Is there a role of faecal microbiota transplantation in reducing antibiotic resistance burden in gut? A systematic review and Meta-analysis

Abstract Objectives The aim of current systematic review and meta-analysis is to provide insight into the therapeutic efficacy of fecal microbiota transplantation (FMT) for the decolonization of antimicrobial-resistant (AMR) bacteria from the gut. Methods The protocol for this Systematic Review was prospectively registered with PROSPERO (CRD42020203634). Four databases (EMBASE, MEDLINE, SCOPUS, and WEB of SCIENCE) were consulted up until September 2020. A total of fourteen studies [in vivo (n = 2), case reports (n = 7), case series without control arm (n = 3), randomized clinical trials (RCT, n = 2)], were reviewed. Data were synthesized narratively for the case reports, along with a proportion meta-analysis for the case series studies (n = 102 subjects) without a control arm followed by another meta-analysis for case series studies with a defined control arm (n = 111 subjects) for their primary outcomes. Results Overall, seven non-duplicate case reports (n = 9 participants) were narratively reviewed and found to have broad AMR remission events at the 1-month time point. Proportion meta-analysis of case series studies showed an overall 0.58 (95% CI: 0.42-0.74) AMR remission. Additionally, a significant difference in AMR remission was observed in FMT vs treatment naïve (RR = 0.44; 95% CI: 0.20-0.99) and moderate heterogeneity (I2=65%). A subgroup analysis of RCTs (n = 2) revealed FMT with further benefits of AMR remission with low statistical heterogeneity (RR = 0.37; 95% CI: 0.18-0.79; I2 =23%). Conclusion More rigorous RCTs with larger sample size and standardized protocols on FMTs for gut decolonization of AMR organisms are warranted. KEY MESSAGE Existing studies in this subject are limited and of low quality with moderate heterogeneity, and do not allow definitive conclusions to be drawn. More rigorous RCTs with larger sample size and standardized protocols on FMTs for gut decolonization of AMR organisms are warranted.


