“GLP-1 supplement” is a marketing category, not a pharmacological one. It covers capsules of berberine, probiotic strains such as Akkermansia, soluble fibers like psyllium, and botanicals such as yerba mate, sold on the claim that they nudge the body's own GLP-1 or mimic what the prescription drugs do. None of them is a GLP-1 receptor agonist, and the distinction between raising a hormone and activating its receptor for a week at a time is set out in how the drugs work. What follows is narrower: what randomized trials measured on the scale, and how large that number is next to the drugs the products borrow their name from.
What the FDA does and does not review
Start with the regulatory floor, because it explains why the claims and the evidence drift apart. Under the Dietary Supplement Health and Education Act, the FDA does not have the authority to approve dietary supplements before they are marketed, and it generally does not approve supplement claims or labeling before use. Unlike drugs, which must be proven safe and effective for their intended use first, the agency is largely limited to enforcement after a product is already on sale.[1]
So a bottle that says “supports healthy GLP-1 levels” has not been checked by anyone for whether it does. That phrasing is a structure/function claim, and the law requires such a claim to carry a disclaimer that the FDA has not evaluated it.[1] The prescription drugs sit on the other side of that line, which is the subject of what “not FDA-approved” means.
Berberine: real, significant, and under a kilogram
Berberine carries the “nature's Ozempic” nickname, and it is also the supplement with the deepest trial base. A 2026 systematic review pooled 23 randomized controlled trials in adults. Berberine reduced body weight by a mean difference of −0.88 kg (95% CI, −1.36 to −0.39), body-mass index by 0.48 and waist circumference by 1.32, with no significant effect on waist-to-hip ratio.[2] The authors asked future trials to report the purity, potency and gram amounts of what they tested, and to fix common biases such as missing blinding and randomization.
An earlier 2020 meta-analysis of 10 studies found a smaller waist and a lower body-mass index, but no significant change in body weight itself: −0.11 kg (95% CI, −0.99 to 0.76, p = 0.79).[3] Read together, the two analyses agree that berberine moves the scale by somewhere between nothing and about a kilogram. The newer one reached significance by adding trials, not by finding a larger effect.
The same axis as the drugs
In STEP 1, 1,961 adults without diabetes took semaglutide 2.4 mg or placebo for 68 weeks. Mean body weight fell 14.9% against 2.4%, and the placebo-adjusted difference in absolute terms was −12.7 kg (95% CI, −13.7 to −11.7).[4] Set beside the 2026 berberine estimate, that is a ratio of roughly 14 to 1. The trial itself is walked through in the STEP 1 article, and how the figure holds up across the rest of the program is in the semaglutide weight-loss article.
Tirzepatide widens the gap. In SURMOUNT-1, 2,539 adults on the 15 mg dose lost a mean 20.9% of body weight at 72 weeks against 3.1% on placebo, and 57% of that group lost a fifth of their weight or more.[5] The full dose-by-dose picture is in the tirzepatide weight-loss article and the SURMOUNT-1 article. No supplement trial in the literature reviewed here reports a percentage of body weight in that range, or a 20% responder rate at all.
The comparison is not perfectly matched, and the mismatch deserves a sentence. The drug trials ran 68 and 72 weeks with thousands of participants; most supplement trials are smaller and shorter. Duration alone does not close the gap, though. The 2020 berberine analysis found its body-mass index and waist effects tracked treatment duration, while body weight still did not separate from control.[3] And at the 5% threshold that clinicians use as a first marker of response, STEP 1 put 86.4% of the semaglutide group over the line against 31.5% on placebo.[4]
Fiber: the largest supplement number has a catch
Soluble fiber produces the biggest pooled figure on this page. A 2017 meta-analysis of 12 randomized trials (n = 609) lasting 2 to 17 weeks found isolated soluble fiber reduced body weight by 2.52 kg (95% CI, −4.25 to −0.79) against placebo, along with lower fasting glucose and fasting insulin. The authors flagged considerable between-study heterogeneity and asked for caution in reading it.[6]
Psyllium is the fiber most often sold under the GLP-1 label, and it has its own meta-analysis. Across 22 randomized trials, psyllium did not significantly change body weight (−0.28 kg, 95% CI, −0.78 to 0.21), body-mass index or waist circumference.[7] The pooled fiber number, in other words, does not transfer to the one fiber product a shopper is most likely holding. Fiber has other documented uses, including for the constipation that the drugs themselves cause, which is a separate question from weight.
