Berberine Supplements: 9 of 15 Failed a Basic Potency Test

If you are shopping for a berberine supplement, you have already been sold on the science, and fairly so. Berberine has a stronger clinical file than most of what surrounds it on the shelf, and unlike a lot of that shelf its trials measured things that matter, like HbA1c and LDL. That argument has been made, including here, at length.

This article is about the part that comes after you are convinced. Between a good trial result and a bottle in your kitchen sit two questions nobody puts on a label: is the compound in the capsule, and does any of it get anywhere. Both have been measured. Neither answer is comfortable.

Somebody bought fifteen bottles and weighed what was inside

In 2018, a group from the University of Kansas Medical Center published the study the supplement aisle deserved and mostly ignores. They bought fifteen unique berberine products from US dietary supplement vendors, pulled berberine out of three separate capsules from each one, and quantified it by ultra-high-performance liquid chromatography tandem mass spectrometry. Not a survey of labels. An assay.

The headline number: average content across the products was 75% of the label claim, with a standard deviation of 25%. Individual products ranged from 33% to 100% of what they said they contained.

Nine of the fifteen, which is 60%, failed to meet what the authors describe as “the potency standards of 90% to 110% of labeled content claim, as commonly required of pharmaceutical preparations by the U.S. Pharmacopeial Convention.” That is not an exotic bar. It is the ordinary tolerance a manufactured tablet is expected to hit.

Read the range again, because the shape of it matters more than the average. It runs from 33% to 100%. The ceiling is exactly the label claim, which means not one product overshot. Every deviation went the same direction: less than promised. If products were merely sloppy you would expect scatter on both sides of the line. Scatter that only ever lands short is a different kind of finding.

And then the part that removes the obvious defence. The team checked whether price rescued you, and it did not. Their evaluation of cost against measured potency “failed to demonstrate an association between quality and cost,” and they close the abstract with a sentence worth taping to the shelf: “the quality of the berberine product cannot be inferred from its cost.” The expensive bottle was not the honest bottle. It was a bottle.

The authors declared no conflicts of interest, which in a literature this crowded with manufacturer-funded work is worth saying out loud.

One honest limit: fifteen products, purchased at one point in time, in one market. It is not a census of every berberine supplement sold, and a specific brand today might assay perfectly. What it establishes is the base rate you are betting against when you buy blind, and that base rate is roughly one in three products holding under 90% of its stated dose.

The dose you are aiming at is not one capsule

Now compound that error with a second one, because the potency problem only matters relative to a target, and most people have the target wrong.

The trial that made berberine famous is a 2008 pilot in Metabolism by Yin, Xing and Ye. Study A randomised 36 adults with newly diagnosed type 2 diabetes “to treatment with berberine or metformin (0.5 g 3 times a day) in a 3-month trial,” and reported that “the hypoglycemic effect of berberine was similar to that of metformin,” with HbA1c falling from 9.5% to 7.5%.

Half a gram, three times a day, for three months. That is the regimen those numbers came out of.

So run the arithmetic on a real purchase. Target 1,500 mg a day in three divided doses. Take a product assaying at the study average and you are at roughly 1,125 mg. Take one from the bottom of the observed range and you are near 500 mg a day, a third of the trial dose, while doing everything the label told you. Now suppose the label itself said once daily. You are not running a slightly weaker version of the trial. You are running something the trial never tested, and if it does nothing you will conclude berberine does not work for you, which is the wrong conclusion drawn from the wrong dose of an unverified product.

Two more things from that same paper, since people quote its HbA1c figure and skip its methods. The authors call it a pilot study, in 36 people, over three months, which is nowhere near the footing metformin stands on. And “during the trial, 20 (34.5%) patients experienced transient gastrointestinal adverse effects,” with no functional liver or kidney damage observed. A third of people in a supervised trial had digestive trouble. Splitting doses across meals is the standard mitigation, and it is also, conveniently, the trial protocol.

Almost none of it reaches your blood

Here is where berberine stops behaving like a normal supplement.

In rats, the absolute oral bioavailability of berberine has been measured at 0.68% in one pharmaceutics study and 0.36% in another. Under one percent. The second paper worked out where it goes: “approximately half of berberine ran intact through the gastrointestinal tract and another half was disposed of by the small intestine,” and the authors concluded that “intestinal first-pass elimination of berberine is the major barrier of its oral bioavailability.” They also found a 70-fold higher concentration-time curve for berberine in liver than in plasma.

Those are rats, and rat pharmacokinetics do not transfer automatically. There is one human measurement, and it needs its disclosure before its numbers.

