Efficacy and safety of fecal microbiota transplantation for treatment of ulcerative colitis: A post-consensus systematic review and meta-analysis

Abstract

BACKGROUND

Numerous studies have assessed the efficacy and safety of fecal microbiota transplantation (FMT) as a therapy for ulcerative colitis (UC). However, the treatment processes and outcomes of these studies vary.

AIM

To evaluate the efficacy and safety of FMT for treating UC by conducting a systematic meta-analysis.

METHODS

The inclusion criteria involved reports of adult patients with UC treated with FMT, while studies that did not report clinical outcomes or that included patients with infection were excluded. Clinical remission (CR) and endoscopic remission (ER) were the primary and secondary outcomes, respectively.

RESULTS

We included nine studies retrieved from five electronic databases. The FMT group had better CR than the control group [relative risk (RR) = 1.53; 95% confidence interval (CI): 1.19-1.94; P < 0.0008]. ER was statistically significantly different between the two groups (RR = 2.80; 95%CI: 1.93-4.05; P < 0.00001). Adverse events did not differ significantly between the two groups.

CONCLUSION

FMT demonstrates favorable performance and safety; however, well-designed randomized clinical trials are still needed before the widespread use of FMT can be recommended. Furthermore, standardizing the FMT process is urgently needed for improved safety and efficacy.

Key Words: Fecal microbiota transplantation, Randomized clinical trials, Remission, Ulcerative colitis, Meta-analysis

Core Tip: We conducted a meta-analysis on the use of fecal microbiota transplantation (FMT) for treating ulcerative colitis (UC), marking the first meta-analysis following the Rome consensus by experts in inflammatory bowel disease. This study stands out as it contributes to the establishment of standard procedures for FMT in UC treatment and facilitates its clinical application. Through a comprehensive analysis of existing research data, we found that FMT holds significant potential in UC treatment and has shown promising efficacy to a certain extent. This finding provides robust support for expanding clinical practices while also suggesting further avenues for research to elucidate the mechanisms and optimal therapeutic strategies of FMT in UC treatment.

INTRODUCTION

Ulcerative colitis (UC) is an inflammatory bowel disease (IBD) characterized by chronic inflammation of the colon and rectum, accompanied by an alternating pattern of relapse and remission. Common symptoms include diarrhea, abdominal pain, and the presence of blood in the stool[1]. Since the 1990s, the incidence of UC has increased in newly industrialized regions such as Asia and South America[2]. The pathogenesis of UC is not fully understood but is currently believed to be genetically and environmentally driven, leading to immune dysfunction and abnormal effects on intestinal microbes[3]. Dysbiosis observed in UC is characterized by specific changes in the intestinal bacterial makeup including a decrease in the population of Bacteroidetes and certain groups of Firmicutes (such as Clostridium IXa and IV groups, Bifidobacteria, Lactobacillus, and Faecalibacterium prausnitzii) and an increase in Proteobacteria and Actinobacteria. The dysbiosis of microbes with related metabolic pathways and molecular mechanisms is crucial in intestinal immunity in patients with UC[4]. Fecal microbiota transplantation (FMT) is the process of transplanting feces from a healthy donor into an unhealthy recipient, used clinically as a therapy for Clostridium difficile infection (CDI). FMT can restore microbial diversity and correct dysbiosis[5]. The success of FMT in treating CDI has attracted much attention, and many have hypothesized its effectiveness for IBD treatment.

