Here is a pattern we see constantly. Someone starts a GLP-1 injection — Ozempic, Mounjaro — and the weight comes off. Then life happens. Supply runs out, the budget tightens, or they simply feel ready to stop. And the weight comes back.
One of our own cases shows it starkly: a 24-year-old woman came down to 82 kg on Ozempic, then had a 26-week gap without treatment. She regained 7 kg in that window — arriving at 89 kg, heavier than where the second round of treatment would begin.
This is not a willpower problem. It is a biology problem, and it is the single biggest unsolved issue with GLP-1 therapy. Which is why the most interesting gut research right now is not about losing weight at all. It is about keeping it off.
The bacterium that lives in your gut wall
Most probiotics — Lactobacillus, Bifidobacterium — live in the open channel of your gut and mostly pass straight through. Akkermansia muciniphila does something different. It lives inside the mucus layer coating your intestinal wall, feeding on mucin. That feeding signals your gut to produce more mucus, so the bacterium maintains the very barrier it lives in.
That barrier matters more than it sounds. When it thins, bacterial debris leaks into the bloodstream and drives the low-grade inflammation that sits underneath insulin resistance. People with obesity and type 2 diabetes consistently have less Akkermansia than people without.
Akkermansia also makes a protein called P9, which has been shown to stimulate the gut's own GLP-1 release — the same hormone pathway the injections target, approached from the opposite direction. To be clear about scale: this is nothing like the effect of a prescription drug, and it has been shown in laboratory models rather than in people. But it is the reason researchers became interested in combining the two ideas at all.
Why the dead version works better
Now the genuinely counterintuitive part. When researchers heated Akkermansia until it was dead, it did not stop working — it worked better.
The reason is a protein on its outer membrane, Amuc_1100, which signals to your gut lining through a receptor called TLR2. That protein survives to around 70°C, the temperature used for pasteurisation. So heat destroys the organism but leaves the active component intact.
This is why the category is properly called a postbiotic rather than a probiotic — an inactivated microbe, or its parts, that still delivers a benefit. It also solves unglamorous real-world problems: no cold chain, no potency lost on a warm Hong Kong shelf, and no possibility of infection from a living culture.
What the human trials show
The landmark study, published in Nature Medicine in 2019, gave pasteurised Akkermansia to adults with overweight and insulin resistance for three months. Against placebo, insulin sensitivity improved 28.6%, fasting insulin fell 34.1%, and total cholesterol fell 8.7%. Body weight and fat mass also drifted down, though those particular changes were too small to rule out chance — the study was designed to test safety, not to prove weight loss.
The more relevant trial came in 2026, and it asked the question that matters after a GLP-1: can you hold the line? Ninety adults dieted down by at least 8%, then ate freely for 24 weeks with pasteurised Akkermansia or placebo.
- Placebo group regained 3.2 kg
- Akkermansia group regained 1.2 kg
Roughly two-thirds less regain, with no serious side effects. That is the finding worth knowing if you have ever come off a GLP-1 — or expect to.
What we see alongside GLP-1 therapy
Three of our own cases added AH39 on top of GLP-1 treatment they were already taking.
The most informative one is the third, because nothing else changed. Her Ozempic dose stayed at 0.5 mg throughout — no increase, no switch — and AH39 was added on 2 April 2026. Over the following twelve weeks her weight moved 64.6 kg → 63.4 kg, and her liver fat fell from CAP 310 to 261 dB/m, a drop of 49 — out of the range usually described as significant fatty liver and into the mild range.
The other two, both already stable on Mounjaro:
- 24-year-old: 79.6 kg → 76.4 kg in two months (−3.2 kg), reaching −11.2 kg from where she began in 2023
- 62-year-old: 67.9 kg → 65.8 kg (−2.1 kg), after a post-Ozempic rebound had pushed her back to 70.4 kg
We want to be straight about what this is and is not. These are three individual people, not a controlled trial, and all three were on GLP-1 medication that produces weight loss by itself — so the weight numbers cannot be credited to AH39 alone. The liver reading is the more interesting one, precisely because the drug dose was held constant while it moved. Even there, losing weight reduces liver fat on its own, so some of that 49-point drop belongs to the 1.9 kg she lost.
