraftment” was ascertained according to
studies with a comparator arm, only prospective the following key concepts:
cohort studies were included in order to assess
temporality. Reviews, abstracts, conference papers, ● Was a change in recipient microbiota observed
and posters were excluded. following FMT?
Two investigators (JG and JD) independently
performed the searches using Rayyan software. JG If a change in recipient microbiota was observed,
performed initial screening to identify potentially then:
eligible studies. Articles were first screened by title
and abstract. Remaining articles were further scru ● Was this change toward the donor microbiota?
tinized by full-text review. JD acted as a secondary ● To what extent/how significant was that
reviewer and was blind to JG’s screening outcomes. change?
Where there was a lack of consensus between the ● For how long did the microbiota changes per
two reviewers, the senior author (WM) acted as sist following cessation of FMT?
a third reviewer to make a final decision on whether
the study met inclusion criteria.
Results
Risk of bias assessment Study selection
Methodological heterogeneity was evaluated by
comparing included data using the ‘risk of bias’ The systematic search identified 5,495 de-duplicated
tables. The Cochrane Risk of Bias tool was used to studies, of which 26 met eligibility criteria for inclu
assess the risk of bias in randomized trials. JG and sion (see Figure 1).
JD independently evaluated risk of bias.36
Study characteristics
Statistical analysis
As per Table 1, of the 26 articles that were included
Data from individual trials were to be combined, in the final review, 20 were Randomized Controlled
and a meta-analysis performed only if the data were Trials (RCTs), two of which were open-label, three
amenable. Patient groups, disease entity, and out single-blind and 15 double-blind. The remaining
come measures needed to be sufficiently similar in six studies encompassed non-randomized-
order for synthesis to occur. Our meta-analyses controlled studies (n = 3), case–control studies
were conducted in Comprehensive Meta-Analysis (n = 2), and cohort studies (n = 1). Eight studies
(version 3.3.070)37 using a Mantel-Haenszel ran investigated the use of FMT for Inflammatory
dom-effects model to account for heterogeneity Bowel Disease (IBD), six for functional gut disor
between studies. The I2-statistic was used as an ders, four for hepatic disorders, three for metabolic
indicator of heterogeneity. A value of 0% indicates syndrome, two for antibiotic-resistant organisms,
no observed heterogeneity, and larger values indi and one each for “pouchitis,” obesity without meta
cate increasing heterogeneity. Due to the limited bolic syndrome, and Human Immunodeficiency
number of studies included in each meta-analysis, Virus (HIV).
no sensitivity or subgroup analyses were per A total of 1149 participants were enrolled in the
formed. Furthermore, no test of publication bias included studies, with a mean of 44 participants per
was performed due to the limited number of study (sample sizes ranged from 6 to 165). When
trials.38 broken down by disorder, there were a total of 463
e1854640-4 J. E. GREEN ET AL.
participants in studies relating to IBD, 424 for func (n = 2), Norway (n = 2),14,40 France (n = 2),45,46
tional gut disorders, 104 for metabolic syndrome/ India (n = 2), Denmark (n = 1), Japan (n = 1),55
obesity, 109 for hepatic disorders, 60 for antibiotic- Austria (n = 1),56 Canada (n = 1),57 Sweden
resistant organisms, and 14 for other disorders. (n = 1),58 and one was an international multi-site
Study follow-up periods varied from 2 weeks to collaborative study between Switzerland, the
12 months. The largest sub-groups by disorder Netherlands, Israel, and France.59
were UC (n = 6) and irritable bowel syndrome
(IBS) (n = 5). These groups were large enough to
Methodological factors for FMT manufacture
allow for meta-analyses.
process
Seventeen studies included both males and
females, eight included males only, and one did The FMT manufacture process varied significantly
not provide demographic data. Studies were con between studies (see Table 1); indeed, no two studies
ducted in US (n = 6),41,44,47–51 The Netherlands used the same process. Of the 26 included FMT
(n = 4),39,42,43,52 Australia (n = 2),53,54 China studies, 11 delivered FMT via colonoscopy, enema,
Table 1. Summary of FMT methodological factors.
Dosing regimen
Study (ie. dosage, frequency,
details Methodological factors duration) Donor factors Adjunctive treatment Control intervention
Saidani et al, Route of administration: NGT Duration of intervention: Number of donors (i.e. single vs. multiple): not stated Antibiotic pre-treatment of recipients: Description of control intervention:
2019 Aerobic vs. anaerobic once-off Type of donor (e.g. Relative vs. non-related): not stated yes (colistin 6-MUI and treatment as usual
preparation: aerobic Initial dosage: 50 g stool Donor screening for mental illness or metabolic risk aminoglycoside (either gentamycin
Fresh vs. frozen: fresh Maintenance regimen: N/A factors: or amikacin) if sensitive, or other
Unclear, stated used “French authorities’ antibiotics if resistant
recommendations” Bowel cleanse: bowel preparation
5 days prior, and day before FMT
Other: nasopharangeal decolonization
using chlorhexidine gluconate gargle
and nose swab, and pantoprazole
Huttner Route of administration: NGT Duration of intervention: Number of donors (i.e. single vs. multiple):single Antibiotic pre-treatment of recipients: Description of control intervention:
et al, (two sites – Israel and The NGT as once off, Type of donor (e.g. Relative vs. non-related): not related yes (colistin sulfate and neomycin treatment as usual
2019 Netherlands) or encapsulated followed by daily Donor screening for mental illness or metabolic risk sulfate)
FMT (two sites – Switzerland encapsulated FMT for 2 factors: Bowel cleanse: no
and France) consecutive days Screened for “chronic diseases or medical history” but Other: N/A
Aerobic vs. anaerobic Initial dosage: 40 g stool not mental health or metabolic risk factors specifically
preparation: aerobic via NGT, unclear dosing
Fresh vs. frozen: frozen for encapsulated FMT
Maintenance regimen: N/A
Sokol et al, Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple):single Antibiotic pre-treatment of recipients: Description of control intervention:
2020 colonoscopy once-off Type of donor (e.g. Relative vs. non-related):non-related no same but placebo (physiological
Aerobic vs. anaerobic Initial dosage: 50–100 g Donor screening for mental illness or metabolic risk Bowel cleanse: bowel preparation with serum) used in sham group
preparation: aerobic fresh stool factors: states excluded if “chronic disease, long term polyethylene glycol
Fresh vs. frozen: fresh Maintenance regimen: N/A treatment” or BMI > 27, but screening for mental Other: N/A
illness or risk factors not specifically described.
Bajaj et al, Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
2017 retention enema once-off Type of donor (e.g. Relative vs. non-related): non-related yes (metronidazole, ciprofloxacin, treatment as usual
Aerobic vs. anaerobic Initial dosage: 90 mL frozen Donor screening for mental illness or metabolic risk amoxicillin)
preparation: aerobic aliquot of fecal factors: Bowel cleanse: no
Fresh vs. frozen: frozen suspension metabolic risk factors – yes Other: N/A
Maintenance regimen: N/A mental illness – yes
Bajaj et al, Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
2019 encapsulated FMT once-off Type of donor (e.g. Relative vs. non-related): non-related no same but placebo capsules used
Aerobic vs. anaerobic Initial dosage: 15 capsules Donor screening for mental illness or metabolic risk Bowel cleanse: no (containing a sterile solution of
preparation: aerobic (unclear dosage per factors: Other: N/A cocoa butter, glycerol and brown
Fresh vs. frozen: frozen capsule) metabolic risk factors – yes food color).
Maintenance regimen: N/A mental illness – yes
Aroniadis Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
et al, encapsulated FMT 3 days Type of donor (e.g. Relative vs. non-related): non-related no same, but placebo capsules used
2019 Aerobic vs. anaerobic Initial dosage: 25 capsules Donor screening for mental illness or metabolic risk Bowel cleanse: no (containing nontoxic brown
GUT MICROBES
preparation: Not specified containing 9.5 g fresh factors: Other: esomeprazole pigment)
Fresh vs. frozen: frozen stool each metabolic risk factors – yes
Maintenance regimen: two mental illness – yes
further doses of 25
capsules over the next
2 days
e1854640-5
(Continued)
Table 1. (Continued).
Dosing regimen
Study (ie. dosage, frequency,
e1854640-6
details Methodological factors duration) Donor factors Adjunctive treatment Control intervention
El-Salhy Route of administration: NDT Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
et al, Aerobic vs. anaerobic once-off Type of donor (e.g. Relative vs. non-related): non-related no same but autologous FMT used
2019 preparation: aerobic Initial dosage: 30 or 60 g Donor screening for mental illness or metabolic risk Bowel cleanse: no
Fresh vs. frozen: frozen feces factors: Other: N/A
Maintenance regimen: N/A metabolic risk factors – screened for “metabolic
disorders”
mental illness – no
Holster et al, Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple):single Antibiotic pre-treatment of recipients: Description of control intervention:
J. E. GREEN ET AL.
2019 colonoscopy once-off Type of donor (e.g. Relative vs. non-related): non-related no same but autologous FMT used
Aerobic vs. anaerobic Initial dosage: 30 g feces Donor screening for mental illness or metabolic risk Bowel cleanse: bowel preparation prior
preparation: aerobic Maintenance regimen: N/A factors: to procedure
Fresh vs. frozen: frozen metabolic risk factors – yes Other: N/A
mental illness – yes
Vrieze et al, Route of administration: NDT Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
2012 Aerobic vs. anaerobic once-off Type of donor (e.g. Relative vs. non-related): non-related no same but autologous FMT used
preparation: aerobic Initial dosage: 500 mL Donor screening for mental illness or metabolic risk Bowel cleanse: bowel lavage the night
Fresh vs. frozen: fresh aliquot of FMT factors: before
suspension metabolic risk factors – screened for diabetes and Other: fasting the night before
Maintenance regimen: N/A obesity, but did not specify other metabolic risk
factors
mental illness – no
Smits et al, Route of administration: NDT Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
2019 Aerobic vs. anaerobic once-off Type of donor (e.g. Relative vs. non-related): non-related no same but autologous FMT used
preparation: aerobic Initial dosage: 500 mL Donor screening for mental illness or metabolic risk Bowel cleanse: bowel lavage
Fresh vs. frozen: fresh aliquot of FMT factors: Other: N/A
suspension metabolic risk factors – yes
Maintenance regimen: N/A mental illness – no
Allegretti Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
et al, encapsulated FMT 8 weeks (3 doses given Type of donor (e.g. Relative vs. non-related):non-related no same but placebo capsules used,
2019 Aerobic vs. anaerobic 4 weeks apart) Donor screening for mental illness or metabolic risk Bowel cleanse: no containing normal saline, food
preparation: aerobic Initial dosage: 30 capsules, factors: Other: N/A color and glycerol)
Fresh vs. frozen: frozen containing 0.75 g stool metabolic risk factors – yes
per capsule mental illness – yes
Maintenance regimen: 12
capsules given at 4 and
8 weeks
Costello Route of administration: Duration of intervention: 3 Number of donors (i.e. single vs. multiple): pooled (3–4 Antibiotic pre-treatment of recipients: Description of control intervention:
et al, colonoscopy followed by doses given in 1 week donors, mixed) no same but autologous FMT used
2019 retention enemas Initial dosage: 50 g stool Type of donor (e.g. Relative vs. non-related): non-related Bowel cleanse: bowel preparation with
Aerobic vs. anaerobic Maintenance regimen: two Donor screening for mental illness or metabolic risk polyethylene glycol the night before
preparation: anaerobic enemas containing 25 g factors: Other: loperimide immediately prior to
Fresh vs. frozen: frozen stool within first week metabolic risk factors – yes procedure
mental illness – screened for depression, but not other
mental illnesses
Sood et al, Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
2019 colonoscopy 42 weeks (6 doses, Type of donor (e.g. Relative vs. non-related): non-related no same but placebo FMT (normal
Aerobic vs. anaerobic 8 weeks apart) Donor screening for mental illness or metabolic risk Bowel cleanse: polyethylene glycol saline with food color) used
preparation: aerobic Initial dosage: 100 g feces factors: bowel lavage the night before
Fresh vs. frozen: frozen Maintenance regimen: metabolic risk factors – no Other: N/A
same mental illness – no
(Continued)
Table 1. (Continued).
