Berberine
hydrochloride

Berberine is a plant-derived substance taken as a supplement to improve blood-sugar and cholesterol control, with a smaller effect on weight. Its benefits are real but modest, and were measured above all in people with some metabolic problem.

At the molecular level, almost everything berberine does hangs on the same switch: it activates AMPK, an enzyme the cell uses as an energy sensor. From there its effects on glucose and lipid metabolism follow, in a cascade.

It has been studied on three fronts: blood sugar, cholesterol and weight.

Where it works best is on blood sugar: in people with type 2 diabetes or prediabetes it lowers glucose and glycated hemoglobin appreciably in small studies. On cholesterol it reduces LDL and triglycerides fairly consistently. On weight, the effect is small (a couple of kilos versus placebo) and seems more a consequence of improving metabolism than a direct fat-burning effect.

The mechanism makes clear why. On glucose, AMPK activation improves insulin sensitivity and curbs hepatic glucose production. On lipids, it does what made it famous: it stabilizes the messenger RNA of the LDL receptor by a pathway (ERK) different from the one statins use (SREBP), and it also inhibits PCSK9. That is why it adds to the effect of a statin instead of overlapping with it. In Kong's foundational series (2004), with no placebo group, three months of berberine lowered LDL by 25%, total cholesterol by 29% and triglycerides by 35%. But an elegant pathway does not guarantee a large magnitude: the placebo-controlled RCTs (below) return berberine to its real size — modest and consistent.

A separate case is polycystic ovary syndrome, where it is studied for its action on insulin resistance. Here an honest caveat: in a trial that compared it with letrozole for achieving pregnancy, berberine fell behind (around 22% live births versus 36% for letrozole). It improves metabolic parameters, but it does not replace fertility treatments.

On cognition there is barely anything: a single trial in people with schizophrenia and no evidence that it prevents cognitive decline in a healthy population. Today it is an unknown, not a promise.

The trial is Pu et al. (2023): 106 patients with chronic schizophrenia in whom berberine improved negative symptoms and cognitive function, attributed to an anti-inflammatory effect. Outside that clinical context there is nothing: no RCT on preventing cognitive decline in a healthy population, only preclinical animal work. The Alzheimer's Drug Discovery Foundation concludes that there is no clinical evidence that it prevents cognitive decline or Alzheimer's. Confidence: very low.

It makes sense above all if your blood sugar or cholesterol are high; in a healthy person the expected benefit is marginal. It deserves respect: it has important interactions and contraindications (you will see them below). If you take medication or have an underlying condition, decide it with a doctor.

The studies used berberine HCl, 1 to 1.5 g a day split into two or three doses, with meals. Starting with a low dose the first week reduces digestive discomfort. This is what was done in the trials, not a regimen for treating a disease: for that, a professional sets the dose.

Splitting it into several doses with meals smooths the post-meal glucose spikes and spreads the digestive load. With dihydroberberine, which is absorbed considerably better, lower doses are used; but, as the section on forms shows, that does not mean its effect is equally well proven.

The body absorbs less than 1% of the active ingredient you consume, and at the usual doses that can cause digestive discomfort.

That “pharmacokinetic paradox” is one of its most curious points: with an oral bioavailability below 1% and very low plasma concentrations, it should not do much, and yet it does. The dominant hypothesis is that it acts locally in the intestine and through its metabolites and the microbiota, rather than circulating in the blood.

No. Almost all the research was done with berberine HCl. Other forms (dihydroberberine, phytosomes) promise better absorption, but have far less backing. Two bottles both labeled “berberine” may not behave the same. The detail, right below.

The difference between forms is real and has consequences. Here is the picture:

The practical message: when a “berberine” study actually uses dihydroberberine or a branded compound, its results do not automatically carry over to an ordinary HCl capsule. Lumping all forms together as if they were one inflates the impression that “berberine always works”.

Two forms with a sensitive conflict of interest illustrate this. Dihydroberberine (Glucovantage®) reached, in a small crossover study (Moon et al., Nutrients 2022; n=5 healthy men), a plasma berberine concentration higher than that of 500 mg of HCl; but it is only pharmacokinetic, with no clinical outcomes, and comes from the patent holder. The phytosome (Berberine Phytosome™) improved glycemia in an RCT of 49 subjects with impaired fasting glucose (Rondanelli et al., 2023), with authors linked to Indena, the patent holder. Legitimate signals, but not transferable to an HCl capsule or to each other.

