We thank Bogatic et al. 1 for their interest in our work 2 and for highlighting determinants, beyond donor and dosing, for clinical effect of faecal microbiota transplantation (FMT). The mechanisms of action for FMT remain incompletely understood and may vary across disease indications. Our findings, together with increasing evidence, suggest that clinical effectiveness reflects the combined contribution of several interacting factors, underscoring the need for quality improvement with monitoring of clinical outcomes and safety. Donor effects may be more prominent in diseases such as ulcerative colitis than in Clostridioides difficile infection, indicating a need for disease-specific optimisation strategies. While metagenomic approaches to donor selection are emerging, their clinical utility remains limited 3. Our study indicates a dose effect, as repeated FMT doses improved FMT outcomes across all patient subgroups after stratification by disease severity, antibiotic response, and comorbidities. We did not identify any subgroup that failed to benefit from repeated FMT doses, supporting its use as standard care rather than limiting it to selected patients. As clinicians applying FMT, we rely on repeated donations of intestinal microbiota from living donors, and we apply an inherently heterogeneous and variable product. Recently regulated in Europe by the SoHO Regulation 4 and guided by expert bodies such as the European Directorate for the Quality of Medicines & HealthCare (EDQM), this is increasingly acknowledged by regulating authorities, in parallel with tissues and cells. Importantly, regulatory practices and consequences for patients' access to FMT differ massively between for example, the US and Europe 5. The use of FMT in patients with fulminant, refractory or recurrent C. difficile infection (rCDI) may remain the backbone of FMT services worldwide. Regarding FMT versus bacterial consortia for cure of rCDI, Rode and coworkers reported a head-to-head randomised controlled trial with same-route administration and found that a 12-strain bacterial mixture was vastly inferior to FMT 6. A network meta-analysis found that FMT was superior to live biotherapeutics products (LBP) for cure of rCDI, although the encapsulated live biotherapeutic product Vowst was only slightly less effective than FMT 7, 8. The short and long-term safety profiles following treatment of rCDI by FMT or LBPs seem equal, as Cold and coworkers' cohort study over 5-years following FMT or bacteriotherapy found no difference in survival, risk of hospital admission or onset of new diseases such as cancer, hypertension, diabetes mellitus or inflammatory bowel diseases 8, 9. Whether donor intestinal microbiota should be applied as a SoHO or used as a starting material for manufacturing medicinal products remains unresolved. With our current fragmented understanding of how microbiome manipulation affects human health and treatment responses, both paths may be pursued with mutual synergies. It remains undetermined whether defined consortia will eventually replace FMT in any of these conditions, providing equal efficacy and safety. With expanding access to FMT, treated patients will increasingly differ from trial populations, with greater comorbidity and severity 10. This may lead to lower real-world effectiveness and calls for a more dynamic therapy planning, including repeated dosing as standard care and treatment of refractory and fulminant disease. The authors' declarations of personal and financial interests are unchanged from those in the original article 2. Simon Mark Dahl Baunwall: writing – review and editing. Sara Ellegaard Paaske: writing – review and editing. Jens Frederik Dahlerup: writing – review and editing, writing – original draft. Christian Lodberg Hvas: writing – review and editing, writing – original draft. The authors have nothing to report. This article is linked to Paaske et al. papers. To view these articles, visit https://doi.org/10.1111/apt.70395 and https://doi.org/10.1111/apt.70625. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Paaske et al. (Tue,) studied this question.