Repeated Fecal Microbial Transplantations and Antibiotic Pre-Treatment Are Linked to Improved Clinical Response and Remission in Inflammatory Bowel Disease: A Systematic Review and Pooled Proportion Meta-Analysis

The response of patients with inflammatory bowel disease (IBD) to fecal microbial transplantation (FMT) has been inconsistent possibly due to variable engraftment of donor microbiota. This failure to engraft has resulted in the use of several different strategies to attempt optimization of the recipient microbiota following FMT. The purpose of our study was to evaluate the effects of two distinct microbial strategies—antibiotic pre-treatment and repeated FMT dosing—on IBD outcomes. A systematic literature review was designed and implemented in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A medical librarian conducted comprehensive searches in MEDLINE, Embase, Scopus, Web of Science Core Collection, and Cochrane Library on 25 November 2019 and updated on 29 January 2021. Primary outcomes of interest included comparing relapse and remission rates in patients with IBD for a single FMT dose, repeated FMT dosages, and antibiotic pre-treatment groups. Twenty-eight articles (six randomized trials, 20 cohort trials, two case series) containing 976 patients were identified. Meta-analysis revealed that both repeated FMT and antibiotic pre-treatment strategies demonstrated improvements in pooled response and remission rates. These clinical improvements were associated with increases in fecal microbiota richness and α-diversity, as well as the enrichment of several short-chain fatty acid (SCFA)-producing anaerobes including Bifidobacterium, Roseburia, Lachnospiraceae, Prevotella, Ruminococcus, and Clostridium related species.


Introduction
Inflammatory bowel disease (IBD) is a chronic inflammatory condition of the gastrointestinal tract categorized by Crohn's disease (CD), ulcerative colitis (UC), and indeterminate colitis [1,2]. The incidence of IBD is steadily increasing worldwide [3], as are its extensive healthcare and economic burdens. While IBD is believed to involve a host's genetic predisposition, environmental factors, and an imbalanced gut microbial community, the etiology of IBD has yet to be fully elucidated [4][5][6][7][8]. The complex pathophysiology underlying IBD has led to the current implementation of non-specific therapeutic strategies centered on systemic immunosuppression [9,10]. Despite the significant complications associated with these strategies, ongoing high rates of refractory disease remain [11][12][13] suggesting that alternative targeted approaches are needed to enhance the clinical efficacy and safety of modern IBD therapies [14].
Accumulating evidence suggests that imbalances in the gut microbiome, a highly diverse community of microorganisms that inhabits the gastrointestinal tract of humans, plays a causative role in the pathogenesis of IBD [15][16][17]. In general, gut microbial communities of patients with IBD are characterized by reduced microbial diversity, an increased abundance of aerobic pro-inflammatory bacteria, and a reduction in anaerobic bacteria that generate beneficial anti-inflammatory metabolites, such as short-chain fatty acids (SCFA). These findings have fostered growing interest in adopting microbiota-targeted strategies into the forefront of modern IBD therapeutics [18][19][20] in order to reduce the need for long-term immunosuppressants and their associated adverse complications.
Fecal microbial transplantation (FMT) is one such microbiota-targeted strategy that has shown initial promise for the management of IBD by implanting members of microbiota from healthy donors in an attempt to restore imbalances in host-microbial ecology [21]. However, clinical response of IBD to FMT has shown extensive inter-study heterogeneity [22], which might stem from the variable engraftment of donor derived microbes and the high or persistent populations of unfavorable pathobionts in the host [23][24][25][26]. In this regard, both antibiotic pre-treatments (to lessen competitive interactions) and increased frequency of FMT delivery may both enhance the engraftment of putatively beneficial microbes, correcting dysbiotic populations, and promoting clinical response and disease remission [27][28][29][30]. While several trials utilizing either antibiotic pre-treatments [31][32][33][34] or repeated FMT regimens [35,36] have been conducted in patients with IBD, no pooled analyses of these findings exist, therefore hindering the optimization of FMT-based IBD therapies.
The purpose of our study was to address this important gap in knowledge by conducting a systematic review and meta-analysis to characterize the effects of antibiotic pre-treatment and repeated FMT dosing on IBD response and remission. Our primary outcome was to compare differences in pooled relapse and remission rates between antibiotic pre-treatment and repeated FMT dosing strategies. Secondary outcomes included comparing differences in fecal microbiota composition associated with disease response and remission for these two approaches.