What we can say honestly: three people continued in the right direction after adding AH39, one showed a marked liver-fat improvement on an unchanged drug dose, and none had trouble tolerating the combination. Individual results vary, and a supplement is not a substitute for treatment.
The liver angle — promising, and genuinely unproven
Liver fat is where this gets interesting, because fatty liver travels with exactly the problems Akkermansia is thought to touch: insulin resistance, a leaky gut barrier, and low-grade inflammation. Bacterial debris crossing a thinned gut wall arrives at the liver first.
Two human trials have reported movement in liver markers. The 2019 Nature Medicine study noted reductions in blood markers of liver dysfunction and inflammation, and the 2025 yogurt trial found a significant fall in the liver enzyme AST.
But we should be clear about the gap: no randomised trial has yet measured whether Akkermansia reduces liver fat itself. Enzymes are not the same as fat, and one person's scan is not evidence. Our case is a signal worth following, not a claim — and we will say so plainly until somebody runs the trial.
AH39 or classic Akkermansia — which one?
Something worth understanding before you buy any Akkermansia product: evidence attaches to strains, not species. Two strains can differ like two dog breeds, and the number of billions on the front of the box tells you far less than which strain is inside.
AH39 — the strain in nubu's own AH39 Pro BioBalance Complex — is a patented, pasteurised strain with FDA sGRAS status and 99.87% genomic similarity to the reference strain used in the trials above. It was isolated and developed in Asia rather than adapted from the European reference strain.
It is also not a single-ingredient capsule. The formula pairs 30 billion units of pasteurised AH39 with two supporting ingredients:
- XOS (xylo-oligosaccharides) — a prebiotic fibre. Where the postbiotic delivers the signalling protein directly, XOS feeds the Akkermansia and other mucus-layer bacteria you already carry. Supplying the signal and feeding the resident population are two different jobs.
- Green tea polyphenols — catechins with their own metabolic literature, and independently associated with higher Akkermansia levels in the gut.
Practically, being pasteurised makes it robust: 36 months at room temperature, no refrigeration, and no live culture to cause an infection.
Its own published research remains an immunology study showing an AH39 surface protein acting on regulatory T-cells — work done in cells and mice, not in people. That is the strain's genuine, specific finding, and it is also the honest limit of it.
Classic Akkermansia — the reference strain, and the one carrying the bulk of the human trial data cited above, including the weight-maintenance result. Coming soon to nubu for those who want the strain the published studies were run on.
Neither is a lesser version of the other. One is the most-studied strain in humans; the other is a patented, room-temperature-stable strain with its own immune research. If your priority is following the published trials, take the classic. If it is stability, safety margin and a strain developed specifically for supplement use, take AH39.
Getting it right
The recommended course is 2–4 capsules a day, taken before a meal, for 12 weeks. Every trial above ran for months rather than weeks, and measured effects in single kilograms — this is a slow, compounding tool, not a switch.
Roughly what to expect, and when:
- Weeks 2–4 — appetite and fullness are usually the first thing people notice
- Weeks 4–8 — digestive comfort and energy
- Weeks 8–12 — weight, waist and metabolic markers, if they move at all
Treat that as a typical pattern rather than a promise. It reflects what tends to be reported, not a guaranteed schedule, and plenty of people will sit outside it.
It also works with your diet, not instead of it. Fibre, plants and polyphenols feed the Akkermansia you already have, and in every study cited the supplement was added on top of normal eating and activity.
If you are tracking your own numbers, the ones most likely to shift are fasting glucose Blood Glucose (Fasting) and Glycated Hemoglobin (HbA1c), with Cholesterol and Low-Density Lipoprotein (LDL) moving more slowly, and hs-CRP (High-Sensitivity C-Reactive Protein) as a rough read on background inflammation. If fatty liver is part of your picture, the liver enzyme Aspartate Aminotransferase (AST/GOT) is the cheap marker to watch, while Liver Fat on a FibroScan Score is the direct measure — the one our own case moved.
This article is general health education, not medical advice, and AH39 Pro is not a treatment for diabetes or obesity. If you are taking a GLP-1 medication such as Ozempic or Mounjaro, are pregnant or breastfeeding, are immunocompromised, or manage a diagnosed condition, speak with your doctor before adding any supplement — and never adjust prescribed medication on your own.
