In plain words

Two small studies gave people stool bacteria from a donor and compared them with people who got a dummy treatment or their own stool. Neither found a clear change in how well the body responds to insulin. A combined analysis reported that it may help, but it may include those same two studies.

The short answer

The two randomized trials on this page both found no significant improvement, so the evidence leans toward no effect. Craven and colleagues (2020) followed 21 people with fatty liver disease for 6 months after donor or own-stool FMT; the ledger has no dose for it. The FMT-TRIM trial (Yu and colleagues, 2020) gave adults with obesity weekly oral capsules for 6 weeks; the ledger has no n for it. A meta-analysis by Qiu, Liang and Li (2023), a pooled estimate that may include the trials above, found FMT may improve insulin sensitivity in metabolic syndrome. Certainty is low, because both trials are small or short and studied different groups.

The 60-second version

Every study in the video is listed below, with its quote.

Watch on YouTube

Where the evidence sits

Insulin sensitivity improved

◪ Pooled analysis, Qiu, 2023 · may overlap studies counted here · not seen in long-term group

No significant change

■ Randomized trial, Craven et al., 2020 · 21 people · 6 months
■ Randomized trial, Yu et al., 2020 · ledger has no n · possible gain if low diversity

■ measured by the source's own study · ◪ pooled analysis, may overlap studies counted here · □ reported, reviewed or argued, no new data

Evidence weight3232 / 100?

Who says what

SourceFindingDesignPeopleDose and lengthFunding and independence
Craven L et al., 2020No significant change in HOMA-IRRandomized trial, donor vs own stool21 with fatty liver diseaseNot in ledger; 6 monthsNot in ledger
Yu et al., 2020No significant change vs placeboRandomized, placebo-controlledAdults with obesity; n Not in ledgerWeekly oral capsules for 6 weeksNot in ledger
Qiu, Liang & Li, 2023May improve; not seen long termMeta-analysisPeople with metabolic syndromeNot in ledgerNot in ledger
Craven L et al., 2020

No significant change in HOMA-IR

Design
Randomized trial, donor vs own stool
People
21 with fatty liver disease
Dose and length
Not in ledger; 6 months
Funding
Not in ledger
Yu et al., 2020

No significant change vs placebo

Design
Randomized, placebo-controlled
People
Adults with obesity; n Not in ledger
Dose and length
Weekly oral capsules for 6 weeks
Funding
Not in ledger
Qiu, Liang & Li, 2023

May improve; not seen long term

Design
Meta-analysis
People
People with metabolic syndrome
Dose and length
Not in ledger
Funding
Not in ledger

What did the trials measure?

Craven and colleagues compared donor stool (allogenic) with each person's own stool (autologous) in 21 people with fatty liver disease, measuring insulin resistance with HOMA-IR over 6 months:

There were no significant changes in HOMA-IR or hepatic PDFF in patients who received the allogenic or autologous FMT.

The FMT-TRIM team gave adults with obesity lean-donor capsules by mouth once a week for 6 weeks, or placebo, and measured insulin sensitivity with a clamp test. The ledger has no n for this trial:

There were no statistically significant improvements in insulin sensitivity in the FMT group compared to the placebo group

The same trial reported no significant difference in adverse events between the groups and no serious adverse events related to FMT.

Did anyone respond?

The FMT-TRIM authors looked further at their own data, in analyses they call exploratory. They suggest a possible benefit in one subgroup, with no n for it in the ledger:

Exploratory analyses suggest possible improvement in metabolism after FMT among study participants with low baseline microbiome diversity

What does the pooled analysis add, and did any benefit last?

Qiu, Liang and Li pooled studies of FMT used alongside other treatment in metabolic syndrome; the ledger has no n, dose or duration for it, and it may include the two trials above. They wrote that FMT:

may be useful in the treatment of metabolic syndrome, especially in improving HbA1c, insulin sensitivity, and HDL cholesterol.

In the same pooled estimate, which may include the trials above, the authors split results by follow-up length and found the short-term gains did not carry over:

The previously described benefits were not observed in the long-term group

Why the sources disagree

The null results come from individual randomized trials; the positive one is a pooled estimate that may include those trials, so it is not a separate measurement. The groups differ: fatty liver disease, obesity, metabolic syndrome. So do the measures: HOMA-IR in one trial, a clamp test in the other. Follow-up ran 6 weeks in FMT-TRIM and 6 months in Craven's trial, and Qiu's own analysis saw no benefit in its long-term group. Qiu describes FMT as an add-on to other treatment, and FMT-TRIM hints that baseline microbiome diversity may matter. The ledger has no funding details for any source.

What would settle it

A pre-registered, placebo-controlled trial of donor FMT in adults with metabolic syndrome, large enough to compare people with low and high baseline microbiome diversity, measuring insulin sensitivity by clamp at 6 weeks, 6 months and 12 months. A matching result from a second, independent group would raise the certainty.

How this page was made

The 3 sources came from a citation-ranked literature search on FMT. Studies in disease groups such as C. difficile infection and ulcerative colitis, and mouse studies, were left out because they did not measure metabolic outcomes in people. Every claim was copied word for word from its source into the Longevity Circle ledger and checked against the original text, and a person approved each one. Certainty follows the GRADE levels used in medical reviews. Drafting was assisted by AI.

The ledger holds only 3 studies on this question.

Changelog: 23 Sep 2026, first version, 3 sources.