Reply to Galperine et al and Jansen

To the Editor—We thank Jansen and Galperine and colleagues for their comments [1, 2] regarding our recently published study [3] and we respectfully respond below. Jansen expresses concern about the validity of our a priori assumption for fecal microbiota transplantation (FMT) effectiveness by enema (92%) and questions its impact on our reported results. This estimate came from a metaanalysis we performed in 2008 (when we began designing our study), based on the observational studies reported at the time, the majority of which used enema administration. It was used to calculate the trial sample size and had no direct effect on the futility analysis. Even if we had calculated a larger final sample size using a smaller difference in effectiveness between vancomycin and FMT, the overall conclusions of the study would not have changed. In our intention-to-treat analyses, we investigated the probability of showing benefit (conditional power) for FMT for final sample sizes that range from the planned 57 per arm to 1000 per arm. As the final sample size increased, there was an increase in the probability of showing benefit for FMT, but it reached a maximum near 28% (equivalently, a 72% chance of not showing FMT superiority) for sample sizes greater than 800 per arm. Jansen also questions the clinical utility of our comparison of randomized and nonrandomized patients as listed in Table 1 of our article. Our purpose was to present the characteristics of all patients who consented to the study, including the nonrandomized patients who unexpectedly did not have Clostridium difficile recurrences after enrollment. We felt it was important to draw attention to the differences between the randomized and nonrandomized patients in order to better understand why patients did not have recurrent disease. Comparison of these 2 groups had no influence on the analyses of the primary outcome. The differences between the FMT and vancomycin groups that Jansen notes are very small and unlikely to have hidden an important advantage of FMT. Regarding Jansen’s concern about the longer vancomycin exposure in the vancomycin tapering group compared with the FMT group, the intention was to compare standard of care with FMT as it would be applied to a patient with acute recurrence of C. difficile infection. Setting day 0 after 2 weeks of full-dose vancomycin treatment enabled us to measure recurrences during the vancomycin tapering period, which is a known risk. We included a conservative follow-up period of 120 days to ensure that we captured recurrences in both groups; indeed, there were no recurrences past day 48 in either arm. Galperine and colleagues indicate that the allowance of donors to provide stools up to 48 hours prior to FMT may have resulted in loss of microbiota in the administered FMT. Prior to our study, we demonstrated, using semiquantitative culture, that stools stored at 2°C –5°C for up to 24 hours maintained aerobic and anaerobic viability, and stools stored for up to 48 hours had a 8% loss of fecal flora [4]. For pragmatic reasons, we requested that donors begin collecting stools 48 hours prior to FMT; however, for the procedure, we used their most proximal stool collection. Our mean time from donor stool collection to FMT administration was 14.79 hours (range, 0.25–34.75 hours), with no apparent difference between cases of successful FMT and FMT failure (mean, 14.72 and 14.84 hours, respectively). The amount of donor stool infused during the FMT procedure, as measured by donor stool mass, enema volume, and retention time, is raised by all authors as a possible reason for FMT failure. The precise targets for these indices have not yet been established. Therefore, we cannot assume that recipients who did not retain the entire FMT in our study had a suboptimal infusion. Indeed, Kelly et al showed positive results using colonoscopic administration of FMT with as little as 20 g of donor stool [5]. Galperine and colleagues propose that the absence of bowel preparation is a possible reason for low FMT success. We do note emerging evidence that suggests a possible association between use of bowel preparation and positive FMT outcome [6]. At the time our study was designed, there was no empirical evidence to support such an association. The more recent trials cited by Galperine and colleagues that demonstrate differing FMT success rates according to presence or absence of bowel preparation also had many other methodological differences, including route of FMT administration, which makes it difficult to attribute the outcomes to bowel preparation alone [7–9]. The role of bowel preparation should be further evaluated. Finally, Galperine and colleagues assert that the lower success rate in our study compared to other studies may be attributed to the administration of a single FMT infusion. We agree that our results suggest that single FMT by enema is not the optimal approach. At the time that our study was designed, the relevance of repeat FMT administration was not apparent. As researchers begin to report on the effectiveness of first and subsequent FMT administrations, the importance of multiple FMT administrations has emerged [10]. Unfortunately, the data on effectiveness of first FMT administration are not always readily accessible within publications and, until our study, little attention had been directed to this consideration. There are many donor, patient, and methodological factors that may influence FMT success. One of the biggest challenges in the field is that comparative research on these factors has not been prioritized, perhaps because such research is onerous and time consuming. As a start, we urge others who are publishing clinical research on FMT to be transparent on the details of their methodologies, so that results can be interpreted in the full context of the research. Potential conflicts of interest. S. M. P. reports receiving advisory board honoraria from Cubist/Merck and speaker honoraria from Merck outside the submitted work. All other authors: no potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.