Dosing regimen
Study (ie. dosage, frequency,
details Methodological factors duration) Donor factors Adjunctive treatment Control intervention
Herfarth Route of administration: Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
et al, endoscopic route (unspecified 14 days (endoscopic Type of donor (e.g. Relative vs. non-related): non-related not specified same but inert placebo FMT used
2019 whether NGT, NDT or NJT), FMT, followed by daily Donor screening for mental illness or metabolic risk Bowel cleanse: not specified
followed by encapsulated FMT encapsulated FMT) factors: Other: not specified
Aerobic vs. anaerobic Initial dosage: 24 g stool metabolic risk factors – Screened for BMI but unclear if
preparation: not specified Maintenance regimen: 6 screened for other metabolic risk factors
Fresh vs. frozen: not specified capsules daily for mental illness – yes
14 days (total dose 4.2 g
stool/day)
Philips et al, Route of administration: NDT Duration of intervention: Number of donors (i.e. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
201814 Aerobic vs. anaerobic daily for 7 days Type of donor (e.g. Relative vs. non-related): related antibiotics were continued if the control groups consisted of alternative
preparation: aerobic Initial dosage: 100 mL of Donor screening for mental illness or metabolic risk person was already on them treatments (steroids, nutrition or
Fresh vs. frozen: fresh strained and filtered factors: Bowel cleanse: no pentoxifylline therapy)
stool metabolic risk factors – yes, mental illness – no Other: fasting 4 hours before and after
Maintenance regimen:
same
Ishikawa Route of administration: Duration of intervention: Number of donors (ie. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
et al, colonoscopy once off Type of donor (eg. Relative vs. non-related): related yes (amoxicillin, fosfomycinand 2 weeks antibiotics only
201739 Aerobic vs. anaerobic Initial dosage: 150–250 g (spouses or relative) metronidazole) (Amoxicillin, Fosfomycin,
preparation: aerobic fresh donor stool Donor screening for mental illness or metabolic risk Bowel cleanse: bowel preparation metronidazole)
Fresh vs. frozen: fresh Maintenance regime: N/A factors: not comprehensively described – states given, in addition to bowel lavage
“medical history” with poloyethylene glycol given prior
to treatment
Other: Scopolamine given post-
treatment to slow gastric transit time
Tian et al, Route of administration: NJT Duration of intervention: Number of donors (ie. single vs multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
201740 Aerobic vs. anaerobic daily for 6 days Type of donor (eg. Relative vs. non-related): non-related no Rx as usual (12 weeks of education,
preparation: aerobic Initial dosage: 100 mL fresh Donor screening for mental illness or metabolic risk Bowel cleanse: nil specified behavioral strategies and oral
Fresh vs. frozen: frozen stool factors: Other: N/A laxatives), avoidance of probiotics.
Maintenance regime: same screened for metabolic syndrome and “any ongoing
diseases”
Vujkovic- Route of administration: Duration of intervention: Number of donors (ie. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
Cvijin colonoscopy once-off Type of donor (eg. Relative vs. non-related): non-related no treatment as usual
et al, Aerobic vs. anaerobic Initial dosage: 250 mL of Donor screening for mental illness or metabolic risk Bowel cleanse: bowel preparation used
201730 preparation: aerobic stool suspension factors: yes (Golytely)
Fresh vs. frozen: frozen Maintenance regime: N/A (Openbiome protocol used) Other: N/A
Johnsen Route of administration: Duration of intervention: Number of donors (ie. single vs. multiple): pooled (2 Antibiotic pre-treatment of recipients: Description of control intervention:
et al, colonoscopy once-off donors) and these were mixed. no same but with autologous FMT.
201741 Aerobic vs. anaerobic Initial dosage: not Type of donor (eg. Relative vs. non-related): non-related Bowel cleanse: Picoprep bowel
preparation: aerobic described Donor screening for mental illness or metabolic risk preparation used
Fresh vs. frozen: fresh or frozen Maintenance regime: same factors: screened for metabolic syndrome, and chronic Other: loperamide used prior to
GUT MICROBES
(1:1 ratio) Frequency of doses aerobic fatigue but not specifically for mental illness procedure
Kootte et al, Route of administration: NDT Duration of intervention: Number of donors (ie. single vs. multiple): single Antibiotic pre-treatment of recipients: Description of control intervention:
201736 Aerobic vs. anaerobic once-off Type of donor (eg. Relative vs. non-related): non-related no same plus autologous FMT prepared
preparation: partially Initial dosage: not clearly Donor screening for mental illness or metabolic risk Bowel cleanse: bowel lavage prior to in same way. Taken 6 hours prior to
anaerobic described factors: procedure FMT
Fresh vs. frozen: fresh Maintenance regime: N/A not screened for mental illness but screened for Other: fasting prior to procedure
metabolic risk factors
e1854640-7
(Continued)
e1854640-8
Table 1. (Continued).
Dosing regimen
Study (ie. dosage, frequency,
details Methodological factors duration) Donor factors Adjunctive treatment Control intervention
Rossen et al, Route of administration: NDT Duration of intervention: Number of donors (ie. single vs. multiple): single, except Antibiotic pre-treatment of recipients: Description of control intervention:
201537 Aerobic vs. anaerobic two FMTs given 3 weeks for 6 recipients no same plus autologous FMT prepared
preparation: aerobic apart Type of donor (eg. Relative vs. non-related): non-related Bowel cleanse: bowel preparation in same way. Taken on morning of
Fresh vs. frozen: fresh Initial dosage: 120 g fresh Donor screening for mental illness or metabolic risk (Moviprep) evening before and FMT
J. E. GREEN ET AL.
or both,14,45,47,48,53–58,60 nine were delivered endosco used related donors only,55,62 one used either,56
pically, either via nasojejunal tube,61nasoduodenal and four did not specify.46,57,63,64
tube39,40,42,43,52,62,63 or nasogastric tube;46 four were
delivered orally via encapsulated FMT;41,49,50,64 and Screening protocol
two studies used a mixed methodology of endo Donor screening protocols overall were incomple
scopic delivery or encapsulated FMT,59 or endo tely and poorly described. Where screening was
scopic route followed by encapsulated FMT.51 stated as occurring, the methods for screening
Twenty two studies used an aerobic preparation were frequently not provided. However, the more
of FMT,14,40,42,43,45–50,52,53,55–63 one used an anae recent studies tended to have better reporting of
robic preparation,54 and two did not specify.41,51 screening protocols and more comprehensive
Two used a semi-anaerobic preparation of FMT screening. Fourteen studies specifically screened
(33, 37) in which feces was exposed to some oxy for metabolic risk factors, but only seven specifi
gen during the procedure but attempts were made cally described screening for mental illnesses.
to minimize this; for example, prior to prepara
tion, feces was stored in oxygen-depleted saline
solution (36). Five studies used fresh feces, four Study results
used frozen feces, and two protocols allowed for Efficacy
use of fresh or frozen feces. One protocol did not Results were categorized by disorder and are summar
describe whether feces were fresh or frozen.63 ized in Table 2. Of the 26 included studies, 10 reported
significant results for their primary outcome mea
Dose sures, where these related to clinical efficacy. These
Dosing was inconsistently described. Twelve stu 10 studies related to functional gut disorders,14,40,61,64
dies did not provide clear information regarding Hepatitis B,63 IBD,53,54,57 antibiotic-resistant
amount of stool used. Fourteen studies reported on organisms,46 and metabolic syndrome.52 The evalu
the initial sample of fresh stool, whilst eight ated conditions were highly heterogeneous, even
described the amount of “suspension” used, which within groups. Nonetheless, it was possible to perform
consisted of filtered stool diluted with normal saline meta-analyses for two groups of disorders: IBS, and
and sometimes mixed with a cryoprotectant such as active UC.
glycerol. Doses of 12 g-250 g of fresh stool were
reported in the 14 studies that did provide these Inflammatory bowel disease
data. There were eight studies of IBD, six of which were
of active UC, the remaining two being of Crohn’s
Adjunctive treatments Disease (CD),45 and maintenance of remission in
A wide range of adjunctive treatments were UC.60 Sokol et al45 conducted a randomized, single-
employed. Fourteen studies used bowel blind, controlled trial comparing colonoscopic
preparation,14,39,42,43,45,47,52–56,58,60 and six studies FMT with placebo in 17 adults with CD. There
used antibiotics.44,46,48,55,56,59,62 Ren et al63 did not was a significant decrease in CD symptoms in the
state whether bowel preparation was used and FMT group compared with placebo (p = .03).45
Herfarth et al51 did not report whether any adjunc Sood et al60 conducted a double-blind, rando
tive treatments were used. mized-controlled trial (RCT) of colonoscopic
FMT compared with placebo as maintenance treat
Donor methods ment for inactive UC. The study did not find
Four studies used multiple donors (i.e. a pooled a significant difference in the primary outcome
sample),14,53,54,64 two did not adequately describe measure (steroid-free clinical remission) between
whether single or multiple donors were used,46,63 groups (p = .111); however, significant between-
and the remaining 20 studies used single donors. group differences were reported in endoscopic
Nineteen studies used non-related donors, two remission (p = .026), histological remission
Table 2. Summary of primary outcomes of included studies.