Eligibility Criteria
A systematic literature search strategy was designed using the Population, Intervention, Comparison, Outcome, and Study Design (PICOS) framework and implemented in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. FMT was defined as the administration of a fecal matter solution from a healthy donor to the gastrointestinal tract of a recipient to confer a health benefit. Our inclusion criteria included studies with adults (age ≥ 18 years) that had a diagnosis for IBD and received FMT. All modalities of FMT delivery, such as colonoscopy, nasogastric tube, oral capsules, or enemas, and any regimen of antibiotic pre-treatment were included. Studies were excluded if disease was localized to the surgical pouch (i.e., pouchitis), patients had concurrent Clostridioides difficile infection, less than six patients were enrolled, or in a pediatric population. Duplicate studies, kin studies, studies using animal models, and non-English studies were also excluded.

Search Strategy
A medical librarian (JK) systematically searched the MEDLINE (via Ovid), Embase (Ovid), Scopus, Web of Science Core Collection, and Cochrane Library (via Wiley) databases on 25 November 2019 (see Supplemental Table S1 for full-text search strategy) and updated on 29 January 2021. No language or date limits were applied. To complement this approach, the research team also screened the first 200 results from Google Scholar for inclusion. Manual searches of references from included studies were further performed to identify potentially missed articles.

Study Selection
Titles and abstracts of relevant articles were first manually screened for inclusion by two independent reviewers (VM, SR). Studies meeting initial screening criteria by at least one reviewer were selected for a full text review by two independent reviewers (VM, SR) using pre-specified inclusion and exclusion criteria. Disputes were resolved by a third reviewer (JD). Data were extracted independently by two reviewers (VM, SR) into separate Excel spreadsheets and cross-examined for accuracy. Studies were then assessed for methodological quality and bias using the Newcastle-Ottawa [37] tool for cohort studies and the Cochrane Risk of Bias [38] evaluation for randomized controlled trials (RCTs).

Data Extraction
Study characteristics were evaluated for study design, year, and country of origin. Primary outcomes of interest included relapse and remission rates following FMT. Secondary outcomes included differences in fecal microbiota composition, and adverse events. Patient characteristics included age, sex, mean disease duration, type of IBD, histology disease scoring, and current medications. FMT strategy-specific variables included donor stool processing, mode of delivery, type of FMT regimen, and type and duration of antibiotic pre-treatments.

Data Synthesis
Continuous data were expressed as mean ± standard deviation (SD). For the purpose of meta-analysis, data extracted as medians and interquartile ranges were converted to mean ± SD using methods outlined by Hozo et al. [39]. Meta-analyses of pooled proportions were conducted using a random effects models by the DerSimonian-Laird method [40]. Estimates of heterogeneity were obtained from inverse-variance fixed-effect models. Pooled estimate variances were stabilized using the Freeman-Tukey Double Arcsine Transformation. Heterogeneity was assessed using the Chi-squared test with significance set at p < 0.10 and the amount of heterogeneity quantified by the I 2 statistic as low <50%, moderate 50-75%, or high >75% [41]. Categorical data were assessed using either Chi-squared or Fischer's exact tests. A two-sided α of less than 0.05 was considered statistically significant. Meta-analysis was conducted using the metaprop function in STATA (v15.1; StataCorp, College Station, TX, USA).