Promising results have been published by various retrospective trials on FMT for IBD treatment, particularly in patients diagnosed with UC. Bennett was the first to perform FMT for UC treatment[6]. Several trials on FMT for UC treatment, mostly one-armed cohort studies or case series, have yielded satisfactory results. The first two randomized controlled trials (RCTs) evaluating the efficacy and safety of FMT for the treatment of UC were published in 2015[7,8]. To date, several relevant RCTs have been conducted. However, the treatment methods used varied in these trials, with results that are not easily generalizable. Therefore, the non-standardization of FMT processes must be resolved. The first international Rome consensus on the use of FMT in IBD treatment was recently published, recommending an optimal FMT framework to promote future quality management[9]. Notably, previous systematic reviews and meta-analyses have not proposed to standardize the process of FMT for the treatment of IBD, as the first meta-analysis after this consensus, this study aimed to evaluate the efficacy and safety of FMT for UC treatment to advance FMT process standardization, help implement large-sample and multi-center RCTs, and, ultimately, apply the technology to clinical practice.

MATERIALS AND METHODS

Search strategy

All studies published in PubMed, Cochrane, EMBASE, Wanfang Data, and China National Knowledge Infrastructure by November 2023 were searched using specific search strategies adjusted to individual databases. Table 1 presents the search strategies for these databases. This study considered only original research published in English or Chinese. Two investigators conducted a literature search and assessed the findings, with any discrepancies discussed and resolved by a senior researcher.

Table 1 PubMed search strategy.

Query search term
1	Colitis, Ulcerative [Mesh]
2	Ulcerative Colitis [Title/Abstract] OR Inflammatory Bowel Disease, Ulcerative Colitis Type [Title/Abstract]
3	1 OR 2
4	Fecal Microbiota Transplantation [Mesh]
5	Fecal Microbiota Transplant [Title/Abstract] OR Fecal Transplant [Title/Abstract] OR Donor Feces Infusion [Title/Abstract] OR Intestinal Microbiome Transplant [Title/Abstract] OR Intestinal Microbiota Transplantation [Title/Abstract] OR Microbiota Transfer [Title/Abstract] OR Fecal Fecal Transplantation [Title/Abstract]
6	4 OR 5
7	randomized controlled trial [Publication Type] OR randomized [Title/Abstract] OR placebo [Title/Abstract]
8	3 AND 6 AND 7

Inclusion and exclusion criteria

We included: (1) Randomized controlled trials; (2) studies with adult participants aged 18–70 years with UC; (3) studies on FMT-based intervention administered orally via colonoscopy, nasogastric tube, nasoduodenal tube, and nasojejunal tube or by enema; and (4) studies with explicitly described endpoints. We excluded: (1) Animal or in vitro studies; (2) studies in languages other than English or Chinese; (3) studies that included patients with co-infections; (4) studies without a separate report data on patients with UC or studies containing data from multiple overlapping studies; and (5) reviews and meta-analyses.

Quality assessment

Two authors independently assessed the quality of each study, and divergences were resolved through discussion. Methodological quality was evaluated using the Cochrane risk-of-bias tool, which assesses seven areas: (1) Generating random sequences (potentially introducing selection bias); (2) concealing allocation (potentially introducing selection bias); (3) blinding participants and personnel (potentially introducing performance bias); (4) blinding outcome assessment (potentially introducing detection bias); (5) incomplete outcome data (potentially introducing attrition bias); (6) selective reporting (potentially introducing reporting bias); and (7) other biases. The level of risk of bias was evaluated and categorized as “low,” “high,” or “unclear.”

Data extraction

Duplicates of the retrieved articles were imported and removed using Zotero (version 6.0.30). Two authors independently read the titles and abstracts of these papers, and eligible papers were screened according to the inclusion criteria. Finally, all papers were reviewed in full. The information obtained included specific details such as the name of the author, year of publication, country of origin, type of patients involved, FMT method, control mode used, method of delivery, type of donor, time of evaluation, number of clinical remissions (CR) observed, number of endoscopic remissions (ER) observed, and number of adverse reactions reported. Subsequently, the data were arranged in tables for convenience.

