In plain words

In two of the three trials, taking akkermansia did not change the overall mix of bacteria in the gut. One of those trials reported more akkermansia itself in stool, especially with the live form. Another trial found it settled in well in people who started with little of it.

The short answer

Of three randomized trials, two measured the overall makeup of gut bacteria and found no overall change. One tested live or pasteurized akkermansia (2019); the ledger has no n for it. The other gave a heat-killed form to 150 people for 12 weeks (2024). A 2021 paper that appears to describe the 2019 trial reports that 10 billion cells a day for 3 months in 32 people was associated with more akkermansia in stool. The third trial (2025, 12 weeks, the ledger has no n) found high colonization in people starting with low akkermansia. Certainty is low, because the trials used different forms and measures.

Where the evidence sits

No overall change

■ Randomized trial, Depommier et al., 2019 · ledger has no n · also reported by Depommier et al., 2021 · stool akkermansia rose
■ Randomized trial, Lee et al., 2024 · ledger has no n · 12 weeks

Depends on starting level

■ Randomized trial, Zhang et al., 2025 · ledger has no n · 12 weeks

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

Evidence weight4646 / 100?

Who says what

SourceFindingDesignPeopleDose and lengthFunding and independence
Depommier et al., Nature Medicine, 2019Overall structure of the gut microbiome unchangedRandomized trial, live or pasteurizedOverweight and obese volunteers; n not in ledgerNot in ledgerNot in ledger
Depommier et al., Cells, 2021Associated with more akkermansia in stool, more with the live formRandomized trial; appears to describe the 2019 trial3210 billion cells a day for 3 monthsNot in ledger
Lee et al., 2024No difference in diversity or composition against placeboRandomized, placebo-controlled, heat-killed150One 500 mg capsule a day for 12 weeksNot in ledger
Zhang et al., 2025High colonization in people starting lowRandomized trialNot in ledgerDose not in ledger; 12 weeksNot in ledger
Depommier et al., Nature Medicine, 2019

Overall structure of the gut microbiome unchanged

Design
Randomized trial, live or pasteurized
People
Overweight and obese volunteers; n not in ledger
Dose and length
Not in ledger
Funding
Not in ledger
Depommier et al., Cells, 2021

Associated with more akkermansia in stool, more with the live form

Design
Randomized trial; appears to describe the 2019 trial
People
32
Dose and length
10 billion cells a day for 3 months
Funding
Not in ledger
Lee et al., 2024

No difference in diversity or composition against placebo

Design
Randomized, placebo-controlled, heat-killed
People
150
Dose and length
One 500 mg capsule a day for 12 weeks
Funding
Not in ledger
Zhang et al., 2025

High colonization in people starting low

Design
Randomized trial
People
Not in ledger
Dose and length
Dose not in ledger; 12 weeks
Funding
Not in ledger

Does akkermansia change the overall mix of gut bacteria?

Researchers at UCLouvain, publishing in Nature Medicine in 2019, gave overweight and obese volunteers live or pasteurized akkermansia. The ledger has no n, dose or duration for this row. On the whole community, they wrote:

the supplementation with either Pasteurized A.muciniphila or Alive A.muciniphila did not affect the overall structure of the gut microbiome

Lee and colleagues (2024) gave 150 people one 500 mg capsule of a heat-killed strain, or a placebo, every day for 12 weeks. They compared stool samples between the groups:

we could not find significant differences in the diversity and composition of human fecal microbiota between the ETB-F01 and placebo groups

The same trial reports no serious adverse events in either group.

Does the bacterium itself build up in the gut?

A 2021 paper in Cells by Depommier and colleagues appears to describe the same trial as the 2019 report. In 32 people taking 10 billion cells a day for 3 months, it looked at akkermansia alone:

both supplementations were associated with a significant increase of the abundance of A. muciniphila recovered in the feces of the participants, with a more pronounced effect for the alive form

Zhang and colleagues (2025) ran a 12-week trial of a strain called AKK-WST01. The ledger has no n or dose for it. They split results by how much akkermansia people carried at the start:

In participants with low baseline A. muciniphila, AKK-WST01 supplementation showed high colonization efficiency

Why the sources disagree

No independent measurements in the ledger contradict each other on this question. The results differ in what was measured. The 2019 and 2024 trials looked at the whole community, while the 2021 paper and the 2025 trial tracked akkermansia itself. The forms differ too: live, pasteurized and heat-killed bacteria were each tested, and the 2021 paper reports a larger rise with the live form. The 2025 trial reports its colonization result for people who started with low levels.

What would settle it

A pre-registered, placebo-controlled trial randomizing adults to live, pasteurized or heat-killed akkermansia at 10 billion cells a day for 12 weeks. It would sequence stool at baseline, 6 weeks, 12 weeks and 4 weeks after stopping, report both overall diversity and akkermansia abundance, and analyze results by starting level. A matching result from a second group would raise certainty.

How this page was made

The 4 sources come from a search of PubMed-indexed studies on akkermansia given as a supplement. Studies were left out when they tested other supplements, used only animals, or enrolled people with liver disease. The ledger holds 5 gut microbiome composition claims from 5 sources; this page uses four of those sources, plus one safety claim. 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.

Changelog: 29 Sep 2026, first version, 4 sources.