NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Drekonja D, Reich J, Gezahegn S, et al. Fecal Microbiota Transplantation for Clostridium Difficile Infection: A Systematic Review of the Evidence [Internet]. Washington (DC): Department of Veterans Affairs (US); 2014 Jul.
Fecal Microbiota Transplantation for Clostridium Difficile Infection: A Systematic Review of the Evidence [Internet].
Show details| REVIEWER COMMENT | RESPONSE |
|---|---|
| 1. Are the objectives, scope, and methods for this review clearly described? | |
| Yes | No response needed |
| Yes | No response needed |
| Yes. This review aims to answer questions regarding the efficacy of fecal microbiota transplantation (FMT) for initial, recurrent or refractory C difficile infection (CDI) compared to standard therapy as well as potential harms and patient acceptability. The Methods (source of data and data abstraction/assessment/analysis) are clearly described. | Thank you |
| Yes; they were well spelled-out. | Thank you |
| Yes | No response needed |
| Yes. The purposes of the review, as indicated by the “Key Questions,” are specific and are clearly important. | Thank you |
| Yes | No response needed |
| 2. Is there any indication of bias in our synthesis of the evidence? | |
| No | No response needed |
| No | No response needed |
| No. In general your synthesis of the evidence was accurate and did not appear overly supportive or opposed to FMT. The overwhelmingly positive results from these numerous series and the clinical experiences of many physicians who’ve used FMT are admittedly “low quality” evidence. Unfortunately, RCTs of this therapy are difficult to perform and, as the use of FMT has expanded rapidly over the past few years, finding subjects willing to be enrolled in a placebo-controlled study will become increasingly difficult. Meanwhile, the large number of patients with recurrent C. difficile infection who’ve failed standard therapies should be offered FMT based on the evidence we have so far. Inclusion of some of the data on MECHANISMS of effect would be a valuable addition to this systematic review. Researchers have done sequencing of the fecal microbiome of FMT treated patients before and after FMT. Recurrent C. difficile patients have characteristic low diversity of species and lack normally dominant populations of anaerobes. Post-FMT, this dysbiosis is remedied and this coincides with clinical cure. Papers that I suggest including discussion of: Khoruts A. et al. J Clin Gastroenterol; 2010 & Song Y, et al. PLoS One 2013 | We have included a brief discussion on the mechanism of FMT in the introduction, including the suggested references. |
| No. The reviewers took care to point out limitations/possible sources of bias in the available literature | Thank you. |
| No. There is no bias, although systematic review is obligatorily affected by publication bias. | No response needed |
| No. Very cautious. | No response needed |
| Possibly. I believe that some of the authors have previously submitted a letter of intent for a VA CSP on FMT. This raises some concern that the authors might use the ESP process to try to impact funding decisions for their grant proposal. Therefore, I suggest that the ESP leadership consider whether or not further review or action is needed. Specific lines that suggest bias on the part of the authors include: | Regarding potential COI. We thank the reviewer for highlighting this point and agree that we should disclose the information noted. We take real or potential COI very seriously as ESP reports must be free of significant conflict to be credible and useful for practice and policy. After discussion with ESP leadership we agree that transparency would be enhanced if we include notation that Drs. Drekonja and Shaukat have submitted a proposal to VA-CSP for a randomized trial assessing FMT: “The Veterans Affairs Fecal Microbiota Therapy Trial for Recurrent Clostridium difficile Infection: A Planning Request for a VA Cooperative Study” The LOI proposes a blinded randomized controlled trial to compare FMT (via enema) to a placebo enema, both administered as an adjunctive treatment after a standard course of oral fidaxomicin (10 days) for recurrent CDI, for efficacy in preventing subsequent episodes of recurrent CDI. We now include this disclosure information in the Preface. Of note, Dr. Drekonja had previously submitted a similar proposal that was not approved for VA planning. Subsequent to our being awarded the contract for, and during the conduct of, the FMT ESP review Drs. Drekonja and Shaukat were informed of an opportunity for additional FMT studies through VA-CSP and encouraged by this reviewer to submit a revised proposal for consideration. Drs. Drekonja and Shaukat did this. and have not yet received reviewer comments or funding decisions. |