Yerba mate: one small trial
The yerba mate evidence for weight is thin enough to state in full. A 12-week double-blind trial in Korean adults with obesity gave 3 g a day to 15 participants and placebo to 15. Body fat mass, percent body fat and waist-to-hip ratio fell significantly against placebo, with no clinically significant safety signal.[8] Thirty people over three months is a reason to run a larger trial, not a result that can be compared with anything above.
Does any supplement actually raise GLP-1?
One has a randomized trial saying so, and its details matter. In a double-blind trial of 142 adults with metabolic syndrome, four months of pasteurized Akkermansia muciniphila did not change the primary endpoint, whole-body insulin sensitivity, against placebo. Exploratory analyses found a larger post-glucose-load GLP-1 excursion over placebo (p < 0.01), and the clearest benefits were confined to participants who started with low gut Akkermansia; in that subgroup the weight reduction did not reach significance (p = 0.06).[9] The trial was run by the company developing the strain.
A second trial from the same strain's developer tested weight maintenance rather than loss. Ninety adults first lost at least 8% of body weight on an eight-week low-energy diet, then took the bacterium or placebo for 24 weeks. Regain was 1.2 kg against 3.2 kg (P = 0.012), with no serious treatment-related adverse events.[10] That is a two-kilogram difference in what came back after a diet did the work, which is a real finding and a different claim from replacing a drug.
Even granting the GLP-1 result, the scale of it is the issue. Native GLP-1 released after a meal is broken down within minutes, which is why the approved drugs are engineered molecules built to persist for days; the pharmacology is covered in the half-life article. A statistically larger post-meal spike of the body's own hormone has not been shown, in any trial reviewed here, to produce weight loss resembling a receptor agonist's.
Safety: berberine is not inert
“Natural” is not a pharmacokinetic category. In a randomized crossover study in healthy men, two weeks of berberine 300 mg three times a day decreased the activity of three major drug-metabolizing enzymes. A CYP2D6 probe ratio rose about ninefold, a CYP2C9 ratio doubled, and exposure to the CYP3A4 probe midazolam rose by roughly 40%, with oral clearance down 27%.[11] The authors concluded that drug-drug interactions should be considered when berberine is taken.
Those three enzymes handle a long list of common prescriptions, which makes a supplement shelf a medication-list question. Gastrointestinal effects are the other recurring signal: in a 2008 pilot in type 2 diabetes using 0.5 g three times a day, 20 patients (34.5%) reported transient gastrointestinal adverse effects.[12] Stacking that on top of a GLP-1 drug, whose own nausea and constipation are well documented, adds to the same organ system. How the drugs already change the absorption of oral medicines is covered in the oral medications article.
What none of this decides
None of these products is worthless on its own terms. Berberine has documented effects on glucose and lipids, fiber on fasting glucose and regularity, and the Akkermansia maintenance trial is a genuine signal worth replicating. The claim the category is sold on is a different one, and on the scale it is off by about an order of magnitude. The older generic comparison that most resembles berberine's profile is in the metformin article.
The prescription route has its own verification problem. Many telehealth sellers dispense compounded semaglutide or tirzepatide, and compounded drugs are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before they are dispensed. What to check before paying is in the telehealth red-flags article.
The same botanicals turn up in a patch form, which raises a separate legal and absorption problem; see GLP-1 patches.