Introduction
Currently, antimicrobial resistance (AMR) has been identified as one of the major threats to global health, food production, and economic development [1]. The US Centres for Disease Control and Prevention has estimated that each year, >2.8 M patients are infected with antibiotic-resistant (AR) bacteria and >35,000 dies of these infections. Also, nearly 223,900 people in the United States required hospital care for C. difficile and at least 12,800 people died in 2017 [2]. AMR is most often conferred through the expression of antimicrobial resistance genes that reduce a microbe's susceptibility to the effects of antibiotics. AMR bacteria are stratified to, Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species, known to be "ESKAPE", are responsible for the majority of hospital infections with higher mortality rates [3]. Data on AMR in Europe are reported by the European Antimicrobial Resistance Surveillance Network (EARS-Net) 2018 report, stated that more than half (58.3%) of the E. coli and third (37.2%) of K. pneumoniae isolates responsible for invasive diseases were resistant to at least one of the antimicrobial groups under regular surveillance (i.e. aminoglycosides, aminopenicillins, fluoroquinolones, third-generation cephalosporins, and carbapenems) [4].
The human gastrointestinal (GI) tract is colonized by different kinds of bacteria, archaea, fungus, and viruses, consensually termed as gut microbiota [5]. The intestinal microbiome of healthy patients often consists of wellbalanced diversified microbiota members that predominantly belong to just four phyla-the Bacteroidetes, Firmicutes, Actinobacteria, and Proteobacteria and known to pose colonization resistance (CR) [6]. Gut microbiota can produce a variety of compounds that play key roles in the colon micro-ecology and host homeostasis. Delicate contemporary approaches to unravel the importance of the symbiosis of gut microbiota lead to its identification as a potential target of many chronic diseases ranging from gastrointestinal inflammatory and metabolic conditions to neurological, respiratory, and cardiovascular illnesses [7]. Apart from that, gut microbiota plays a beneficial role in maintaining human health via producing short-chain fatty acids (SCFA), vitamins and acting as a shield to protect the host from colonization by pathogenic bacteria [8].
However, under certain circumstances, the patient may develop a compromised microbiota by declining alpha-composition and it causes CR vulnerability and eventually leads to exogenous bacterial colonization. The deleterious effects of antibiotics on gut microbiota have been extensively studied [9,10]. Other than antibiotics, proton pump inhibitors (PPIs) are one of the most commonly prescribed drugs in western medicine [11], and they lead to a profound and prolonged reduction of gastric acid production. The association of PPIs and the risk of some enteric infections namely, Clostridium difficile, Campylobacter, Salmonella are well documented [12][13][14]. Importantly, antipsychotic drugs such as olanzapine [15], pimozide [16], fluphenazine [17], and flupenthixol dihydrochloride [18] were proven to pose in vitro antibacterial properties against a broad spectrum of bacteria alone and in combination with other antibiotics. A significant population of the community consumes these drugs and they may have disturbed gut microbiota composition, and consequently put their CR at risk. Hence, their intestinal microbiota dysbiosis provides favourable conditions for AMR bacteria to colonize and eventually act as a reservoir for horizontal resistant gene transfer [19,20]. Horizontal gene transfer has been documented as an important mechanism for the transfer and acquisition of antimicrobial resistance genes within and between gut bacterial species (Figure 1) [21][22][23].
Research on the bilateral relationship between gut microbiota and human health has been in the spotlight during the last decade. This has been mainly driven forward by improved next-generation sequencing (NGS) technologies [24] and novel proteomic [25] approaches, allowing the profiling of entire microbial communities with high efficacy and low cost.
The conservation effect of gut microbiota from AR has encouraged scientists to study faecal microbiota transplantation (FMT) to restore a healthy gut microbiome by eliminating AR colonization. The concept behind FMT is to directly alter the gut microbial composition of the recipient to establish the alpha-diversity [26]. This is achieved via the administration of a frozen, encapsulated or fresh solution of faecal matter from a donor into the intestinal tract of a recipient to confer a health benefit via altering the gut microbial composition [27]. The initial step of the process involves a thorough screening procedure to identify a suitable donor. This includes a questionnaire of the donors' family health history, contemporary exposure to any medication and a series of laboratory tests to ensure there is no transmittable disease or pathogens [28]. According to the contemporary guidelines from the Infectious Diseases Society of America (IDSA) and the Society of Healthcare Epidemiology of America (SHEA), as well as the European Consensus Guidelines, FMT is recommended as a second-line treatment modality against recurrent C. difficile infection, due to over 90% efficacy in randomized control trials [26]. But the application of FMT for the decolonization of AMR microorganisms other than C. difficile is controverisal due to a limited number of reports [29][30][31][32][33][34][35][36].
The main objective of our review is to descriptively analyze the non-duplicate data and pool all published data to ascertain a conclusive statistical picture of the primary outcome, set as the decolonization of AMRs (except recurrent C. difficile) in adults (>18 years old) by FMT intervention at the 1-month time point. Secondary outcomes are to provide systematic descriptive analyses of case studies on decolonization of AMR via FMT, identifying adverse effects associated with FMT procedures, and outline key steps to be followed in future rigorous clinical trials.

Protocol development
We registered our review protocol (CRD42020203634) in the International Prospective Register of Systematic Reviews (PROSPERO) that is available at https://www. crd.york.ac.uk/prospero/display_record.php?ID=CRD420 20203634. We adhered to the recommendations of the PRISMA-P 2015 statement in developing this protocol and conducting the review [37].

Search strategy
This global search was performed using four bibliographic databases, namely, EMBASE and MEDLINE through PubMed, SCOPUS, and WEB of SCIENCE. The following search terms were used: ("Faecal microbiota transplantation" OR "FMT" OR "bacteriotherapy" OR "gut microbiota transplantation" OR "fecal transplantation" OR "intestinal microbiota transfer") AND ("intestinal antibiotic-resistant bacteria" OR "antibioticresistant bacteria" OR "intestinal antimicrobial resistance" OR "AMR"). The same search term was used in all databases to retrieve the data for this review. Apart from that, we have searched the following clinical trial registries; United States (www.clinicaltrials.org), Australia-New Zealand (www.anzctr.org.au) United Kingdom (www.isrctn.com), Germany (www. drks.de), and China (www.chictr.org.cn) The final database update was carried out on 15 September 2020.

