Berberine: Is it really "nature's Ozempic"?

Berberine is not "nature's Ozempic", because it does not belong to the GLP-1 drug family and works through a different switch in the body. The nickname sells well, but the two have little in common, starting with the hormone Ozempic is built on.

Ozempic (semaglutide) and Mounjaro (tirzepatide) are GLP-1 drugs. They copy a gut hormone you release after eating, the one that slows your stomach and tells your brain you are full. Berberine does not copy that hormone or plug into its receptor. In rat and cell studies it nudges the gut to release a little more of your own GLP-1.^[18] In people, one small two-week trial with a branded low-dose form found roughly a quarter more GLP-1 in the blood of adults whose fasting blood sugar was in the prediabetes range.^[39] Two of its authors work for the maker, though, and it did not show that the extra GLP-1 explains any effect on blood sugar.^[39] Either way, that is a trickle next to what a GLP-1 drug delivers.

What berberine does in lab studies is switch on a cellular fuel gauge called AMPK. It works like a low-battery alarm that tells cells to burn energy instead of storing it. Metformin works through the same switch, and studies in mice, rats and cell lines showed that berberine flips it too.^[9] A later cell study found that berberine reaches the switch by a different route and flips it more weakly than metformin.^[6] Its authors stress that the two work differently.

That makes "plant metformin" the fairer nickname, and even that one needs a footnote, since the switch is shared but the wiring is not. The difference matters because people buy berberine picturing Ozempic-style results. Ozempic is a prescription drug that acts on your hunger hormones, while berberine is a supplement that gives cell metabolism a small push. Both touch blood sugar, which is also where berberine's trial record is strongest.

Berberine lowers blood sugar and cholesterol modestly, and the people who benefit most already have type 2 diabetes or metabolic syndrome. If you are healthy and take it as a longevity supplement, the evidence has much less to offer, because the effects show up mainly in people with those conditions. The effects are real, though smaller and less well tested than those of metformin or a GLP-1 drug.

In type 2 diabetes, a large review of 37 randomized trials found HbA1c about 0.63 percentage points lower.^[1] HbA1c is your three-month blood-sugar average, and the unit matters: an HbA1c of 8.0% would fall to roughly 7.4%, not by 0.63% of its value. Fasting blood sugar fell by about 0.82 mmol/L, or roughly 15 mg/dL, in the two units labs use for blood sugar.^[1] Two other reviews land in the same zone of 0.5 to 0.7 points.^{[8, 13]} On paper that is less than the roughly 1.1 points metformin alone achieves against placebo.^[23] That comparison crosses different trials and different people, though, and the direct comparisons tell a less tidy story.^[2]

Those direct comparisons come from a 2024 review that split berberine taken on top of a diabetes drug from berberine taken alone.^[2] As an add-on, berberine lowered HbA1c by about 0.69 points more. Taken alone, berberine did not lower HbA1c clearly across the small trials, and that result shifted when a single trial was left out.^[2] A few small trials against placebo looked better.^[2] The head-to-head trials against metformin found no clear difference, but they are far too small to show the two are equal.^[2] So as an add-on berberine gives a modest extra push, while on its own its effect is not yet pinned down.

Cholesterol is the other place where the reviews agree. Four reviews find a real drop in LDL cholesterol (the "bad" kind) of around 0.3 to 0.5 mmol/L.^{[5, 2, 10, 14]} Triglycerides, another type of blood fat, fall usefully too.^{[5, 2, 10, 14]} HDL, the "good" kind, barely moves, and LDL also fell a little in a large recent placebo-controlled trial.^[30] In one of those reviews, though, combination products drove most of the drop, and berberine alone had a weaker LDL effect that was not clearly significant.^[14]

All of these numbers share one limit: most trials are small, short (one to three months) and from a single country. Read the pooled results as the optimistic end of the range, not as a promise of what you will see.

