Efficacy of Fecal Microbiota Transplantation for Clostridium difficile Infection in Children

Abstract

BACKGROUND & AIMS:

Fecal microbiota transplantation (FMT) is commonly used to treat Clostridium difficile infection (CDI). CDI is an increasing cause of diarrheal illness in pediatric patients, but the effects of FMT have not been well studied in children. We performed a multi-center retrospective cohort study of pediatric and young adult patients to evaluate the efficacy, safety, and factors associated with a successful FMT for the treatment of CDI.

METHODS:

We performed a retrospective study of 372 patients, 11 months to 23 years old, who underwent FMT at 18 pediatric centers, from February 1, 2004, to February 28, 2017; 2-month outcome data were available from 335 patients. Successful FMT was defined as no recurrence of CDI in the 2 months following FMT. We performed stepwise logistic regression to identify factors associated with successful FMT.

RESULTS:

Of 335 patients who underwent FMT and were followed for 2 months or more, 271 (81%) had a successful outcome following a single FMT and 86.6% had a successful outcome following a first or repeated FMT. Patients who received FMT with fresh donor stool (odds ratio [OR], 2.66; 95% CI, 1.39–5.08), underwent FMT via colonoscopy (OR, 2.41; 95% CI, 1.26–4.61), did not have a feeding tube (OR, 2.08; 95% CI, 1.05–4.11), or had 1 less episode of CDI before FMT (OR, 1.20; 95% CI, 1.04–1.39) had increased odds for successful FMT. Seventeen patients (4.7%) had a severe adverse event during the 3-month follow-up period, including 10 hospitalizations.

CONCLUSIONS:

Based on the findings from a large multi-center retrospective cohort, FMT is effective and safe for the treatment of CDI in children and young adults. Further studies are required to optimize the timing and method of FMT for pediatric patients—factors associated with success differ from those of adult patients.

Clostridioides difficile, formerly known as Clostridium difficile, is the most common cause of antibiotic-associated diarrhea and a substantial public health threat.^1,2 Previously thought to be a disease of the elderly and infirm, C difficile infection (CDI) has been increasingly recognized as a significant cause of diarrheal illness in pediatric patients. A population-based cohort study in children demonstrated a 12.5-fold increase in CDI incidence from 1991 to 2009.³

One of the unique challenges associated with CDI is its propensity to recur. After an initial infection, 12%–22% of pediatric patients have a recurrence of disease,^3–5 similar to rates reported in adults.^6,7 The reasons for recurrence are believed to be multifactorial, including dysbiosis of the host microbiome, continued C difficile exposures, and an inadequate host immune response.^8,9

Fecal microbiota transplantation (FMT), the transfer of stool from a healthy individual to a patient, has demonstrated great efficacy in the treatment of severe, refractory, and recurrent CDI in adults. A systematic review, including 7 randomized controlled trials, demonstrated overall clinical resolution of CDI in 92% of adults undergoing FMT.¹⁰ Notably, safety concerns remain, because the downstream consequences of changes in the microbiome induced by FMT are poorly understood. Alterations in the microbiome have been associated with the development of autoimmune, metabolic, and psychiatric diseases.^11–13 The implications of altering the microbiome at such an early age in development make these safety concerns particularly relevant to pediatric patients and warrant further investigation. In addition, infectious and noninfectious complications, including the development of flares of inflammatory bowel disease (IBD), have been reported after FMT in adults.^14–16

FMT has been successfully used in pediatric patients for CDI, with the first reported use in the literature in 2010.¹⁷ All pediatric FMT data currently come from small case series and case reports, with the largest study to date including 47 pediatric patients treated at a single center.^18–22 Recently North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and European Society for Pediatric Gastroenterology, Hepatology, and Nutrition published the first position paper on the use of FMT for CDI in children.²³ However, with the rising incidence of CDI in children and the high rates of recurrent disease, additional data to support the use of FMT in children are urgently needed. This is the largest and first multicenter retrospective cohort study in pediatric and young adult patients to evaluate the efficacy, safety, and factors associated with a successful FMT for the treatment of CDI.

Methods

Setting and Participants

This multicenter retrospective study included pediatric and young adult patients (ages 11 months–23 years) who underwent FMT at 18 pediatric centers across the United States for a diagnosis of CDI from February 1, 2004, to February 28, 2017. Centers were recruited through the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition FMT Special Interest Group (Supplementary Table 1). The institutional review boards of all institutions approved the study.

