PROMISE: phase 3 RCT protocol of encapsulated FMT in cirrhosis due to alcohol-related and metabolic liver disease
This document is a clinical trial protocol — no efficacy or clinical outcome data are yet available; direction of effect is insufficient.
| Population | 300 adults with ALD, MASLD, or MetALD cirrhosis, MELD-Na 8–16, recruited from June 2023 |
|---|---|
| Intervention | Lyophilised encapsulated FMT capsules every 91 days for 21 months |
| Comparator | Matched placebo capsules, identical administration |
| Outcome | Time-to-infection or decompensating event requiring ED/hospitalisation; All-cause mortality; Incidence of antimicrobial resistance; Antibiotic usage; Quality of life; Liver disease severity (MELD-Na); Hospitalisation rate |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| Time-to-infection or decompensating event requiring ED/hospitalisation | not yet reported | — | Very low | — | |
| All-cause mortality | not yet reported | — | Very low | — | |
| Incidence of antimicrobial resistance | not yet reported | — | Very low | — | |
| Antibiotic usage | not yet reported | — | Very low | — | |
| Quality of life | not yet reported | — | Very low | — | |
| Liver disease severity (MELD-Na) | not yet reported | — | Very low | — | |
| Hospitalisation rate | not yet reported | — | Very low | — |
Context
Cirrhosis is associated with severe dysbiosis, increased intestinal permeability, and bacterial translocation, driving infection, decompensation, and death. Heavy antimicrobial use in this population further amplifies antimicrobial resistance (AMR) risk. The prior feasibility trial (PROFIT, NCT02862249) established safety and tolerability of endoscopic jejunal FMT but lacked power for clinical endpoints.
What the study showed
The study reports NO results — this is exclusively a published protocol. No efficacy, safety, or clinical outcome data are presented. Recruitment began in June 2023 and results will be disseminated in future publications.
How it was done
Phase 3, multicentre, randomised, double-blind, placebo-controlled, parallel-group trial with planned n=300 and 24-month follow-up. Randomisation is 1:1 (FMT vs. placebo). The protocol includes extensive mechanistic endpoints: faecal microbiome, resistome, metabolome, proteome, cytokines, and MAIT cell assessments.
Effect magnitude
No effect size is available — the trial has not reported results. The sample size calculation assumes power to detect a clinically meaningful difference in the primary endpoint, but exact parameters are not detailed in the provided text.
Risk of bias
As a protocol, RoB 2 assessment is not yet applicable. Potential limitations include: restriction to ALD/MASLD/MetALD aetiology (excluding viral and other cirrhosis); MELD-Na 8–16 range limits applicability to more advanced disease; standardisation across lyophilised FMT batches represents a methodological challenge not fully addressed.
What this study does NOT prove
This protocol does not prove efficacy, superiority over placebo, or any clinical benefit of FMT in cirrhosis. It does not apply to viral, biliary, or autoimmune cirrhosis, nor to patients with MELD-Na outside the 8–16 range.
In clinical practice
No change in clinical practice is supported by this document. The protocol provides no evidence of efficacy. Clinicians should await trial results before considering encapsulated FMT for cirrhosis.
Limitations
As a protocol, RoB 2 assessment is not yet applicable. Potential limitations include: restriction to ALD/MASLD/MetALD aetiology (excluding viral and other cirrhosis); MELD-Na 8–16 range limits applicability to more advanced disease; standardisation across lyophilised FMT batches represents a methodological challenge not fully addressed.
What is still missing
Primary and secondary results of PROMISE have not yet been published. The central question — whether lyophilised FMT reduces infection, decompensation, and mortality in metabolic/alcohol-related cirrhosis — remains unanswered.
Technical appendix
Version history
- 1.0 · 2026-11-04 — Auto-generated under Evidence Standard v1.0