If you try berberine with your doctor, a before-and-after check of HbA1c or fasting glucose, plus blood fats, shows whether your numbers moved. A continuous glucose monitor adds detail, and so does a HOMA-IR reading, an insulin-resistance score worked out from fasting glucose and insulin. A change in one person still does not prove berberine caused it, because diet, weight and season move these numbers too. If you take diabetes drugs, do the monitoring with your doctor, since the combination can push blood sugar too low.^[28]

Berberine's strength lies in blood sugar and blood fats rather than weight: on average, reviews find between no loss and about 2 kg.^{[4, 20, 29]} The largest placebo-controlled trial in people with obesity found no difference at all.^[30] Next to the GLP-1 drugs, that near-zero is the heart of the "natural Ozempic" problem, and that one trial shows it most clearly.

The trial ran in China with 337 adults who had obesity and fatty liver disease but no diabetes.^[30] For six months they took either 1 g of berberine a day or a placebo. Both groups lost about 2 kg, which is a respectable result for a placebo, and there was no difference between them. The fat around the organs and in the liver, the trial's main targets, did not change either.^[30]

The largest review, with 23 trials, comes out slightly kinder, with a small extra loss of just under 1 kg.^[29] Its authors warn, though, that many of those trials did not report purity or dose properly and were not blinded.^[29] Older reviews mainly disagree about how small the effect is, ranging from no clear change,^[4] to about 2 kg^[20] with the rest in between.^{[15, 5, 8]}

The drugs berberine gets compared with play in another league. Semaglutide 2.4 mg is the Wegovy dose, higher than Ozempic's diabetes doses. In the STEP 1 trial it took off about 15% of body weight over 68 weeks, against about 2.4% on placebo.^[21] In SURMOUNT-1, tirzepatide took off about 15 to 21% over 72 weeks, depending on dose, against 3.1% on placebo.^[22] The comparison is indirect, though. The berberine trials ran one to six months and report the difference to a control group, while the GLP-1 trials ran well over a year.

Even with that caveat, the gap shows in every row:

Sources: berberine reviews and trials^{[4, 20, 29, 30]} STEP 1,^[21] SURMOUNT-1.^[22]

Even allowing for the longer trials, the GLP-1 drugs take off many times more weight than berberine's best pooled estimate. My call: if you take berberine for its metabolic effects, treat any weight you lose as a small bonus, never as the reason.

Most trials gave 500 mg of berberine two to three times a day, so 1,000 to 1,500 mg in total, split across meals.^[3] That is the dose tested in short trials of one to three months, not a safe or recommended intake. In its 2026 draft, the EU food safety authority EFSA could not set any safe intake level.^{[25, 34]}

The official numbers sit far below the trial dose. France's food agency ANSES sees drug-like effects from about 400 mg a day and side effects from about 600 mg a day.^[28] Its provisional toxicity value works out at only 0.1 mg a day for a 60 kg adult, a tiny fraction of a single trial capsule.^[28] Belgium had capped berberine and related plant compounds (isoquinoline alkaloids) in supplements at 10 mg a day, as ANSES reported in 2019.^[28] EFSA also found no adequate long-term toxicity studies,^[34] so the trial dose tells you what was tested, not what is safe.

Behind those big trial doses sits a very small number. In rats, well under 1% of a swallowed dose reaches the bloodstream intact, about 0.4%.^[31] Human figures are not established, but blood levels after a dose are also very low.^[7] Studies, mainly in rats, point to three reasons.^{[12, 31]} A pump in the gut wall pushes some berberine back out, the gut wall and liver break much of it down on the first pass, and it dissolves poorly.

That low uptake is the selling point of dihydroberberine (DHB), marketed as the "better absorbed" form. It probably does absorb better, but the evidence is one tiny pilot.^[7] Five healthy young men each tried placebo, 500 mg berberine and two DHB doses, each for about a day. Over the two hours after the last dose, 100 mg of DHB gave about 6.7 times the blood berberine of 500 mg regular berberine. The authors call this preliminary, and blood levels already differed between conditions at the start.^[7]

Glucose and insulin did not differ between any of the conditions in that pilot, placebo included.^[7] That says nothing either way about blood sugar, because the trial was far too short and the men had normal blood sugar to begin with. No trial has yet shown that DHB lowers blood sugar more than plain berberine.