Types of studies
This review included studies that investigated the effectiveness of FMT in eliminating AMR colonization confined to the gut. As a fact, FMT is an emerging therapy for the decolonization of AMR for intestinal carriage, hence the data on randomized clinical trials Figure 1. Concept illustration of colonization resistance due to alpha diversity of gut microbes and the effect of antimicrobials in destabilization of symbiotic stage. The disturbed gut microbiota could be colonized with AMRs and leads to dysbiosis. FMT is an alternative therapeutic modality to restore the alpha diversity by decolonizing AMRs. are scanty. Thus, we included different types of clinical trials including, cohort studies, case studies, case series studies, and case reports. Apart from clinical investigations, we also reviewed in vivo studies dealing with the same purpose. We excluded all the review articles, news, conference proceedings, editorials, and letters to the editor, expert opinions, or commentaries as they did not provide adequate information for review. Except for case series studies (used only for the metaanalysis), we also excluded studies (case reports) that were previously reviewed.

Types of participants and eligible AMR bacteria
The adult population (>18 years) was included in the review. Studies with paediatric patients were excluded since their gut microbiome is in a dynamically developing stage and the study outcome could not be compared with the adults [38], and those who are carrying rather stable gut microbiome [39]. Patients who received antibiotics concurrently at the time the FMT were also excluded, due to its direct effect on the composition of the transplanting bacteria [40]. However, patients whose concomitant antibiotic therapy was discontinued at least 24 h before the FMT procedure were included.
AMRs involving viruses and fungi and AMR bacteria colonized outside the gut were also excluded. Furthermore, FMT on the patients who had recurrent or refractory C. difficile infections alone was also excluded from our review process, since there are other reviews solely focussed on the topic [41][42][43].

Types of interventions and outcome measures
The single-arm intervention trials were abundant in numbers in this review, but studies with control groups including placebo, antibiotics, or treatmentnaive were also included, where necessary. Different FMT administration routes (upper and lower gastrointestinal routes) were considered including; caecum through colonoscopy, oral gavage, naso-duodenal tube and delivered via enema. The fresh, frozen or encapsulated samples were used in FMT along with related or unrelated donors, as well as, single or multiple FMTs were included in the review.
The primary outcome measured was the decolonization of AMR within 1 month (30 days) upon FMT with two consecutive negative results of rectum swabs or confirmed by PCR. However, the inter-study variation could be observed in the outcome for decolonization testing and are included in this review.

Assessment of risk of bias of eligibility criteria for the articles
This assessment was performed in several ways including, the use of https://www.covidence.org online platform. At the beginning, all the citations (from 4 databases) were uploaded to the software in "RIS" file format. A preliminary abstract screening was carried out by NR and AL for inclusion and the selected studies were subjected to scrutiny by full-text review by NR, AL, and PD to avoid the risk of bias. Any disparities were resolved upon the consultation of senior author MI.
Moreover, the selected studies were further verified by NR and PD, according to the Joanna Briggs Institution (JBI) standardized critical appraisal checklists for case series and case reports for the possibility of bias in its design, conduct, and analysis [44,45]. The completed JBI forms for the included studies were given in Supplementary information, Section 1. The studies with selection "No" for at least 3 questions in the appraisal form considered to pose a lack of integrity in the study design, hence excluded from data abstraction and reviewing.
Overall completeness and transparency of the included case reports were verified using CARE guidelines [46] and completed forms for the included studies are given in Supplementary information, Section 2. Similarly, PROCESS guidelines [47] were used to verify the quality and risk of bias of the case series studies (Supplementary information, Section 3). Quality of evidence for two randomized control trials and the remaining 3 case series studies were assessed following the GRADE recommendations [48] and completed on the GRADEpro online software (Supplementary information, Section 4).

Data abstraction
We developed a data abstraction spreadsheet using Excel version 2016 software (Microsoft Corporation, Redmond, Washington, USA). We conducted the data abstraction for the included full-text articles, and the data were independently assessed by two review authors. We extracted the following information: title, DOI number, type of study, acceptance or rejection of topic, type of faecal material, methods of infusion, characteristics or composition of stool, types of patients, the role of FMT on the antibiotic-resistant bacteria, mode of action, and side-effects with potential remarks.

Data synthesis
Two separate meta-analyses were carried out for the case series studies with and without control arms for their primary objective (percentage decolonization at the 1-month time point). However, in two studies [49,50], the decolonization rate was considered at the 35 to 48-day time point. The case reports were excluded from the meta-analysis, since the single data is not adequate to provide the estimation of effect size and for which, we have provided only a narrative synthesis [51].
The outcome for the case series studies is dichotomous with only two responses, such as AMR decolonization was successful or not. The meta-analysis for the case series studies without a control arm was analysed as proportions of decolonization [52,53] under the random-effects model using StatsDirect v3 statistical software. Analyses of the case series studies with the inclusion of a proper treatment naïve group were conducted via RevMan 5.4 using a random-effects model. For both meta-analyses, the combined effect was illustrated via creating a forest plot.
A statistically significant p-value was based on p < .05. Clinical and methodological heterogeneity across the included studies were expressed descriptively, while the statistical heterogeneity between the studies included for the two meta-analyses was expressed by using the I 2 statistic, while I 2 values interpreted as; low, moderate, and high levels of heterogeneity where, I 2 < 50%, I 2 50-75%, and I 2 > 75%, respectively [54].