Search Results and Study Characteristics
Comprehensive search of the five databases yielded a total of 4220 results, and after duplicate records were removed, 3624 articles remained ( Figure 1). After initial screening of the titles and abstracts, the text of 45 articles were fully reviewed. Following full text review, 28 articles were eligible for inclusion in the final systematic review. No prior systematic reviews examining FMT outcomes with respect to antibiotic pre-treatment or repeated FMT regimens were identified. Of the included articles, six were randomized controlled trials, 20 were prospective cohort trials, and two studies were case series.
Of the 28 studies reviewed, 22 included patients with UC, four included patients with CD, and two studies assessing both UC and CD. Most studies examined disease response in patients with mild to moderate disease (n = 9 studies), with twelve studies assessing patients with severe disease (Table 1). Study duration and follow-up ranged from 4 weeks to 13 years with most studies having a follow up ≤12 weeks (n = 17). Five studies utilized pre-operative antibiotics prior to FMT, with only two studies utilizing the same antibiotic regiments. Nearly half of the studies included a single FMT delivery (n = 12), while the remaining trials use varied regimens.

Risk of Bias Assessment
Risk of bias for cohort studies was characterized using an adjusted 7-point Newcastle-Ottawa scale of selection, comparability, and study outcome categories (Supplemental Table S2). The 19 included cohort studies demonstrated low to moderate risk of bias due to a lack of long-term follow-up greater than three months (n = 7 studies), and inadequate description or evaluation fecal microbiota changes (n = 8 studies). The six randomized trials were assessed for bias using the Cochrane Risk of Bias tool and together demonstrated low risk of bias (Supplemental Tables S3).

Baseline Demographics
A total of 976 patients were identified from the 28 studies included (Table 2). Twentytwo studies included only patients with UC (n = 767), while three studies included patients with CD (n = 87) alone. The mean weighted age of all patients was 40.0 years, of which 59% were on average male with a mean weighted disease duration of 6.2 years. The proportion of patients receiving concurrent corticosteroids varied extensively from 7% to 100%. Patients with a diverse spectrum of IBD severity were included although the majority of included patients had mild-moderate disease (n = 439; 9 studies). Prior to FMT, total Mayo scores for UC activity ranged from 6.1 to 11.1 and CD activity index ranged from 275 to 345. No significant differences in clinical characteristics were observed between CD and UC patients prior to FMT.

Risk of Bias Assessment
Risk of bias for cohort studies was characterized using an adjusted 7-point Newcastle-Ottawa scale of selection, comparability, and study outcome categories (Supplemental Table S2). The 19 included cohort studies demonstrated low to moderate risk of bias due to a lack of long-term follow-up greater than three months (n = 7 studies), and inadequate description or evaluation fecal microbiota changes (n = 8 studies). The six randomized trials were assessed for bias using the Cochrane Risk of Bias tool and together demonstrated low risk of bias (Supplemental Table S3).

Baseline Demographics
A total of 976 patients were identified from the 28 studies included ( Table 2). Twentytwo studies included only patients with UC (n = 767), while three studies included patients with CD (n = 87) alone. The mean weighted age of all patients was 40.0 years, of which 59% were on average male with a mean weighted disease duration of 6.2 years. The proportion of patients receiving concurrent corticosteroids varied extensively from 7% to 100%. Patients with a diverse spectrum of IBD severity were included although the majority of included patients had mild-moderate disease (n = 439; 9 studies). Prior to FMT, total Mayo scores for UC activity ranged from 6.1 to 11.1 and CD activity index ranged from 275 to 345. No significant differences in clinical characteristics were observed between CD and UC patients prior to FMT.

FMT Administration, Dosing, and Donor Characterization
FMT methodologies varied substantially across all studies. The most frequent mode of FMT was via colonoscopy (n = 19 studies), followed by nasoduodenal/naso-jejunal tube (n = 4 studies), enemas (n = 4 studies), gastroscopy (n = 3 studies), and oral capsules (n = 1 study). The dosage of FMT ranged from 12 g to 300 g of stool per administration with 50% (n = 10 studies) of all studies delivering multiple doses. Antibiotic pre-treatment regimens ranged from three to 14 days prior to FMT (n = 5 studies), with most studies using a combination of antibiotics (n = 4 studies) and specifically vancomycin (n = 3 studies). FMT donors of included studies were typically healthy donors unrelated to the recipients. Nine studies utilized donors that were either relatives or specifically chosen by the patients.
Taken together, pooled response and remission rates were more favorable for patients receiving repeated FMT regimens than single FMT alone. Heterogeneity for all pooled analyses was high with all I 2 values greater than 70%.
Taken together, pooled response and remission rates were more favorable for patients receiving repeated FMT regimens than single FMT alone. Heterogeneity for all pooled analyses was high with all I 2 values greater than 70%.
Similar to the repeated FMT analysis, heterogeneity for the pooled proportion analyses of antibiotic pre-treatment was high.
Similar to the repeated FMT analysis, heterogeneity for the pooled proportion analyses of antibiotic pre-treatment was high.

