Fecal Microbiota Transplantation for Clostridium

1 downloads 0 Views 209KB Size Report
Clostridium difficile infection (CDI) is the leading cause of anti- ... Fecal microbiota transplantation (FMT) has been used as a treatment to reconstitute the normal microbial homeo- ...... ditions from depression to inflammatory bowel disease ( 50,51 ). .... human fecal microbiome after bacteriotherapy for recurrent Clostridium.
Fecal Microbiota Transplantation for Clostridium difficile Infection: Systematic Review and Meta-Analysis Zain Kassam, MD, FRCPC1, Christine H. Lee, MD, FRCPC, FIDSA2,3,4, Yuhong Yuan, MD, PhD1,5 and Richard H. Hunt, MB, FRCP, FRCPC, MACG, AGAF1,5 OBJECTIVES:

The clinical and economic burden of Clostridium difficile infection (CDI) is significant. Recurrent CDI management has emerged as a major challenge with suboptimal response to standard therapy. Fecal microbiota transplantation (FMT) has been used as a treatment to reconstitute the normal microbial homeostasis and break the cycle of antibiotic agents that may further disrupt the microbiome. Given the lack of randomized-controlled trials (RCTs) and limitations in previous systematic reviews, we aimed to conduct a systematic review with robust methods to determine the efficacy and safety profile of FMT in CDI.

METHODS:

An electronic search was conducted using MEDLINE (1946–March 2012), EMBASE (1974–March 2012) and Cochrane Central Register of Controlled Trials (2012). The search strategy was not limited by language. Abstract data were excluded and only completed studies that underwent the full, rigorous peer-review process were included. Studies that used FMT via any delivery modality for laboratory or endoscopically proven CDI with clinical resolution as primary outcome were included. A sample size of 10 or more patients was a further criterion. Elements of the Centre for Reviews and Dissemination checklist and the National Institute of Clinical Excellence quality assessment for case series checklist were employed to determine study quality. Eligibility assessment and data extraction were performed by two independent researchers. Both unweighted pooled resolution rates (UPR) and weighted pooled resolution rates (WPR) were calculated with corresponding 95% confidence intervals (CI) for overall studies, as well as predefined subgroups.

RESULTS:

Eleven studies with a total of 273 CDI patients treated with FMT were identified; no RCTs were found as none have been published. Two-hundred and forty-five out of 273 patients experienced clinical resolution (UPR 89.7%; WPR 89.1% (95% CI 84 to 93%)). There was no statistically significant heterogeneity between studies (Cochran Q test P = 0.13, I2 = 33.7%). A priori subgroup analysis suggested that lower gastrointestinal FMT delivery (UPR 91.4%; WPR 91.2% (95% CI 86 to 95%)) led to a trend towards higher clinical resolution rates than the upper gastrointestinal route (UPR 82.3%; WPR 80.6% (95% CI 69–90%)) (proportion difference of WPR was 10.6% (95% CI −0.6 to 22%)). No difference in clinical outcomes was detected between anonymous vs. patient selected donors. There were no reported adverse events associated with FMT and follow-up was variable from weeks to years.

CONCLUSIONS: FMT holds considerable promise as a therapy for recurrent CDI but well-designed, RCTs and

long-term follow-up registries are still required. These are needed to identify the right patient, efficacy and safety profile of FMT before this approach can be widely advocated. SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg

Am J Gastroenterol advance online publication, 19 March 2013; doi:10.1038/ajg.2013.59

INTRODUCTION Clostridium difficile infection (CDI) is the leading cause of antibiotic-associated diarrhea and infection rates are increasing. In

the United States, the incidence of CDI almost tripled between 1996 and 2005 (31/100,000 vs. 84/100,000) (1). The emergence of the hypervirulent NAP1/ribotype 027 strain has been associated

1

Division of Gastroenterology, Department of Medicine, McMaster University Health Science Centre, Hamilton, Ontario, Canada; 2Department of Medicine, Division of Infectious Diseases, McMaster University, Hamilton, Ontario, Canada; 3Department of Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, Canada; 4Hamilton Regional Laboratory Medicine Program, Hamilton, Ontario, Canada; 5Farncombe Family Digestive Health Research Institute, Hamilton, Ontario, Canada. Correspondence: Richard H. Hunt, MB, FRCP, FRCPC, MACG, AGAF, Division of Gastroenterology, Department of Medicine, McMaster University Health Science Centre, Room 4W8A, 1200 Main Street West, Hamilton, Ontario, Canada L8N 3Z5. E-mail: [email protected] Received 21 November 2012; accepted 14 February 2013 © 2013 by the American College of Gastroenterology

The American Journal of GASTROENTEROLOGY

1

REVIEW

CLINICAL AND SYSTEMATIC REVIEWS

nature publishing group

REVIEW

2

Kassam et al.