Study selection
The search strategy from EMBASE and MEDLINE via PubMed, SCOPUS, and WEB of SCIENCE yielded one thousand five hundred and eighty-five (n ¼ 1585) studies, none of the studies were retrieved from the grey literature. Upon removal of duplicates (n ¼ 612), the title and abstracts of nine hundred and seventy-three (n ¼ 973) studies were screened, and nine hundred and twenty-one (n ¼ 921) studies did not meet the selection criteria. The full texts of the remaining fiftytwo (n ¼ 52) studies were assessed against the predefined inclusion and exclusion criteria, but thirtyeight (n ¼ 38) studies did not meet the ultimate inclusion criteria. Additionally, the studies already included in previous literature reviews [38][39][40]50] were also excluded in abstracting data or review in detail. However, previously published case series studies were included in 2 meta-analyses to get an overall statistical conclusion on the effect of FMT on the decolonization of AMRs at the 1-month time point. Finally, fourteen (n ¼ 14) studies were eligible for this detailed data abstraction and the broad discussion. The process of the selection of studies is summarized in a PRISMA flow diagram ( Figure 2). Three reviewers, NR, AL, and PD, screened the studies for inclusion and exclusion in the systematic review and meta-analysis and all the review authors mutually agreed on the final set of articles to be included.

Characteristics of included studies
The eligible articles had a publication date from September 2015 to September 2020. Out of fourteen studies, the majority are case reports or case studies (n ¼ 7) followed by different types of case series (n ¼ 5) and two (n ¼ 2) in vivo studies involving mice models. Out of the five case series studies, two (n ¼ 2) were carried out in multiple centres (Switzerland, Netherland, France, and Israel) [49,50], while the remaining (n ¼ 3) are single-center studies [55][56][57]. Most of the case reports/studies (n ¼ 5/7) were carried out in Europe; namely, Denmark (n ¼ 2) [58,59], Poland (n ¼ 2) [60,61], Netherland (n ¼ 1) [62], and one study each in USA [63] and South Africa [64]. AMR colonization in the gut was varied among studies, particularly in CREs. Except in one case study where the gut was colonized by MDR Salmonella infantis, multiple MDR bacteria colonized participants across all other studies [63]. The clinical diagnostics of the subjects who underwent FMT across the case reports/studies included patients with; immunocompromised [61], diabetic [62], renal transplant [58], different types of leukemia [60,63] and critically ill with multiple organ failure [59,64]. We have observed an inter-study variability in using PPIs. None of the case studies/reports used PPIs in the pre-treatment step, except in Bilinki et al. [61]. This observation was in contrast to that in case-series studies, where PPIs (omeprazole and pentaprazole) were used in all studies as a pre-treatment in different time points, except in Leo et al. [49] Bowel cleansing was not conducted in any of the case studies/reports, but it has been applied in many case series studies [55][56][57]. Except in one case study [64], the FMTs in all included studies were procured from unrelated healthy donors. Among most case studies/ reports (n ¼ 5/7), fresh stool samples were used and diluted with sodium chloride solutions (saline). In contrast, FMT was applied as capsules from frozen stool samples (n ¼ 3/5) with a dosage of 15 capsules/day for two consecutive days, in case series studies [49,50,55]. Among capsule-based FMTs, one study used 80% of glycerol instead of 10% in preparation of the capsule, to increase the stability [50].