| Page 3, line 2: why do they describe the RCT as “one small high risk of bias RCT”? Later, they study is described as “moderate quality” (page 14, line 6). | Page 3: The discrepancy was an error. We apologize and have corrected this so that the study is appropriately rated as moderate quality and as moderate risk of bias. Study quality and strength of evidence is assessed solely by our ESP core staff who are methodologically trained in these assessment methods. Our methodology is consistent with that widely used in the ESP and the AHRQ-EPC program. Content collaborators (in this case Drs. Shaukat, Drekonja, Reich and Gezahegn) do not perform these assessments. We make them aware of our findings and discuss our rationale. We routinely employ 2 individuals (Dr. Greer and Mr. MacDonald to independently assess study quality and overall strength of evidence). The ESP director reviews and resolves through discussion any discrepancies. |
| Page 4, line 36: they call for more RCTs. This may be completely appropriate, but is also self-serving and leads the reader to wonder if they would have made the same recommendation if they weren’t trying to get funding for an RCT. | Page 4: Please see our comment above regarding assessment methods for study quality and strength of evidence. We continue to believe that the current evidence is insufficient to fully address the effectiveness and comparative effectiveness of FMT. Randomized controlled trials are the highest quality studies to accurately assess effectiveness and comparative effectiveness. We have further refined our future research needs section to identify specific gaps that might be addressed by RCTs as well as other potential study designs that are needed to close additional gaps in knowledge. |
| Page 10, lines 14-17: they say that the strength of evidence is insufficient or low at best but seem to be ignoring the RCT here. It is not clear why the RCT is not included in this section. | Page 10: We have included mention of the 2 RCTs in this sentence but strength of evidence would typically be considered low even taking into account the contributions of the 2 RCTs. |
| Page 33, lines 7-9: again, they call for more RCT’s. The description of the RCT here and earlier in the review focus on the low response to antibiotics. The study was stopped early because a review by the data monitoring committee felt that they should (or perhaps I don’t fully know the story). If it was scientifically inappropriate to stop the study, then they have a case. But if it was deemed unethical to continue the study due to a dramatic difference, shouldn’t the focus be on that fact? | Page 33: Information about the early stopping of the RCT has been added. |
| Page 34, lines 9-14: it is unlikely that the emergence of more virulent or aggressive strains would lead to lower responses with standard therapy, right? So the argument about historical controls is not bolstered by this statement. | Page 34: We have clarified the statement about historical controls. |
| Page 34, line 22: The urgent call for RCTs can be self-serving here. | Page 34: We have deleted the word “urgent” and modified the research needs sections of both the Executive Summary and the full report. |
| 3. Are there any published or unpublished studies that we may have overlooked? | |
| No | |
| The review is complete for the time period studied. However, there is a paper published subsequent to the review period, which is a small (n=20), randomized, nonblinded trial of FMT using frozen stool comparing delivery via NG tube to delivery via colonoscopy (Youngster I et al. Clin Infect Dis 2014 Apr 23; epub ahead of print). | Thank you for the suggested reference. We have added this paper. |
| Yes. | Thank you for the suggestions. |
| Key Question #1 (recurrent CDI by colonoscopy) Rohlke F, Surawicz C, Stollman N. Fecal Flora Reconstitution for Recurrent Clostridium difficile infection: Results and Methodology. J Clin Gastroenterol 2010;44:567-70. | KQ#1 We have added the Rohlke data (and deleted the Brandt paper which included some patients from the Rohlke series). |
| Key Question #4 (harms) Our paper was recently accepted and would be an important addition to your safety discussion. In this multicenter series of 80 immunocompromised patients treated with FMT and at least 12 weeks of post FMT follow up, there were no infections transmitted. Kelly C. Ihunnah C. Fischer M, et al. Fecal Microbiota Transplant for Treatment of Clostridium difficile Infection in Immunocompromised Patients. Am J Gastro Accepted 2014. This data was also presented orally at the American College of Gastroenterology 2013 meeting in San Diego | KQ#4 We have added the Kelly 2014 paper to the report. |
| No | No response needed |
| Yes. Yoon et al. J Clin Gastroenterol 2010; 44; 562-66 | We have added the Yoon data (and deleted the Brandt paper which included some patients from the Yoon series). |
