Berberine

Berberine works like Ozempic

  • Built on 15 Sources
  • Last checked

No trial has tested berberine against Ozempic (semaglutide), so the only comparison available sets separate trials side by side, and on those numbers berberine's effects are smaller. In the largest placebo-controlled berberine trial in type 2 diabetes, HbA1c, a measure of average blood sugar over the previous few months, fell 0.4 percentage points more on berberine than on placebo over 12 weeks. In the 30-week trial reported on Ozempic's US label, it fell 1.2 to 1.4 points more than on placebo. On weight, that Ozempic trial recorded losses of 3.8 to 4.7 kg against 1.2 kg on placebo. A 2026 review pooling 23 berberine trials found body weight about 0.9 kg lower than in control groups, while the largest double-blind berberine trial, in people with obesity, found no difference from placebo in body weight, belly fat or liver fat after six months.

The GLP-1 connection behind the comparison comes from animal work: berberine restored GLP-1 production in mice on a high-fat diet, and none of the sources here measured GLP-1 in people taking berberine. Ozempic is a prescription drug that acts on the GLP-1 receptor directly. No trial in these sources tested replacing a prescribed diabetes drug with berberine. NIH's complementary health center reports that goldenseal, a berberine-containing herb, lowered blood levels of the diabetes drug metformin by about 25 percent in healthy adults and that berberine can be harmful to newborns. NCCIH advises anyone taking medicine to consult a health care provider before using herbal products. Three case reports published in 2026 describe torsades de pointes, a dangerous heart rhythm, in adults taking berberine, two of them people with type 2 diabetes taking it for blood sugar, and France's food safety agency lists people with diabetes or heart disorders among those at higher risk from berberine supplements.

Blood sugar is where the berberine evidence is most consistent. PREMOTE, a double-blind trial at 20 centres in China, gave 409 adults with newly diagnosed, untreated type 2 diabetes berberine at 1.2 g a day, probiotics, both, or placebo for 12 weeks after a week of antibiotic pretreatment. HbA1c fell 0.99 points on berberine and 0.59 on placebo, and berberine caused more digestive side effects. A 2021 review pooling 46 trials in type 2 diabetes and prediabetes, many open-label and many giving berberine alongside standard diabetes drugs, found HbA1c 0.73 points lower than in control groups. A 2023 review limited to 20 placebo-controlled trials in 1,761 people found a smaller difference, about 0.4 points across the seven trials that measured HbA1c. A 2025 review limited to 12 placebo-controlled trials in 889 people with features of metabolic syndrome found fasting glucose about 0.5 mmol/L lower, waist circumference about 3.3 cm smaller and BMI about 0.4 lower than on placebo, with adverse events no more common.

Weight is where the results agree least. The 2026 review of 23 randomized trials found body weight 0.88 kg lower and BMI about 0.5 lower with berberine than in control groups, and its authors asked future trials to report the purity and dose of what was tested and to address the lack of blinding and randomization. BRAVO, published after that review, randomized 337 adults with obesity and fatty liver disease but no diabetes, at 11 hospitals in China, to berberine 1 g a day or placebo for six months. Visceral fat, liver fat, body weight, BMI and waist size did not differ between the groups. LDL cholesterol fell 7.7 mg/dL more on berberine, and adverse events, including episodes of low blood sugar, were similar in both arms.

Ozempic's label describes the injectable drug as a once-weekly GLP-1 receptor agonist for adults with type 2 diabetes, also used to lower the risk of heart and kidney events in some of them. The weight figures above come from a diabetes trial of 388 adults with an average BMI of 33. The label carries a boxed warning about thyroid C-cell tumors seen in rodents, and states that it should be used in pregnancy only if the benefit justifies the risk to the fetus.

NIH's lactation database explains the newborn concern: berberine can displace bilirubin from the blood protein that carries it, and bilirubin can then build up in an infant's brain; how much berberine passes to a breastfed infant is unknown. NCCIH advises against goldenseal in pregnancy, breastfeeding and infancy, and states that very little berberine may be absorbed when goldenseal is taken by mouth. NIH's liver injury database reports that berberine has not been linked to liver enzyme rises or to clinically apparent liver injury. The two large trials here used plain berberine; dihydroberberine, a form sold as more absorbable, was not tested in them.

Heart rhythm is a separate concern. In laboratory studies, berberine blocks a potassium channel that helps the heart's electrical system reset between beats, and blocking it can lengthen the QT interval on a heart tracing. Three case reports published in 2026 describe torsades de pointes, a fast, irregular rhythm that can lead to cardiac arrest, in adults taking berberine: a 63-year-old woman with type 2 diabetes who had taken 450 mg twice a day for six months, a 92-year-old man with type 2 diabetes two weeks after he started 500 mg a day, and a 61-year-old man taking berberine and dihydroberberine together with turmeric and quercetin, whose blood magnesium was slightly low. In all three the QT interval was long and shortened after the supplements were stopped. Case reports cannot show how often this happens. France's food safety agency reviewed berberine supplements in 2019. It found effects like a drug's from 400 mg a day, adverse effects from 600 mg a day, and risks of digestive upset, low blood sugar and low blood pressure, and it listed interactions with drugs including metformin, other diabetes drugs, statins and some antibiotics. Because the toxicology studies were of poor quality, it could set only a provisional limit, about 0.1 mg a day for a 60 kg adult, a small fraction of the 1 to 1.2 g a day used in the two large trials here. It advised pregnant and breastfeeding women, people with diabetes, people with liver or heart disorders, and children and adolescents not to take berberine supplements.

The signal grid

Formal Research

Evidence Volume How much formal human research exists.
Consistency Whether the studies point the same way.
Study Quality How well the research was designed.
Recency Whether the picture reflects recent work.

Broader Signal

Traditional Use Whether there is a documented history of use.
Mechanism How well the biology is understood.
Momentum Which way the research is moving.

See the full grid, every Source and what each one found in the Substantia app.

Coming soon to the App Store for iPhone

Plain English, with the Sources shown. No verdicts, no scores, nothing for sale.

Join the launch list

All claims