Berberine has no quality seals of its own. It may carry generic certifications (USP, NSF), which verify purity, not efficacy.

The parallel does not hold beyond the figure: semaglutide mimics a gut hormone (GLP-1) that regulates appetite and satiety; berberine acts through other metabolic pathways and does not produce that effect on appetite. Different mechanisms, different results.

A selection of the most solid placebo-controlled RCTs, with their figures and their limits. Beside each main study we note the design, the population, the intervention, the result, the conflict of interest and the citation.

Context (not main; no placebo group or no clinical outcome): Yin et al. 2008 (berberine compared with metformin in type 2 diabetes, similar glucose-lowering effect, but without placebo); Kong et al. 2004 (defines the lipid mechanism); and, as quantitative context on weight, the Asbaghi et al. 2020 meta-analysis (12 RCTs): −2.07 kg, −0.47 kg/m² BMI and −1.08 cm of waist versus placebo, with no effect on ALT/AST.

There are barely any studies longer than six months, so its long-term safety is not well established. And the real size of its effect on weight is still debated.

Three concrete shadows. The heterogeneity of forms and doses: many meta-analyses mix HCl, dihydroberberine and branded compounds, so the apparent consistency is greater than the real one. The provenance: a large share of the trials are Chinese, with populations and practices that do not always carry over. And the discrepancy between meta-analyses: on weight, they range from almost zero to −2 kg depending on which are included.

Beneath all that there is an underlying unknown: the signals of genotoxicity in vitro and of carcinogenicity in rodents that concern EFSA are pending in vivo confirmation, and safety beyond six months still lacks a solid body of evidence. It is the gray area that explains both the regulatory caution and the difficulty of closing the debate.

It is worth reading this part calmly, because berberine is not harmless.

It is contraindicated in pregnancy and breastfeeding, and must never be given to newborns or infants: it can cause them a serious problem with bilirubin.

It interacts with many medications, because it alters the way the liver processes them. Special care with immunosuppressants, anticoagulants, some statins, antidiabetics and digoxin. If you take medication regularly, consult your doctor or pharmacist first.

The detail of the interactions is quantified in humans, and that is why it carries weight:

The best-documented case is cyclosporine: in kidney transplant recipients, adding berberine raised its trough concentration. Here a nuance that should not be exaggerated: the “+88.9%” that is usually cited is the increase from baseline within the treated group; the net effect attributable to berberine, comparing between groups, was about 29% (Wu et al., 2005). It remains relevant for a narrow-margin drug, but the honest number is the second. Added to this is an additive hypoglycemia risk when combining it with antidiabetics (metformin, insulin, sulfonylureas).

The neonatal contraindication has a concrete primary basis: in vitro, berberine displaces bilirubin from albumin with a potency about 10 times greater than phenylbutazone and about 100 times greater than papaverine (Chan, 1993). In pregnancy it is contraindicated because it crosses the placenta and is associated with a risk of uterine contractions; during breastfeeding, because it is excreted in milk.

As for its legal situation: in January 2026, the relevant EFSA panel (the European food safety authority) endorsed a draft opinion concluding that, with the available data, no amount of berberine can be established as safe, owing to concerns about genotoxicity and liver toxicity. There was a public consultation that closed in May 2026. The nuance matters: as of today it remains a draft, not an adopted final opinion, and the European Commission has not made any decision yet. But the direction points to growing scrutiny.

The detail: the draft is processed under Article 8(2) of Regulation (EC) 1925/2006, at the request of France's ANSES, based on external preparatory work of systematic review. With the consultation closed on 11 May 2026, as of this article's date (June 2026) EFSA has not yet adopted the final opinion, nor has the European Commission decided anything; the possible outcomes range from a ban (Annex III, part A) to “Union scrutiny” (part C, with four years to provide data). France had already taken a position earlier: ANSES documents berberine's pharmacological effects from about 400 mg a day, without ruling out that lower doses produce them, and, given the scarcity of toxicological studies, set a very low reference value (equivalent to about 0.1 mg a day for a 60 kg person, which many supplements exceed). It therefore advises against its use by pregnant women, breastfeeding women, diabetics and people with liver or heart disorders. The opinion, moreover, is not unanimous: experts linked to the industry argue that the safety signals could come from other compounds in the plants, not from pure berberine, and criticize relying on in vitro data and very prolonged animal studies. The debate remains open; this article reports it without taking sides.

None of this is medical advice: it is information to take to someone who can actually advise you on your case.