with a rise in disease severity, with mortality reported in up to 6.9% of cases (2). The associated economic burden is also significant. Nosocomial CDI increases the cost of otherwise matched hospitalizations by fourfold, translating to a cost reported up to $4.8 billion/year in the United States (3,4). There is a growing concern regarding reduced efficacy of antimicrobial therapy for CDI. Since 2000, failure rates of metronidazole for uncomplicated CDI have increased from 2.5% to more than 18% (1,5). After 2 or more episodes of recurrence, the risk of subsequent recurrence exceeds 60% with antibiotic therapy (1,6,7). Given the unacceptable failure rate of antibiotic treatment and in particular recurrence with standard therapy, alternative treatment approaches have been explored for those who have failed standard antibiotic therapy. Fecal microbiota transplantation (FMT) is a strategy that restores the diversity of the gut microflora, which may confer protection against toxigenic C. difficile (8,9). The FMT target patients are those who have traditionally failed conventional metronidazole and/or vancomycin therapy, where there is no subsequent standard therapy for clinical cure. Spontaneous resolution rates have not been described outside of ongoing clinical studies, and symptomatic patients are conventionally treated with long-term vancomycin in an attempt to control symptoms and minimize complications, including the need for hospitalization or colectomy. Given this suboptimal approach, FMT treatment was introduced with the first reported in 1958 by Eiseman et al. (10). Over the subsequent 50 years, promising reports of FMT have suggested an ~90% clinical cure rate for recurrent CDI. However, there have been no published randomized-controlled trials (RCT) (11–14). In the absence of randomized-controlled evidence, two previous systematic reviews have summarized the FMT literature; however, both have methodological limitations (13,14). Gough et al. (13) included both unpublished and abstract data, which has not been subjected to peer-review and thus compromises the quality of their results. Moreover, the largest case series (n = 65) included was an unpublished report and the second largest case series (n = 45) included was from an abstract (13). Accordingly, we are concerned that these reports may be driving the results, especially as the authors did not apply any methodological quality assessment. Additionally, Gough et al. (13) included individual case reports in their review calling into question the inherent publication bias that accompanies such studies. A more recent systematic review by Guo et al. (14) also had limitations . The authors restricted their search to English-language journals and contemporary publications, thus potentially threatening generalizability. Also, from a methodology perspective, the title/abstract search and data extraction were performed by a single reviewer with only a cross-check for the data extraction step instead of having two independent reviewers. Neither systematic review placed a priori inclusion criteria for case series with a sample size threshold to minimize the heterogeneity and bias attached to individual case reports and small case series. To our knowledge, no studies have completed appropriate subgroup meta-analysis. In order to improve on the previous two systematic reviews, we aimed to conduct a systematic review with robust methods to The American Journal of GASTROENTEROLOGY

evaluate the evidence for the efficacy of FMT in CDI and to explore the areas of research, which may be required before establishing its general use.

METHODS Search strategy

An electronic search was conducted using MEDLINE (1946– March 2012), EMBASE (1974–March 2012) and Cochrane Central Register of Controlled Trials (Issue 3, 2012). The search strategy was not limited by language. Abstract data were excluded and only completed studies that underwent the full, rigorous peer-review process were included. Search terms, both free text and medical subject headings (MeSH), included Clostridium, C. difficile, CDI, CDAD, pseudomembranous, enterocolitis, antibiotic diarrhea, etc. Studies identified were then combined using the set operator “AND” with studies identified by the following terms: fecal, stool, transplant, donor, flora, enema, infusion, instillation, reconstitution, microbiota, homologous, bacteriotherapy. Variations of root word were also searched alone or in combination. A recursive search of the bibliographies of all relevant papers was also conducted. Study selection and extraction

Eligibility criteria was determined a priori by the study authors and included FMT via any delivery modality for CDI that had laboratory confirmation, using either enzyme immunoassay or polymerase chain reaction for the presence of toxin and/or gene or endoscopic evidence of pseudomembranes with the clinical criteria of diarrhea. Studies were only included if they specified clinical resolution as the primary outcome and not CDI toxin assay or culture, which is consistent with current guidelines (5,15). Only case series with 10 or more patients were included to minimize bias inherently associated with individual case reports and small cases series. Our primary outcome measure was predefined as clinical resolution of diarrhea regardless of the number of FMT per patient. Two independent reviewers (Z.K. and C.H.L.) reviewed the title and abstract search with inclusion decisions for each paper made independently based on the eligibility criteria. Data extraction for eligible studies was also conducted independently with the use of a standardized, pretested data extraction form. A third party compared data extraction forms (Y.Y.) and any discrepancies were resolved by consensus (with R.H.H.). Information collected included demographic data; community vs. health-care facilityrelated CDI; pre-FMT therapy; FMT delivery modality, FMT dose; FMT timing (from collection to administration); FMT donor (patient-select vs. healthy unrelated volunteer); clinical resolution; adverse events, and follow-up. Methodology quality appraisal

Two investigators independently assessed all studies selected for inclusion in the review for methodological quality using two tools. Core elements of the Centre for Reviews and Dissemination checklist and the National Institute of Clinical Excellence VOLUME 104 | XXX 2013 www.amjgastro.com

Fecal Microbiota Transplantation for Clostridium difficile

Studies excluded (title and/or abstract not appropriate; n = 2679) National Institute of Clinical Excellence (NICE) checklist

1. Were selection/eligibility criteria adequately reported?

1. Case series collected in more than one center?

2. Were patients recruited consecutively?

2. Is the objective of the study clearly described?

3. Were patients recruited prospectively?

3. Are the inclusion and exclusion criteria clearly reported?

4. Was loss to follow-up reported or explained?

4. Is there a clear definition of the outcome reported?

5. Were at least 90% of those included at baseline followed up?

5. Were data collected prospectively?

Studies retrieved for full-text evaluation (n = 30)

Studies excluded (n = 19) -Commentaries without original data (n = 8) -Ineligible sample size (n = 5) -Conference abstracts (n = 4) -Ineligible etiology (n = 2)

6. Is there an explicit statement that the patients were recruited consecutively? 7. Are the main findings of the study clearly described?

Studies eligible for inclusion (n = 11) Figure 1. Flow chart of search.

8. Are outcomes stratified? Total NICE score (out of 8)

(NICE) quality assessment for case series checklist were employed in keeping with previous literature (Table 1) (14,16–18). For subgroup analysis, we used the NICE total score and considered studies with a total score greater or equal to 4 to be “higher quality” and scores < 4 to be “lower quality”. Statistical methods

Clinical resolution rates were used as the primary summary outcome. Both unweighted pooled resolution rates (UPR) and weighted pooled resolution rates (WPR) were calculated with corresponding 95% confidence intervals (CI) for overall studies, as well as predefined subgroups. Weighted pooled rates were analyzed using the random effects model (DerSimonian-Laird method) (19). Heterogeneity across studies was assessed using the I2 statistic and the Cochran Q test, while I2 > 50% or Cochrane Q < 0.10 was considered to indicate significant heterogeneity across studies. Clinical resolution rates between subgroups were compared by calculating the unweighted proportion difference, as well as the weighted proportion difference. Statistical analyses were performed by using statistical software StatsDirect (Version 2.7.8, StatsDirect Ltd., Cheshire, UK).

that did not meet pre-defined sample size eligibility (10,32–34), one study did not concern CDI (35) and in one study 15/16 cases were not clearly CDI in etiology (36). Methodological quality of included studies