Data Collection

Study data were collected and managed using REDCap (Research Electronic Data Capture) tools hosted at Vanderbilt University Medical Center.²⁴ Patients who underwent FMT for severe, refractory, and recurrent CDI were eligible for data abstraction. Data were imported into REDCap from the review of clinical records by individual institutions.

The primary aim of the study was to determine the success rate of FMT in pediatric and young adult patients with a diagnosis of CDI. Successful FMT was defined as no recurrence of CDI within 2 months post-FMT per standard definitions.²⁵ Recurrence required both a return of symptoms and positive testing for C difficile based on laboratory testing preference at the study institution. Patients with less than 60 days of follow-up were excluded from the analysis of this aim because of uncertain recurrence status. Patients who underwent FMT for refractory CDI, defined as CDI not responding to conventional treatment,²⁶ were also excluded because of controversies surrounded confirming this diagnosis in pediatric patients. The secondary study aim was to identify factors associated with a successful FMT.

Additional data were collected on post-FMT adverse events (AEs) and severe AEs (SAEs) for 3 months post-FMT. AEs were defined as an undesirable experience in a patient post-FMT. SAEs included death, life-threatening events, hospitalizations, or important medical events.

Data Analysis

Continuous data were assessed for adherence to the normal distribution by examining quantile-quantile plots and the Shapiro-Wilk test. In all cases, data were not normal and are presented as medians with interquartile ranges. Categorical data are described by frequency count and percentage. Unadjusted analyses were performed to compare patients with and without a successful FMT using the Wilcoxon rank sum test for continuous variables and Fisher exact test for categorical variables. Comparison by hospital involved a 17-degree of freedom test with many cell counts <5, and was therefore assessed by Fisher exact test based on Monte Carlo estimation from 10,000 samples. Stepwise logistic regression was performed to determine independent predictors of a successful FMT using criteria P < .10 for an effect to enter the model and P < .05 for it to remain in the model. Continuous variables were confirmed to be linear in the logit using the Box-Tidwell test. Solid organ transplant and short bowel syndrome were present in only 9 and 10 patients, respectively, and each was determined to be confounded by presence of a feeding tube, and subsequently dropped from the regression analysis. The final model was assessed with the Hosmer-Lemeshow goodness-of-fit test.²⁷ All tests of significance were 2-sided, with P < .05 indicating statistical significance. Data were analyzed using SAS version 9.4 (Cary, NC).

Results

Participants

A total of 372 patients were included. Patients had a median age of 10 years (interquartile range, 3–15), with a range of 11 months to 23 years (Table 1). The most common comorbidity was IBD, which was present in 120 (32%) patients. There were 21 patients newly diagnosed with IBD at the time of colonoscopy for FMT. Most children (71%) underwent a vancomycin taper before undergoing FMT, and the median time from initial CDI until FMT was 7 months (interquartile range, 4–12) (Table 2).

Table 1.

Characteristics of Children Undergoing FMT for CDI at 18 Pediatric Centers (N = 372)^a

Demographics
Age, y	10.0 (3.0–15.0)
Female sex	186 (50.0)
Raceb
White	331 (89.0)
Black	15 (4.0)
Asian	7 (1.9)
American Indian	1 (0.3)
Unknown	21 (5.6)
Comorbidities
Inflammatory bowel disease	120 (32.3)
Crohn’s disease	51 (13.7)
Ulcerative colitis	63 (16.9)
Indeterminate colitis	6 (1.6)
Immunocompromisedc	111 (29.8)
Presence of feeding tube	72 (19.3)
Gastroesophageal reflux disease	37 (10.0)
Short bowel syndrome	10 (2.7)
History of solid organ transplant	9 (2.4)
Solid tumor malignancy	9 (2.4)
History of stem cell transplant	6 (1.6)
Hematologic malignancy	6 (1.6)

CDI, Clostridium difficile infection; FMT, fecal microbiota transplantation.

^aData are presented as median (interquartile range) or n (%).

^bPatients may have data in more than 1 category; percentages may not sum to 100%.

^cImmunocompromised patients includes patients with IBD on oral steroids and/or biologics and patients without IBD with a diagnosis of malignancy, solid organ transplant, stem cell transplant, or primary immunodeficiency.

Table 2.