Statistical analysis

We used Review Manager software (version 5.4.1) to perform a meta-analysis and compare the rates of remission and adverse reactions between the FMT and control groups. We combined the risk ratio (RR) and 95% confidence interval (CI) for the data analysis. To evaluate statistical heterogeneity for each meta-analysis, we used the Cochran Q test (c²) and I² method. A P value < 0.1 in the Q test indicated statistical significance. In such cases, fixed-effect models were employed; otherwise, random-effect models were used. The I² method was used to assess the level of heterogeneity, with scores of 0%–30%, 30%–60%, 50%–90%, and 75%–100% indicating low, moderate, substantial, and considerable heterogeneity, respectively. Furthermore, we conducted subgroup analyses based on different factors such as delivery route, control mode, and pre-FMT treatment. All statistical tests were two-tailed, with a significance level of P < 0.05.

RESULTS

Search results

Our search yielded 484 potentially relevant studies. After removing duplicates, 265 articles were evaluated. After the two authors independently screened the titles and abstracts of the articles, 13 full-text articles were retrieved, four of which were excluded. The meta-analysis included a total of nine RCTs that examined the efficacy of FMT in individuals diagnosed with UC[7,8,10-16]. Figure 1 presents a flowchart illustrating the literature screening process.

Figure 1 Flowchart of the literature screening process. RCT: Randomized clinical trial.

Characteristics of the included studies

An overview of the characteristics of the studies included in this meta-analysis is presented in Tables 2 and 3. These studies, published between 2015 and 2022, were all RCTs, focusing specifically on patients with mild-to-moderate UC, with Mayo scores of 4–10 and an endoscopic Mayo sub-score of 1. Regarding transplantation methods, the studies used several modalities such as colonoscopy, enema, and nasoduodenal tubes. Both single and multiple donors were used in the studies, with control groups receiving placebo, autologous FMT, 5-aminosalicylic acid (5-ASA), and standard drug therapy. However, in most of the studies included, CR (Mayo score of ≤ 2) and ER (endoscopic Mayo score of ≤ 1) were the primary and secondary outcomes, respectively. The follow-up period ranged from 7 to 48 weeks, focusing more on 8 and 12 weeks.

Table 2 Characteristics of the included studies.

Ref.	Study design	Intervention			Outcomes
		FMT		Control mode	CR (n/total)		Adverse events	Follow up (weeks)	Reference standard of CR
		pre-FMT therapy	Delivery route		FMT	Control
Moayyedi et al[7], 2015	Double-blind, RCT	NR	Enema	Water enema	9/36	2/35	5	7	Mayo score ≤ 2, endoscopic Mayo score of 0
Rossen et al[8], 2015	Double-blind, RCT	Bowel lavage	Via naso-duodenal tube	AFM	7/17	5/20	4	12	SCCAI scores ≤ 2, Mayo endoscopic score decrease ≥ 1
Paramsothy et al[15], 2016	Double-blind, RCT	NR	Colonoscopic infusion	NR	18/41	8/40	3	8	Mayo score ≤ 2 points with subscores ≤ 1
Costello et al[16], 2017	Double-blind, RCT	NR	Colonoscopic infusion	AFM	19/35	17/34	5	8	Mayo score of ≤ 2 with an endoscopic Mayo score of ≤ 1
Crothers et al[10], 2018	Double-blind, RCT	7 days of antibiotics	Colonoscopic infusion and daily FMTc	Sham FMT	2/7	0/8	NR	12	> 3 point reduction in Mayo score
Midha et al[14], 2018	Double-blind, RCT	NR	NR	NR	12/14	4/14	0	48	Mayo score = 1
Zhang et al[11], 2019	RCT	Sulfapyridine 0.75 g, qid	Enema	Sulfapyridine 0.75 g, qid	48/50	35/50	9	NR	NR
Kedia et al[13], 2022	Open-labeled RCT	FMT + UC-SAID	Colonoscopic infusion	SMT	21/35	10/31	NR	8	Decline in SCCAI > 3
Tkach et al[12], 2022	RCT	Mesalazine 1 g, tid	Colonoscopic infusion	SMT	21/26	19/27	8	8	Partial Mayo score ≤ 2

NR: No report; AFM: Autologous fecal microbiota; SMT: Standard medical therapy; CR: Clinical remission; SCCAI: Simple clinical colitis activity index; UC-SAID: Ulcerative colitis-specific anti-inflammatory diet.