Primary outcome measure (POM) relating to clinical efficacy
e1854640-10
Johnsen Study design: Double blind, randomized Specific disorder: IBS Intervention: bowel preparation and loperamide given POM: Clinical response, defined as symptom relief of more Favors FMT
et al, placebo controlled parallel group, (excluded prior to frozen FMT via colonoscopy. than 75 points assessed by IBS-SSS, 3 months after FMT. 36/55 (65%) participants in the FMT group compared
201741 single center trial. 2:1 randomization dominating Control: Autologous FMT. with12/28 (43%) in the placebo showed a clinical
Country: Norway constipation group) response at 3 months (p = 0 · 049).
Sample size (n): N = 83 (FMT = 45, Age in years (median):
control = 29) FMT group: 43
Follow-up period: 12-months Control group: 45
Halkjaer Study design: double blind placebo Specific disorder: IBS Intervention: 25 capsules of frozen encapsulated FMT POM: Reduction of IBS-SSS between baseline and 3-month Favors placebo
et al, controlled trial for FMT capsules in IBS. Age in years (mean): daily for 12 days while fasting. Bowel preparation follow up in the treatment group compared with the There was a significant difference in change in IBS-SSS
45
2018 1:1 randomization FMT group: 37.28 given the day prior to first treatment. placebo group. groups between the FMT and placebo groups favoring
Country: Denmark Control group: 35.54 Control: As above, but with placebo capsules used the placebo group (p = .012).
Sample size (n): n = 52 (FMT = 26,
control = 26)
Follow-up period: 6-months
Aroniadis Study design: double blind, randomized, Specific disorder: IBS Intervention: 25 capsules of FMT (daily) for 3 days plus POM: Difference in IBS-SSS between the groups at 12 weeks No significant result
et al, placebo controlled crossover trial Age in years (mean): TAU. Proton pump inhibitor prior to FMT. The difference in IBS-SSS and psychiatric outcome
2019 Country: America FMT group: 33 Control: Treatment as usual, plus placebo. measures (HADS) scores between the FMT group and
Sample size (n): n = 48 (FMT = 25, Control group: 42 the control group were not significant.
control = 23)
Follow-up period: 12–24 weeks
El-Salhy Study design: a double blind, randomized, Specific disorder: IBS Intervention: TAU plus NGT FMT of 30 or 60 g. POM: Reduction in the IBS-SSS total score of ≥50 points at Favors FMT
et al, placebo-controlled study Age in years (mean): Control: As above, but 30 g autologous FMT used. 3 months following transplantation Responses occurredafter 3 months 30 g FMT and 60 g FMT
2019 Country: Norway 30 g FMT group: 39.2 groups (p < .0001).
Sample size (n): N = 165 (30 g FMT = 55, 60 g FMT group: 39.3 There was a significant improvement in the mental health
60 g FMT = 55, control = 55) Control group: 41.2 sub-score of FAS in both groups at 3 months compared
Follow-up period: 3 months with placebo (p < .05).
Holster et al, Study design: a double blind, randomized, Specific disorder: IBS Intervention: Bowel preparation followed by POM: Effect on IBS symptoms using the IBS version of the No significant result
2019 placebo-controlled study Age in years (median): colonoscopic FMT. GSRS-IBS. No significant differences in GSRS-IBS scores, anxiety or
Country: Sweden FMT group: 34 Control: As above, but autologous FMT. depression symptoms (HADS) between the allogenic
Sample size (n): N = 16 (FMT = 8, Control group: 39 and autologous groups were found.
control = 8)
Follow-up period: 6 months
Hepatic disorders
Philips et al, Study design: Retrospective cohort study Specific disorder: Intervention: Fresh FMT daily for 7 days via NDT. Fasting POM: Survival at 90 days. No significant result
201814 Comments on study design: four arms of Severe alcoholic before and after procedure. There was no significant difference in 90-day survival
study and unclear allocation of groups hepatitis Control: Three control groups consisting of alternative between the groups.
Country: India Age in years (mean): treatments were used (steroids, nutrition or
Sample size (n): N = 51 (FMT = 16, steroids FMT group: 47.6 pentoxifylline therapy)
8, nutritional support 17, pentoxifylline Control groups:
10) steroids 44.3, nutrition
Follow-up period: 90 days 49.6, pentoxifylline
48.7
(Continued)
Table 2. (Continued).
Primary outcome measure (POM) relating to clinical efficacy
(or relevant secondary outcome measures where POM did
Author/date Study details Population details Description of Intervention not relate to efficacy) Results for outcome measures of clinical efficacy
Ren et al, Study design: case controlled, single blind, Specific disorder: Intervention: FMT via NDT delivered every 4 weeks (for POM: HbeAg clearance, defined as the loss of HbeAg. Favors FMT
201746 open-label pilot trial Hepatitis B 1–7 treatments) plus TAU 4/5 participants in FMT group achieved clearance of
Comments on study design: Data analyst Age in years (median): Control: TAU HbeAg, compared to 0/13 of the control group
blinded to study design, participants FMT group: 27 (p = .0002).
allocated into groups based on their Control group: 33
treatment preferences (for FMT or not)
Country: China
Sample size (n): N = 18 (FMT = 5,
control = 13)
Follow-up period: 32–36 weeks
Bajaj et al, Study design: An open-label, randomized Specific disorder: Intervention: TAU, plus 5 days of pre-treatment with POM did not relate to clinical efficacy. POM not relevant to clinical efficacy
2017 clinical trial Recurrent hepatic antibiotics (metronidazole, ciprofloxacin and Relevant clinical outcomes included: changes in cognitive A significant improvement was observed in cognitive
Country: America encephalopathy amoxicillin), followed by a 12 hour washout, then FMT function at day 20, cirrhosis severity (MELD score, outcomes for FMT compared with control on both PHES
Sample size (n): N = 20 (FMT = 10, Age in years (mean): via retention enema. albumin) and changes in liver function total score (p = .003) and EncephalApp Stroop (p = .01).
control = 10) FMT group: 64.5 Control: TAU No significant changes were observed in cirrhosis severity
Follow-up period: 12–15 months Control group: 62.9 or liver function for either group.
Bajaj et al, Study design: A Phase 1, Randomized, Specific disorder: Intervention: 15 capsules of frozen, encapsulated FMT. POM did not relate to clinical efficacy. POM not relevant to clinical efficacy
2019 single-blind, Placebo-Controlled Trial Age in years (mean): Control: As above, but placebo capsules used, containing Relevant clinical outcomes included: changes in cognitive Following intervention, a significant improvement in
Country: America FMT group: 63.3 80% cocoa butter, 20% glycerol and brown food testing on PHES and EncephalApp, serum LBP, and cognitive function was observed in FMT group, but not
Sample size (n): N = 20 (FMT = 10, Control group: 64.2 coloring. changes in duodenal mucosal expression of the control group compared to baseline for
control = 10) inflammatory cytokines, barrier proteins, and AMPs EncephalApp score (p = .02), but not PHES score.
Follow-up period: 5 months A significant post-treatment reduction in LBP was also
observed in the FMT group (P < .009), but not the
control group.
No significant change in duodenal expression of
inflammatory cytokines, barrier proteins or AMPs was
observed.
Inflammatory Bowel Disease
Ishikawa Study design: open-label, non-randomized Specific disorder: UC Intervention: 2 weeks of antibiotics followed by fresh POM: Clinical response (CAI of <10, and decrease of 3 or less) No significant result
et al, prospective control study Age in years (mean): colonoscopic FMT (bowel preparation given prior), and clinical remission (CAI of 3 or less) at baseline No significant difference was observed either in clinical
201739 Country: Japan FMT group: 40.4 followed by scopolamine. compared with the four week follow up. response or clinical remission between the FMT and
Sample size (n): N = 41 (FMT = 21, Control group: 44.7 Control: antibiotics only control groups.
control = 20)
Follow-up period: 4 weeks
Rossen et al, Study design: Single center, randomized Specific disorder: Active Intervention: 2 doses fresh FMT via NDT 3 weeks apart, POM: Clinical remission at 12 weeks (SCCAI score < or = 2, No significant result
201537 (1:1), double blind trial UC each preceded by bowel preparation. and > or = 1 point improvement on the combined Mayo There was no significant difference between the FMT group
Country: The Netherlands Age in years (mean): Control: As above, but autologous FMT used. endoscopic score of sigmoid and rectum) compared to and control groups regarding clinical remission at
Sample size (n): N = 50 (FMT = 25, Data not provided baseline. 12 weeks.
control = 25)
Follow-up period: 12 weeks
Kump et al, Study design: Open label, prospective, Specific disorder: Intervention: Pre-treatment with antibiotics followed by POM: Mayo Score at the 90-day follow up point between the P-values/significance not described.
GUT MICROBES
201742 non-randomized controlled study Therapy-refractory, 5 treatments of fresh FMT given 14 days apart. Bowel two groups, wherein a reduction in total Mayo score by 10/17 (59%) of the FMT group achieved a clinical response
Comments on study design: means of active UC preparation given prior to first FMT only. First FMT three or more was considered a clinical response, and and four participants (24%) achieved clinical remission
allocation not described Age in years (mean): delivered colonoscopically, and subsequent FMTs via a score of two or less was considered remission compared to 1/10 (10%) in the control group achieving
Country: Austria FMT group: 44 sigmoidoscopy. partial response. P-values not provided, hence
Sample size (n): N = 27 (FMT = 17, Control group: 36 Control: Antibiotics only. significance unclear.
control = 10)
Follow-up period: 90 days
e1854640-11
(Continued)
Table 2. (Continued).
Primary outcome measure (POM) relating to clinical efficacy
(or relevant secondary outcome measures where POM did
e1854640-12
Author/date Study details Population details Description of Intervention not relate to efficacy) Results for outcome measures of clinical efficacy
Moayyedi Study design: double blind, placebo Specific disorder: Active Intervention: Fresh or frozen FMT given as retention POM: UC remission (Mayo score <3 and endoscopic Mayo Favors FMT
et al, controlled, parallel design study (1:1) UC enema weekly for 6 weeks. score 0) at week 7 compared with baseline. At week 7, UC remission in the FMT group was 9/38 (24%)
201543 Country: Canada Age in years (mean): Control: As above, but water used as FMT placebo. compared with 2/37 (5%) in the placebo group
Sample size (n): N = 75 (FMT = 38, FMT group: 42.2 (p = .03).
control = 37) Control group: 35.8
Follow-up period: 12 months
Paramsothy Study design: multicentre, double-blind, Specific disorder: Active Intervention: Frozen FMT delivered via colonoscopy POM: Steroid free clinical and endoscopic remission (total Favors FMT
et al, randomized, placebo-controlled parallel UC on day 1, followed by daily self-administered enemas Mayo score ≤2, with all sub-scores ≤1, and ≥1 point At week 8, remission rates in the FMT group were 27%
201744 design (1:1) Age in years (median): of frozen FMT delivered 5 times per week for 8 weeks. reduction from baseline in endoscopy sub-score) at week compared with 8% in the placebo group, a RR of 3.6,
J. E. GREEN ET AL.
Country: Australia FMT group: 35.6 Control: As above, but placebo colonoscopy and enema 8 compared with baseline. (95% CI 1.1–11.9, p = .021).