The five important criteria in the Centre for Reviews and Dissemination quality assessment checklist were used in keeping with previous literature in this field (14). These addressed selection bias (selection criteria, consecutive cases), attrition bias (follow-up loss rationale with at least 90% follow-up rate) and detection bias (prospective design) (14). None of the eleven studies met all the five criteria (Supplementary Table online). We also utilized the NICE quality assessment tool for evaluating case series studies. Using the NICE criteria, only 2 of the 11 studies were multicenter (37,38) and reviewers felt that 4 studies did not have clear description of the objectives (39–42). Additionally, it was deemed that 4 studies did not state clear inclusion/exclusion criteria (40,42–44) and 1 study did not clearly define the outcome (44). None of the studies was a prospective study or stratified the outcomes. Only 4 studies explicitly stated that there was consecutive case recruitment (39,43,45,46) although all but 1 study had a clear main finding according to reviewers (40). Overall, 7 studies had a NICE total score of greater than or equal to 4 and 4 studies had a score of < 4. No study had a maximum NICE total score of 8, which would be a “high-quality” study (Supplementary Table online).

RESULTS

Patient demographics

The initial search strategy for FMT in CDI yielded 2,709 publications. Of those, 2,679 were excluded after screening titles and abstracts. Subsequently, 30 papers were retrieved in full text and of these, 11 studies met our eligibility criteria (Figure 1). Excluded studies included eight commentaries without original data (20–27), four conference proceedings (28–31), five studies

As is highlighted in Table 2, the 11 selected studies yielded a total of 273 patients (37–47). The majority of studies included both in-patient and ambulatory patients (6/11 studies); 8 of 11 studies reported a mean or median age over 65 years; 65% (178/273) cases were female. Most studies did not report prior relevant comorbid conditions. The majority of studies did not report whether

© 2013 by the American College of Gastroenterology

The American Journal of GASTROENTEROLOGY

REVIEW

Studies identified in search (n = 2,709)

Table 1. Centre for Reviews and Dissemination partial checklist and the National Institute of Clinical Excellence checklist for appraising quality of case series studies Centre for Reviews and Dissemination checklist: core domains

3

4

Kassam et al.

REVIEW

Table 2. Study characteristics

Author (reference)

Sample size

Patient type (in-patient, out-patient, mixed)

CDI type (recurrent, refractory, both)

Donor (patient selected, anonymous, both)

Delivery modality

Stool sample dose/solution

Follow-up data

Total NICE score

Kassam et al. (47)

27

Mixed

Both

Anonymous

Enema

150 g Stool/300 ml sterile water

Mean “427.3 days”

4

Mattila et al. (37)

70

Mixed

Both

Both

Colonoscopy

20–30 ml Stool/ 100 ml water

3 Month and 12 month

5

Kelly et al. (43)

26

Out-patient

Recurrent

Patient selected

Colonoscopy

6–8 Tablespoon stool/ 1,000 ml sterile water or saline; total aliquoted dose 500–960 ml

Mean “10.7 months (range 2–30 months)”

4

Polak et al. (44)

15

NR

Recurrent

Patient selected

Nasojejunal tube

20–50 g Stool/dilute in 50 ml saline

NR

2

Mellows et al. (39)

13

Mixed

Both

Patient selected

Colonoscopy

Stool amount NR/saline (amount NR); Total aliquoted dose 300–600cc

Mean “5 months (range 1–10 months)”

4

Garborg et al. (40)

40

Mixed

Recurrent

Patient selected

Gastroscopy, Colonoscopy

50–100 g Stool/ 250 ml saline; total aliquoted dose ~200 ml

NR

1

Rohlke et al. (38)

19

Out-patient

Recurrent

Patient selected

Colonoscopy

Variable (full quantity-“several ounces”)/ saline (amount: NR); total aliquoted dose 200–300cc

Mean “27.2 months (range 6–65 months)”

5

Yoon et al. (45)

12

NR

Both

Patient selected

Colonoscopy

Stool amount NR/1,000 ml saline; total aliquoted dose ~250–400cc

Mean NR; range “3 weeks to 8 years”

5

MacConnachie et al. (41)

15

Mixed

Recurrent

Patient selected

Nasogastric tube

30 g Stool/150 ml saline; total aliquoted dose 30 ml

Mean NR; median 16 weeks (range 4–24 weeks)

2

Aas et al. (46)

18

Mixed

Recurrent

Both

Nasogastric tube

30 g Stool/50–70 ml saline; total aliquoted dose 25 ml

90 Days

5

LundTonnesen et al. (42)

18

In-patient

Unclear

Anonymous

Colonoscopy, Gastrostomy tube

5–10 g Stool/milk (amount: NR)

“2–3 Weeks”

2

NICE, National Institute of Clinical Excellence; NR, not reported.

CDI was acquired in a health-care facility or in the community (9/11 studies) and only 1 study reported CDI strain type (37). Six studies focused on recurrent CDI as defined by the authors, (38–40,43,44,46) whereas four studies looked at both recurrent and refractory disease and 1 study was unclear on the CDI classification (42). Fecal microbiota transplant characteristics

Two studies utilized unrelated screened FMT donors (42,47) while Mattila et al. (37) used unrelated donors for 9/70 cases and Aas et al. (46) used anonymous donors for 4/18 cases (37,46) and the rest were related donors. The remaining studies used patientselected donors, predominantly spouse or close relative. There were a number of different FMT delivery modalities reported with colonoscopy being used exclusively in five studThe American Journal of GASTROENTEROLOGY

ies (37–39,43,45) and in 17/18 cases in another study (42) with the other case being delivered via a gastrostomy tube. Colonoscopy was used in 2/40 cases in one study (40) with the remainder delivered by gastroscope. Three studies used nasogastric/nasojejunal delivery of FMT (41,44,46) and one study used a retention enema (47) The FMT dose and FMT protocol were also variable, including the FMT suspension, which was reported to contain sterile water, saline, or milk. The timing of FMT was not well-reported but ranged from 6 to 24 h after stool donation. FMT was predominantly delivered as a freshly prepared specimen (10/11 studies) with one study using frozen samples that were subsequently thawed before delivery (42). Five studies used a second FMT in the case of clinical failure, either routinely or in selected cases (38,40,41,44,47). Rates of clinical resolution after a second FMT were variable and ranged from 60% (3/5) to 100% (1/1) (38,47). VOLUME 104 | XXX 2013 www.amjgastro.com

