Fecal Microbiota Transplantation (FMT) in Clostridium Difficile Infection (CDI) Not Responding to Antibiotics

This is a prospective, descriptive cohort study. Patients will receive current standard of
care.

Once a potential subject is planned for FMT, they will be approached by study coordinators or
investigators for participation in the study. If they agree to participate in the study, then
patient health, medical and demographic information will be collected from the patient and/or
from the medical record and entered into the research database, and associated with a
research code with which the stool samples will also be labeled.

The study coordinator will be responsible for ensuring completion of the study questionnaire,
which consists of pre- and post-FMT sections. The questionnaire will be attempted to be
administered in person by the physician or study coordinator at a clinic visit, but may on
occasion be completed at home and mailed in to the study coordinator or physician, or
completed over the phone with the study coordinator or physician.

Stool preparations will be obtained by a nonprofit third party, OpenBiome, with whom the
University of North Carolina has collaborated for over one year. Preparations include a 250mL
preparation for lower administration, a 30mL preparation for upper administration, and frozen
capsules (30 in one treatment) for oral delivery.

The actual FMT procedure will be performed in the majority of cases by colonoscopy. Other
possible routes of administration are via nasogastric or nasoduodenal tube (with upper
formulation), enema and frozen oral tablets. The current preferred route of administration is
via colonoscopy, based on observation (no head-to-head) trials showing improved efficacy, but
standards of care may change in the future. In all cases, the route of administration will be
based on patient characteristics and preferences, not for purposes related to the study. In
all cases, informed consent for the treatment will be documented as guided by the Food and
Drug Administration (FDA). The patient will stop all antibiotics (including vancomycin or
other antibiotics against C. diff) at least 2 days prior to FMT. For all administrations
other than oral capsules, patients will perform a bowel preparation with 4 liters of Golytely
or equivalent osmotic laxative split between the evening prior to the procedure and the
morning of the procedure. In a minority of instances, especially for pediatric patients, the
patient may be electively admitted to the hospital the night before, to monitor medical
stability during the preparation, facilitate bowel preparation, and provide intravenous
hydration. The bowel cleansing preparation is generally taken orally, though in rare
instances must be administered via nasogastric or nasoduodenal (NG/ND) tube. Hospital
admission and bowel preparation via NG/ND tube will only be performed if indicated as
standard-of-care, not for purposes specifically related to the study.

For colonoscopic FMT, the procedure will be performed with anesthesia support at the UNC
Memorial Hospital endoscopy unit. Sedation will be at the discretion of the treating
anesthesiologist in agreement with the endoscopist, but generally consists of intravenous
propofol, with monitored anesthesia care. For colonoscopic FMT, the patient is given two 2mg
doses of loperamide orally, two hours prior to FMT, and is then to take an additional two
capsules after recovery from anesthesia (recommended immediately upon arrival home, when an
outpatient procedure).

Donor stool is obtained from OpenBiome, a third-party stool bank that screens donors for
medication use, allergies, and health-related conditions and prepares stool per a standard
protocol including suspension in normal saline. 250ml of donor stool is used for colonoscopic
FMT in patients with over one meter squared of body surface area. For small children this
quantity may be reduced proportionately. The entire 250ml's are placed into the ileum, cecum
and/or ascending colon, or at the extent reached by the colonoscope, via the operative
channel of the colonoscope. Patients are asked to maintain a right lateral recumbent position
as much as possible for the hour after infusion, to encourage permanence of the material in
the proximal colon. Of note, the FMT colonoscopy does not include meticulous mucosal
examination if not otherwise indicated, and as such is not a substitute for routine
colorectal cancer screening. Patients are monitored by post-anesthesia care unit (PACU)
personnel until ready for discharge, per standard unit protocol.

For nasogastric or nasoduodenal administration, 30ml of donor stool is infused slowly over 1
hour. Before infusion, NG/ND tube placement is confirmed by X-ray per standard hospital
protocol, according to standard hospital consent form.

In rare cases where procedural sedation, colonoscopy or NG/ND tube is contraindicated, frozen
oral capsules or enema may be used to administer FMT. With capsule administration, in
addition to stopping antibiotics two days prior to the procedure, the patient takes proton
pump inhibitor on the day before treatment and the day of treatment. There is no bowel
preparation. The patient takes the 30 capsules orally under direct physician supervision
within 90 minutes. Metoclopramide 5-10 mg x i may be taken one hour prior to the treatment.
With enema administration, bowel preparation will be performed prior to administration. Bowel
cleansing preparation, loperamide administration, and right lateral recumbency will be
attempted as per colonoscopic FMT. The enema will be administered under direct physician
supervision, as well.