Sample size (n): N = 81, FMT = 41, Control group: 35.4 used (saline and glycerol).
control = 40
Follow-up period: 8 weeks
Sokol et al, Study design: a multicentre, randomized, Specific disorder: Intervention: Bowel preparation followed by POM did not relate to clinical efficacy. POM not relevant to clinical efficacy
2020 single-blind placebo-controlled trial Crohn’s disease colonoscopic FMT. Relevant clinical outcomes included: clinical flare rate, The CDEIS decreased significantly 6 weeks after FMT
Country: France Age in years (mean): Control: As above, but “physiological serum” used as change in CDEIS, CRP level, leukocyte level, or fecal (p = .03) but not after sham.
Sample size (n): N = 17 (FMT = 8, FMT group: 31.5 placebo FMT. calprotectin. There was no significant difference in clinical fare rate, fecal
control = 9) Control group: 34 calprotectin, leukocyte level or CRP level between
Follow-up period: 24 weeks groups.
Costello Study design: a multi-center, double blind, Specific disorder: Active Intervention: 3 L polyethlene glycol bowel preparation POM: Steroid-free remission at week 8 defined as 1. Total Favors FMT
et al, randomized, controlled trial UC given the night before, and 2 mg loperimide Mayo score of ≤ 2 AND 2. Mayo endoscopic score of ≤ 1 The primary outcome was achieved in 12/38 (32%) of the
2019 Country: Australia Age in years (median): immediately prior. FMT consisted of frozen pooled donor FMT group compared with 3/35 (9%) of
Sample size (n): N = 73 (FMT = 38, FMT group: 38.5 donor stool via colonoscopy followed by 2 enemas autologous FMT group (p = .03).
control = 35) Control group: 35 on day 3 or 4 and one on day 6 or 7.
Follow-up period: 12 months Control: As above, but with autologous FMT
Sood et al, Study design: a pilot double blind, Specific disorder: Intervention: FMT delivered via colonoscopy every POM: Maintenance of steroid-free clinical remission (Mayo No significant result for POM
2019 randomized, placebo controlled study Inactive UC 8 weeks for 6 treatments plus TAU. Bowel preparation score ≤2, all sub scores ≤ 1) at week 48. Nil significant difference in maintenance of steroid free
Country: India (maintenance) with polyethylene glycol lavage the night prior. Relevant secondary clinical end points included: clinical remission between groups.
Sample size (n): N = 61 (FMT = 31, Age in years (mean): Control: As above, but placebo colonoscopy given (saline achievement of endoscopic remission (endoscopic Mayo However, significant results were achieved for endoscopic
control = 30) FMT group: 33 with food dye). score 0), histological remission (Nancy grade 0, 1) and remission (p = .026), histological remission (p = .033),
Follow-up period: 48 weeks Control group: 34.6 change in inflammatory markers (ESR and CRP) at week and change in inflammatory markers (p < .001).
48.
Antibiotic-resistant organisms
Saidani et al, Study design: A matched case-control Specific disorder: CPE Intervention: FMT via NGT, plus a nasopharangeal POM: Delay in negativation of rectal-swab cultures. Favors FMT
2019 retrospective study (2 controls per case) Age in years (mean): decolonization (8 days prior), bowel wash (5 days prior At day 14 post-FMT, 8/10 treated patients (80%) achieved
Country: France FMT group: 59.2 and 1 day prior), antibiotics for 5 days prior, and the POM, compared with 2/20 (10%) of the control
Sample size (n): N = 30 (FMT = 10, Control group: 60.3 proton pump inhibitor (1 day prior and day of FMT). group (p < .001) in the clearance rate between both
control = 20) Control: TAU groups.
Follow-up period: 6 months
Huttner Study design: A multi-center Specific disorder: CPE Intervention: Colistin sulfate and neomycin sulfate tablets POM: Detectable intestinal carriage of ESBL/CPE by stool No significant result for POM
et al, (International) randomized, open-label, and ESBL for 5 days followed by FMT (either capsules or NGT). culture 35–48 days after randomization Nil significant difference between groups in intestinal ESBL
2019 superiority trial Age in years (median): Control: TAU or CPE rates following treatment.
Country: International (Swizerland, France, FMT group: 70
Israel, The Netherlands) Control group: 64
Sample size (n): N = 39 (FMT = 22,
control = 17)
Follow-up period: 150–210 days
Obesity or metabolic syndrome
(Continued)
Table 2. (Continued).
Primary outcome measure (POM) relating to clinical efficacy
(or relevant secondary outcome measures where POM did
Author/date Study details Population details Description of Intervention not relate to efficacy) Results for outcome measures of clinical efficacy
Kootte et al, Study design: Double blind, randomized Specific disorder: Intervention: Participants were fasted and received bowel POM did not relate to clinical efficacy. POM not related to clinical efficacy
201736 controlled trial of obese metabolic Metabolic Syndrome preparation prior to fresh FMT via NDT. Relevant clinical outcomes included: metabolic changes, Significant improvement in insulin sensitivity was observed
syndrome subjects Age in years (median): Control: As above, but participants received autologous insulin sensitivity and plasma metabolites at 6 and in FMT group at 6 weeks (p < .05), but not at 18 weeks.
Country: The Netherlands FMT group: 54 FMT. 18 weeks following FMT.
Sample size (n): N = 44 (FMT = 26, Control group: 54
control = 12)
Follow-up period: 18 weeks
Vrieze et al, Study design: a double blind, randomized Specific disorder: Intervention: Participants were fasted from the night POM: Change in insulin sensitivity at 6 weeks. Favors FMT
2012 controlled pilot study metabolic syndrome before. Bowel lavage with polyethylene glycol Peripheral insulin sensitivity improved at week 6 compared
Country: The Netherlands (insulin sensitivity) solution given prior to FMT via NDT. with baseline for the FMT group (p < .05), but not the
Sample size (n): N = 18 (FMT = 9, Age in years (mean): Control: As above, but autologous FMT used. control group.
control = 9) FMT group: 47 There was no significant change in hepatic insulin
Follow-up period: 6 weeks Control group: 53 sensitivity at week 6, diet composition, resting energy
expenditure, or counter-regulatory hormones.
Smits et al, Study design: a double blind, randomized Specific disorder: Intervention: Bowel lavage, followed by FMT via NDT. POM: TMAO production (as a possible indicator for No significant change in POM
2019 controlled pilot study metabolic syndrome Control: As above, but autologous FMT used. cardiovascular disease risk) At 2 weeks, there was no significant difference from
Country: The Netherlands (TMAO production) baseline in fasting plasma TMAO, 24 hour urinary TMA
Sample size (n): N = 20 (FMT = 10, Age in years (mean): excretion, 24 hour urinary TMAO excretion, plasma d3-
control = 10) FMT group: 52.3 carnitine appearance or 24 hour urinary d3-TMA
Follow-up period: 2 weeks Control group: 57.7 excretion for FMT or control groups.
Allegretti Study design: a double blind, randomized, Specific disorder: Intervention: Initial dose of 30 FMT capsules and POM did not relate to clinical efficacy. POM not related to clinical efficacy
et al, placebo-controlled, pilot study obesity without a maintenance dose of 12 capsules at week 4 and Relevant clinical outcomes included: Obesity related There was a significant between group change in area
2019 Country: America metabolic syndrome week 8. biomarkers such as change in weight or short-chain fatty under the curve for leptin, with a larger increase in the
Sample size (n): N = 22 (FMT = 11, Age in years (mean): Control: As above, but placebo capsules used. acids, and change in area under the curve for GLP1 or placebo group at week 12 compared with baseline
control = 11) FMT group: 44.5 leptin at week 12 (p = .001).
Follow-up period: 26 weeks Control group: 43.2 There was no significant change in mean BMI, or area
under the curve for GLP1, or short chain fatty levels at
week 12 for either group.
Other disorders
Vujkovic- Study design: open label, non-randomized, Specific disorder: HIV Intervention: Bowel preparation given prior to frozen POM is not clearly stated. Relevant clinical outcomes No significant result
Cvijin prospective controlled study Age in years (median): colonoscopic FMT. included: HIV-associated markers of inflammatory Nil significant changes in any of the inflammatory markers
et al, Comments on study design: Participants FMT group: 61 Control: TAU activation (CD38, HLA-DR, CD8 + T-cells, plasma rations between FMT and control groups was observed at any
201730 were selected for having low CD4 Control group: 64 of kynurenine to tryptophan, IL-6, sCD14) over time (from of the follow up points compared with baseline.
counts on ART (but excluded if they baseline up to 24 week follow up).
were too low)
Country: America
Sample size (n): N = 8(FMT = 6, control = 2)
Follow-up period: 24 weeks
Herfarth Study design: a prospective, placebo Specific disorder: Intervention: Two boluses of endoscopic FMT followed by POM did not relate to clinical efficacy. POM not related to clinical efficacy
et al, controlled, double blind randomized, antibiotic- daily dosing of 6 FMT capsules for 14 days. Relevant clinical outcomes included: clinical remission All patients experienced relapse (ie. remission rate of zero
GUT MICROBES
2019 controlled trial dependant pouchitis Control: As above, but placebo FMT used. (defined as an mPDAI < 4 and no need for antibiotics in for both groups).
Country: America Age in years (mean): weeks 4, 8, and 16) and change in fecal calprotectin level There was no significant change in fecal calprotectin levels
Sample size (n): n = 6 (FMT = 4, control = 2) FMT group: 39.25 as data were only available for 5 participants.
Follow-up period: 16 weeks Control group: 33.5
POM – Primary Outcome Measure, TAU – treatment as usual, UC- Ulcerative Colitis, ASD – Autism Spectrum Disorder, RCT – Randomized Controlled Trial, NDT – Nasoduodenal Tube, NJT – Nasojejunal Tube, NGT – Nasogastric Tube, FMT – Fecal Microbiota Transplant, ESR –
Erythrocyte sedimentation rate, CRP – C-Reactive Protein, IBS – Irritable Bowel Syndrome, CDEIS-Crohn’s Disease Endoscopic Index of Severity, CPE – Carbapenemase-Producing Enterobacteriaceae, ESBL – Extended spectrum beta-lactamase, HIV – Human
Immunodeficiency Virus, HADS – Hospital Anxiety and Depression Scale, FAS – Fatigue Assessment Scale, GSRS – Gastrointestinal Symptom Rating Scale, DSR – Daily Stool Records, SCCAI – Simple Clinical Colitis Activity Index, CAI – Clinical Activity Index, IBS-SSS – Irritable
e1854640-13
Bowel Syndrome Severity Scoring System, HbeAg – Hepatitis B-e Antigen, mPDAI – modified pouch activity
e1854640-14 J. E. GREEN ET AL.