Garborg et al., 2000

0.83 (0.67, 0.93)

MacConnachie et al., 2009

0.73 (0.45, 0.92)

Polak et al., 2011

0.87 (0.60, 0.98)

Lund-Tonnesen et al., 1998

0.83 (0.59, 0.96)

Kassam et al., 2010

0.93 (0.76, 0.99)

Kelly et al., 2012

0.92 (0.75, 0.99)

Mellow and Kanatzar, 2011

0.92 (0.64, 1.00)

Mattila et al., 2012

0.94 (0.86, 0.98)

Rohlke et al., 2010

1.00 (0.82, 1.00)

Yoon and Brandt, 2010

1.00 (0.74, 1.00)

Aas et al., 2003

0.83 (0.59, 0.96)

Combined

0.89 (0.84, 0.93)

0.4

0.6

0.8

1.0

1.2

Clinical resolution rates (95% confidence interval) Figure 2. Meta-analysis plot of weighted clinical resolution rates of fecal microbiota transplantation in Clostridium difficile.

Also, MacConnachie et al. (41) only performed a second (successful) FMT because of relapse after receiving an inadequate initial donor stool sample. Treatment effect

As summarized in Figure 2, 245/273 patients experienced clinical resolution based on the study authors’ definitions (UPR 89.7%; WPR 89.1%, 95% CI 84.0%, 93.3%) with no statistically significant heterogeneity between studies (Cochran Q test P = 0.13, I2 = 33.7%). The data were also examined using our interpretation of clinical resolution after assessing individual cases within each study, when available. Individual cases that were empirically treated with FMT for suspected CDI without a positive CDI test were excluded ((40)—two patients; (45)—one patient). Additionally, cases that achieved clinical resolution after FMT but were retreated with another course of antibiotics at a subsequent date and experienced a further episode of CDI were deemed likely “reinfection” instead of a lack of clinical resolution ((43)—two patients (41); —one patient). Specifically, these cases were independently reviewed by three authors based on the clinical details from the manuscript. Consensus was achieved based on the timing of symptoms (ranged between 29 days and 11 months after clinical resolution post-FMT) and documented reintroduction of antibiotics before diarrhea. Lastly, a case with ongoing symptoms after receiving an incorrectly low volume FMT but experienced clinical resolution after a second correct dose of FMT was deemed clinical resolution, based on our criteria (41). Success rates were similar with 246/270 patients having clinical resolution (UPR 91.1%; WPR 90.9%, 95%CI 85.7%, 94.9%; test for heterogeneity Cochran Q test P = 0.07, I 2 = 41.9%). Selected a priori subgroup analysis was performed when data was available with biological © 2013 by the American College of Gastroenterology

plausibility. None of the subgroups had significant heterogeneity across studies. Our first analysis compared the efficacy of lower gastrointestinal (colonoscopy and enema) to upper gastrointestinal delivery (nasogastric/nasojejunal tube and gastroscopy). When studies had mixed upper and lower gastrointestinal FMT, we were able to separate and analyze individual cases, as raw data for this was available. Eight studies, including two reporting mixed delivery, utilized lower gastrointestinal delivery (40,42). The rate of clinical resolution was 203/222 cases treated with lower gastrointestinal delivery (UPR 91.4%) with a WPR of 91.2% (95%CI 86.0%, 95.2%) with no significant heterogeneity between studies (Cochran Q test P = 0.18; I2 = 31%). Five studies, including two from mixed FMTs, used upper gastrointestinal delivery (40,42). The rate of clinical resolution was 42/51 cases (UPR 82.3%) with a WPR of 80.6% (95%CI 69.3%, 89.8%) with no significant heterogeneity between studies (Cochran Q test P = 0.89, I2 = 0%). The proportion difference between lower and upper gastrointestinal delivery for unweighted rates was 9.1% (95% CI − 0.1%, 22.1%) favoring colonoscopy/enema FMT (P = 0.046). Similarly, the proportion difference for the weighted rate was 10.6% (95% CI −0.6%, 21.8%) with a non-significant difference between groups favoring FMT by colonoscopy/enema (Table 3). Our second analysis compared the efficacy based on the type of donor: patient-selected vs. anonymous healthy donors. Two studies used mixed donor populations; however, cases were separated and individually assessed in the study, which provided raw data (37). The other study was analyzed in the patient-selected arm as 15/18 cases were in this group (46). Nine studies used patient-selected donors including one study with a mixed donor population. The rate of clinical resolution was 196/219 cases (UPR 89.5%) with a The American Journal of GASTROENTEROLOGY

5

REVIEW

Fecal Microbiota Transplantation for Clostridium difficile

6

Kassam et al.

REVIEW

Table 3. Subgroup analysis for fecal microbiota transplantation in Clostridium difficile

Weighted rate (95% CI)

Proportion difference of unweighted rate (95% CI, P value)

Proportion difference of weighted rate (95% CI, P value)

203/222 (91.4%)

91.2 % (86.0%, 95.2%)

9.1% ( − 0.1%, 22.1%), P=0.046

10.6% (−0.6%, 21.8%), NS

42/51 (82.3%)

80.6% (69.3%, 89.8%)

196/219 (89.5%)

89.2% (83.2%, 94.0%)

− 1.2% ( − 8.5%, 9.9%), NS

− 0.7% ( − 10.5%, 9.1%), NS

49/54 (90.7%)

89.9% (80.3%, 96.6%)

Higher (total NICE score ≥4)

173/185 (93.5%)

92.9% (88.8%, 96.1%)

11.7% (3.7%, 21.5%), P=0.003

12.3% (3.5%, 21.1%), P < 0.05

Lower (total NICE score < 4)

72/88 (81.8%)

80.6% (72.0%, 88.0%)

Unweighted rate n/N (percentage)

Lower gastrointestinal delivery (colonoscopy, enema) Upper gastrointestinal delivery (nasogastric/nasojejunal tube, gastroscopy, gastrostomy tube)

Subgroups Delivery modality

Donor type Patients selected (related family member, partner, spouse, close friend) Anonymous Methodological quality

CI, confidence interval; NICE, National Institute of Clinical Excellence, NS, non-significant.