A 2021 randomised, double-blind crossover pilot in Nutrients gave five men a 500 mg dose of berberine, two doses of dihydroberberine, or placebo, then sampled venous blood over two hours after a standardised test meal. Peak plasma berberine after the 500 mg berberine dose was 0.4 ng/mL, plus or minus 0.17. Placebo was 0.22, plus or minus 0.18. The pairwise comparisons the paper flags as significant are the dihydroberberine ones; it reports no significant separation between 500 mg of berberine and placebo. Glucose and insulin did not differ across the protocol either, at p values of 0.97 and 0.24.

The disclosure: the study “was funded by NNB Nutrition,” the dihydroberberine used “was manufactured by NNB Nutrition (Nanjing, China) and is marketed as Glucovantage,” and the senior author “serves as a paid advisor for NNB Nutrition.” A trial funded by the maker of an alternative form found the alternative form absorbed better. That is exactly the structure you should discount for, and the paper does describe a conflict-of-interest management plan filed with its IRB. It also enrolled five people, and baseline berberine levels already differed between groups before anyone took anything.

So do not treat 0.4 ng/mL as settled human fact. Treat it as a small, interested measurement that happens to point the same way as independent rodent pharmacology, which is the only reason it is here at all.

Why poor absorption is not the gotcha it looks like

It would be easy to end the last section with a flourish about how berberine cannot possibly work. That would be wrong, and it is worth being clear about why.

The diabetes trials did not measure plasma berberine. They measured HbA1c, fasting glucose and lipids, and those moved. A compound can be nearly undetectable in blood and still act, if the place it acts is not the blood. Berberine concentrates in the liver, which is where a lot of glucose handling happens, and about half of an oral dose stays in the gut, where it alters the bacterial population that also has a hand in metabolism.

That reframe changes what you should conclude. Poor absorption is not evidence of failure. It is evidence that berberine behaves like a gut-and-liver agent, and that has three practical consequences: the split dosing schedule is load-bearing rather than a formality, gastrointestinal side effects are the mechanism’s shadow rather than a coincidence, and any product promising to solve absorption is promising to change the thing the clinical evidence was actually generated with. Better absorbed is not the same as better studied.

What to do with all this

Buy on documentation, not on price or design. The one variable tested against potency was cost, and it predicted nothing. Ask for a certificate of analysis for the lot, from a third-party laboratory, showing measured berberine content. A brand that has one will show you. A brand that answers with the word “premium” has told you what it has.

Aim at the protocol, not at the bottle’s convenience. The evidence base is 500 mg three times daily with meals, run for around three months, because HbA1c reflects a roughly three-month average and there is nothing to see at two weeks. A once-daily product is not that. If you would rather understand the general logic of dose timing, that is its own subject.

Expect the gut to complain, and split doses because of it. A third of trial participants had transient gastrointestinal effects under supervision. Starting at one dose a day and building up is sensible, so long as you know where you are heading.

Do not read low absorption as licence to buy a reformulated version. The novel forms have the marketing advantage of a real pharmacological problem to point at, and the disadvantage that the outcome trials were run on ordinary berberine hydrochloride.

Treat it as an adjunct, and tell your doctor. Berberine’s serious issues are not on the potency list. It interacts with medications and carries genuine contraindications, which are covered separately, and the population with the best reason to take it is also the population most likely to be on other drugs. If you are considering it alongside other metabolic supplements, the combination question and the stacking question both deserve their own answers rather than being assumed away.

The bottom line

The strange thing about berberine is that its evidence problem and its product problem run in opposite directions. The clinical file is unusually good for a supplement. The manufacturing file, on the one occasion anybody assayed it, showed an average of 75% of label claim, a floor of 33%, nine of fifteen products outside the standard potency window, and no relationship between what a product cost and what it contained.

Which means the weakest link in taking berberine is not the science. It is that you cannot verify the dose you are taking, aimed at a target most labels state incorrectly, of a compound that barely enters your bloodstream and probably does not need to.

None of that is a reason to skip it. It is a reason to buy it the way you would buy a reagent rather than a wellness product: assay documentation first, trial protocol second, patience third. And if the reason berberine came up at all was a video calling it something it is not, the nickname is worth unpicking before the purchase, not after.

This article is for general information and is not medical advice. Talk to a qualified healthcare professional before starting any supplement, especially if you take medication, are pregnant or breastfeeding, or have a medical condition. Berberine is not a replacement for prescribed treatment for diabetes or any other condition.