Akkermansia and type 2 diabetes: what the 2025 trial actually found

For most of the last decade the human story of Akkermansia muciniphila rested on a single small trial. In 2025 a study in Cell Metabolism took the question into people living with type 2 diabetes (Zhang et al., 2025). It is the kind of result that gets reported badly, because the headline finding was negative and the subgroup finding was not.

So here it is straight.

The thread before 2025

The earliest human signal was not a supplement trial. In 2016 Maria Carlota Dao and colleagues followed overweight adults on a calorie-restricted diet and found that those who started with more Akkermansia in their gut tended to have better insulin sensitivity and a healthier lipid and inflammation profile than those who started low (Dao et al., 2016). Observational. It showed the bacterium travelling alongside better metabolism, without establishing which way the arrow pointed.

The first time anyone actually gave people Akkermansia was 2019, when Clara Depommier and colleagues ran a proof-of-concept trial in 32 overweight, insulin-resistant adults (Depommier et al., 2019). Both forms were safe, and the pasteurised group showed the clearest movement in metabolic markers, which is the pasteurised paradox we cover separately. The authors called it small and exploratory. A door opening, not a conclusion.

What the 2025 trial found

Zhang and colleagues ran a 12-week randomised, double-blind, placebo-controlled trial in 58 participants with overweight or obesity and type 2 diabetes. Everyone received routine lifestyle guidance; half also received a live A. muciniphila strain called AKK-WST01 (ClinicalTrials.gov NCT04797442).

The primary result was null. Both groups lost weight and both saw their HbA1c fall, and there was no statistically significant difference between them. Lifestyle guidance worked. Adding the bacterium, across the group as a whole, did not measurably add to it.

The finding that made the paper interesting came from splitting participants by how much Akkermansia they already had. In those who started low, the supplement colonised well and was followed by significant reductions in body weight, fat mass and HbA1c that the placebo group did not show. In those who started high, it colonised poorly and produced no significant clinical improvement. The authors then reproduced the pattern in germ-free mice given faeces from low- and high-Akkermansia donors.

Their conclusion is a hypothesis, and they frame it as one: the benefit of supplementing this bacterium may depend on how much of it you already carry.

Why that is not proof, and why it matters here

Three things have to be said, and the third is the one most often left out.

A subgroup result is weaker evidence than a primary result. Splitting 58 people into two groups by baseline abundance leaves very small numbers on each side, and subgroup findings are where statistical noise most often looks like signal. This is the standard reason such findings are treated as hypothesis-generating: they tell you what trial to run next, not what is true.

The trial did not fail because the bacterium is inert. It may simply be that a 12-week trial against active lifestyle guidance is a hard place to detect an additional effect. But "the study design may have masked something" is a reason to keep researching, not a reason to report a null result as a positive one.

And the strain was not ours. AKK-WST01 is a live organism. Vekovia · Bilberry uses the pasteurised form, which is a different preparation with a different regulatory status and its own separate evidence, including the Depommier trial above. Findings about a live strain do not transfer to a pasteurised one, and we are not going to borrow them.

The honest part

Vekovia is a food supplement. It is not a treatment for type 2 diabetes, or for anything. A trial conducted in people with a diagnosed condition is not a licence to make a claim about that condition, and a null trial with an interesting subgroup is not an evidence base. We describe research and stop there.

Where that leaves the bacterium

Honestly? About where it was, plus one useful idea.

Akkermansia muciniphila sits at an active intersection of the gut lining and metabolic health. The human evidence is real and still thin: an observation in 2016, a 32-person proof-of-concept in 2019, and now a 58-person trial that missed its primary endpoint but raised a specific, testable question about who might respond.

That question — whether the people who benefit are the ones who start with least — is a genuinely good one. It is also unanswered. We would rather tell you that than tell you the trial worked.

The bacterium anchors Vekovia · Bilberry alongside freeze-dried European bilberries, whose polyphenols the broader literature connects to gut microbial composition. The pairing follows the science. It does not add a promise on top of it.