(0.033), and change in inflammatory markers and response are summarized in Supplementary
(p < .001) favoring FMT.60 Table 2.
Meta-analysis confirmed that FMT was asso
Meta-analysis for ulcerative colitis subgroup ciated with a significant improvement in clinical
Six studies reported on active UC, which was suffi remission rates in UC compared to control condi
cient to perform a meta-analysis for clinical remis tions (OR = 3.634, 95% CI = 1.940 to 6.808, n = 6
sion, clinical response, endoscopic remission, and studies, I2=0%, p < .001) (see Figure 2). FMT was
endoscopic response. Outcome measures were het also associated with a significant improvement in
erogeneous, were collected at different time points clinical response rates in UC compared to control
(between 7 weeks and 90 days), and used differing (OR = 2.634, 95% CI = 1.441 to 4.815, n = 6 studies,
definitions of clinical response/remission and I2=33%, p = .002) (see Figure 3), as well as for
endoscopic remission/response. Five of six used endoscopic remission rates (OR = 4.431, 95%
Mayo score, whilst one used Clinical Activity CI = 1.901 to 10.324, n = 5 studies, I2=0%, p
Index (CAI) score. Definitions and data for clinical = .001) (see Figure 4). However, FMT showed no
remission and response are summarized in significant improvement in endoscopic response
Supplementary Table 1, and endoscopic remission rates in UC compared to controls (OR = 1.065,
95% CI = 0.432 to 2.625, n = 2 studies, I2=0%, p health subscale of the Fatigue Assessment Scale
= .892) (see Supplementary Fig 1). (FAS), and reported a significant difference
between the means of the group who received
Functional gut disorders 30 g FMT (13.3, s.d. 3.1) compared with the placebo
In a trial of nasojejunally delivered FMT given daily group (14.7, s.d. 3.4) at 1 month (p < .05), and the
for six days in adjunct to treatment as usual (TAU) group who received 60 g FMT (13.1, s.d. 3.1), com
for slow-transit constipation, Tian et al61 reported pared with the placebo group (14.5, s.d. 2.7) at
a clinical remission rate of 36.7% for the FMT 3 months (p < .05) in favor of FMT.
group compared with 13.3% for the TAU control
group (p = .04). Hepatic disorders
Five studies reported on IBS, which was suffi Of the four studies in hepatic disorders, three had
cient to perform a meta-analysis for clinical significant results for clinical efficacy favoring FMT
response and average change in IBS-SSS. Different over control, whilst the fourth did not report signifi
studies used different definitions of clinical cant outcomes. In a trial of nasoduodenally delivered
response: four used IBS-SSS, and one used GSRS- FMT every 4 weeks (for 1–7 treatments) plus TAU for
IBS, mostly at 3 months. Definitions and data for Hepatitis B, Ren et al63 reported that four of the five
clinical response and change in IBS-SSS are sum participants achieved clearance of HbeAg, whereas all
marized in Supplementary Table 3. 13 of the TAU controls continued to have a positive
Meta-analysis revealed no significant difference HbeAg titer (p = .0002). Bajaj et al48 conducted an
in IBS-SSS (Hedge’s g = 0.282, 95% CI = −1.373 to open-label RCT investigating FMT for recurrent
1.937, n = 3 studies, I2=97%, p = .739) or clinical hepatic encephalopathy using a retention enema
response (OR = 1.699, 95% CI = 0.273 to 10.588, compared with TAU and reported a significant
n = 5 studies, I2=92%, p = .739) following FMT improvement in two measures of cognitive out
compared to control (see Supplementary material; comes in favor of FMT (p < .01 for both). Similarly,
Figures 2 and 3) in a single-blind RCT of encapsulated FMT com
pared with placebo capsules, Bajaj et al49 found
Psychiatric outcomes a significant improvement in cognitive outcomes
Only three studies assessed psychiatric outcomes for the FMT group but not the placebo group
and all three were conducted in IBS populations. (p = .02). Philips et al62 conducted a retrospective
Two of the studies used the Hospital Anxiety and cohort study comparing FMT with three control
Depression Scale (HADS) to measure depression groups (steroids, nutritional support or pentoxi
and anxiety symptoms and neither reported fylline) for the treatment of severe alcoholic hepa
a significant change in symptoms between groups titis, but found no significant improvement in the
post-intervention. The third study used the mental primary outcome of 90-day survival (p = .179).
e1854640-16 J. E. GREEN ET AL.
Metabolic syndrome or obesity without metabolic over control and the other without significant out
syndrome comes. In a retrospective matched case-control
Four studies evaluated FMT for the treatment of study of nasogastric FMT compared with TAU for
metabolic syndrome or obesity without metabolic Carbapenemase-Producing Enterobacteriaceae
syndrome, and only one of these had significant (CPE), Saidani et al46 reported a significant delay
results for clinical efficacy regarding the primary in negativation of rectal swab cultures 2-weeks
outcome. The remaining three showed significant post-FMT compared with TAU (p < .001).
results for secondary outcome measures, all in favor Huttner et al59 conducted a multicentre, rando
of FMT. Vrieze et al52 conducted a double-blind mized, open-label, superiority trial of nasogastric
pilot RCT of nasoduodenally delivered FMT com or encapsulated FMT (treatment was site depen
pared with autologous FMT for metabolic syn dant), compared with TAU for CPE and Extended
drome and reported a significant improvement in spectrum beta-lactamase (ESBL), but did not iden
week 6 peripheral insulin sensitivity (p < .05) in tify a significant between-group difference in the
favor of FMT, but not in hepatic insulin sensitivity primary outcome measure for clinical efficacy
(p = .08), diet composition, resting energy expen (p-value not provided).
diture, or counter-regulatory hormones. In
a double-blinded RCT examining nasoduodenally
Other conditions
delivered FMT for metabolic syndrome, Kootte et
Two studies were not able to be grouped with the
al39 did not find significant differences in their
others. They evaluated FMT for the treatment of
primary outcome measure (change in intestinal
individuals with HIV and antibiotic-dependant
microbiota in relation to insulin sensitivity at
pouchitis, respectively. Neither showed clinical
18 weeks), nor did they observe a significant change
efficacy.
in BMI or SCFA levels at any study time point. In
terms of secondary outcomes, change in fecal
microbiota composition at 6 weeks associated Safety data
with improved peripheral insulin sensitivity (from
25.8 [19.3–34.7] to 28.8 [21.4–36.9] mmol kg/ There were variable quality and completeness of
1 min/1, p < .05) in the allogenic FMT group, reporting of safety data for both serious adverse
whereas autologous FMT had no effect (from 22.5 events (SAEs) and mild to moderate AEs, across
[16.9–30.2] to 20.8 [17.6–29.5] mmol kg/1 min/ studies (see Table 3). Studies had a follow up period
1, p > .5). ranging from four weeks55 to 1 year .14 SAEs
Smits et al43 conducted a double-blind pilot RCT Of the 26 included studies, 23 provided clear
of nasoduodenally delivered FMT compared with descriptions of SAEs. A total of 69 SAEs were
autologous FMT for TMAO production in partici reported from 12 studies; 26 occurred in partici
pants with metabolic syndrome and did not find pants allocated to receive FMT, and 43 in partici
a significant difference between groups. In pants in the control groups (see Supplementary
a double-blind pilot RCT of encapsulated FMT Table 4). Of the 26 SAEs that occurred in partici
compared with placebo capsules for obesity- pants allocated to receive FMT, all but one was
related biomarkers in participants with obesity but deemed unlikely to be related to the intervention.
without metabolic syndrome, Allegretti et al found
a significant between-group difference in area Table 3. Completeness of reporting of AE.
under the curve at week 12 for leptin compared Number reported (total studies,
Completeness of reporting of AE n = 26)
with baseline (p = .001), but no significant change
Mild-moderate AE
for other biomarkers of obesity.50 Detailed reporting of AE 9 (34.6%)
Generic statement only or limited 9 (34.6%)
reporting of AE
Antibiotic-resistant organisms Not reported at all 8 (30.8%)
Two studies evaluated FMT for the treatment of SAE
Clearly described 23 (88.5%)
colonization of antibiotic-resistant organisms, one Not reported at all or not clearly 3 (11.5%)
demonstrating significant clinical efficacy of FMT reported
GUT MICROBES e1854640-17
Twenty of these SAEs occurred in participants who analysis reported change in microbiome following
received FMT via colonoscopy or enema, and six in FMT. Fourteen of the 23 studies reported whether
those receiving FMT endoscopically or via capsules. the change in microbiota was toward the donor and,
When broken down by specific disorder, 17 of these of these, 11 confirmed that the recipient microbiome
SAEs occurred in participants with inflammatory did move toward the donor microbiome. The
bowel disease, three in participants with hepatic remaining three studies did not report significant
encephalopathy, five in participants with antibiotic results.
resistant organisms and one in a participant Reporting on the extent or significance of micro
with IBS. biota changes was inconsistent across studies and
the complexity of microbiome data analysis has
meant it was not possible to answer the question
Mild to moderate AEs
of the extent to which the recipient microbiome
Due to the inconsistent quality and completeness of changed toward the donor, as no clear quantifica
reporting of mild to moderate AE, it was only tion was provided by the included studies. As such,
possible to pool/summarize data for a small num these data are not reported in Table 4.
ber of included studies (see Supplementary Regarding longevity of the observed changes in
Table 5). These studies were related to IBS, recipient microbiota, it was not possible to answer
(n = 4), UC (n = 2), slow transit constipation this question in this review, as included studies
(n = 1), hepatic encephalopathy (n = 1), and meta either did not follow-up recipients for long enough,
bolic syndrome (n = 1). As such a cross-indication or did not measure microbiota changes frequently
assessment of adverse events was not possible as the enough to be able to state the duration for which
data were insufficiently reported across disorders. any changes were observed. However, with these
Similar rates of mild to moderate AEs were limitations in mind, it appears that the demon
observed in participants allocated to FMT com strated microbiome changes were transient and
pared to the control groups (see Supplementary appeared to last between 2 weeks to 1 year follow
Table 5). However, the following AEs were more ing the intervention.
common in participants receiving allogenic FMT
compared with those allocated to control groups: Correlation of “engraftment” with clinical findings
nausea (reported in 80% of FMT recipients com Fourteen of 23 studies reported on associations
pared with 72% in control groups), constipation between “engraftment” and clinical outcomes, and
(reported in 17.4% of FMT recipients compared of these, 12 studies had statistically significant results
with 2.4% in control groups), diarrhea (reported with 10 reporting a significant association between
in 16.8% of FMT recipients compared with 6.7% successful engraftment and clinical efficacy and two
in control groups), transient, or low-grade fever reporting no association between efficacy and engraft
(reported in 8.4% of FMT recipients compared ment. These data are summarized in Table 4.
with 3.0% in control groups) and vomiting
(reported in 5.9% of FMT recipients compared
Risk of bias assessment
with 2.9% in control groups).
