Does the Donor Matter? Results from PUNCH CD 2, a Randomized Controlled Trial of a Microbiota-based Drug for Recurrent Clostridium difficile Infection

Arnab Ray, MD; Courtney Jones, BS; Bill Shannon, PhD, MBA; Sharina Carter, BS

Download this poster (981 KB)
Download the abstract (92 KB)

ACG 2016 Annual Scientific Meeting
October 15-19, 2016, Las Vegas, NV


  • Questions have been raised about the applicability of a universal donor in regard to microbiota-based therapies for recurrent Clostridium difficile infection (CDI).
  • Conventional procedure has been to use a second donor in case of treatment failure.
  • Analysis of PUNCH CD, an open-label Phase 2 study of RBX2660, a microbiota-based drug targeted at recurrent CDI, found that the specific donor used to manufacture drug product did not affect outcomes.1
  • Donor vs. patient factors in therapy outcomes were further studied in PUNCH CD 2, a randomized, double-blind, placebo-controlled multicenter study of RBX2660.


  • Patients in the blinded phase of the PUNCH CD 2 were randomized into one of 3 arms: 2 doses of RBX2660 (Group A); 2 doses of placebo (Group B); or 1 dose of RBX2660 and 1 dose of placebo (Group C) via enema with doses 7 days apart.
  • Failures could receive up to 2 doses of open-label active treatment, 7 days apart.
  • Patients who received multiple doses of RBX2660 could receive drug manufactured from the same or different donors. Donors could also be used in a different order.
  • Success was defined as the absence of Clostridium difficile associated-diarrhea at 8 weeks following completion of the last treatment.
  • A mathematical model was used to determine whether the specific donor affected the results.


  • Each dose of RBX2660 consisted of 50 gm of human stool/150 mL of suspension in a single-dose ready-to-use enema bag.
  • RBX2660 was manufactured using standardised, quality-controlled processes.

Sourcing and Traceability

  • A total of 20 donors were used to prepare the RBX2660 drug product used in the study.
  • Donations were collected on site at Rebiotix Inc., Roseville, MN.
  • Donations were not pooled across donors.
  • Units of RBX2660 were identified by a unique batch number and traceable to a specific donor and recipient.
  • Patients who received two doses of RBX2660 randomly received product from the next available batch.

Mathematical Model

  • A mixed model was used to predict treatment success taking into account donors and dose numbers.
  • Repeated measures data was included to account for patients who received two doses of RBX2660. In this analysis, a patient was classified as either a success or failure for each dose.
  • P < .05 indicated statistical significance; all analyses were done using the R statistical package.


  • The donor was not significant (P > 0.99).
  • The variance = 0, indicated that no difference in outcome by donor was expected and that the treatment rates for success and failure were the same for each donor.
  • One outcome (patient) may be associated with up to two different donors. Therefore, patients that received two doses of RBX2660 are represented twice.
  • The number of patients per donor varied but the proportion of successes and failures were similar across all donors.


  • The analysis of a randomized double-blinded, placebo-controlled multicenter study, demonstrated that the specific donor does not affect the outcomes achieved with administration of RBX2660 for recurrent CDI.
  • The results are consistent with a previous analysis of an open-label study of RBX2660 and suggest that donor-to-patient matching is not necessary for the treatment of recurrent CDI.


  1. Ray A, Jones C. Does the donor matter? Donor vs patient effects in the outcome of a next-generation microbiota-based drug trial for recurrent Clostridium difficile infection. Future Microbiol. 2016;11:611-6.