Application of FMT in AMR decolonization in gut
In vivo models The study characteristics of the two eligible in vivo studies are given in Table 1. FMT is effective in Figure 2. Flowchart showing the systematic review process with the bibliometric assessment, including article attrition and study selection. Briefly, articles were filtered through an automated bibliographic database search using keywords. Table 1. Characteristics of the in vivo studies eligible for the review.
Mrazek et al. [65] Caballero et al. [66] In  [65,66]. Both included studies have shown that FMT significantly decreased the faecal load (in terms of CFU) of colonizing resistant bacteria when compared to those that did not receive FMT [65,66]. A 2.5 and 4 log 10 average CFU reduction was observed in MDR P. aeruginosa level in mice receiving FMT by murine donors and human donors, respectively, within 7 days post-FMT [65], implying donors of the same species can produce a stronger colonization resistance [65].
KPC was cleared more effectively than VRE in mice receiving FMT (100% and 60% clearance for K. pneumoniae and VRE, respectively) [66]. It may suggest there are different mechanisms of colonization resistance against VRE and K. pneumoniae, or K. pneumoniae is more susceptible to colonization resistance [66].
Except for the FMT donor, the mice model, colonization confirmation, and the route of FMT administration were quite similar in both studies (Table 1). However, the dosage of FMT and target MDR bacterial species vary among studies.

Case reports of FMT
The study characteristics of the seven case studies/ reports included for the review are given in Table 2. In the case reports, the patients were mostly colonized with Enterobacteriaceae (Table 2). Besides, some other MDR pathogens were also confirmed such as; Candida albicans, Candida parapsilosis, Enterococcus faecalis, E. faecium, and S. infantis [58,60,[62][63][64]. Except Ueckermann et al. [64], for the reaming studies, the colonization confirmation was achieved via rectal swab analysis It is apparent from Table 2 that most of the AMR strains with different resistant mechanisms were susceptible to the FMT at the 1-month time point [58,59,61,63,64]. However, Stalenhoef et al. [62], reported that the patient who received single FMT failed to decolonize ESBL þ E. coli resistant to carbapenems, gentamicin, piperacillin/tazobactam, and colistin, respectively but eradicated MDR P. aeruginosa successfully during the 3 months follow-up period. Similarly, Biernat et al. [60], reported that patient-1 has completely resolved the symptoms after having three consecutive doses of FMT with a 1-week interval and decolonized all the pathogens. On the contrary, the patient-2 of the same study who received four doses of FMT with a 1-week gap failed to decolonize multidrug-resistant E. coli, and Citrobacter freundii.
Though most of the studies did not mention on side effects, some patients complained of having loose stool [61,62], abdominal cramps, anorexia, and diarrhoea (Table 2) [61].

Case series studies
The characteristics of the case series studies (studies not discussed elsewhere) included in this review are given in Table 3 [49,50,[55][56][57]. So far only two [49,50] randomized control trial results have been reported (Table 3). For statistical interpretation, we have divided case series studies into 2 cohorts including the case series, with and without a control arm. Additionally, we have pooled the results of all reported case series studies published so far [29][30][31][32][33][34][35][36], including the studies previously discussed elsewhere [42,43,53], to obtain a larger sample set for analysis for the FMT intervention.

listed in
Primary outcome (decolonization of AMRs) for the case series studies with a control arm All 111 participants were included in the intention-totreat analysis of the primary outcome, of whom 57 received FMT and 54 received no treatment. Significantly, more patients receiving donor FMT achieved clinical remission at the 1-month time point compared with those receiving control interventions, with a pooled RR of not achieving remission of 0.44 (95% CI: 0. 20-0.99) (p ¼ .03) with an I 2 ¼ 65% (forest plot, Figure 4). The pooled rate of clinical remission in all 4 trials was 63.2% (n ¼ 36/57) in the group receiving donor FMT and 22.2% (n ¼ 12/54) in those receiving control interventions. Statistical assessment for publication bias was not performed because only 4 included trials were inadequate for funnel plots or regression-based assessments. Altogether, the forest plots obtained from the meta-analysis (Figures 3 and  4) point towards favouring FMT as an alternative treatment modality for the decolonization of AMR bacteria in the gut. Table 2. Characteristics of published case study/reports describing outcomes of FMT against AMR Decolonization.
Regarding the intervention of FMT, the method of sample preparation was somewhat identical. In all cases, the donors were unrelated and the intervention was in the form of capsules, except in two studies where the samples were aseptically prepared in saline and administered (Table 3) [56,57]. In most studies FMT was administered twice for each subject and on consecutive days [49,50,55,56].