Table 3 Efficacy and safety data of the included studies.

Ref.	Patients		ER		AEs
	FMT (n)	Control (n)	FMT (n)	Control (n)	FMT (n)	Control (n)
Moayyedi et al[7], 2015	36	35	NA	NA	3	2
Rossen et al[8], 2015	17	20	NA	NA	2	2
Paramsothy et al [15],2016	41	40	7	3	2	1
Costello et al [16], 2017	35	34	19	6	3	2
Crothers et al [10], 2018	7	8	3	0	0	0
Midha et al [14], 2018	14	14	8	0	0	0
Zhang et al [11], 2019	50	50	29	15	5	4
Kedia et al [13], 2022	35	31	13	3	NA	NA
Tkach et al [12], 2022	26	27	NA	NA	6	2

Efficacy and safety data of the studies. NA: Not available; ER: Endoscopic remission; AEs: Adverse events.

Quality assessment

To evaluate the reliability of the included studies, we used the Cochrane risk-of-bias tool (Figures 2 and 3). A few studies did not indicate how random sequencing was generated during this evaluation, and allocation was concealed. Additionally, blinding was not fully performed in some studies, resulting in an unclear risk of selection bias and a high risk of performance and detection bias. However, we determined that all included studies had a low risk of attrition and reporting biases.

Figure 2 Quality assessment of the studies.

Figure 3 Percentage risk of bias in the studies.

Metaanalysis of CR

All the included studies report data on CR. In the FMT group, CR was achieved in 157 of 261 patients (60.15%) compared with 100 of 259 patients (38.61%) in the control group. To assess the relationship between CR and FMT, we used a random-effects model while considering the heterogeneity of risk (c² = 13.95; P = 0.08; I² = 43%). The combined RRs and corresponding 95%CIs were calculated, as shown in Figure 4. The findings indicated that the FMT group exhibited better CR than the control group (RR = 1.53, 95%CI: 1.19–1.94, P < 0.0008). Publication bias was not assessed because qualitative and quantitative studies have sensitivities < 10[17].

Figure 4 Forest plot of the result of the meta-analysis of clinical remission.

Subgroup analyses based on CR

To further evaluate how the various methods used in the studies affected the efficacy of FMT, we performed subgroup analyses of CR. These analyses targeted various factors including delivery route, control mode, and pre-FMT therapy. Based on these factors, the participants were categorized into three groups. The delivery route determined the composition of the initial group; Moayyedi et al[7] and Zhang et al[11] administered FMT via an enema, Rossen et al[8] administered FMT via nasoduodenal tube, and others administered FMT via colonoscopy. In comparison, the results (Figure 5A) of the enema group were not statistically significant (RR = 2.07; 95%CI: 0.57-7.51; P = 0.27) and were highly heterogeneous (c² = 3.34; P = 0.07; I² =70%). The group that received colonoscopy infusion exhibited favorable outcomes in UC treatment (RR = 1.41; 95%CI: 1.01–1.96; P = 0.006). Based on the control mode, the sham FMT/water method (RR = 4.61; 95%CI: 1.25-16.94; P = 0.02) had better effects than standard medical therapy (SMT) (RR = 1.34; 95%CI: 1.11–1.62; P = 0.002), while the effect of autologous fecal microbiota (RR = 1.17; 95%CI: 0.78–1.77; P = 0.44) was not statistically significant (Figure 5B). These results showed that FMT was more effective than sham FMT/water, but a comparison with SMT requires further justification. Regarding the analysis based on pre-FMT therapy, there was no significant difference between the 5-ASA group (RR = 1.31; 95%CI: 1.11–1.54; P = 0.001) and the non-5-ASA group (RR = 1.85; 95%CI: 1.28–2.68; P = 0.001) are shown in Figure 5C.