WPR of 89.2% (95%CI 83.2%, 94.0%) with no significant heterogeneity between studies (Cochran Q test P = 0.104, I2 = 39.5%). Three studies used healthy, anonymous volunteers. The rate of clinical success was 49/54 (UPR 90.7%) with a WPR of 89.9% (95% CI 80.3%, 96.6%) and no significant heterogeneity between studies (Cochran Q test P = 0.335, I2 = 8.6%). The unweighted proportion difference between patient-selected and anonymous donors was − 1.2% (95% CI − 8.5%, 9.9%) favoring anonymous donors but was not statistically significant (P = 1). Similarly, the proportion difference for the weighted rate was − 0.7 (95% CI − 10.5%, 9.1%) and not statistically significant (Table 3). Our last subgroup analysis determined if the quality of studies had an impact on the rate of FMT-mediated clinical resolution. Studies with a lower NICE score (total < 4) had clinical resolution rate of 72/88 (UPR 81.8%) with a WPR of 80.6% (95% CI 72.0%, 88.0%) with no significant heterogeneity between studies (Cochran Q test P = 0.83, I2 = 0%). Whereas studies with a higher NICE score (total ≥4) had a clinical resolution rate as 173/185 (UPR 93.5%) with a WPR of 92.9% (95% CI 88.8%, 96.1%) with no significant heterogeneity between studies (Cochran Q test P = 0.46, I2 = 0%). The unweighted proportion difference between lower and higher quality studies was 11.7% (95% CI 3.7%, 21.5%) (P = 0.003), the higher-quality studies showed statistically significant improved clinical resolution. Similarly, the proportion difference for the weighted rate was 12.3% (95% CI 3.5%, 21.1%) (P < 0.05) (Table 3). Adverse events and follow-up

All the studies showed that FMT was safe when administered rectally. Eight studies did not report any adverse events related to FMT; however, adverse events were not actively sought in any of the studies. Three studies that used upper gastrointestinal FMT indicated that the nasogastric/nasojejunal tube could not be ruled out in having a role in an adverse event: upper gastrointestinal The American Journal of GASTROENTEROLOGY

bleed (41), peritonitis (46), and possibly enteritis in a nasojejunal delivery (44). Mortality was reported for each study, and none was attributed to the FMT procedure. Death occurred generally in the critically ill and elderly patients with severe comorbidities. Length of follow-up was variable and inconsistent between studies and in one ranged from 3 weeks to 8 years follow-up (45) (Table 2).

DISCUSSION According to the 11 studies in our review, the overall efficacy of FMT was 245/273 (UPR 89.7%; WPR 89.1%, 95% CI 84.0%, 93.3%) in achieving clinical resolution of CDI. Interestingly, the response may differ between delivery modalities with a weighted proportion difference of 10.6% favoring lower gastrointestinal over upper gastrointestinal delivery, although just barely nonsignificant (95% CI −0.6%, 21.8%). However, these results require verification with formal RCTs. While our review again suggests that FMT is a promising intervention for serious recurrent CDI, after conventional treatments have been exhausted, there are still many questions to be addressed before FMT can be recommended as the routine standard of care. Safety challenges

Although patient perceptions of FMT are positive (48), navigating the known and unknown risks remains a major issue for the clinician managing patients with CDI. Despite rigorous donor screening for infectious disease, communicable disease transmission remains an inevitable risk. These risks are well defined in blood transfusion practice, however, they are less clear in FMT especially as donor screening protocols vary widely, despite a recent call for standardization (49). Additionally, for anonymous “universal” donors it remains unclear how often screening needs to occur. There are also a number of, as yet, unexplored risks in VOLUME 104 | XXX 2013 www.amjgastro.com

FMT as the gut microbiome have been implicated in many conditions from depression to inflammatory bowel disease (50,51). Moreover, the literature is not consistent in long-term followup for adverse events, which have been spontaneously reported rather than actively sought. It is well known that this increases the risk for underreporting. RCTs and long-term adverse outcome registries are required to better define the short and long-term risks of FMT. Classification and research definitions

A major obstacle in FMT research is the lack of a standardized CDI terminology. Published reports of FMT in CDI do not clearly define diarrhea and variable terms such as recurrent, refractory, recalcitrant, relapsing and reinfection were often used interchangeably without clarification. The Infectious Diseases Society of America describes CDI as: (i) the presence of diarrhea, defined as passage of three or more unformed stools in 24 or fewer consecutive hours; (ii) a stool test positive for the presence of toxigenic C. difficile or its toxins or colonoscopic or histopathologic findings demonstrating pseudomembranous colitis. However, the Infectious Diseases Society of America then states the “same criteria should be used to diagnose recurrent CDI” without any further guidance (5). More recently, the Fecal Microbiota Transplantation Workgroup has recommended indications for FMT, which include: “recurrent or relapsing CDI” defined as at least three episodes of mild to moderate CDI and failure of 6- to 8-weeks vancomycin with or without an alternative antibiotic (e.g., rifaximin and nitazoxanide) (49). Establishing the definition of recurrent CDI is essential to strengthen future FMT research and the interpretation of results. Gough et al. (13) attempted to establish operational definitions, however, their definition of “resolution” is not consistent with current guidelines because they include “diagnostic confirmation of the absence of disease” as the current guideline recommends against the laboratory test of cure for CDI (5). Standardized terminology is important from both research and clinical perspectives. One challenge of managing CDI is to distinguish between post-infectious irritable bowel syndrome and recurrent infection given the well-established persistence of the CDI toxin (52–53). Research in developing clinical criteria and laboratory detection methods to distinguish these two conditions will be important for the appropriate management of patients and to determine the true burden of recurrent CDI. Delivery modalities

