Faecal Microbiota Transplantation and Glycaemic Outcomes in Adults with Metabolic Disorders: Systematic Review and Meta-analysis of RCTs
FMT did not significantly reduce HOMA-IR in adults with metabolic disorders (MD −1.29; 95% CI [−3.04, 0.47]), with very low certainty of evidence.
| Population | Adults with type 2 diabetes, prediabetes, insulin resistance or related metabolic disorders |
|---|---|
| Intervention | Faecal microbiota transplantation (FMT) |
| Comparator | Control (autologous placebo, diet or other comparators; heterogeneity in comparator type identified) |
| Outcome | HOMA-IR; HOMA-IR + HOMA2-IR (sensitivity analysis); HbA1c; Peripheral insulin sensitivity (clamp); Fasting glucose |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| HOMA-IR | MD -1.29, 95% CI -3.04 to 0.47, I2=72% | — | Low | — | 6 studies |
| HOMA-IR + HOMA2-IR (sensitivity analysis) | MD -1.11, 95% CI -2.50 to 0.29 | — | Low | — | |
| HbA1c | not pooled; data sparse | — | Very low | — | |
| Peripheral insulin sensitivity (clamp) | possible donor-phenotype effect; inconclusive; low certainty | — | Low | — | |
| Fasting glucose | not pooled; insufficient data | — | Very low | — |
Context
Insulin resistance and type 2 diabetes affect hundreds of millions of adults worldwide, and current therapies do not fully restore insulin sensitivity. FMT has been proposed as an additional therapeutic target modulating gut microbiota, but prior RCT results were inconsistent.
What the study showed
FMT did not significantly reduce HOMA-IR versus control (MD −1.29; 95% CI [−3.04, 0.47]; I² = 72%; 6 RCTs, n = 120). Sensitivity analysis combining HOMA-IR and HOMA2-IR was also non-significant (MD −1.11; 95% CI [−2.50, 0.29]). HbA1c data were sparse and not pooled across comparator types. Clamp-derived findings suggested a possible donor-phenotype effect but remained inconclusive.
How it was done
Systematic review and meta-analysis of RCTs (PubMed and Scopus, January 2000–May 2026). Random-effects models using Paule–Mandel estimation with Hartung–Knapp–Sidik–Jonkman confidence intervals. Risk of bias assessed with RoB 2. Sensitivity analyses for comparator type, data reconstruction and individual study influence.
Effect magnitude
The estimated effect on HOMA-IR was MD −1.29 (95% CI [−3.04, 0.47]), non-significant and clinically imprecise; substantial heterogeneity (I² = 72%) undermines interpretation of the point estimate.
Risk of bias
The primary analysis included only 6 RCTs with 120 participants, providing very low statistical power. High heterogeneity (I² = 72%) and diversity in comparator types limit validity of the pooled estimate. RoB 2 was applied, but individual study risk profiles were not detailed in the abstract. HbA1c and fasting glucose data were insufficient for robust meta-analysis.
What this study does NOT prove
This study does not prove that FMT is ineffective for glycaemic control — only that available evidence is insufficient and imprecise to claim benefit. Findings are not generalisable to paediatric populations or unstudied standardised FMT regimens.
In clinical practice
There is no basis to recommend FMT for glycaemic improvement in adults with metabolic disorders outside research settings. Clinicians should not extrapolate glycaemic benefit from current data.
Limitations
The primary analysis included only 6 RCTs with 120 participants, providing very low statistical power. High heterogeneity (I² = 72%) and diversity in comparator types limit validity of the pooled estimate. RoB 2 was applied, but individual study risk profiles were not detailed in the abstract. HbA1c and fasting glucose data were insufficient for robust meta-analysis.
What is still missing
Larger RCTs with standardised glycaemic outcomes, defined FMT protocols and prospective donor-phenotype selection are needed for definitive conclusions.
Technical appendix
Version history
- 1.0 · 2026-11-02 — Auto-generated under Evidence Standard v1.0