Figure 5 Subgroup analysis of clinical remission based on the delivery route (A), the control mode (B) and pre-fecal microbiota transplantation therapy (C).

Metaanalysis of ER

Six of the studies had data on ER. In the FMT group, 77 of 182 patients achieved ER (43.40%) compared with 27 of 177 patients in the control group (15.25%). The result of the analysis (Figure 6) showed a statistically significant difference (RR = 2.80; 95%CI: 1.93–4.05; P < 0.00001) with low heterogeneity (c² = 4.83; P = 0.44; I² = 0%).

Figure 6 Forest plot for meta-analysis of endoscopic remission.

Metaanalysis of adverse reactions

Kedia et al[13] did not report adverse effects; therefore, we only considered the remaining eight studies while analyzing adverse reaction analyses. Among the 221 patients in the FMT group, 21 (9.29%) experienced adverse reactions compared with 13 of the 228 (5.70%) patients in the control group. However, the findings were not statistically significant (RR = 1.64, 95%CI: 0.85-3.17, P = 0.14), indicating no significant difference in the occurrence of adverse reactions between the FMT and control groups. Furthermore, the meta-analysis revealed minimal variation among the studies (c² = 1.07, P = 0.96, I² = 0%) (Figure 7).

Figure 7 Forest plot for meta-analysis of adverse reactions.

DISCUSSION

Although the mechanism by which dysbiosis of the intestinal flora affects UC has not been fully explored, there is evidence that they are correlated[18]. FMT has emerged as a means of modifying the intestinal microbiome and has already been used clinically for CDI, with several studies confirming its potential in treating UC[19]. This study aimed to investigate the efficacy and safety of FMT for UC treatment. We analyzed data from nine RCTs involving 580 patients. The results indicated that the FMT group experienced notably higher rates of CR and ER than the control group. These findings were statistically significant with minimal variations, demonstrating the potential of FMT for UC treatment. However, several issues still require further research including the selection of fecal donors, choice of delivery route, and handling of the pre-FMT.

The choice of FMT donor may significantly impact the efficacy of FMT in UC treatment[20]. Kazerouni and Wein[21] predicted that collecting stools from multiple donors could improve remission rates. Paramsothy et al[15] reported that FMT from multi-donors achieved post-FMT remission rates (FMT, 27% vs placebo, 8%, P = 0.02) similar to those from single-donors. Therefore, further research is needed to determine whether FMT from multi-donors or a single donor is more effective. However, similar genetics and environments may lead to alterations in the microbiota of related donors, even though fecal microbiota transplant donors for UC do not have UC. In this case, alterations in the microbiota of the related donor may lead to a relapse of UC in the patient; therefore, unrelated donors may be preferred for UC treatment. Donors related to the recipients may share similarities because of shared genes or environment, and using their feces may make the treatment well tolerated but potentially less effective. In such cases, the fecal flora of unrelated individuals may be more effective[22]. Certain studies have discovered “super-donors,” whose fecal material results in significantly more successful outcomes than other donors. However, evidence for “super-donors” remains limited, and more research is needed to fully understand and validate this phenomenon[23]. As the number of RCTs included was insufficient to perform a relevant study, we did not obtain results on donor selection.