Overview of Microbiota Reporting of Included Studies
Although FMT aims to shift the gut microbial communities of patients with IBD, only 64% of studies (n = 18 studies) characterized the recipient's fecal microbiota following FMT and only two studies directly assessed associations between IBD remission and fecal microbiota compositional changes (Table 4). Further, no study directly compared microbial changes of antibiotic pre-treatment vs. no pre-treatment or repeated FMT vs. single-dose FMT. Only five studies provided donor microbial characterization. The majority of studies (n = 14 studies) used 16 s rRNA gene amplicon sequencing methods, with three studies using whole-genome sequencing and one using Bacteroides HSP60 sequencing.

Changes in Alpha and Beta Diversity Following FMT
Of these 18 studies, nine (50%) reported an increase in microbial richness and αdiversity following FMT, as estimated by the abundance of operational taxonomic units (OTUs), Chao1, Simpson and Shannon indices. Six studies reported no change in αdiversity after FMT. Changes in β-diversity evaluated using Bray-Curtis dissimilarity were reported in five studies, with the majority (n = 4 studies) showing that the microbial ecology of FMT recipients underwent shifts towards those of their respective donors. Within these four studies, increased engraftment was associated with improved clinical outcomes.
In terms of specific bacterial shifts, the effects of FMT were shown to be highly variable (Table 4). Nonetheless, 15 of the 18 studies (83%) that evaluated for shifts in specific gut microbial taxa reported increases in the abundance of anaerobes purported to produce health promoting anti-inflammatory SCFAs, such as Bifidobacterium, Roseburia, Lachnospiraceae, Prevotella, Ruminococcus, and Clostridium related species.

Recipient and Donor Microbial Ecology Associated with IBD Outcomes
Findings from the two studies that assessed associations between IBD remission and fecal microbiota compositional and functional changes were also variable. Parmsothy et al. provided the best assessment of bacterial taxa and corresponding metabolic pathways related to specific IBD outcomes. Following intensive multi-donor FMT, patients with sustained remission had increased relative abundance of Eubacterium halii, Roseburia inulivorans, and Ruminococcus while those who relapsed had higher proportions of Fusobacterium, Escherichia, and Prevotella. Metabolomics of remission patients further revealed increased activation of metabolic pathways associated with the biosynthesis of SCFAs and secondary bile acids. In addition, only one study by Kump and colleagues explored the role of donor microbiota with respect to IBD outcomes following FMT. Patients that received donor fecal microbiota of greater bacterial richness and α-diversity (assessed by OTU abundance and Shannon diversity) and with increased Ruminococcus and Akkermansia abundances were shown to have higher rates of IBD remission.

Reported Adverse Events
Overall, FMT in patients with IBD was shown to be safe and well tolerated. Frequently reported symptoms related to FMT included a transient self-limiting fever alleviated with paracetamol, and non-specific transient gastrointestinal symptoms such as abdominal discomfort, bloating, nausea, vomiting, and diarrhea (Table 5). Of 26 studies that reported serious adverse events, 13 patients with UC required colectomies and one required hospitalization due to disease progression. One patient also contracted Clostridioides difficile requiring a colectomy and one patient contracted cytomegalovirus infection seven weeks after FMT. Overall, the reported serious adverse events were suggested by the authors to be unrelated to the FMT therapy. No patient receiving FMT intervention in the included studies suffered mortality.

Costello 2019
Single donor FMT (colonoscopy and 2 enemas over a week) Conservative care Anti-TNF alpha blockade therapy or colectomy Table 5. Cont.