If you and your doctor decide plain berberine is worth a try, three practical points help:

• Split the dose and take it with meals; it is gentler on the stomach.
• Start low. The gut side effects mostly appear in the first four weeks, and in one trial about a quarter of patients needed a lower dose.^[3]
• Plain berberine is the tested form: no trial has shown that branded "enhanced" products work better in your body rather than just in your bloodstream.

HTD1801: a berberine drug, not a supplement

A tested berberine-based product does exist, but it is a drug candidate called HTD1801 (berberine ursodeoxycholate), not something you can buy as a supplement. Its developer ran two 24-week phase 3 trials in China, the large final tests before a drug can be approved.^{[32, 33]} On top of metformin, HTD1801 lowered HbA1c about 0.5 percentage points more than placebo, and on its own about 0.7 points more.^{[32, 33]} Diarrhea was its most common side effect. As of October 2026 it is not an approved medicine, and its results say nothing about supplements or DHB.

Berberine carries two kinds of risk: its own side effects, mostly gut trouble, and interactions that can raise the levels of other drugs.^{[28, 34]} Its long-term safety is still an open question.^[34] If you take medication, the interactions are the most immediate danger, because "natural" says nothing about what a substance does to your other drugs.

In people, berberine slows three liver enzymes, CYP3A4, CYP2D6 and CYP2C9, so drugs cleared by them can build up in the blood.^{[19, 11]} Lab and animal data suggest that drugs carried by the gut pump P-glycoprotein may be affected too.^[28] Higher levels matter most for drugs where small changes are risky, which doctors call a narrow therapeutic range.^[28]

Four drug groups need the most care:

• Cyclosporine (a drug that keeps the immune system from rejecting a transplant): in kidney-transplant patients, berberine raised cyclosporine blood levels by about a third compared with patients not taking it.^[11] Cyclosporine has a narrow safe range, so the transplant team must know and may need to change the dose. Never add berberine on your own.
• Statins (simvastatin, atorvastatin): many are cleared by CYP3A4. A week of berberine had no clear effect on simvastatin levels in a small study of healthy volunteers,^[35] but France's food agency still lists statins among the drugs that interact with berberine, based on lab data.^[28] There are no data in patients or for atorvastatin. Simvastatin's risk of muscle damage (myopathy) rises with dose, which is why drugs that strongly block CYP3A4 must not be combined with it.^[40] If you take a statin, ask before you add berberine, and report unexplained muscle pain.
• Blood-sugar drugs (insulin, sulfonylureas, metformin): stacking berberine on top may push blood sugar too low (hypoglycemia). Pooled trials did not show a clear rise,^[1] but France's food agency concluded that berberine carries a risk of low blood sugar, especially combined with diabetes drugs.^[28] Check with your doctor, above all if you take insulin or sulfonylureas.
• Blood thinners and other drugs cleared by CYP2D6 or CYP2C9: in healthy men, two weeks of berberine slowed both enzymes.^[19] Those two enzymes clear the blood thinner warfarin, many antidepressants, beta blockers and sulfonylureas,^[19] so the levels of these drugs can rise too. Ask your doctor or pharmacist before you combine them.

Berberine's own most common side effect is gut trouble. In a small 2008 trial, about a third of patients had short-lived gut problems, mainly gas, diarrhea and constipation, mostly in the first four weeks.^[3] Most cases that needed a lower dose were in people also taking metformin, acarbose or insulin, and none were severe on berberine alone.^[3] EFSA's review of the trial evidence also found gut symptoms to be the most frequent effect.^[34]

Some groups should not take berberine at all:

• Pregnancy, breastfeeding and newborns. Animal studies suggest berberine may increase uterine contractions, and it may pass into breast milk.^[28] Bilirubin, a yellow pigment from worn-out red blood cells, normally travels through the blood bound to a protein called albumin, and a lab and rat study showed that berberine knocks it loose.^[16] In babies, that raises the risk of kernicterus, a type of brain damage from too much bilirubin. The US National Center for Complementary and Integrative Health advises against berberine-containing goldenseal in pregnancy, while breastfeeding and for infants.^[24]
• Children and teenagers, people with liver disease and people with heart disease. France's food agency advises these groups, together with pregnant and breastfeeding women, not to take berberine supplements.^[28] It also advises people with diabetes against them. If you have diabetes and want to try berberine, do it only with your doctor, who can adjust your medication and watch for low blood sugar.^[28]