FMT susceptible AMRs and their decolonization time points
The time frame achieved for the decolonization with respect to their resistance mechanism was also analysed as percentage decolonization at 1, 2, and 4 weeks (1 month) after FMT, and the results are presented in Table 4. The decolonization rates were within 33.3-100% at the one-month time point (Table  4) with the highest (100%) decolonization rate for AMRs in Enterobacter hormechai (n ¼ 1), Klebsiella oxytoca (n ¼ 1), Citrobacter freundi (n ¼ 3), Acinetobacter baumanii (n ¼ 1), and Citrobacter koseri (n ¼ 1) (Table  4) and the lowest against Serratia marcescens with KPC (n ¼ 2) (Table 4). However, it is worth noting that the sample size is inadequate to draw a robust conclusion for the effectiveness of FMT against these strains. Stalenhoef et al. [62] Grosen et al. [58] Soto et al. [64] Biernat et al. [60] Bahl et al. [59] Ueckermann et al. [ Table 3. Characteristics of case-series studies (non-duplicate) with and without control arm included for the review.

Ongoing and completed clinical trials not published
We have extended our background search on different clinical trial registries and found that twenty-two (n ¼ 22) more clinical trials have already been registered to investigate AMR decolonization of the gut via FMT. In that, twelve (n ¼ 12), three (n ¼ 3), six (n ¼ 6), and one (n ¼ 1) studies are randomized, nonrandomized, interventional and observational clinical   trials, respectively ( Table 5). Out of the 12 randomized clinical trials, two studies were (n ¼ 2/12) already completed, but results have not been published. A summary of all the clinical trials is given in Table 5.