Discussion
To our knowledge, we present the first systematic review and meta-analysis evaluating the effects of antibiotic pre-treatment and repeated FMT approaches on improving response in patients with IBD response. Notably, our meta-analysis revealed that repeated FMT and antibiotic pre-treatment were associated with improvements in both pooled IBD response and pooled remission rates. These improvements were associated with key changes in fecal microbial composition such as increased bacterial richness, α-diversity and relative abundance of anaerobes purported to produce SCFAs. Taken together, our findings are novel in that they highlight the potential of these microbiota-targeted strategies to optimize the efficacy of FMT for the management of IBD.
Our findings are in agreement with previous systematic reviews and meta-analyses examining the impact of FMT as a therapy for IBD. In 2014, Colman et al. first identified a lack of literature characterizing FMT treatment efficacy despite publications investigating FMT therapy for IBD more than doubling since 2012 [64]. The systematic review and metaanalysis of 18 studies consisting of 122 IBD patients by Colman and colleagues further revealed that the pooled proportion of patients achieving clinical remission was 36.2% (95% CI 17.4-60.4%). The authors concluded that, while FMT demonstrated variable efficacy, further rigorously designed RCTs were needed to determine efficacy, with a particular need for studies that investigate the effects of FMT frequency and route of administration. More recently, Imdad et al. conducted a 2018 Cochrane review examining FMT therapy on IBD response and remission [65]. Four studies with a total of 277 UC patients were identified and revealed an improved clinical response (RR 1.70; 95% CI 0.98-2.95) and endoscopic remission (RR 2.96; 95% CI 1.60-5.48) for patients receiving FMT vs. placebo. These systematic reviews were, however, limited by a lack of high-quality RCTs and standardized fecal microbiota analysis. Our study addresses a number of these gaps by evaluating both high-quality RCTs and cohort studies, which allowed us to specifically characterize the impact of FMT frequency and antibiotic pre-treatment on IBD outcomes.
Repeated FMT strategies have been employed with variable success in a number of different clinical entities thought to be associated with imbalances in host-microbial ecology [67][68][69]. Perhaps the most compelling evidence for repeated FMT is observed in the Clostridiodes difficile infection (CDI) literature. In a recent systematic review and meta-analysis by Baunwall et al., repeated FMT was found to be superior to single-dose FMT in management of recurrent CDI (91% vs. 84%) [69]. Similarly, El-Salhy et al. demonstrated an increased clinical efficacy for repeated FMT dosing in patients with irritable bowel syndrome, albeit in a small case series of 10 patients [68]. Lastly, in a double-blinded placebo-controlled pilot trial, repeated FMT in patients with obesity and metabolic syndrome demonstrated successful engraftment of donor derived microbes, but without any clinical improvements in host metabolic parameters [67]. These inconsistencies are in large part due to the dramatic study heterogeneity with respect to donor selection, FMT preparation and route of delivery, as well as underlying differences in host-gut microbiome interactions implicated in disease pathophysiology [70]. Notwithstanding, our study findings indeed suggest that repeated FMT dosing provides a promising approach to improve IBD outcomes by facilitating donor microbe engraftment, increase α-diversity, and promote SCFA producing taxa.
Ongoing debate exists regarding the pre-treatment of recipients with antibiotics prior to FMT to increase efficacy [71,72]. Conceptually, antibiotic pre-treatment helps provides a proverbial ecological clean slate for the engraftment of donor microbes by freeing up otherwise occupied niches. Elegant work by Ji et al. compared antibiotic pre-treatment versus bowel cleansing or no pre-treatment in mice prior to FMT. The authors demonstrated that FMT efficacy was dependent on the number of niches available for donor microbe engraftment [73]. Further, they found that antibiotic pre-treatment proved to be the most effective strategy for enhancing host gut microbiota reprogramming by increasing donor microbe colonization. Work by Freitag et al., on the other hand, demonstrated that antibiotic pretreatment prior to FMT in mice had only minor effects on overall donor microbial engraftment [71]. Antibiotics disrupted pre-FMT host microbial communities, yet only select donor-derived bacterial taxa such as Bifidobacterium were increased and no improvements in overall similarity to the donor microbiota were noted. Indeed, questions remain regarding the optimal antibiotic regimens required to make niches accessible, which niches should be targeted for FMT re-colonization, and whether the potential benefit surpasses the potential harm associated with antibiotic resistance and CDI. While our findings are promising as they show improvements in IBD remission and relapse for groups receiving antibiotic pre-treatment prior to FMT, further studies are needed that evaluate the mechanisms and implications of similar approach on IBD.