Characteristics of CDI and FMT in Children

CDI history	N = 372a
Number of CDI episodes before FMT (n = 360)	3.0 (3.0–4.0)
Time from initial CDI diagnosis until FMT, mo (n = 345)	7.0 (4.0–12.0)
Hospitalization for a CDI-related event (n = 328)	95 (29.0)
Indication for FMTb
Recurrent CDI	363 (97.6)
Refractory CDI	32 (8.6)
Severe or complicated CDI	11 (3.0)
Antibiotics used before FMTb
Vancomycin (standard course)	337 (90.6)
Metronidazole	309 (83.1)
Vancomycin taper	264 (71.0)
Fidaxomicin	36 (9.7)
Nitazoxanide	34 (9.1)
Rifaximin	28 (7.5)
FMT Procedure
Donor stool selectionb
Patient-selected	161 (43.3)
Commercial stool bank	110 (29.6)
Local stool bank	99 (26.6)
Cleanout before FMT (n = 363)	328 (90.4)
Stool type (n = 370)
Fresh	161 (43.5)
Thawed, previously frozen	209 (56.5)
Type of administration
Colonoscopy	285 (76.6)
Nasogastric/gastronomy tube	34 (9.1)
Nasoduodenal/nasojejunal/duodenalc/jejunostomy tube	33 (8.9)
Capsule	14 (3.8)
Enema	4 (1.1)
Sigmoidoscopy	2 (0.5)
Median volume (mL) of FMT (n = 362)	240 (120–250)
Loperamide used post-FMT (n = 365)	131 (35.9)

CDI, Clostridium difficile infection; FMT, fecal microbiota transplantation.

^aN = 372 unless otherwise specified. Data presented as median (interquartile range) or n (%).

^bPatients may have data in more than 1 category; percentages may not sum to 100%.

^cDuodenal infusions occurred via esophagogastroduodenoscopy.

Outcomes

Efficacy.

Of the 372 patients, 31 were excluded from analysis because they had less than 60 days of follow-up and 6 were excluded because of refractory CDI. The outcome analysis included 335 patients, of which 271 (80.9%) had a successful outcome. In the 64 patients that did have a recurrence of CDI, 34 (53.1%) underwent repeat FMT, which was successful in 19 of the 34 (55.9%). Therefore, the overall success rate of 1 or 2 FMT was 86.6%.

Factors associated with successful fecal microbiota transplantation.

The unadjusted association of factors for primary (first treatment) FMT are shown in Table 3. In addition, we examined age as a categorical variable (0–6 years, 7–12 years, 13–18 years, and 19–23 years) and found no relationship between age and response to FMT (P = .59) by Pearson chi-square test. Response to FMT did not differ by hospital site (P = .07 by Fisher exact test based on Monte Carlo estimation). There was a change in FMT success over time with an increase in the odds of success by 15% for every additional year since 2004 (odds ratio, 1.15; 95% confidence interval [CI], 1.00–1.34; P = .06) (Figure 1).

Table 3.

Unadjusted Predictors of Primary (First Treatment) FMT Failure for the Treatment of CDI in children (N = 335)^a

	FMT
Predictor	Failure (n = 64)	Success (n = 271)	P valueb
Demographics
Age, y	8.0 (4.0–13.5)	10.0 (3.0–15.0)	.38
Female sex, n (%)	28 (43.8)	135 (49.8)	.41
Comorbid disease, n (%)
Immunocompromised	28 (43.8)	83 (30.6)	.04
Inflammatory bowel disease	26 (40.6)	85 (31.4)	.18
Presence of feeding tube	21 (32.8)	44 (16.2)	.005
Short bowel syndrome	5 (7.8)	5 (1.8)	.03
Solid organ transplant	5 (7.8)	4 (1.5)	.01
Medical history, n (%)
History of bowel surgery	12 (18.8)	38 (14.0)	.33
Hospitalization within 3 mo before CDI	9 (14.1)	34 (12.6)	.68
Surgery within 3 mo before CDI	8 (12.5)	13 (4.8)	.04
Hospitalization within 1 y before FMT	16 (25.0)	66 (24.4)	1.00
Prednisone (in patients with IBD)	17 (25.6)	39 (14.4)	.03
Immunomodulator use (in patients with IBD)	15 (23.4)	37 (13.7)	.06
CDI history
Number of CDI episodes before FMT (n = 323)	4.0 (3.0–5.0)	3.0 (3.0–4.0)	.04
Time from CDI diagnosis until FMT, mo (n = 313)	10.0 (5.0–17.0)	7.0 (4.0–12.0)	.004
Fidaxomicin use before FMT, n (%)	11 (17.2)	22 (8.1)	.04
Symptom response to antibiotics before FMT, n (%)			.14
Full improvement	36 (56.3)	176 (64.9)
Partial improvement	26 (40.65)	77 (28.4)
No improvement	2 (3.1)	8 (3.0)
Uncertain/unknown	0 (0)	10 (3.7)
FMT indication, n (%)
Recurrent CDI	64 (100)	268 (98.9)	1.00
Refractory CDI	7 (10.9)	17 (6.3)	.19
Severe CDI	2 (3.1)	7 (2.6)	.68
FMT procedure
Fresh (vs frozen) donor stool (n = 334), n (%)	16 (25.0)	125 (46.3)	.002
Bowel lavage before FMT (n = 328), n (%)	52 (81.3)	245 (92.8)	.008
Delivery: colonoscopy (vs other), n (%)	40 (62.5)	219 (80.8)	.003
Median volume of FMT and interquartile range, mL (n = 325)	225 (60–250)	250 (125–250)	.005
Loperamide use post-FMT (n = 329), n (%)	14 (23.0)	107 (39.9)	.01