If you take any prescription drug, clear berberine with your doctor or pharmacist before you start, not after. That goes double for statins, immune-suppressing drugs, blood thinners and anything for blood sugar. France's food agency goes further: it advises against taking berberine supplements alongside any drug treatment, and against combining several supplements or products with many ingredients.^[28]

EU status (as of October 2026)

As of October 2026, berberine is not formally banned in the EU, but its status is unsettled. The EU food safety authority EFSA could not set a safe intake level for any berberine plant preparation in its 2026 draft opinion.^{[34, 25]} Germany's consumer centres say isolated berberine is most likely an unapproved novel food.^[26] You can still find it online in Germany, Austria and Switzerland, but that may not last.

Two separate EU processes are at work, and for the capsules themselves the novel-food one matters more. Isolated or synthetic berberine, the form in most capsules, falls under the EU novel-food rules.^[26] EFSA's Article 8 review covers 13 berberine-containing plant preparations.^[34] A novel food is a new ingredient that needs EU approval before sale, and an application for berberine has been with EFSA since May 2024.^[26] An extract of barberry fruit, which contains much less berberine, may still be allowed.^[26] Bark and root preparations, in the view of EU authorities, do not belong in food.^[26] Switzerland is looser: bark or root preparations are allowed in supplements there, usually at 500 mg one to three times a day.^[26]

The plant review is the older of the two processes, and it reached its current draft in steps:

• 2019: France's food agency ANSES concluded that berberine supplements can cause gut problems, low blood sugar and low blood pressure.^[28] A member state then asked the European Commission to act.^[34]
• June 2022: the Commission asked EFSA for a safety assessment.^{[34, 25]}
• 5 July 2023: EFSA opened a call for data under Article 8(2) of Regulation (EC) 1925/2006, the procedure that can put a substance added to food on a banned, restricted or watch list.^[37]
• Early 2026: the draft opinion could not set a safe intake for any of the preparations. Lab tests of berberine showed signs of gene damage that studies in living animals have so far neither confirmed nor ruled out, and adequate long-term toxicity studies are missing. Liver tumors in rodents came from goldenseal root, and rare but serious liver injury in people came from greater celandine, where EFSA says berberine's part is unclear.^{[34, 25]}
• 3 March to 11 May 2026: the public consultation on the draft ran.^[36]
• Next: as of October 2026 EFSA had not published a final opinion. After it does, the Commission decides whether to ban berberine plant preparations, restrict them or put them under scrutiny.^{[25, 34]}

Whatever the legal outcome, quality is a separate problem. Supplements are regulated as foods, so nobody checks a product's effect or side effects before sale the way medicines are tested and approved.^[26] France's food agency found that many berberine products do not say how they are made or how pure they are.^[28] Some mix berberine with plant powder without stating the total amount, so what you actually take is uncertain.^[28] Statutory health insurers in Germany do not pay for supplements, because they count as foods, so berberine comes out of your own pocket.^[38]

Verdict: who might consider it

Berberine is a real option for people with type 2 diabetes or metabolic syndrome, where it helps most as an add-on.^{[1, 2, 5]} For them it may be worth a talk with the doctor, alongside current care and not instead of it.^[28] Its interaction risks may mean the medication needs adjusting.^[28]

If you are healthy and hoping for Ozempic-style weight loss, berberine is the wrong tool, because the trials found little to no weight loss. Pregnant or breastfeeding women, children and teenagers, and people with liver or heart disease should skip it entirely.^[28] My view: if blood sugar is your real concern, metformin is the better-tested version of the same idea and worth asking your doctor about first. If you and your doctor do try berberine, get an HbA1c and blood-lipid test beforehand and again afterwards, so you can see whether your numbers moved.