We acknowledge that our systematic review and meta-analysis has a number of important limitations. Pooled analysis of our primary outcomes demonstrated a high degree of heterogeneity and does not allow for direct comparison of effect size associated with either repeated FMT or antibiotic pre-treatment regiments. The heterogeneity of our results was extensive and, in a large part, due to differences in study design, FMT regimens and individualized responses to FMT. In general, the administration and preparation of FMT is not standardized with practice patterns varying dramatically. Major differences in route of delivery, donor selection, dosing rationale, and antibiotic pre-treatment regimen are all likely to promote inter-study heterogeneity in our review. Follow-up timeframes also ranged from two weeks to 13 years, with nearly half of the studies having a follow up <3 months. This may have introduced a bias towards more favorable clinical response and remission rates following FMT therapy. Therefore, arguments can be made that, given the immense variability of such disparate study interventions, more focused inclusion criteria are warranted in future studies. As this is the first IBD review to evaluate repeated FMT and antibiotic pretreatment concepts, we elected a priori to broadly include all potentially relevant literature in order to highlight current limitations and to allow for explorative hypothesis generation.
Correlations regarding outcomes and antibiotic pre-treatment should also be interpreted with caution given the small proportion of patients within included studies and the lack of direct comparison with patients receiving FMT alone. Histologic assessments preand post-FMT were also not consistently reported across studies hindering our ability to evaluate the histologic effects of FMT on disease activity, or the effects of FMT on mucosal adherent bacterial communities. The findings of our review also heavily favored patients with UC and are therefore less generalizable to CD. Additionally, consistent reporting and analysis of fecal microbiota compositional data for both donors and patients were not reported across all studies, which limits the ability to elucidate potential underlying features of the gut microbiome important for optimizing clinical efficacy. Finally, our literature search revealed a number of abstracts and protocols not ultimately published as final manuscripts, which is indicative of publication bias in the FMT literature.
Despite these limitations, our study provides the first systematic review and metaanalysis that evaluates the impact of two key microbial-based strategies which optimize the efficacy of FMT on IBD outcomes. Results of this study may have a number of important implications. Firstly, we demonstrate that repeated FMT dosing and antibiotic pre-treatment approaches have a promising role in optimizing IBD remission and response rates following FMT. Second, results of this study also highlight a need for standardization of FMT therapy protocols (donor, dose, delivery, and pre-treatment) and reporting of microbial data as the lack of this data seen in current practices preclude meaningful meta-analysis of microbial ecology. Lastly, additional high quality randomized trials are needed which directly compare these two strategies in order to help overcome the high degree of heterogeneity in present studies and to elucidate the mechanisms through which these improved outcomes occur. Only through such standardization practices can we eventually bring tailored microbial transplant therapies from the forefront of current IBD research to standard clinical practice.

Conclusions
Repeated fecal microbial transplantation and antibiotic pre-treatment engraftment strategies in patients with IBD were associated with improvements in pooled response and remission rates following FMT. These improvements were associated with an increase in fecal microbiota richness, α-diversity, and several SCFA-producing anaerobic taxa. Further standardization of FMT therapies is required to bring microbial-targeted therapies based on FMT from the forefront of current IBD research to modern clinical practice.
Supplementary Materials: The following are available online at https://www.mdpi.com/2077-038 3/10/5/959/s1, Table S1: Full-text search strategy of included databases; Table S2: Newcastle-Ottawa scale for assessing risk of bias for included cohort studies; Table S3: Cochrane risk of bias assessment for included randomized trials. Institutional Review Board Statement: Ethical review and approval were waived for this study due to the systematic review study design.
Informed Consent Statement: Informed consent was waived for this study as this was a systematic review.

Conflicts of Interest:
The authors have no conflict to disclose.