Incomplete reporting precluded comparison of According to the Cochrane Risk of Bias tool36 (see
AE rates between different routes of FMT; however, Supplementary Table 6), nine studies were evaluated
encapsulated FMT appears to have been the best- as “low risk,”14,40–42,49,53,54,58,60,64 six as “some
tolerated route. concerns,”39,42,43,45,50,52,57 and five as “high
risk.”48,57–59,61 Studies rated “high risk” were: Tian
Successful microbial “engraftment” et al,61 due to incomplete reporting across most
Microbiome analysis pre- and post-FMT was per domains and inadequate randomization processes;
formed in 23 of the 26 included studies. All micro Moayyedi et al57 due to likely inadequacy of blinding
biome analyses used 16s RNA sequencing. The data of participants as water enemas were used as placebo,
relating to “engraftment” are summarized in Table 4. which would likely be easily differentiated from true
All 23 of 23 studies which measured microbiome FMT by recipients; Bajaj et al48 and Huttner et al59 as
e1854640-18
El Salhy et al, 201949 Yes Not described Appeared changed at one month (only data point). Yes
Aroniadis et al, 201934 Yes Yes Changes appeared to persist for 12 weeks (final data point). No
Costello et al, 201948 Yes Yes Changes appeared to persist for 8 weeks, reduced by 12 months. Yes
Sood et al, 201954 Not measured N/A N/A N/A
Herfarth et al, 201935 Yes Only in one of six Not clearly described. Yes
recipients
Allegretti et al, 201965 Yes Yes Changes appeared to persist for 12 weeks (final data point). Not described
Vrieze et al, 201247 Yes Not described Appeared changed at six weeks (only data point). Yes
Smits et al, 201838 Yes Yes, “in some but not Appeared changed at two weeks (only data point). Not described
all participants” –
further detail not
provided.
Bajaj et al, 201732 and long term Yes Yes Changes appeared to persist for over one year (reported in long term Not described
data reported in Bajaj et al, paper).
201956
Bajaj et al, 201933 Yes Not measured Appeared changed at day 30 (only data point). Not described
Sokol et al, 202051 Not significant overall. However when data from Not significant Not significant overall, however when data was corrected for 2 participants Yes
2 participants was removed, a significant who were considered “treatment failures”, duration of changes were
change was observed in the remaining 8 significant at 6 weeks for the FMT group, but microbiota were
participants considered back to baseline at week 14.
30
Vujkovic-Cvijin et al, 2017 Yes Yes Changes in microbiota were most significant between 2–4 weeks, and less Not described
significant by week 8.
Ren et al, 201746 Yes Not described Not described. Not described
Philips et al, 201814 Yes Not described Changes appeared to persist for up to 90 days (study duration). Not described
Kump et al, 201742 Yes Yes Changes appeared to persist for 90 days (final data point). No
Ishikawa et al, 201739 Yes Not described Appeared changed at 4 weeks, but not measured beyond that. Yes
Tian et al, 201740 Not measured N/A N/A N/A
Kootte et al, 201736 Yes Not described Changes demonstrated at 6 weeks. No changes apparent at 18 weeks. Yes
Paramsothy et al, 201744 Yes Yes Changes persisted for 8 weeks after intervention finished (final data point). Yes
Johnsen et al, 201741 Not measured N/A N/A N/A
Halkjaer et al, 201845 Yes Yes 3 months (final data point). Not significant
Rossen et al, 201537 Yes Yes, but only in Changes were observed at 12 weeks (final data point). Yes
responders
Moayyedi et al, 201543 Yes Not significant Changes were observed at 6 weeks (only data point). Not significant
GUT MICROBES e1854640-19
the studies were open label, with an absence of blind other than CDI. This review aimed to recruit higher
ing; and Herfarth et al,51 due to an absence of quality studies by excluding uncontrolled studies,
a statistical pre-analysis plan, and the fact that the which represent a majority of studies in this field.
trial was ceased after only six participants were Whilst other reviews have been conducted with
randomized. respect to safety of FMT for indications other
than CDI, these reviews are either not recent,65 or
were restricted to a single indication such as
Discussion
IBS29–31 and IBD.32–34 With respect to efficacy,
Statement of principal findings whilst other reviews have been published for single
indications, such as IBS29–31 and IBD,32–34 there
This review identified FMT trials for conditions other
have been no holistic reviews looking at all indica
than CDI, with promising, albeit mixed, outcomes
tions other than CDI. As far as the authors of this
regarding efficacy and safety. Meta-analysis of UC
review are aware, this review also represents the
studies found FMT to be superior to control condi
most up to date systematic review and meta-
tions for active disease in terms of endoscopic remis
analysis of the safety and efficacy of FMT for IBS.
sion, clinical remission, and clinical response. In
However, the 26 studies included were hetero
contrast, meta-analysis of the five IBS studies did
geneous and of mixed quality, with several using
not yield significant results regarding symptoms or
open-label designs and small samples.
clinical response. Regarding clinical efficacy in other
Encouragingly, more recently published studies
applications of FMT, studies were too heterogeneous
appear to be of higher quality, using more robust
to perform meta-analyses, but four yielded evidence
study designs (such as double-blinded RCTs), and
of clinical efficacy in slow-transit constipation,
with clearer and more complete descriptions of
Hepatitis B, colonization of CPE, and insulin sensitiv
study methodology.
ity in metabolic syndrome. The impact of FMT on
It was possible to conduct meta-analyses for both
psychiatric outcomes was assessed in three studies of
IBS and active UC. However, due to the lack of
IBS patients, with one of these finding significant
consensus regarding outcome measures and small
improvements.
number of included studies, the results of these
This review also found that FMT was safe and well
meta-analyses should be considered preliminary at
tolerated. Similar rates of mild to moderate AEs were
this stage. Further, due to the low numbers, tests for
observed in participants who received FMT compared
publication bias (e.g. eggers regression and funnel
to those allocated to control groups, while SAEs were
plots) were unable to be carried out.
more commonly reported in participants allocated to
We were unable to undertake a quantitative ana
control/placebo groups.
lysis on the level of engraftment, given the gaps in
Not all studies assessed or reported whether FMT
data in the included studies. Future studies should
results in successful engraftment of the donor micro
evaluate microbial engraftment as a result of FMT,
biome into the recipient, but a majority of those that
allowing for a systematic assessment.
did report it confirmed a move toward the donor
This paper evaluated safety data across a range of
microbiome following FMT and that these changes
indications, finding broadly that FMT is well toler
persisted for up to 1 year. Furthermore, four of the
ated and safe. However, due to the poor quality and
five studies that reported on association between
incompleteness of reporting in several papers,
microbiome changes and clinical efficacy, four of
a cross-indication analysis of safety data was not
five confirmed such an association. This suggests
possible. We recommend future FMT studies
that FMT alters the recipient microbiome, and that
report more clearly on mild to moderate and SAE.
it is possible that this change is a contributing factor to
With respect to SAE, these were observed more
clinical efficacy.
frequently in control group participants than those
allocated to receive FMT, and of the SAE observed
Strengths and weaknesses of the review
in FMT recipients, most considered to be unrelated
This review is the first systematic review to evaluate to FMT. In understanding this finding, it should be
both safety and efficacy of FMT for all disorders noted that most SAE were likely due to the
e1854640-20 J. E. GREEN ET AL.
underlying disease process rather than the FMT Beta Lactamase-producing Escherichia coli (E.coli))
procedure, a majority of reported SAE were flares were transferred via FMT, resulting in one death. In
of the disease in inflammatory bowel disease parti these cases, donor feces were not screened for anti
cipants. Thus, FMT may have prevented disease biotic-resistant organisms.66 A further warning was
flares. issued on 12th March 2020 advising of six cases of
additional transmission of antibiotic-resistant organ
isms via FMT provided by a US-based stool bank
Implications for clinicians and policymakers
(enteropathogenic Escherichia coli in two cases and
With respect to active UC, our meta-analysis Shigatoxin-producing Escherichia coli in four cases)
revealed that FMT appears to be clinically effica and two deaths that occurred in recipients of FMT,
cious compared to control conditions. Four of the but in which FMT may not have been the cause of
six included studies used a gold-standard double- death.67 It is now standard across widely accepted
blind placebo-controlled RCT design, and all six protocols in the United Kingdom, United States, and
included studies favored FMT over control condi Australasia to screen thoroughly for antibiotic-
tions regarding clinical efficacy, notwithstanding resistant organisms. These recent serious incidents
limitations described above. Thus, the quality and highlight the importance of adhering to rigorous
consistency of outcomes appear to favor FMT in screening protocols, such as the Openbiome
the treatment of active UC, making this is Protocol in the US,68 the British Guidelines for
a promising area for research attention. donor screening,69 or the Australasian guidelines.70
Evidence supporting the application of FMT in
the treatment of IBS is more equivocal. Five studies
Unanswered questions, challenges for the field of
included in this review showed mixed outcomes,
FMT research, and future directions
with three reporting that FMT was favorable, and
two finding that control conditions were more Of all the uses of FMT for conditions other than
effective than FMT. Possible reasons for these CDI, the most promising at this stage is for active
mixed findings are discussed in depth in other flares of UC. Further large scale, high-quality stu
review papers,29–31 which note, inter alia, that dies, utilizing consistent data points to measure
route of delivery, choice of placebo (i.e. inert vs primary outcomes, are urgently called for. Other
autologous FMT), and patient group may have indications with some promise include metabolic
contributed.29–31 Suffice to say, there is no strong syndrome/obesity, antibiotic-resistant organisms,
evidence at this stage that FMT is efficacious for the and certain hepatic disorders. Whilst published
treatment of functional gut disorders. outside of the time range of the search performed
This review also identified two additional studies for this review, a recent Phase I study investigating
relating to inflammatory bowel disease yielding FMT for Alcohol Use Disorder showed safety and
significant outcomes regarding clinical efficacy, efficacy.71 Thus, we watch with interest the growing
and four studies that evaluated FMT for metabolic field of FMT for hepatic disorders.
syndrome or obesity that focused on biological out However, on the whole, studies in these emer
comes rather than clinical efficacy. As such, these ging areas are heterogeneous and generally of poor
are identified as conditions of interest for further quality, with most using open-label designs and
research only. only one study using clinical efficacy as a primary
Regarding safety, FMT appears generally to be outcome measure. Again, further high-quality
a safe and well-tolerated treatment, with orally admi research, using larger sample sizes and double-
nistered FMT appearing to be the best-tolerated route. blinded, placebo-controlled designs, and that use
However, it is also important to note that on 13th clinical efficacy endpoints as a primary outcome,
June, 2019 the American-based Food and Drug are needed for these emerging indications.