A number of delivery modalities have been described for FMT: nasogastric/nasojejunal tube, gastroscopy, colonoscopy, and retention enemas. Only two studies have assessed two different delivery modalities and there are no formal studies comparing the outcome of various FMT routes of administration (40,42). Delivery into the small bowel via nasogastric tube or gastroscope is pathophysiologically different to the administration into the colon by colonoscopy or retention enema. Although the possibility of bacterial overgrowth, particularly in the elderly with hypoor achlorhydria has been postulated, the impact and relationship of this and the small bowel on the colonic microbiome is not well established (14). Retention enema is safe, simple and inexpensive, © 2013 by the American College of Gastroenterology

however, the theory that rectal repopulation of the microflora will lead to proximal colonization has yet to be validated and diminished rectal sphincter tone in the elderly may compromise retention of the FMT infusate, and decrease success (47). Colonoscopy, with its ability to deliver large volume FMT throughout the entire colon is appealing, however, the procedure may be unsafe in active C. difficile colitis, increasing the risk of perforation. Moreover, colonoscopy is time consuming, resource intensive and carries other risks associated with the procedure. Certainly, the concept of bowel preparation decreasing the burden of CDI and its spores has been proposed as a potentially beneficial cointervention but there is no data. Although our analysis suggests that colonic FMT may be superior to upper gastrointestinal FMT, RCTs comparing instillation in the upper vs. lower GI tract and enema compared with colonoscopy are needed to identify the best modality. Managing hypervirulent strains

The emergence of the NAP-1/ribotype 027 strain may have decreased overall response to standard antibiotic treatment and changed CDI outcomes. Mattila et al. (37) were the only group to type the CDI strain and report 100% (34/34) clinical resolution with a single FMT via colonoscopy in non-ribotype 027 CDI compared with 89% (32/36) with ribotype 027 CDI (37). Future research should include CDI typing to gain insight into the impact of FMT on hypervirulent strains. Fecal microbiota transplant donor

One key unresolved question is the selection of FMT donor. There is variability in donor types used across the existing reports: unrelated vs. patient-selected (traditionally close relative or spouse). A healthy prescreened, unrelated donor allows prompt treatment as it can take up to 2 weeks to obtain screening laboratory results; this also reduces overall laboratory screening costs and can provide more standardized treatment to patients, as well as across health-care facilities. Additional potential benefit may provide a completely different microbiome compared with that which may be partially shared by a household donor, thus enhancing diversity, and “resetting” the microbiome. This concept, although intuitively attractive, must be confirmed. Petrof et al. (54) have reported success in their design of synthetic stool, which could minimize the theoretical transmission of unknown pathogens. Although using metagenomic data to administer targeted bacteriotherapy holds promise in the approach to recurrent CDI treatment, this is still in an early phase requiring much more research. Future research challenges and directions

As this time, there are no published prospective, randomized trials assessing the efficacy of FMT in comparison with a comparator treatment. However, one RCT has recently been completed (55) and two are currently recruiting patients (56,57). Although these trials promise to answer some of the questions around FMT, there remain serious challenges for trial design. For example, in the FECAL trial only one of three arms receives FMT and as delivery by the upper gastrointestinal route appropriate blinding is not The American Journal of GASTROENTEROLOGY

7

REVIEW

Fecal Microbiota Transplantation for Clostridium difficile

REVIEW

8

Kassam et al.

possible and a strong placebo effect is common with invasive procedures. The concept of reconstituting with the patients own stool (placebo) compared with a donor has been proposed, however, the storage and processing may alter the microbiome and affect outcomes (58). Furthermore, the idea of irradiating fecal supernatant has been proposed but this is expensive and will also likely alter the microflora (59). Despite the challenges moving forward with FMT, for some patients it may be the only treatment option. In this review, we have highlighted several important areas that require further research. For example, is there clinical equipoise for a RCT comparing FMT with vancomycin for first recurrence of CDI, recognizing the high rates of recurrence and high cost of oral vancomycin. Additionally, should there be a third arm to such a RCT with comparison to fidaxomicin, as there is no formal evaluation of its use in recurrent CDI. The implications on resources may be a factor in this type of study given the high costs of vancomycin and fidaxomicin. Accordingly, cost-effectiveness will need to be explored. There is also the need for an accurate laboratory test to determine appropriate treatment response and cure. The advantages to establishing a reliable test are noteworthy including allowing delineation between post-infectious irritable bowel syndrome and recurrent CDI, as well as minimizing unnecessary contact isolation in health-care facilities.

Financial support: None. Potential competing interests: Christine Lee is the principle investigator for a clinical trial, comparing fresh versus frozen-and-thawed fecal transplant for management of recurrent CDI. She is an advisory board member for MikrobEX Inc.

