Statistics from Altmetric.com
We read with interest the article by Furrie and colleagues (Gut 2005;54:242–9). While we believe this approach represents a very interesting advance in our understanding of aspects of the mucosal response to synbiotic therapy in ulcerative colitis (UC), we would like to raise some questions about the design of the study, which relate in particular to the conclusion that the synbiotic cocktail produces some improvement in disease activity in UC.
Five patients were taking steroids, and six patients were taking immunosuppressants in each of the active treatment and placebo groups (see table 2). While the study design states that no treatment changes were made once the patients were started on test therapy, no information is given as to whether either the steroids or immunosuppressants were started, or had their dose changed, in the period immediately before the test therapy began. Given that the lag between recruitment and initiation of the test treatment was up to two months, we need reassurance that no treatment changes were made during this period that could have contributed to the later clinical and histological changes associated with the test therapies.
Two of the outcome measures seem to have been scores of sigmoidoscopic appearance and microscopic disease activity, which have not been previously validated formally. Can we be reassured that the conclusions drawn from these results would have been the same had the authors used an unmodified Baron sigmoidoscopic score,1 and a more widely used histological activity index?2 Indeed, we note that there were in fact no significant changes after the symbiotic therapy in either the simple colitis activity index, sigmoidoscopic score (p = 0.06, using a t test which assumes normal distribution), bowel habit index, or histological score.
During the lag period between enrolment in the trial and initiation of the test treatment, one patient in the placebo group went into spontaneous remission (SCAI 1, modified Baron score 0) and so no longer fulfilled the entry criteria for the study. However, this subject still appears to have been included in the evaluation of the response to placebo and hence may have skewed the results for this group.
The authors reported a significant reduction in expression of mRNA for human beta- defensins 2–4 and the inflammatory cytokines tumour necrosis factor α and interleukin 1α in mucosal biopsies. It is of course possible that these changes might be associated with subsequent clinical, sigmoidoscopic, and/or histological improvement, but we would question whether the data presented convincingly show initiation of the resolution of inflammation stated in the title. We agree with the authors that a much larger scale randomised controlled clinical trial of this synbiotic cocktail is needed, using conventional and well validated measures of response, before we can draw firm conclusions about its efficacy (or safety).
Conflict of interest: None declared.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.