CDI, Clostridium difficile infection; FMT, fecal microbiota transplantation; IBD, inflammatory bowel disease.

^aAmong N = 372 patients with data, 37 had primary response unknown. An additional 13 are missing data on at least 1 predictor. N = 335 unless otherwise specified.

^bP value from Wilcoxon rank sum test or Fisher exact test.

Figure 1.

Percentage of successful FMT by year. Shown above each bar is the total number of subjects with FMT per year. The predicted probability and odds of successful FMT are also shown. Each additional year was associated with a 1.15-fold increase in the odds of a successful FMT (95% confidence interval, 1.00–1.34; P = .06).

There were 4 independent predictors of FMT success in multivariable regression analysis (Table 4). The odds of a successful outcome using fresh donor stool were 2.66 times the odds of success using thawed, previously frozen donor stool (95% CI, 1.39–5.08; P = .003). The odds of a successful FMT when delivered by colonoscopy was 2.41 times the odds of all the other modes of delivery combined (95% CI, 1.26–4.61; P = .008). The lack of a feeding tube was associated with an increase in the odds of a successful FMT (odds ratio, 2.08; 95% CI, 1.05–4.11; P = .04). A decrease in the total number of recurrences of CDI before FMT was associated with an increase in the odds of a successful outcome; specifically, 1 less CDI recurrence correlated with a 20% increase in success (odds ratio, 1.20; 95% CI, 1.04–1.39; P = .02).

Table 4.

Logistic Regression of Predictors of Primary FMT Failure for the Treatment of CDI in Children (N = 322)^a

Predictorsb	Odds ratio (95% CI)	P value
Fresh (vs frozen) donor stool	2.66 (1.39–5.08)	.003
Delivery by colonoscopy (vs other means)	2.41 (1.26–4.61)	.008
Absence of feeding tube (vs feeding tube)	2.08 (1.05–4.11)	.04
1 less CDI episode before FMT	1.20 (1.04–1.39)	.02

CDI, Clostridium difficile infection; CI, confidence interval; FMT, fecal microbiota transplantation.

^aAmong N = 372 patients with data, 37 had primary response unknown. An additional 13 are missing data on at least 1 predictor.

^bCandidate predictors included all variables shown in Table 3, except short bowel syndrome and solid organ transplant that were determined to be confounded with presence of a feeding tube.

Additional candidate predictors included in the model were female sex, age, diagnosis of IBD, immunocompromised diagnosis, history of bowel surgery, hospitalization within 3 months before initial CDI, surgery within 3 months before initial CDI, hospitalization within 1 year before FMT, prednisone use, immunomodulatory use, number of months from CDI diagnosis until FMT, fidaxomicin use before FMT, use of a clean out, volume of FMT, response to antibiotic use before FMT, loperamide use post-FMT, and year of FMT. After adjusting for the 4 independent predictors previously mentioned, none of these additional candidates were statistically significant.

Safety.

Of the 335 patients with >60 days of follow-up, 19 (5.7%) had an AE related to FMT with abdominal symptoms including diarrhea, pain, and bloating being the most common. Seventeen patients had an SAE in the 3-month follow-up period (Table 5); 2 were believed to be related to FMT and 5 were possibly related. One patient who had FMT delivery into the proximal jejunum via esophagogastroduodenoscopy had vomiting post-procedure and developed aspiration pneumonia. An additional patient was admitted immediately following the procedure for vomiting and diarrhea. No deaths occurred in the 3-month follow-up period. One patient died from heart failure 6 months post-FMT, but this was not related to the FMT.