CONCLUSION

1. Kelly CP, LaMont JT. Clostridium difficile—more difficult than ever. N Engl J Med 2008;359:1932–40. 2. Loo VG, Poirier L, Miller MA et al. A predominantly clonal multi-institutional outbreak of Clostridium difficile-associated diarrhea with high morbidity and mortality. N Engl J Med 2005;353:2442–9. 3. Vonberg RP, Reichardt C, Behnke M et al. Cost of nosocomial Clostridium difficile-associated diarrhoea. J Hosp Infect 2008;70:15–20. 4. Dubberke ER, Olsen MA. Burden of Clostridium difficile on the healthcare system. Clin Infect Dis 2012;55(Suppl 2): S88–92. 5. Cohen SH, Gerding DN, Johnson S et al. Clinical practice guidelines for Clostridium difficile infection in adults: 2010 update by the society for healthcare epidemiology of America (SHEA) and the infectious diseases society of America (IDSA). Infect Control Hosp Epidemiol 2010;31: 431–55. 6. Pepin J, Saheb N, Coulombe MA et al. Emergence of fluoroquinolones as the predominant risk factor for Clostridium difficile-associated diarrhea: a cohort study during an epidemic in Quebec. Clin Infect Dis 2005;41:1254–60. 7. Petrella LA, Sambol SP, Cheknis A et al. Decreased cure and increased recurrence rates for Clostridium difficile infection caused by the epidemic C difficile BI strain. Clin Infect Dis 2012;55:351–7. 8. Chang JY, Antonopoulos DA, Kalra A et al. Decreased diversity of the fecal microbiome in recurrent Clostridium difficile-associated diarrhea. J Infect Dis 2008;197:435–8. 9. Khoruts A, Dicksved J, Jansson JK et al. Changes in the composition of the human fecal microbiome after bacteriotherapy for recurrent Clostridium difficile-associated diarrhea. J Clin Gastroenterol 2010;44:354–60. 10. Eiseman B, Silen W, Bascom GS et al. Fecal enema as an adjunct in the treatment of pseudomembranous enterocolitis. Surgery 1958;44:854–9. 11. Borody TJ, Khoruts A. Fecal microbiota transplantation and emerging applications. Nat Rev Gastroenterol Hepatol 2011;9:88–96. 12. Bakken JS. Fecal bacteriotherapy for recurrent Clostridium difficile infection. Anaerobe 2009;15:285–9. 13. Gough E, Shaikh H, Manges AR. Systematic review of intestinal microbiota transplantation for recurrent Clostridium difficile infection. Clin Infect Dis 2011;53:994–1002. 14. Guo B, Harstall C, Louie T et al. Systematic review: faecal transplantation for the treatment of Clostridium difficile-associated disease. Aliment Pharmacol Ther 2012;35:865–75. 15. Bauer MP, Kuijper EJ, van Dissel JT et al. European Society of Clinical Microbiology and Infectious Diseases (ESCMID): treatment guidance

Study Highlights WHAT IS CURRENT KNOWLEDGE

3The clinical and economic burden of Clostridium difficile infection (CDI) is significant. 3The overall efficacy of current antibiotics for treatment of recurrent CDI is suboptimal. 3Case reports and case series have suggested that fecal microbiota transplantation (FMT) may be a possible treatment for recurrent CDI.

WHAT IS NEW HERE

3Based on the uncontrolled observational data, FMT appears to be effective in treating recurrent CDI but there are no published randomized-controlled trials.

3Lower gastrointestinal FMT delivery via colonoscopy or enema may be preferred to upper gastrointestinal delivery although randomized controlled trials are required to confirm.

REFERENCES

Overall, FMT holds considerable promise as a therapy for recurrent CDI but despite being reported over 50 years ago many serious questions still remain to be answered before this approach can be widely advocated. Much quality research is still needed because we lack the right evidence to delineate the ideal patient to target for this treatment. Many questions remain about the administration of FMT to ensure the most safe and effective modality. Well-designed RCTs and long-term follow-up registries are needed to capture the efficacy and safety profile of FMT. In the interim, patients who do undergo FMT, when no alternative treatment exists, must be counseled on both the known and potential unknown risks during the informed consent process. ACKNOWLEDGMENTS

We would like to thank Dr Premysl Bercik and Karin Dearness for their assistance in translation support. CONFLICT OF INTEREST

Guarantor of the article: Richard H. Hunt, MB, FRCP, FRCPC, MACG, AGAF. Specific author contributions: Zain Kassam: acquisition of data; analysis and interpretation of data; wrote the manuscript; Christine Lee: acquisition of data, critical revisions of the manuscript for important intellectual content; Yuhong Yuan: statistical analysis and critical revisions of the manuscript for important intellectual content; Richard Hunt: study concept and design; critical revisions of the manuscript for important intellectual content; study supervision. All authors discussed the results and implications, as well as commented on the manuscript at all stages. The American Journal of GASTROENTEROLOGY

VOLUME 104 | XXX 2013 www.amjgastro.com

document for Clostridium difficile infection (CDI). Clin Microbiol Infect 2009;15:1067–79. 16. Chambers D, Rodgers M, Woolacott N. Not only randomized controlled trials, but also case series should be considered in systematic reviews of rapidly developing technologies. J Clin Epidemiol 2009;62:1253–60. 17. Moga C, Guo B, Schopflocher D et al. Development of a quality appraisal tool for case series studies using a modified delphi technique. Inst Health Econ 2012; www.ihe.ca/documents/caseseriesstudiesusingamodifiedDelphiTechnique.pdf (accessed October 2012). 18. National Collaborating Centre for Acute Care.. Preoperative tests: the use of routine tests for elective surgery: evidence, methods & guidance. National Institute for Clinical Excellence (NICE) 2003. 19. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88. 20. Martin L. Modified fecal transplantation. J Clin Gastroenterol 2011;45:742. 21. Ho N, Prasad V. Clostridium difficile diarrhea and fecal transplantation. J Clin Gastroenerol 201;45:742–3. 22. Rubin TA, Gessert CE, Aas J. Stool transplantation for older patients with Clostridium difficile. J Am Geriatr Soc 2009;57:2386. 23. Midtvedt K. Clostridium difficile-associated diarrhea treated with homologous feces. Tidsskr Nor Laegeforen 1998;30:118:1758. 24. Tvede M, Rask-Madsen J. Bacteriotherapy for Clostridium difficile diarrhea. Lancet 1990;335:110. 25. Wilson KH. Bacteriotherapy for Clostridium difficile colitis. Lancet 1989;2:1096. 26. Seal DV, Borriello SP, Piper M. Bacteriotherapy for Clostridium difficile diarrhoea. Lancet 1989;2:558. 27. Collins DC. Pseudomembranous enterocolitis. Further observations on the value of donor fecal enemata as an adjunct in the treatment of pseudomembranous enterocolitis. Am J Proctol 1960;2:389–91. 28. Wilcox GM. Early experience with Fecal Bacteriotherapy (FB) program for recurrent C difficile infection (CDI). Gastroenterology 2011;140:S361. 29. Kassam Z, Hundal R, Marshall JK et al. Fecal transplantation via retention enemas is effective for recurrent or refractory Clostridium difficileassociated diarrhea. Gastroenterology 2010;138:S207–8. 30. Arkkila PE, Uusitalo-Seppala R, Lehtola L et al. Fecal bacteriotherapy for recurrent Clostridium difficile infection. Gastroenterology 2010;138:S5. 31. Keller JJ, Van Nood E, Speelman P et al. Antonie van Leeuwenhoek. 2009;95:S40–1. 32. Hellemans R, Naegels S, Holvet J. Fecal transplantation for recurrent Clostridium difficile colitis, an underused treatment modality. Acta Gastroenterol Belg 2009;72:269–70. 33. Harkonen N. Recurrent pseudomembranous colitis treated with the donor feces. Duodecim 1996;112:1803–4. 34. Schwan A, Ajolin S, Trottestam U et al. Relapsing Clostridium difficile enterocolitis cured by rectal infusion of homologous faeces. Lancet 1983;2:845. 35. Grehan MJ, Torody TJ, Leis SM et al. Durable alteration of the colonic microbiota by the administration of donor fecal flora. J Clin Gastroenterol 2010;44:551–61. 36. Bowden TA, Mansberger AR, Lykins LE. Pseudomembraneous enterocolitis: mechanism for restoring floral homeostasis. Am Surg 1981;47:178–83. 37. Mattila E, Uusitalo-Seppala R, Wuorela M et al. Fecal transplantation, through colonoscopy, is effective therapy for recurrent Clostridium difficile infection. Gastroenterology 2012;142:490–6. 38. Rohlke F, Surawicz CM, Stollman N. Fecal flora reconstitution for recurrent Clostridium difficile infection: results and methodology. J Clin Gastroenterol 2010;44:567–70. 39. Mellow MH, Kanatzar A. Colonoscopic fecal bacteriotherapy in the treatment of recurrent Clostridium difficile infection—results and follow up. J Okla State Med Assoc 2011;104:89–91.