| Not sure. This web site (section on C diff in this author’s briefing) may be worth considering: http: | Thank you for the suggestion. Our review protocol specifies that we include data from peer-reviewed journal publications. |
| The authors may wish to include this paper: Kelly et al. “Fecal Microbiota Transplant for Treatment of Clostridium difficile Infection in Immunocompromised Patients.” The American Journal of Gastroenterology. advance online publication 3 June 2014; doi: 10.1038/ajg.2014.133. | Thank you for the suggested reference. We have added this paper. |
| No | |
| 4. Please write any additional suggestions or comments below. If applicable, please indicate the page and line numbers from the draft report. | |
| Well done and thorough review and interpretation of peer reviewed published literature on FMT. I would recommend that the authors provide distinguishing definitions for recurrent CDI and refractory CDI. | Thank you for the suggestion. We have added definitions in the Methods section where we describe our Population. |
| It is unclear to me why the event rate for the VanNood trial is cited at 0.81 since the overall rate of cure was >0.90 when patients who received >1 FMT were included. All of the patients had received multiple courses of antibiotics, so it could be argued that FMT is being held to a higher standard. This may be important because unlike a drug, which is highly standardized, human stool is not, and it may be that a small fraction of donors do not have the flora necessary to cure C. difficile, and subsequent transplant with a different donor may be curative. This phenomenon is seen in clinical practice. | Patients in the control arms of the van Nood trial were considered to have failed after a single course of antimicrobial therapy, and were then offered treatment with donor feces off protocol. Thus, we believe that it is most appropriate to apply the same criteria for FMT (assessment of the primary endpoint after one treatment course). Otherwise, the comparison becomes biased, in that one treatment (FMT) is allowed to have 2 attempts at success, whereas the control treatments are allowed only one. We do separately report the success rates observed after multiple attempts of FMT, which we think conveys more information to our readers. |
| Page 18 lines 24-28. Note: Patients included in the Brandt series from 2012 (long-term follow up study) INCLUDED overlap of cases reported in Mellow 2011, Kelly 2012, Yoon 2012 AND Rohlke 2010. It would be better to analyze each of these series individually (which have more complete data on these patients) rather than Brandt’s long term f/u study (which pools all of these other studies). | Page 18 Thank you for the suggestions. We have deleted the Brandt series and now report each of the series individually. |
| The question of FMT for severe/complicated CDI comes up very frequently. There is much less literature to support use of FMT in these severe and/or complicated cases. You included some of this in with the “refractory” efficacy discussion (Weingarden 2013), but it should really be a separate discussion and analysis. Weingarden 2012, Neemann K, et al. Transpl Infect Dis 2012 You DM, et al. Ann Intern Med 2008 and Aroniadis O, et al. DDW 2013 (oral presentation and abstract) are other cases/series that specifically were performed for severe disease. | We agree that the question of FMT for severe/complicated CDI is common. We did not specifically create such a category for analysis, but note that the reports identified also are described as being “refractory to medical therapy” (You 2008), and “this patient’s refractory CDI” (Neeman 2012). Thus, we believe that this is an appropriate place to discuss these cases. The cases presented in You 2008 and Neeman 2012 are briefly discussed, but since they are single case reports they did not meet our inclusion criteria. Similarly, the Aroniadis abstract is not included since we did not include unpublished data. We were unable to locate the citation for Weingarden 2012. |
| In Research gaps/future research (page 34) you could include list of ongoing clinical trials (clinicaltrials | We have added a list of ongoing trials from clinicaltrials |
| - Is there a specific microbiota population dynamics reference for p6/line 41-end of paragraph (post-CDI treatment microbiome perturbations)? - It may be useful to include references regarding resilience of microbiota post-antibiotic treatment or in setting of C. difficile e.g. a C. difficile/FMT microbiota dynamics study by Song et al (PLOSOne 2013)) - It could also be helpful to specifically invoke the term microbial diversity in discussing alteration of the gut microbiome. | The issue of microbial diversity has been added to the executive summary. |