Table 5.

Serious Adverse Events in Children Undergoing FMT for CDI

Serious Adverse Eventa	Comorbidity	Days post-FMT event occurred	FMT relationship
Hospitalizations
Aspiration pneumonia	Leukemia	0	Related
Fever with central line, + rhinovirus	Leukemia	7	Unrelated
Lower back fracture	IBD	104	Unrelated
IBD flare (n = 3), colectomy (n = 2)	IBD	2–51	Possibly related
CDI with toxic megacolon	Short bowel syndrome	7	Unrelated
Urinary tract infection	Imperforate anus, atrophic kidney, and hydronephrosis	53	Unrelated
Vomiting and dehydration	Congenital heart disease, short bowel syndrome	0	Related
Small bowel obstruction	GERD, feeding tube	56	Unrelated
Aspiration pneumonia; pancreatitis	IBD, asthma	40	Unrelated
Heart rejection	Heart transplant	59	Unrelated
Other serious adverse events
New diagnosis IBD (n = 2)	None	37–40	Unrelated
Perianal abscess	IBD	84	Unrelated

CDI, Clostridium difficile infection; FMT, fecal microbiota transplantation; GERD, gastroesophageal reflux disease; IBD, inflammatory bowel disease.

^aAll serious adverse events are a single episode unless otherwise noted.

Discussion

This large multicenter cohort study demonstrated very good efficacy of FMT for the treatment of CDI in 335 children and young adults with no episodes of recurrent CDI in 81% of the patients. In the 64 patients that experienced CDI recurrence, 34 (53%) underwent repeat FMT, which was successful in 19 (56%). Thus, the overall success rate of 1 or 2 FMT in the treatment of recurrent CDI was 87%. This is similar to what has been previously reported in adults,^28–31 and demonstrates that FMT is comparably efficacious in children and young adults. Notably, however, the pediatric population with CDI differs from the adult population by sex (adults with CDI are more likely to be female than pediatric patients),³² which may suggest additional differences in the pathophysiology of CDI.

Independent predictors of FMT success identified included the lack of a feeding tube, the use of fresh stool versus thawed and previously frozen stool, delivery via colonoscopy, and a lower number of CDI episodes before undergoing FMT. The presence of a feeding tube was previously identified as a risk factor for primary and recurrent CDI.^4,33 It has been theorized that enteral feeding provides a portal for C difficile spores to access the gut and that the use of formulas alters the intestinal microbiome because of changes in gut absorption and motility.³⁴ It is also plausible that the presence of a feeding tube is a surrogate for other variables, including medical complexity or antibiotic use.

Prior adult studies have compared fresh versus frozen donor stool for CDI. Notably, Lee et al³⁵ did not find a significant difference between the use of frozen-and-thawed versus fresh FMT in their randomized, double-blind, noninferiority trial of adults undergoing FMT for CDI. In the per-protocol population, clinical resolution was 83.5% for the frozen FMT group and 85.1% for the fresh FMT group.³⁵ Jiang et al²⁹ demonstrated the highest cure rates (25/25) for FMT with fresh product, lowest for the lyophilized product (16/23), and intermediate for the frozen product (20/24); however, the difference between fresh and frozen FMT differences did not reach statistical significance.

Even after accounting for multiple other effects in our model, patients in our cohort who received FMT via fresh stool had superior outcomes compared with those administered frozen stool. The reasons that these data differ from prior adult data is unclear. However, the microbial shifts that occur during a freeze-thaw cycle have not been fully delineated and there may be alterations in the viability of critical taxa for our pediatric population. Prospective pediatric trials of FMT for CDI comparing fresh with previously frozen stool are needed to better delineate efficacy and allow for appropriate treatment algorithms. In addition, donor characteristics (ie, age, sex) were not collected and may represent important variables in outcome analysis.

We found colonoscopic delivery of FMT to be superior to other routes of administration. In adult studies, trends have suggested that FMT via colonoscopy is slightly more effective. A systematic review with meta-analysis of demonstrated a significant difference between lower gastrointestinal and upper gastrointestinal routes of delivery, with clinical resolution in 95% versus 88%, respectively (P = .02).¹⁰ In our study, 21 patients were diagnosed with IBD during colonoscopy for FMT. Colonoscopy may be preferable because of the ability to identify additional comorbidities that often confound the diagnosis of CDI. Conversely, colonoscopy does assume some inherent risk, (including risks of anesthesia and perforation), so upper gastrointestinal delivery may be warranted in the appropriate patient.