© 2013 by the American College of Gastroenterology

40. Garborg K, Waagsbo B, Stallemo A et al. Results of faecal donor instillation therapy for recurrent Clostridium difficile-associated diarrhoea. Scand J Infect Dis 2010;42:857–61. 41. MacConnachie AA, Fox R, Kennedy DR et al. Faecal transplant for recurrent Clostridium difficile-associated diarrhoea: a UK case series. QJM 2009;102:781–4. 42. Lund-Tonnesen S, Berstad A, Schreiner A et al. Clostridium difficileassociated diarrhea treated with homologous feces. Tidsskr Nor Laegeforen 1998;118:1027–30. 43. Kelly CR, de Leon L, Jasutkar N et al. Fecal microbiota transplantation for relapsing Clostridium difficile infection in 26 patients: methodology and results. J Clin Gastroenterol 2012;46:145–9. 44. Polak P, Freibergerova M, Jurankova J et al. First experiences with faecal bacteriotherapy in the treatment of relapsing pseudomembranous colitis due to Clostridium difficile. Klin Mikrobiol Infekc Lek 2011;17:214–7. 45. Yoon SS, Brandt LJ. Treatment of refractory/recurrent C. difficile-associated disease by donated stool transplanted via colonoscopy: a case series of 12 patients. J Clin Gastroenterol 2010;44:562–6. 46. Aas J, Gessert CE, Bakken JS. Recurrent Clostridium difficile colitis: case series of 18 patients treated with donor stool administered via a nasogastric tube. Clin Infect Dis 2003;36:580–5. 47. Kassam Z, Hundal R, Marshall JK et al. Fecal transplant via retention enema for refractory or recurrent Clostridium difficile infection. Arch Intern Med 2012;172:191–3. 48. Zipursky JS, Sidorsky TI, Freeman CA et al. Patients attitudes towards the use of fecal microbiota transplantation in the treatment of recurrent Clostridium difficile infection. Clin Infect Dis 2012 ; Oct 22 (e-pub ahead of print). 49. Bakken JS, Borody T, Brandt LJ et al. Treating Clostridium difficile infection with fecal microbiota transplantation. Clin Gastroenterol Hepatol 2011;9:1044–9. 50. Collins SM, Surette M, Bercik P. The interplay between the intestinal microbiota and the brain. Nat Rev Microbiol 2012;10:735–42. 51. Sartor RB. Microbial influences in inflammatory bowel disease. Gastroenterology 2008;134:577–94. 52. Piche T, Vanbiervliet G, Pipau FG et al. Low risk of irritable bowel syndrome after Clostridium difficile infection. Can J Gastroenterol 2007;21:727–31. 53. Dai C, Jiang M. The incidence and risk factors of post-infectious irritable bowel syndrome: a meta-analysis. Hepatogastroenterology 2012;59:67–72. 54. Petrof E, Gloor GB, Vanner S et al. RePOOPulating the gut: use of synthetic stool to cure recurrent C. difficile infection. Can J Gastroenterol 2012;A042. 55. van Nood E. The FECAL trial, fecal therapy to eliminate Clostridium difficile associated longstanding diarrohoea. Netherlands Trial Register. NTR Number 1177. http://www.trialregister.nl/trialreg/admin/rctview. asp?TC=1177(accessed November 2012). 56. Lee CH, Petrof E . Multi-centre trial of fresh vs frozen-and-thawed HBT (Fecal Transplant) for recurrent CDI. U.S. National Institutes of Health Trial Register. NCT Number 01398969. http://www.clinicaltrials.gov/ct2/ show/NCT01398969?term=fecal+transplant&rank=2 (accessed October 2012). 57. Hota S. Oral vancomycin followed by fecal transplant vs. tapering oral vancomycin. U.S. National Institutes of Health Trial Register. NCT Number 01226992. http://www.clinicaltrials.gov/ct2/show/NCT01226992?term=feca l+transplant&rank=1(accessed October 2012). 58. Kelly CR . Fecal transplant for relapsing C. difficile infection. U.S. National Institutes of Health Trial Register. NCT Number 01703494. http://www. clinicaltrials.gov/ct2/show/NCT01703494?term=fecal+transplant&rank=4 (accessed October 2012). 59. Cuzzolin L, Zambreri D, Donini M et al. Influence of radiotherapy on intestinal microflora in cancer patients. J Chemother 1992;4:176–9.

The American Journal of GASTROENTEROLOGY

9

REVIEW

Fecal Microbiota Transplantation for Clostridium difficile