| In general, the document is very well-written and provides an exhaustive (perhaps a bit too exhaustive) review of the literature. Specific comments follow: - The review should compare and contrast its findings more clearly with the meta-analysis by Kassam et al. Am J Gastro 2013. In this study, the RCT by van Nood was not included but the systematic review included 273 patients with pooled resolution rate for CDI of 89%, and lower route FMT performing better than upper route FMT (91% vs 80.6%). | Since the only trial that directly compared to methods of transplantation demonstrated no significant difference, we are hesitant to endorse one route of administration over another based on overall success rates. We have provided the overall success rates for each route, but again caution that direct comparisons are difficult to make between studies. We have clarified this in the executive summary and in the results section. |
| - Key question 3 probably does not merit equal footing as a “key” question such as FMT for recurrent or refractory CDI. The scarcity of pertinent literature illustrates this, and standard of care is specific antibiotic therapy. | The Key Questions were developed a priori and therefore we present the available evidence on FMT for recurrent or refractory CDI |
| - The van Nood RCT, despite its limitations, is still the only RCT available, and is referred variably in the review as low quality, moderate quality, and high risk of bias. The quality description should be consistent, and the RCT provides valuable information despite its limitations and the fact it was stopped early because most patients in the control arms developed CDI recurrence. Furthermore, the RCT used upper route FMT in its active arm, which is probably less effective than lower route FMT based on Kassam meta-analysis results and less desirable by patients. Overall, I think the review needs to state more explicitly that the pooled literature shows that lower route FMT particularly colonoscopy is associated with very high-about 90%- cure rate for recurrent CDI. | We have corrected the discrepancy on the reporting of the risk of bias of the van Nood RCT. We rated the study as moderate risk of bias. See comment above regarding upper route vs. lower route. |
| - Data regarding long-term safety of FMT are lacking, and this should be better highlighted in KQ4, particularly in the summary statements. This is probably a more relevant question than short-term adverse events (which are otherwise well detailed in the document). | We have added a statement regarding the lack of long-term safety data to the relevant section. |
| - A relevant point is to address is effectiveness of FMT based on CDI severity/C difficile strain (specifically the virulent NAP1/027 strain). This is addressed in the study by Mattila et al. | We have added a statement about C. Difficile strain to the results from the Matilla et al. study. |
| - FMT has been used for a variety of indications unrelated to CDI which are probably beyond the scope of this review. It could be worth mentioning this in the preamble. | We agree that it is beyond the scope of this review, and since CDI is the only indication for FMT for which the FDA does not require an investigational new drug application, we opted to not discuss such investigational uses |
| Figures 2, 3, and 4 should probably be changed to go from 0 to 1 not from -1 to 1 because the data represent rates, and rates can’t be less than zero. If I’m misunderstanding the data please ignore this comment. | Your understanding of the data is correct. We are unable to correct this with the software we have available to generate the plots. |
| - The authors do a nice job describing the response rates with FMT. However, there is little to no discussion of the expected rate of response with additional courses of antibiotics. There is only a brief mention of this toward the very end of the document, other than some mention that the response rate to antibiotics in the RCT was lower than expected. It seems that some additional data on the response rate with antibiotics would be appropriate to put in context the results with FMT in the case series. For example, if studies of antibiotics show 75% response rates, then the FMT results are not very impressive. But if the response rates are around 30%, then the FMT results seem more impressive. | We have included data regarding the expected rate of response to additional antibiotic courses in the introduction. |