Discussion
According to the contemporary guidelines from the Infectious Diseases Society of America (IDSA) and the Society of Healthcare Epidemiology of America (SHEA), as well as the European Consensus Guidelines, FMT is recommended as a second-line treatment modality against recurrent C. difficile infection, due to over 90% efficacy in randomized control trials [6]. Our analysis in this study points towards support of decolonization of ESKAPE pathogens with FMT intervention (Table 6). However, RCTs and sample sizes are still limited, and in addition, lack of standardized protocol and their demonstration of improvement in clinical endpoints has been inconsistent. Therefore, RCTs involving larger sample size and consensus on standardized protocols are warranted. The human gut contains up to 3.8 Â 10 13 bacteria that represent around 55% of stool mass [67]. Therefore, selecting a healthy donor without an intact gut microbial composition plays an important role. Similarly, the direct correlation between the effectiveness of the FMT with the pre-bowel preparation of the receptor is an equally important component for successful transplantation. Based on the studies included in this review, it is considered that an optimised FMT protocol should include: (i) rigorous donor screening procedures including, not having taken antibiotics, immunosuppressants, chemotherapy, and PPIs in recent <3 months [50,68], (ii) decontamination of nasopharyngeal colonized sites prior to FMT [56] (iii) !5 days prolonged treatment with high dose of appropriate antibiotic/s regimen to reduce the diversity of gut microflora and discontinue 48 h prior to prime the gut for FMT [49,50,56,58,59,61,62] (iv) two bowel cleansing regimens (one before antibiotic treatment and the other before FMT) [56] or at least (one day before FMT) [57,61] to cleanse the intestinal Figure 6. Important steps to be followed in future rigorous randomized clinical trials.  residues (v) adhere to strict post-FMT decolonization measures (isolation, disinfect the environment and remove catheters or other non-essential medical parts connected) and follow up (at least 3-6 months) with regular stool testing to confirm the sustainable decolonization ( Figure 6). It is worth notice that several studies [50,[55][56][57]61] have used PPIs just before FMT, in view of protecting the transplanting microbiota from gastric acids. However, its evidence for the benefit as a pre-treatment for FMT is controversial [69], since it has been reported that PPIs are associated with an increased risk of CDI [70] and other enteric infections [71]. Also, they can alter the gut microbiota by expanding the Enterococcus and Streptococcus genera and increasing other bacterial genes associated with epithelial invasion [72,73]. Furthermore, a meta-analysis from Hong et al. [74] proved there was no statistically significant benefit from the routine use of pre-treatment with PPIs in the FMT protocol compared to no PPIused protocols.
Even though the mechanistic studies are scanty to reveal the exact mechanism of action of FMT towards decolonization of drug-resistant bacteria, some important observations made during the studies help to speculate that FMT may have a strain-specific mechanism of action apart from general criteria like; bile-salt metabolism, GI luminal pH, and competition for resources.
Caballero et al. [66] conducted an FMT on mice cocolonized with VRE and K. pneumoniae those who are residing in the same region of the GI tract (lower GI tract). Unlike in most closely related species [75][76][77][78], these two strains did not show colonization resistance to each other due to nutrient competition. This may be attributed generally to their different metabolic requirements as well as their ability to switch nutrient precursors in the presence of competing strains [66]. However, investigators specifically observed that the antibiotic pre-treated control group demonstrated a thickening of the mucus layer by K. pneumoniae compared to VRE, indicating their high permeability towards the mucus layer, during the weak expression of host antimicrobial molecules such as RegIIIc upon pre-antibiotic treatment [79]. This mucus infiltration, consequently increased K. pneumoniae translocation to mesenteric lymph nodes (mLNs) relative to VRE. Cocolonization and K. pneumoniae eventually increased VRE translocation too, via opening up the barrier towards mLNs. Now, the more diverse microbiota in FMT can largely colonize the intestine, while the K. pneumoniae and VRE are translocated towards mLNs and the complete eradication was observed in both pathogens upon FMT, which means their translocation is continuous and no replication is taking place within mLNs. Even though the exact mechanism is yet to be identified, previous studies have speculated that [80,81] cells within the colonic lamina propria, including CD103 þ and CX3CR1 þ dendritic cells, are believed to capture K. pneumoniae and carry them to mLNs.
Clostridia, whereas the numbers of Lactobacilli, Bifidobacteria, and mouse intestinal Bacteroides were higher in the faeces derived post-FMT as compared to the pre-FMT status. Therefore, investigators hypothesised the mechanism behind the reduction of MDR P. aeruginosa burden was due to the higher loads of Lactobacilli and Bifidobacteria and for their pronounced production of bacteriocins and short chain fatty acids [82,83].
In a case study [59] of a 69 year-old woman referred with severe recurrent CDI and complicated by intestinal co-colonization with KPC-producing bacteria, XDR K. pneumoniae was successfully decolonized by a single FMT treatment. Interestingly, 16S rRNA amplicon profiling revealed Enterobacterales constituted 18% of the microbiota before FMT and subsequently dropped to lower levels after the FMT by increasing butyrateproducing Faecalibacterium prausnitzii. This species has been associated with treatment success and a mechanism for butyrate-induced reduction of intestinal inflammation and bacterial translocation of the C. difficile pathogen [84]. A similar observation was also made by Billinski et al. [60], where VRE decolonization was observed in a case study of an immunocompromised male patient (51 years). On the same note, the mechanism behind this decolonization was explained previously through in vivo mice model [85], in that the eradication may be associated with direct inhibition of VRE by a single component of healthy gut microbiota belonging to Barnesiella species. However, extensive studies with long-term follow-ups are warranted to determine the stability of this effect.
By using 16S rRNA metagenome sequencing, a substantial increment of Bifidobacterium bifidum was noted at post-FMT of a case series study conducted to decolonize CPE [55]. This species was shown to possess anti-Enterobacteriaceae effects [86] and anti-lysozyme activity [87], decrease biofilm formation [88], and modulate virulence gene expression [89] alone and with the combination of lactitol and Lactobacillus acidophilus to reduced OXA-48-producing Enterobacteriaceae [90]. These might explain the mechanism of action of B. bifidum in the decolonization of CPE in the gut.
Tavoukjian [53], have included 5 case series studies alone and according to the meta-analysis, a higher remission rate upon FMT was observed against P. aeruginosa while the lowest was against K. pneumoniae with NDM-1 and ESBL-producing strains, at the 1month time point. These findings are in line with the analysis outcome of our systematic review, and with a comparatively lower decolonization rate for E. coli with NDM-1 (33.3%) and ESBL (45.5%) and K. pneumoniae with KPC (33.3%) ( Table 4). The literature review by Amrane and Lagier [43], similarly and qualitatively, updated the recent literature.
The application of FMT for the decolonization of MDR bacteria is still in the early stage of development.
Most of the eligible studies are not randomized control trials, except Huttner et al. [50] and Leo et al. [49] The effect of FMT alone is still undefined, whether the role of antibiotics and PPIs prior to FMT or other forms of priming was essential, and the possibility of spontaneous decolonization might also play a role. Therefore, further rigorous randomized control trials with the inclusion of suitable control arms are warranted to establish the therapeutic efficacy of FMT.

Conclusion
In summary, there are limited existing studies which are generally of low quality with moderate heterogeneity, and do not allow definitive conclusions to be drawn. More rigorous RCTs with larger sample size and standardized protocols on FMTs for gut decolonization of AMR organisms are warranted.