In our study, patients with fewer episodes of CDI before FMT were more likely to have a successful outcome. This is further supported by Varier et al,³⁶ who demonstrated a cost savings of FMT when compared with vancomycin for the treatment of recurrent CDI.

We also found that FMT was more likely to be successful in more recent years. The reasons for this are not clear, but may include more careful donor selection, scrupulous antibiotic avoidance post-FMT, and the use of more standardized protocols.

Although uncommon, SAEs occurred in 17 patients in the 3 months post-FMT, and 3 (2.5%) of the 120 patients with IBD had a severe flare requiring hospitalization. Prior adult studies have reported a risk of flare in 15% of patients with IBD after FMT.³⁷ Risk versus benefit needs to be carefully considered in pediatric patients undergoing FMT, particularly those with a diagnosis of IBD. In addition, 2 patients were diagnosed with IBD during the follow-up period. It is unlikely that IBD was the result of FMT, especially given the short duration of the follow-up period, but rather the diarrhea attributed to CDI was likely an early presentation of IBD.

Our study is limited by its retrospective nature and by the collection of available data within the clinic record. Because no systematic evaluation of the symptoms after FMT was conducted, the incidence of non-SAEs is likely underestimated and details regarding these events are limited. Abdominal pain and diarrhea are common post-FMT AEs, reported to occur in up to 70% of adult patients.³⁵ Prospective pediatric studies would provide a more comprehensive evaluation of the rates of non-severe AEs. In addition, some of our analysis was done with a small sample size. Other variables that may have impacted the results include the amount of FMT delivered. As a better understanding of FMT is developed, it will be crucial to characterize not only the optimal volume of FMT needed, but the actual dose of bacteria needed for success. We were unable to include clinic site in the logistic regression model because of numerical instability caused by the large number of sites. Finally, we chose a 2-month window for recurrence based on standard definition. However, the true window of recurrence, particularly for patients with comorbidities, may be longer and not captured by this study design.

One of the major concerns about the use of FMT in pediatric patients is the unknown impact of FMT on the developing microbiome. There is the potential for downstream infectious and noninfectious consequences, such as autoimmune disease and metabolic syndrome. Fortunately, through the use of FMT, we are often able to avoid protracted antibiotic courses. Antibiotics, particularly those received early in life, have been associated with increased risk of atopy, IBD, and obesity.^38–40 However, it remains unknown how specific changes in a host microbiome might change long-term health outcomes. Unfortunately, our 2-month follow-up period was not sufficiently long enough to address this important question.

In conclusion, we identified a successful outcome in 81% of pediatric and young adult patients undergoing a single FMT for CDI and 87% when FMT was repeated. SAEs were uncommon. Those patients who had FMT with fresh stool, FMT via colonoscopy, did not have a feeding tube, or had a lower number of CDI episodes before FMT were more likely to have a successful outcome. Our study suggests that FMT is effective and safe for children with CDI and represents a therapy that may be worth consideration earlier in their disease course. Future prospective controlled studies of FMT are essential to fully evaluate the efficacy, safety, optimal timing, dose, delivery route, and preparation of FMT in children with CDI.

What You Need to Know.

Background

CDI is an increasing cause of diarrheal illness in pediatric patients, but the effects of FMT have not been well studied in children. We performed a multicenter retrospective cohort study of pediatric and young adult patients to evaluate the efficacy, safety, and factors associated with a successful FMT for the treatment of CDI.

Findings

In an analysis of data from 335 patients, we found FMT to be effective and safe for the treatment of CDI in children and young adults. Patients who received FMT with fresh donor stool, had the procedure performed by colonoscopy, did not have a feeding tube, or had fewer episodes of CDI before FMT had increased odds for successful FMT. Seventeen patients (4.7%) had a severe adverse event during 3 months of follow up, including 10 hospitalizations.

Implications for patient care

Further studies are required to optimize the timing and method of FMT for pediatric patients—factors associated with success differ from those of adult patients.

Abbreviations used in this paper:

AE

adverse event

CDI

Clostridium difficile infection

CI

confidence interval

FMT

fecal microbiota transplantation

IBD

inflammatory bowel disease

SAE

severe adverse event