| - The authors state that the RCT on FMT was stopped due to a low rate of response to antibiotics in the control group. This statement only tells half the story. The study was not stopped simply due to a low response rate in the control group. The study was stopped because a planned interim analysis showed a significant difference, which required the data safety monitoring board to terminate the study early. The statistical methods used in this study are rigorous and the authors of the ESP should not downplay the dramatically significant results here. The authors need to be careful in how they describe this study so as not to bias the readers, or the conclusions of the ESP analysis. | -As described in the online appendix to the article by van Nood et al., the interim analysis was not planned, but rather was requested by the principle investigator when multiple individuals involved with the trial became aware of an “(unexpected) extremely low response rate in the 2 control arms, which seemed much lower than the 60% used in the sample size calculation. The principle investigator subsequently requested the data safety monitoring board (DSMB) for advice.” We believe that our characterization of the study is accurate. |
| - The section on side-effects doesn’t review the potential adverse events that were summarized earlier in the document. I wonder why the authors didn’t simply put all of the adverse events in this section, rather than make the reader look through the entire document. | We have clarified that this discussion is regarding potential side effects beyond those observed and reported in the section on harms. Since much is written about these potential harms, and they also may be relevant with regard to future FDA regulations, we believe that a discussion of these potential harms is appropriate in the discussion section. |
| One major comment is that the document doesn’t review the data on the expected “cure” rates with antibiotic therapy. Given that there is only one RCT of FMT vs. antibiotics and the rest of the data is case-series, it would seem appropriate to include a discussion of the effectiveness of antibiotics for recurrent CDI. This can put the case-series data in perspective (acknowledging the limitations of historical controls). The data on this topic are only briefly mentioned on page 33, lines 11-15. | We have added information on expected “cure” rates with antibiotic therapy. Please see the 2nd paragraph of the introduction in the executive summary (page 1), the introduction to main report (page 7), and the summary and discussion (page 42). |
| Page 2, line 30-34: why did the initial search miss those 15 articles? That is >50% of the included articles and is concerning. | Page 2: We, too, were concerned about articles missed during our search. We looked at MeSH terms used to index the missed articles and concluded that the issue was use of different terminology (eg, instillation instead of transplantation) over the years. We searched multiple recent systematic reviews and reference lists of all included studies in an attempt to not miss any eligible publications. |
| Minor comments: | Minor Comments: Thank you for your careful read of the report. |
| ) Page 1, line 22: add “cases” after “500” | ) change has been made |
| ) Page 1, line 23: add “been reported” after “have” | ) sentence has been modified |
| ) Page 11, figure 1: under “Excluded”, I would suggest changing the order of the third bullet to state “Not case report with adverse event or case series” since case series don’t need adverse events to be included | ) change has been made |
| ) Page 17, line 19: suggest expanding on the case of possible peritonitis and pneumonia | ) we have noted that little information was provided by the authors of the series reporting this possible adverse event |
| ) Page 22, line 31: should it read “with only two OTHER of the six...”? | ) we believe this sentence is clear as is |
| ) Page 29, line 24: why not move all discussion of harms into this section so they are all described in one place in the review? | ) we have moved a summary of harms to KQ4 |
| ) Page 38, line 34: “toward” is misspelled | ) change has been made |
| 5. Please provide any recommendations on how this report can be revised to more directly address or assist implementation needs. | |
| No specific recommendations beyond correcting a few typos (eg. C. Diff vs. C. diff used interchangeably in the text). | We have attempted to locate and correct all typographical errors. |
| None at this time, as the review was not designed to specifically support or not support implementation. However, as acknowledged in the review, there is likely to be ongoing trepidation about the ‘unknowns’ of manipulating the gut microbiome, particularly as new research is constantly emerging about its varied roles in our physiology. It’s definitely helpful to present this as an opportunity for expediting high-quality research (as opposed to a potential source of liability). | Thank you. |
| None. Best to get a final copy edit, check references, etc. and release, ASAP than to make a lot of minor changes. Timeliness is important. | We agree. |
- PEER REVIEW COMMENTS/AUTHOR RESPONSES - Fecal Microbiota Transplantation for Clo...PEER REVIEW COMMENTS/AUTHOR RESPONSES - Fecal Microbiota Transplantation for Clostridium Difficile Infection: A Systematic Review of the Evidence
Your browsing activity is empty.
Activity recording is turned off.
See more...