The effect of the delivery method of FMT on the treatment of UC remains inconclusive. Notably, there are two main ways to deliver FMT for the treatment of UC: Through the upper gastrointestinal (GI) tract, which involves methods such as gastroscopy and the use of nasogastric or nasojejunal tubes, and through the lower GI, involving techniques such as colonoscopy and retention enemas. Previous studies on FMT for CDI concluded that patients who received FMT via the lower GI route received larger amounts or concentrations of FMT than those who received FMT via the upper GI route[24]. Some hypotheses indicate that administration via the upper GI tract exposes the FMT to gastric acid, potentially affecting its efficacy[25]; however, another study revealed that this effect may not be significant[26]. Our study revealed that subgroup analyses of enema were not statistically significant (RR = 2.07; 95%CI: 0.57-7.51; P = 0.27) with high heterogeneity (c² = 3.34; P = 0.07; I² = 70%). We hypothesized that the source of heterogeneity stemmed from dissimilar treatment methods among researchers. Zhang et al[11] used 5-ASA in the experimental and control groups. whereas Moayyedi et al[7] did not. Although analysis of the group using colonic infusion showed positive results in treating UC (RR = 1.40; 95%CI: 1.05-1.87; P = 0.006), the results of Rossen et al[8] using nasoduodenal tubes were not significant (RR = 1.65; 95%CI: 0.64-4.25; P = 0.30). Therefore, it is necessary to conduct additional trials and studies to determine the optimal delivery route.

Pre-FMT included antibiotic or 5-ASA treatments. To compare the effects of 5-ASA treatment on efficacy, we performed a subgroup analysis, which showed no significant difference between the 5-ASA and non-5-ASA groups. Combining different treatments may be more effective than FMT alone. Although dysbiosis correction alone can benefit certain patients, it may not sufficiently cure the disease owing to its complex pathogenesis[27]. UC has a complex pathogenesis, and dysbiosis is not directly associated with it[28]. Ishikawa et al[29] combined FMT with a triple-antibiotic therapy. This method of using antibiotic pretreatment to reshape the gut flora to match the donor’s flora showed a higher CR rate (43.8%) than the conventional FMT (42.1%). Further research based on pre-antibiotics is needed to confirm their effectiveness.

The altered gut flora in patients who underwent FMT is a concern. Several studies have found differences in the composition and diversity of the gut flora between patients with UC and healthy individuals. These studies showed that patients with UC have fewer Bacteroides, Clostridium XIV, and Firmicutes and higher numbers of Clostridium cluster IX, Bacteroides, and Proteobacteria[8,30,31]. A previous study confirmed that individuals who responded to FMT showed significant changes in bacterial profiles similar to those of the donor microbiota[20]. However, another study demonstrated that most species diversity observed post-transplantation was initially linked to species initially present in the recipient[32].

However, the adverse effects of FMT remain a crucial issue. A meta-analysis of 129 studies showed that 19% of the adverse reactions (diarrhea, bloating, and abdominal pain) associated with FMT were mild, and 1.4% of patients who received FMT developed severe adverse reactions (including infection and death)[33]. Our analysis revealed that the occurrence of adverse reactions was not significantly different between the two groups. To improve the safety of FMT, clinicians should rigorously screen donors to prevent fecal transmission of infectious pathogens to recipients[34]. In addition, we believe that the standardization of FMT processes is a way to improve its safety.

Compared with previous meta-analyses, our study incorporated the most recent relevant studies, most importantly, our study will promote the establishment of a standardized protocol for FMT in the treatment of IBD, which will lead to the implementation of more multi-center, large-scale RCTs and provide more evidence-based evidence for clinical practice; however, it still has limitations. One crucial aspect is the subjective nature of assessing methodological quality, reporting quality, and assessing the quality of the evidence. Consequently, the outcomes may differ based on the judgment calls made by individual researchers when evaluating each factor. Furthermore, there is a lack of standardization in the definitions of FMT-related variables. Data on these variables are often not reported in publications, and no universal trial design exists. Factors such as patient and donor preparation, dosage, route of administration, and follow-up duration varied considerably. Finally, we did not perform a funnel plot analysis, Begg’s test, or Egger’s test because the included studies were few, which would have led to inadequate testing for publication bias.

CONCLUSION

In conclusion, FMT is a promising treatment for UC that enhances CR and ER rates. However, the limited number of insufficient RCTs with large sample sizes and inconsistencies in trial standards need to be addressed.

ACKNOWLEDGEMENTS

We thank Dr. Cui-Ping Bao for reviewing the manuscript.