Administration (FDA) released a statement warning Furthermore, no studies evaluated the use of FMT
of the risks of FMT. They reported two cases (both for psychiatric conditions, an area of great impor
immunocompromised patients) in whom antibiotic- tance given growing interest and data supporting
resistant organisms, (specifically, Extended Spectrum the relationship between mental health and the gut
GUT MICROBES e1854640-21
microbiome (“the microbiota-gut-brain axis”).7,72 Woolworths, Avant and the Harry Windsor Foundation, has
Much also remains unknown about the ideal meth been a speaker for Astra Zeneca, Lundbeck, Merck, Pfizer, and
odological design for studies of FMT for conditions served as a consultant to Allergan, Astra Zeneca, Bioadvantex,
Bionomics, Collaborative Medicinal Development, Lundbeck
other than Clostridium difficile infection. For one, Merck, Pfizer and Servier – all unrelated to this work. Felice
an important question remains around choice of Jacka has received: (1) competitive Grant/Research support
placebo. Nine studies selected an inert placebo, from the Brain and Behaviour Research Institute, the National
such as water mixed with glycerol and food dye, Health and Medical Research Council (NHMRC), Australian
whilst eight opted for autologous FMT as a placebo. Rotary Health, the Geelong Medical Research Foundation, the
There is some evidence to suggest that even auto Ian Potter Foundation, The University of Melbourne; (2)
industry support for research from Meat and Livestock
logous FMT may have an impact on gut
Australia, Woolworths Limited, the A2 Milk Company, Be
microbiota,58 which may confound results. This Fit Foods; (3) philanthropic support from the Fernwood
needs to be further explored but presents an argu Foundation, Wilson Foundation, the JTM Foundation, the
ment against using autologous FMT as a placebo in Serp Hills Foundation, the Roberts Family Foundation, the
future studies aimed at determining efficacy of Waterloo Foundation and; (4) travel support and speakers
FMT compared with an inactive control. Follow- honoraria from Sanofi-Synthelabo, Janssen Cilag, Servier,
Pfizer, Health Ed, Network Nutrition, Angelini
up periods to assess long-term safety, engraftment, Farmaceutica, Eli Lilly and Metagenics. Wolfgang Marx is
and metagenomics are also an important consid currently funded by an Alfred Deakin Postdoctoral Research
eration for study design. The authors of this review Fellowship and a Multiple Sclerosis Research Australia early-
suggest a follow up period of at least 6 months to career fellowship. Wolfgang has previously received funding
adequately monitor for safety and long-term AE. from the Cancer Council Queensland and university grants/
fellowships from La Trobe University, Deakin University,
University of Queensland, and Bond University, received
Conclusions industry funding and has attended events funded by Cobram
Estate Pty. Ltd, received travel funding from Nutrition Society
This systematic review and meta-analysis provide of Australia, received consultancy funding from Nutrition
preliminary data that FMT may be safe and effec Research Australia, and has received speakers honoraria
tive for several conditions other than CDI. from The Cancer Council Queensland and the Princess
Preliminary meta-analyses suggest efficacy for out Alexandra Research Foundation. The Food & Mood Centre
has received Grant/Research support from Fernwood
comes related to active ulcerative colitis but not Foundation, Wilson Foundation, the A2 Milk Company, and
IBS. Hepatic disorders, metabolic syndrome/obe Be Fit Foods.
sity, and antibiotic-resistant organisms were also
identified as emerging areas of interest for FMT
research. Regarding safety, there was little differ
ence in SAEs between participants allocated to Author contributions statement
receive FMT and those allocated to control groups. JG wrote the protocol document, performed the primary and
With respect to mild to moderate adverse events, secondary searches, data extraction, data analysis, and drafted
similar rates were also observed in treatment and and edited the manuscript. JD was the second reviewer for
control groups. These encouraging pilot outcomes primary and secondary searches and data extraction. WM was
senior author, provided oversight, edited the manuscript, and
provide preliminary support for further high- completed data analysis for the meta-analyses. FJ was a second
quality research in these areas. senior author and provided oversight and input into the
manuscript. All authors read, edited, and approved the final
draft.
Disclosure of Potential Conflicts of Interest
Michael Berk is supported by an NHMRC Senior Principal
Research Fellowship (1156072). MB has received Grant/ ORCID
Research Support from the NIH, Cooperative Research
Centre, Simons Autism Foundation, Cancer Council of Michael Berk http://orcid.org/0000-0002-5554-6946
Victoria, Stanley Medical Research Foundation, Medical Amy Loughman http://orcid.org/0000-0002-0257-1443
Benefits Fund, National Health and Medical Research David Castle http://orcid.org/0000-0002-3075-1580
Council, Medical Research Futures Fund, Beyond Blue, John F. Cryan http://orcid.org/0000-0001-5887-2723
Rotary Health, A2 milk company, Meat and Livestock Board, Felice Jacka http://orcid.org/0000-0002-9825-0328
e1854640-22 J. E. GREEN ET AL.
25. Kelly JR, Borre Y, Patterson E, El AS, Deane J, Deane J, 37. Englewood. Comprehensive meta-analysis version 3. 3
Kennedy PJ, Beers S, Scott K, Moloney G, et al. ed.: Biostat 104; 2011.
Transferring the blues: depression-associated gut microbiota 38. Collaboration C. Recommendations on testing for fun
induces neurobehavioural changes in the rat. J Psychiatr Res. nel plot asymmetry. Cochrane handbook for systematic
2016;82:109–118. doi:10.1016/j.jpsychires.2016.07.019. reviews of interventions version. Cochrane handbook
26. Zheng P, Zeng B, Zhou C, Liu M, Fang Z, Xu X, Zeng L, for systematic reviews of interventions version
Chen J, Fan S, Du X, et al. Gut microbiome remodeling [Internet]. Cochrane Collaboration. 5; 2011.
induces depressive-like behaviors through a pathway 39. Kootte RS, Levin E, Salojarvi J, Smits LP, Hartstra AV,
mediated by the host’s metabolism. Mol Psychiatry. Udayappan SD, Hermes G, Bouter KE, Koopen AM,
2016;21(6):786–796. doi:10.1038/mp.2016.44. Holst JJ, et al. Improvement of insulin sensitivity after
27. Zheng P, Zeng B, Liu M, Chen J, Pan J, Han Y, Liu Y, lean donor feces in metabolic syndrome is driven by
Cheng K, Zhou C, Wang H, et al. The gut microbiome baseline intestinal microbiota composition. Cell Metab.
from patients with schizophrenia modulates the 2017;26(4):611–9.e6. doi:10.1016/j.cmet.2017.09.008.
glutamate-glutamine-GABA cycle and 40. El-Salhy M, Hatlebakk JG, Gilja OH, Brathen
schizophrenia-relevant behaviors in mice. Sci Adv. Kristoffersen A, Hausken T. Efficacy of faecal microbiota
2019;5(2):eaau8317. doi:10.1126/sciadv.aau8317. transplantation for patients with irritable bowel syndrome
28. Leshem A, Horesh N, Elinav E. Fecal microbial trans in a randomised, double-blind, placebo-controlled study.
plantation and its potential application in cardiometa Gut. 2020;69(5):859–867. doi:10.1136/gutjnl-2019-319630.
bolic syndrome. Front Immunol. 2019;10:1341. 41. Aroniadis OC, Brandt LJ, Oneto C, Feuerstadt P,
29. Xu D, Chen VL, Steiner CA, Berinstein JA, Eswaran S, Sherman A, Wolkoff AW, Kassam Z, Sadovsky RG,
Waljee AK, Higgins P, Owyang C. Efficacy of fecal Elliott RJ, Budree S, et al. Faecal microbiota transplan
microbiota transplantation in irritable bowel syndrome: tation for diarrhoea-predominant irritable bowel syn
a systematic review and meta-analysis. Off J Am Coll drome: a double-blind, randomised, placebo-controlled
Gastroenterol. 2019;114(7):1043-1050. trial. Lancet Gastroenterol Hepatol. 2019;4(9):675–685.
30. Ianiro G, Eusebi LH, Black CJ, Gasbarrini A, doi:10.1016/S2468-1253(19)30198-0.
Cammarota G, Ford AC. Systematic review with 42. Rossen NG, Fuentes S, van der Spek MJ, Tijssen JG,
meta-analysis: efficacy of faecal microbiota transplanta Hartman JH, Duflou A, Löwenberg M, van den Brink GR,
tion for the treatment of irritable bowel syndrome. Mathus-Vliegen EM, de Vos WM, et al. Findings from
Aliment Pharmacol Ther. 2019;50(3):240–248. a randomized controlled trial of fecal transplantation for
doi:10.1111/apt.15330. patients with ulcerative colitis. Gastroenterology. 2015;149
31. Myneedu K, Deoker A, Schmulson MJ, Bashashati M. (1):110–118. doi:10.1053/j.gastro.2015.03.045.
Fecal microbiota transplantation in irritable bowel syn 43. Smits LP, Kootte RS, Levin E, Prodan A, Fuentes S,
drome: a systematic review and meta-analysis. United Zoetendal EG, Wang Z, Levison B, Cleophas M,
Eur Gastroenterol J. 2019;7:1033–1041. Kemper E, et al. Effect of vegan fecal microbiota trans
32. LdF C, Borba HH, Tonin FS, Wiens A, Fernandez- plantation on carnitine- and choline-derived
Llimos F, Pontarolo R. Fecal microbiota transplantation trimethylamine-N-oxide production and vascular
in inflammatory bowel disease patients: a systematic inflammation in patients with metabolic syndrome.
review and meta-analysis. Plos One. 2020;15(9): J Am Heart Assoc. 2018;7(7):e008342.
e0238910. doi:10.1371/journal.pone.0238910. 44. Kang DW, Adams JB, Gregory AC, Borody T,
33. Tang L-L, Feng W-Z, Cheng -J-J, Gong Y-N. Clinical Chittick L, Fasano A, Khoruts A, Geis E, Maldonado J,
remission of ulcerative colitis after different modes of McDonough-Means S, et al. Microbiota transfer ther
faecal microbiota transplantation: a meta-analysis. apy alters gut ecosystem and improves gastrointestinal
Int J Colorectal Dis. 2020;35(6):1025–1034. doi:10.1007/ and autism symptoms: an open-label study.
s00384-020-03599-7. Microbiome. 2017;5(1).
