Almost every claim in this market rests on one idea: that your body makes GLP-1 already, so something could help it make more. The first half is true. The second half runs into a number printed on an FDA-approved drug label, and the number is small enough to settle most of the argument.
Every figure below carries its source and the date we read it. This page describes what is known about a hormone; it does not tell anyone what to eat or take, which belongs with a clinician, as our editorial standards set out.
What the hormone is
GLP-1 — glucagon-like peptide-1 — is produced by L cells in the lining of the intestine and released in response to food. It does four things that matter here: it prompts insulin release when glucose is present, it suppresses glucagon, it slows the rate at which the stomach empties, and it increases fullness.
This is ordinary physiology. It happens after every meal, in everyone.
The number that decides the argument
The FDA label for Saxenda, effective 2026-02-25, contains this sentence in its Clinical Pharmacology section:
"Endogenous GLP-1 has a half-life of 1.5 to 2 minutes due to degradation by the ubiquitous endogenous enzymes, dipeptidyl peptidase 4 (DPP-4) and neutral endopeptidases (NEP)."
Two enzymes, present throughout the body, take the hormone apart almost as fast as the gut releases it. The word the label uses is ubiquitous.
The very next sentence on the same label is the contrast:
"Unlike native GLP-1, liraglutide is stable against metabolic degradation by both peptidases and has a plasma half-life of 13 hours after subcutaneous administration."
Here is the whole class, read from the labels on 2026-09-16:
| Half-life | Source | |
|---|---|---|
| Your own GLP-1 | 1.5 to 2 minutes | Saxenda label, eff. 2026-02-25 |
| Liraglutide (Saxenda, Victoza) | 13 hours | Saxenda label, eff. 2026-02-25 |
| Tirzepatide (Zepbound, Mounjaro) | ~5 days | Zepbound label, eff. 2026-08-28 |
| Semaglutide (Wegovy, Ozempic, Rybelsus) | ~1 week | Wegovy label, eff. 2024-04-23 |
A week is 10,080 minutes. Against a half-life of 1.75 minutes, semaglutide persists on the order of 5,700 times longer than the hormone it copies.
The drugs are not different molecules. They are the same molecule that does not go away
This is the part most coverage leaves out, and it is what makes the comparison fair rather than dismissive.
From the Ozempic and Wegovy labels: "Semaglutide is a GLP-1 analogue with 94% sequence homology to human GLP-1."
From the Saxenda and Victoza labels: "Liraglutide is an acylated human GLP-1 receptor agonist with 97% amino acid sequence homology to endogenous human GLP-1(7-37)."
Ninety-four and ninety-seven per cent. These are not synthetic chemicals that happen to hit the same receptor; they are the body's own hormone with small deliberate edits. The Saxenda label even names what the edit is for — the profile comes from "self-association that delays absorption, plasma protein binding, and stability against metabolic degradation by DPP-4 and NEP."
So the engineering was not aimed at making a stronger signal. It was aimed at making the signal survive.
That reframes the "natural" claim entirely. The question was never whether the natural hormone works — it obviously does, and the drugs are proof, because they are copies of it. The question is whether anything can hold the natural hormone in circulation, and the clearance is the obstacle.
Why "boost your natural GLP-1" is the wrong lever
Interventions sold on this premise work on release: more L-cell secretion after a meal. Nothing sold over the counter works on clearance.
That asymmetry is the problem. DPP-4 and neutral endopeptidases do not wait. A larger pulse, cleared on the same two-minute clock, is still a pulse. Producing the effects these medicines produce would require holding receptor activation for hours or days — which is the specific engineering problem Novo Nordisk and Eli Lilly solved with acylation, protein binding and structural edits, and it took them years.
There is a class of prescription drug that does work on the clearance side: DPP-4 inhibitors, such as sitagliptin, which block one of the two enzymes named on the Saxenda label. They are approved for type 2 diabetes, they are prescription-only, and their weight effects are not in the same range as the GLP-1 receptor agonists. That the clearance-side approach exists as a licensed drug class, and still does not reproduce these results, is worth sitting with.
None of this says eating well is pointless. Diet does a great deal, most of it through pathways that have nothing to do with sustained GLP-1 receptor activation. What the half-life figure rules out is narrower and specific: the claim that a food or a capsule reproduces what a GLP-1 medicine does, via the same hormone.
What we could not source, and are not going to guess
Readers land here wanting a ranked list of foods and a number attached to each — "protein raises GLP-1 by X per cent". We are not publishing one, and the reason is worth stating rather than hiding.
Postprandial GLP-1 responses vary enormously with the assay used, whether total or active GLP-1 is measured, the meal composition, the timing of sampling and the population studied. Numbers pulled from individual small studies and presented as a food ranking are not a reliable guide, and we could not find a source that would let us publish one honestly on 2026-09-16.
What can be said, and is not in dispute: eating raises GLP-1, and protein, fat and fermentable fibre are all stimuli for L-cell secretion. That is textbook physiology. What does not follow from it is that any particular food produces a clinically meaningful, sustained effect — because of the clearance figure above.
If a trial is published that measures a food intervention against a sustained receptor-activation endpoint in people, this page will report it and say what it found.
The pages underneath this one
If the question is about products sold on this premise, the evidence for each ingredient — including the 337-person berberine trial that found no effect on visceral or liver fat, and the one positive randomised supplement result, which is about regain rather than loss — is on GLP-1 supplements.
If the question is about patches, the half-life is not even the binding constraint; the molecular weight is. Semaglutide is 4,114 daltons against a skin-penetration threshold usually put at 500. That is on why no patch can deliver a GLP-1, with a dated brand census on GLP-1 patch brands.
And if the question is what the approved medicines actually did in their trials, the head-to-head evidence is on tirzepatide vs semaglutide, and what they cost is on the GLP-1 cost page.