DOP48 Faecal microbiota transplantation in active Ulcerative Colitis: Key lessons from a randomized controlled trial halted for futility

Abstract

Abstract Background Rigorous donor preselection on microbiota level, strict anaerobic processing, and repeated FMT administration were hypothesized to improve FMT outcomes for induction of remission in UC in the RESTORE-UC trial, for which we here report on the clinical results and observed microbial changes. Methods The RESTORE-UC trial was a multi-centric double-blind, sham-controlled randomized trial. Patients with moderate to severe UC (total Mayo 4-10, endoscopic subscore ≥2) were randomly allocated to receive 4 weekly anaerobic-prepared allogenic or autologous donor FMT. Allogenic healthy donors were selected after a rigorous screening by microbial cell count, enterotype, and the abundance of specific genera. The primary endpoint was steroid-free clinical remission (total Mayo ≤2, no subscore >1) at week 8. A pre-planned futility analysis was performed after 66% (n=72) of intended inclusions (n=108). Quantitative microbiota profiling (n=44) was performed at weeks 0 and 8. Results In total, 72 patients were included of which 66 received at least one FMT (allogenic-FMT n=30 and autologous-FMT n=36, Fig. A). At week 8, resp. 3 and 5 patients reached the primary endpoint (p=0.72), indicating no treatment difference of at least 5% in favour of allogenic-FMT. Hence, the study was stopped due to futility. The procedure was well tolerated, and only 2 SAE were observed. Microbial analysis showed no significant differences in baseline microbiome composition between treatment groups. However, numerically more enterotype transitions upon allogenic-FMT compared to autologous-FMT (Fig. B), and more transitions were observed when patients were treated with a different enterotype than their own at baseline (p=0.01). Primary response was associated with lower total Mayo scores, lower bacterial cell counts (Fig. C,D) and a tendency to more Bact 2 prevalence at baseline, independently from FMT origin. At the allogenic donor side, a positive association was suggested between stool moisture and treatment success (Wilcoxon test, p=0.057; Fig. F). However, no indications pointed towards a highly effective ‘superdonor’ profile. Conclusion The RESTORE-UC trial comparing preselected allogenic to autologous FMT did not meet its primary endpoint of steroid-free clinical remission at week 8. Notably, participants exhibited more severe disease, aligning with recent guidelines recommending FMT for mild to moderate cases (Lopetuso & Deleu et al., 2023). In addition, our findings support adding microbial load and dysbiosis to the patient inclusion criteria based on a pre-treatment microbiome screening. Next to patient-selection, further research should additionally consider sterilized sham-control, increased frequency, density, and viability of FMT prior to administration.