34. Zhao HL, Chen SZ, Xu HM, Zhou YL, He J, Huang HL, 45. Sokol H, Landman C, Seksik P, Berard L, Montil M,
Xu J, Nie YQ. Efficacy and safety of fecal microbiota Nion-Larmurier I, Bourrier A, Le Gall G, Lalande V, De
transplantation for treating patients with ulcerative coli Rougemont A, et al. Fecal microbiota transplantation to
tis: a systematic review and meta-analysis. J Dig Dis. maintain remission in Crohn’s disease: a pilot rando
2020;21(10):534–548. doi:10.1111/1751-2980.12933. mized controlled study. Microbiome. 2020;8(1):12.
35. Aroniadis OC, Brandt LJ. Fecal microbiota transplanta doi:10.1186/s40168-020-0792-5.
tion: past, present and future. Curr Opin Gastroenterol. 46. Saidani N, Lagier J-C, Cassir N, Million M, Baron S,
2013;29(1):79–84. doi:10.1097/MOG.0b013e32835a4b3e. Dubourg G, Eldin C, Kerbaj J, Valles C, Raoult D, et al.
36. Higgins JPT, Altman DG, Gøtzsche PC, Jüni P, Faecal microbiota transplantation shortens the coloni
Moher D, Oxman AD, Savovic J, Schulz KF, Weeks L, sation period and allows re-entry of patients carrying
Sterne JAC, et al. The Cochrane collaboration’s tool for carbapenamase-producing bacteria into medical care
assessing risk of bias in randomised trials. BMJ. facilities. Int J Antimicrob Agents. 2019;53(4):355–361.
2011;343:d5928. doi:10.1136/bmj.d5928. doi:10.1016/j.ijantimicag.2018.11.014.
e1854640-24 J. E. GREEN ET AL.
47. Vujkovic-Cvijin I, Rutishauser RL, Pao M, Hunt PW, 56. Kump P, Wurm P, Gröchenig HP, Wenzl H,
Lynch SV, McCune JM, Somsouk M. Limited engraft Petritsch W, Halwachs B, Wagner M, Stadlbauer V,
ment of donor microbiome via one-time fecal microbial Eherer A, Hoffmann KM, et al. The taxonomic compo
transplantation in treated HIV-infected individuals. sition of the donor intestinal microbiota is a major
Gut Microbes. 2017;8:440–450. factor influencing the efficacy of faecal microbiota
48. Bajaj JS, Kassam Z, Fagan A, Gavis EA, Liu E, Cox IJ, transplantation in therapy refractory ulcerative colitis.
Kheradman R, Heuman D, Wang J, Gurry T, et al. Fecal Aliment Pharmacol Ther. 2018;47(1):67–77.
microbiota transplant from a rational stool donor doi:10.1111/apt.14387.
improves hepatic encephalopathy: a randomized clini 57. Moayyedi P, Surette MG, Kim PT, Libertucci J,
cal trial. Hepatology. 2017;66(6):1727–1738. Wolfe M, Onischi C, Armstrong D, Marshall JK,
doi:10.1002/hep.29306. Kassam Z, Reinisch W, et al. Fecal microbiota trans
49. Bajaj JS, Salzman NH, Acharya C, Sterling RK, plantation induces remission in patients with active
White MB, Gavis EA, Fagan A, Hayward M, ulcerative colitis in a randomized controlled trial.
Holtz ML, Matherly S, et al. Fecal microbial transplant Gastroenterology. 2015;149(1):102–109. doi:10.1053/j.
capsules are safe in hepatic encephalopathy: a phase 1, gastro.2015.04.001.
randomized, placebo-controlled trial. Hepatology. 58. Holster S, Lindqvist CM, Repsilber D, Salonen A, de
2019;70(5):1690–1703. doi:10.1002/hep.30690. Vos WM, Konig J, et al. The effect of allogenic versus
50. Allegretti JR, Kassam Z, Mullish BH, Chiang A, autologous fecal microbiota transfer on symptoms, visc
Carrellas M, Hurtado J, Marchesi JR, McDonald JAK, eral perception and fecal and mucosal microbiota in
Pechlivanis A, Barker GF, et al. Effects of fecal micro irritable bowel syndrome: a randomized controlled
biota transplantation with oral capsules in obese study. Clin Transl Gastroenterol. 2019;10(4):e00034.
patients. Clin Gastroenterol Hepatol. 2020;18(4):855– 59. Huttner BD, de Lastours V, Wassenberg M,
63.e2. doi:10.1016/j.cgh.2019.07.006. Maharshak N, Mauris A, Galperine T, Zanichelli V,
51. Herfarth H, Barnes EL, Long MD, Isaacs KL, Leith T, Kapel N, Bellanger A, Olearo F, et al. A 5-day course
Silverstein M, Gerardin Y, Kassam Z. Combined endo of oral antibiotics followed by faecal transplantation to
scopic and oral fecal microbiota transplantation in eradicate carriage of multidrug-resistant
patients with antibiotic-dependent pouchitis: low clin Enterobacteriaceae: a randomized clinical trial. Clin
ical efficacy due to low donor microbial engraftment. Microbiol Infect. 2019;25(7):830–838. doi:10.1016/j.
Inflammatory Intestinal Diseases. 2019;4(1):1–6. cmi.2018.12.009.
doi:10.1159/000497042. 60. Sood A, Mahajan R, Singh A, Midha V, Mehta V,
52. Vrieze A, Van Nood E, Holleman F, Salojarvi J, Narang V, Singh T, Singh Pannu A. Role of faecal
Kootte RS, Bartelsman JF, Dallinga–Thie GM, microbiota transplantation for maintenance of remis
Ackermans MT, Serlie MJ, Oozeer R, et al. Transfer of sion in patients with ulcerative colitis: a pilot study.
intestinal microbiota from lean donors increases insulin J Crohns Colitis. 2019;13(10):1311–1317. doi:10.1093/
sensitivity in individuals with metabolic syndrome. ecco-jcc/jjz060.
Gastroenterology. 2012;143(4):913–6.e7. doi:10.1053/j. 61. Tian H, Ge X, Nie Y, Yang L, Ding C, McFarland LV,
gastro.2012.06.031. Zhang X, Chen Q, Gong J, Li N. Fecal microbiota trans
53. Paramsothy S, Kamm MA, Kaakoush NO, Walsh AJ, van plantation in patients with slow-transit constipation:
den Bogaerde J, Samuel D, Leong RWL, Connor S, Ng W, A randomized, clinical trial. PLoS ONE. 2017;12(2):
Paramsothy R, et al. Multidonor intensive faecal micro e0171308.
biota transplantation for active ulcerative colitis: 62. Philips CA, Phadke N, Ganesan K, Ranade S,
a randomised placebo-controlled trial. Lancet. 2017;389 Augustine P. Corticosteroids, nutrition, pentoxifylline,
(10075):1218–1228. doi:10.1016/S0140-6736(17)30182-4. or fecal microbiota transplantation for severe alcoholic
54. Costello SP, Hughes PA, Waters O, Bryant RV, hepatitis. Indian J Gastroenterol. 2018;37(3):215-225.
Vincent AD, Blatchford P, Katsikeros R, 63. Ren YD, Ye ZS, Yang LZ, Jin LX, Wei WJ, Deng YY,
Makanyanga J, Campaniello MA, Mavrangelos C, et al. Chen XX, Xiao CX, Yu XF, Xu HZ, et al. Fecal micro
Effect of fecal microbiota transplantation on 8-week biota transplantation induces hepatitis B virus e-antigen
remission in patients with ulcerative colitis: (HBeAg) clearance in patients with positive HBeAg
a randomized clinical trial. Jama. 2019;321(2):156–164. after long-term antiviral therapy. Hepatology. 2017;65
doi:10.1001/jama.2018.20046. (5):1765–1768. doi:10.1002/hep.29008.
55. Ishikawa D, Sasaki T, Osada T, Kuwahara-Arai K, Haga K, 64. Halkjær SI, Christensen AH, Lo BZS, Browne PD,
Shibuya T, Hiramatsu K, Watanabe S. Changes in intest Günther S, Hansen LH, Petersen AM. Faecal microbiota
inal microbiota following combination therapy with fecal transplantation alters gut microbiota in patients with
microbial transplantation and antibiotics for ulcerative irritable bowel syndrome: results from a randomised,
colitis. Inflammatory Bowel Dis. 2017;23(1):116–125. double-blind placebo-controlled study. Gut. 2018;67
doi:10.1097/MIB.0000000000000975. (12):2107–2115. doi:10.1136/gutjnl-2018-316434.
GUT MICROBES e1854640-25
65. Wang S, Xu M, Wang W, Cao X, Piao M, Khan S, Yan F, H, Jonathan S, Marina Aloi, et al. European consensus
Cao H, Wang B. Systematic review: adverse events of conference on faecal microbiota transplantation in clin
fecal microbiota transplantation. PLoS One. 2016;11(8): ical practice. Gut. 2017;66(4):569–580. doi:10.1136/
e0161174. doi:10.1371/journal.pone.0161174. gutjnl-2016-313017.
66. Administration FaD. Important safety alert regarding use 70. Trubiano JA, Cheng AC, Korman TM, Roder C,
of fecal microbiota for transplantation and risk of serious Campbell A, May ML, Blyth CC, Ferguson JK, Blackmore
adverse reactions due to transmission of multi-drug resis TK, Riley TV, et al. Australasian society of infectious diseases
tant organisms. FDA; 2019 [updated 13/06/19. updated guidelines for the management of Clostridium diffi
67. FDA US. Safety alert regarding use of fecal microbiota cile infection in adults and children in Australia and New
for transplantation and risk of serious adverse events Zealand. Intern Med J. 2016;46(4):479–493. doi:10.1111/
likely due to transmission of pathogenic organisms: imj.13027.
FDA; 2020 [accessed 2020 Mar 12]. https://www.fda. 71. Bajaj JS, Gavis EA, Fagan A, Wade JB, Thacker LR,
gov/vaccines-blood-biologics/safety-availability- Fuchs M, Patel S, Davis B, Meador J, Puri P, et al.
biologics/safety-alert-regarding-use-fecal-microbiota- A randomized clinical trial of fecal microbiota trans
transplantation-and-risk-serious-adverse-events-likely. plant for alcohol use disorder. Hepatology. 2020.
68. Openbiome. Openbiome quality and safety program doi:10.1002/hep.31496. Epub ahead of print. PMID:
openbiome; 2019. [accessed 2019 Dec 19] https://www. 32750174..
openbiome.org/safety. 72. Dinan TG, Cryan JF. Gut microbiota: a missing link in
69. Cammarota G, Ianiro G, Tilg H, Rajilic-Stojanovic M, psychiatry. World Psychiat. 2020;19(1):111–112.
Kump P, Satokari R, Harry S, Perttu A, Cristina P, Ailsa doi:10.1002/wps.20726.
Copyright of Gut Microbes is the property of Taylor & Francis Ltd and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.