For the study, subjects who are undergoing the above treatment per standard of care will be
asked to submit four (4) total stool samples. The first will be pre-FMT, at any time from 2
weeks prior to FMT (possibly at a clinic visit), up until directly before the pre-FMT bowel
preparation. Post-FMT stool samples at 2 weeks (range 1-3 weeks), 2 months (range 7-9 weeks)
and 6 months post-FMT.

All stool samples will be collected according to the same protocol, described in a handout
(entitled "Stool Collection Instructions") provided to participants. There are several
options for submitting the sample, basically according to patient convenience. If they can
produce a sample at a clinic visit, they can submit at that time. Alternatively, they will be
provided supplies to safely collect and then store samples they collect at home in their own
freezer. They will then have to deliver these samples to UNC (or possibly North Carolina
State University [NCSU] if that is more convenient for a particular subject), which can be
done at one visit with all of the samples, or in segments as they collect them. Before
submission to the biobank, the samples will be labeled with the research code and any other
identifying labels on the stool sample will be removed and destroyed. The de-identified stool
specimens will be aliquoted in the research laboratory for storage at -80° Celsius.

They will be transported by the study coordinators from UNC to a storage unit at NCSU.
Research samples will be stored in a manner that will permit subsequent microbial DNA
extraction, microbial cultivation and metabolite analysis.

They will also be asked to complete a post-FMT section of the questionnaire, from which the
main outcomes of FMT will be determined. They will be evaluated at two post-FMT clinic
visits, which may correspond to first two post-FMT stool samples, where evidence for any
possible adverse events or recurrences will be personally queried and evaluated. If patients
do not attend any scheduled post-FMT visits, the study coordinator will attempt to contact
them by phone and then by mail, both to encourage them to reschedule their clinic visit, of
if this is not possible, then to at least obtain the questionnaire data.

Primary aim (clinical outcomes): Descriptive statistics of the clinical outcomes, such as
mean and median number of days of symptom duration, primary cure rate, percentage of subjects
with resolution of diarrhea, abdominal pain, fatigue, etc will be calculated. Standard
deviations and/or confidence intervals will be provided where appropriate. If statistical
power allows, multivariate logistic regression models may be used to identify patient-level
risk factors predictive of primary cure (such as sex, age, independent-living status, etc.).

Secondary aims: Differences in pre- and post-FMT stool composition, and between early and
late post-FMT samples, will be calculated with McNemar's test for significant differences in
proportions of rRNA (ribosomal ribonucleic acid) sequences pertaining to the relevant
bacterial phyla (Bacteroidetes, Firmicutes, and Proteobacteria), and paired t-tests and
ANOVA's (or their non-parametric counterparts) for significant differences in Shannon
diversity indices (a measure of bacterial diversity) and in concentrations of multiple bile
acids. If these measures are different from pre- to post-FMT samples, then the relationship
of percent differences in each stool parameter to clinical outcome will be evaluated using
logistic regression with primary outcome of "primary cure" within two months of FMT, and a
Cox proportional hazards model with the time variable as "days to primary cure" and the
censor variable as "primary cure." Secondary analyses may be conducted with "any cure" as the
outcome variable if repeat FMT is frequently needed. The investigators will also compare a
sample of donor stool to the post-FMT samples with the same statistical tests, under the
hypothesis that the closer that post-FMT stool resembles donor stool (and the more durable
this resemblance), the better the clinical response. All statistical analyses will be
performed at an alpha of 0.05.

Inclusion criteria:

1. As subjects must meet standard clinical criteria for candidacy for FMT, all patients
will have suffered symptomatic diarrhea and have documented stool c. difficile toxin
positivity. Almost all patients will have failed standard therapy for CDI, which
includes at least one course of treatment with a primary agent (metronidazole,
vancomycin or fidaxomicin), as well as an extended vancomycin taper following either
the second or third recurrence. There may be extenuating circumstances that could
justify FMT for CDI outside of these parameters, such as severe allergy to multiple of
the above agents, though this would be expected to be exceedingly rare.

2. All patients must be able to tolerate either colonoscopy (including anesthesia and
sedation) or placement of a nasogastric/nasoduodenal tube, without a contraindication
to the procedure.

Exclusion criteria:

1. Pregnancy

2. Prisoners (for concerns of vulnerability and logistics)

3. Adult patients who are themselves unconsentable at time of FMT.

4. Life expectancy judged to be <3 months

5. Need for extended antibiotics at baseline, for an indication other than treatment for
CDI

6. Any patient who the clinician judges to be a poor candidate for FMT, for reasons of
comorbidities or not being able to tolerate the procedure for administration, would by
definition be excluded from the study.