DDW 2024: Treatment of Fistulizing Crohn's Disease: Not an NMA

The session discussed the factors to consider for conducting a network meta-analysis (NMA) and its utility in informing practice guidelines. The meta-analysis, published in Journal of Crohn’s & Colitis addressed the challenge of managing perianal fistulae in Crohn's disease, which affects about one in five patients and often requires a multidisciplinary approach with medications and surgery. The study included adult patients with any form of fistulizing Crohn's disease, such as perianal, enterocutaneous, enteroenteric, enterovesical, enterovaginal, or enterouterine fistula with predominant patient being affected by perianal type. The interventions considered were pharmacological treatments alone or in combination with surgery, compared to placebo or active comparators. The primary outcome was fistula response, with fistula remission (complete closure) as a secondary outcome The systematic review identified 35 studies reporting on 38 randomized controlled trials (RCTs) for fistulizing Crohn's disease, excluding those primarily focused on surgical therapies. Nineteen RCTs were exclusively for perianal fistula, while others included a subset of patients with perianal fistula. Most of the trials (30) were placebo-controlled, and only six had a low risk of bias. The interventions studied included: 2 trials of antibiotics, 7 trials of immunosuppressives, 8 trials of Tumor Necrosis Factor (TNF) antagonists, 4 trials of Ustekinumab, 1 trial of vedolizumab, 4 trials of oral small molecules and 2 trials of mesenchymal stem cell therapy. These trials were primarily designed for luminal Crohn's disease, with post hoc analyses conducted on subsets of patients with perianal fistulae to evaluate the efficacy of the treatments.

The review suggested that antibiotics showed uncertain benefit based on two randomized trials, while immunosuppressives demonstrated probable benefit with moderate certainty of evidence. However, the analysis did not specifically focus on absolute effect sizes or minimal clinically important differences (MCIDs) as MCIDs are more relevant for guideline development unlikely for systemic reviews. TNF antagonists also showed probable benefit with moderate certainty, and combining TNF antagonists with antibiotics was probably more beneficial than monotherapy, based on two trials. Vedolizumab and Ustekinumab may be beneficial compared to placebo, each with low certainty of evidence. Janus kinase (JAK) inhibitors, based on pooled analysis of Upadacitinib trials, were probably beneficial with moderate certainty of evidence. Mesenchymal stem cells were initially considered probably beneficial, but the results from the Adipose derived mesenchymal stem cells for induction of remission in perianal fistulizing Crohn's disease (ADMIR CD2) trial, presented at the IMIBD plenary, may not be consistent with previous observations. 

A NMA for perianal Crohn's disease was deemed inappropriate for several reasons. The primary trials were designed to study luminal disease activity, and the presence or absence of perianal disease was not a stratification variable within these trials. Without stratification by this variable, the benefits of a RCT were lost, reducing these studies to well-conducted observational studies rather than true RCTs suitable for NMA. NMA inherently involves many assumptions and is recommended to be conducted with best performed with RCTs. Conducting an NMA with observational studies compounds these assumptions thus, increasing the potential for error. Additionally, the study showed a variability in luminal disease activity among the trials; with some trials such as the primary randomized controlled trials involving mesenchymal stem cells, included patients who were in luminal remission, indicating minimal or no luminal disease activity. In contrast, post hoc analyses of phase three clinical trials for most other biologic therapies involved patients with luminal disease activity, as these trials were designed to evaluate induction therapy for moderate to severely active disease. Additionally, patients in these trials exhibited variable fistula activity, with some trials including patients with non-draining perianal fistulas. Moreover, the pooled analysis of upadacitinib included a subset of patients without draining fistulas at baseline who were still included in the analysis comparing drug to placebo. The variability in co-interventions, prior treatments due to the use of setons, surgical co-interventions, prior failure of biologics, etc indicated significant heterogeneity, rendering a NMA to be inappropriate. Additionally, the studies also show variability in definition and timing of the outcomes assessment. The study suggests TNF antagonists, JAK inhibitiors, Ustekinumab and vedolizumab may be effective for achieving fistula response and remission. Combining TNF antagonists with antibiotics likely enhances efficacy compared to TNF antagonist monotherapy. Future trials should focus on standardized protocols, including surgical co-interventions, surgical treatments and relevant comparators to inform comparative efficacy for patients with draining perianal fistulae in Crohn's disease. 

Digestive Disease Week (DDW) 2024, May 18-21, 2024, Washington, D.C.