Berberine, what the research says

By Dr Conor Kerley, PhD, cardiovascular and nutrition researcher.

The most-searched ingredient in this library was a dye before it was anything else, and almost none of it survives the trip from your mouth to your blood.

Berberine comes from the roots and bark of plants that aren't closely related to one another: barberry, goldenseal, several Phellodendron species, and the huanglian of Chinese herbal practice. What they share is that they are yellow inside, a hard saturated yellow that transfers to anything it touches, and for most of the time humans have known about the stuff that was the entire point of it. It dyed cloth and leather for centuries, and the stain on your fingers was how you knew you had the real root rather than something sold to you as the real root. The plants had a second job in Chinese and Ayurvedic practice, mostly for gut complaints, which has almost nothing to do with why six and a half thousand people a month now type the word into a search box.

What changed is that somebody started measuring. The foundational paper is Yin, Xing and Ye, published in Metabolism in 2008, and its authors called it a pilot, which is the correct word and the one most people quoting it leave out [1]. They gave newly diagnosed adults with type 2 diabetes 1,500 mg a day across three meals for three months, with fifteen people in the berberine arm. What they measured was HbA1c, which is worth understanding before you read a result off it. An ordinary blood-sugar test tells you where you were at the moment the needle went in, which is a number you can flatter by being careful the night before. HbA1c can't be flattered: it reflects the average of the previous three months or so, and it is reported as a percentage. The people in that trial started at 9.5%, which is a long way up, and after three months on berberine they were at 7.5%, with movement in fasting glucose, after-meal glucose, cholesterol and triglycerides alongside it.

Two later poolings gathered up everything published since. Dong and colleagues collected eleven studies and 874 people in 2013 [2], and Guo and colleagues collected forty-six studies and 4,158 people in 2021, at doses running from 600 to 1,500 mg a day [3]. Guo's headline was around three-quarters of a point off HbA1c, which means three-quarters of a percentage point on that same three-month average, so somebody starting at 8% would finish nearer 7.25%. That is a good deal less than Yin's two points, which is what you would expect when one is a single small pilot and the other the average of forty-six trials, and smaller falls in glucose and blood fats came with it. That is the shape of the field, and it is smaller and newer than the internet suggests: one modest trial that got everyone's attention, and two collections that followed the same measures at the same kind of dose.

Almost all of that research used berberine hydrochloride, which is why the term gets searched nearly as often as the plain word does. It sounds like a technicality, and it is the opposite of one, for reasons that only become clear once you follow the capsule.

Under one percent

Less than one percent of the berberine you swallow reaches your bloodstream intact [4].

That figure is not a polite way of saying absorption is modest. Take a 500 mg capsule: almost the entire contents of it, the part you paid for and the part specified in the trial's methods section, does not arrive anywhere. Two obstacles account for most of the loss. The lining of your gut is not a passive sheet but an active border, and part of its equipment is a transporter whose job is to seize foreign compounds that have crossed into the wall and throw them back out again, and berberine is close to a perfect description of what that pump was built to eject [4]. Whatever survives goes to the liver before it goes anywhere else, and the liver takes a large share of the remainder, with further bottlenecks behind those two [4][5]. By the end of the chain, ninety-nine parts in a hundred are gone.

This explains something you notice on reading the trials properly, which is how fussy their methods sections are. The same specifications keep repeating: this salt, at this dose, this many times a day, with meals. Those are not housekeeping details the authors included out of habit. When the margin between arriving and not arriving is that narrow, the form and the schedule are part of the intervention rather than the packaging around it, and a number printed on a label describes what went into the capsule rather than what gets anywhere near your blood. It is also why an entire research field now exists around the problem, with groups working on delivery approaches designed to get more of it across and on identifying which of berberine's breakdown products are doing the work [5]. That literature is live and I read it as it lands, because the conversation about berberine is conducted almost entirely in milligrams and the milligram is the least interesting number in it.

It also explains, I think, why the trials use the doses they use. Nobody has run a proper dose-finding study on berberine, so 600 to 1,500 mg is not a figure anyone calculated from first principles. But set it beside the absorption number and it stops looking arbitrary. If ninety-nine parts in a hundred are going to be discarded before they reach your blood, you have to put a great deal in at the top to get anything measurable out at the bottom. The dose is large because the body is wasteful with it, not because more is better.

The yellow, meanwhile, still does what it always did. It stains the countertop, the mortar and the fingers of anyone who handles the raw root, and for centuries that stain was the proof you had the real thing. It is the one part of berberine that has never had any trouble getting where it is going.

References

1. Yin J, Xing H, Ye J. Efficacy of berberine in patients with type 2 diabetes mellitus. Metabolism. 2008;57(5):712–717. doi:10.1016/j.metabol.2008.01.013. PMID 18442638.
2. Dong H, Zhao Y, Zhao L, Lu F. The effects of berberine on blood lipids: a systemic review and meta-analysis of randomized controlled trials. Planta Med. 2013;79(6):437–446. doi:10.1055/s-0032-1328321. PMID 23512497.
3. Guo J, Chen H, Zhang X, Lou W, Zhang P, Qiu Y, Zhang C, Wang Y, Liu WJ. The Effect of Berberine on Metabolic Profiles in Type 2 Diabetic Patients: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. Oxid Med Cell Longev. 2021;2021:2074610. doi:10.1155/2021/2074610. PMID 34956436.
4. Han Y, Xiang Y, Shi Y, et al. Pharmacokinetics and pharmacological activities of berberine in diabetes mellitus treatment. Evid Based Complement Alternat Med. 2021;2021:9987097. PMID 34471420.
5. Approaching strategy to increase the oral bioavailability of berberine, a quaternary ammonium isoquinoline alkaloid: Part 1 and Part 2. Expert Opin Drug Metab Toxicol. 2023;19(3). doi:10.1080/17425255.2023.2203857 and doi:10.1080/17425255.2023.2203858.

Last reviewed: 10 September 2026. This article is educational and is not medical advice.