Remedies › Berberine

Berberine (berberine hydrochloride)

Also called berberine hydrochloride, berberine HCl, berberine chloride, dihydroberberine, goldenseal / barberry / Coptis (huang lian) extracts, nature's Ozempic.

Berberine is a bitter yellow compound from plants such as barberry, goldenseal, Oregon grape and Chinese goldthread (Coptis, huang lian), long used in Chinese medicine for diarrhoea. It is now sold as a supplement for blood sugar, cholesterol and weight, and social media has nicknamed it 'nature's Ozempic'. It is also being developed as a prescription drug in a modified form (HTD1801).

What it has been studied for

Beyond Google: what a search won’t show you

Places a normal search rarely reaches: trial registries, drug-code names, FDA report databases, patent offices and grant records. Each item links to its primary record.

Drug under another name
HTD1801 / berberine ursodeoxycholate: berberine as a prescription-drug candidate
phase 2 2020-2023, phase 3 completed 2025, papers 2026
A salt of berberine and the bile acid ursodeoxycholic acid, developed by HighTide Therapeutics. Searches for berberine rarely surface it. Two phase 3 trials led from Peking University lowered HbA1c by 0.7 points (alone, 407 people) and 0.5 points (with metformin, 549 people) more than placebo over 24 weeks, with diarrhoea in up to 23.8%. This is the strongest berberine-based evidence anywhere, and it is a drug, not a capsule from a shop.
Completed trial, no results yet
HTD1801 head-to-head against dapagliflozin (phase 3, 367 people)
completed July 2025
The only trial we found comparing a berberine-based drug with a modern diabetes drug. It finished in July 2025; as of 28 October 2026 the registry shows no results and we found no paper.
Unpublished trial
HTD1801 for fatty liver with diabetes or prediabetes (phase 2b, 218 people)
completed March 2025
Completed 19 months ago; no results posted and no paper found. The earlier phase 2 trial (NCT03656744) posted results: liver fat fell 4.8 points vs 2.0 on placebo at the higher dose (P = 0.011), but not at the lower dose (P = 0.199). A 2026 letter from a Cornell statistician questioned the company's later claims of liver-tissue improvement.
Unpublished trial
Does berberine raise GLP-1 in people? (Beijing Tongren Hospital, 16 healthy men)
completed April 2023
The one registered human test of the mechanism behind the 'nature's Ozempic' claim: GLP-1 and GIP hormone levels on berberine vs placebo during a glucose test. No results were posted and we found no paper. Whether it showed an effect is unknown.
Unpublished trial
Berberine to protect the kidneys of people with diabetes during X-ray dye scans (800 patients, Nanjing)
completed December 2021
A large randomized trial in people with diabetes and kidney impairment. Nearly five years after completion there are no posted results and PubMed holds only animal studies on the question (one of them retracted).
Trial in progress
Berberine to prevent diabetes and heart disease (2,024 people at high risk, China National Center for Cardiovascular Diseases)
results due after December 2026
Fully enrolled, placebo-controlled, measuring blood-sugar status after one year. It will be the largest berberine trial ever reported and may settle whether berberine prevents diabetes.
Stalled trial
5,200-person prevention trial in metabolic syndrome
registered November 2021, never updated
Planned to test 500 mg twice daily against placebo for heart attacks, strokes and diabetes. Still listed as not yet recruiting, five years on.
FDA adverse-event reports
Reports to the FDA naming a berberine product
2008-2026
68 reports name a berberine product; 25 of them were filed in 2024-2025 as the 'nature's Ozempic' trend grew. 28 involved a hospital stay and 7 were called life-threatening; in 24 berberine was the only product listed. Problem groups: stomach and bowel (28), allergy or skin (14), heart rhythm or fainting (12, including one QT prolongation with berberine alone and one cardiac arrest with a dihydroberberine product), liver injury or jaundice (6), blood pressure (5) and blood sugar (4, including one hospital stay for low blood sugar with berberine alone). Two reports may describe the same person. A report is not proof that berberine caused anything.
Patent
Berberine patents filed by the scientists behind the 2004 cholesterol study
2004-2011
Inventors from the Institute of Medicinal Biotechnology in Beijing, including authors of the 2004 Nature Medicine study, filed 'Methods and compositions for the treatment of hyperlipidemia' in at least seven patent offices (WO, US, EP, CA, NZ, KR, MX) and 'The use of berberine as insulin sensitizer' (WO2004032924). A patent is a legal claim, not proof; it also shows the early researchers had a commercial stake.
Patent
US 2011/0281852: berberine for antipsychotic weight gain
2011
Filed by G. Davies and Y. Hu; Hu also led the 2012 uncontrolled pilot study reporting about 5 lb of weight loss in obese adults. Trials of berberine for antipsychotic weight gain later came out positive in Hong Kong and China, but from different teams.
Reformulated product
Dihydroberberine: the 'better-absorbed' version
2021-2026
Sold as a stronger form of berberine. Its only published human test measured blood levels in 5 men (PMID 35010998); trials of its effects are being run by a manufacturer (NCT07210684, NCT07322679). A 2026 Mayo Clinic case report of torsades de pointes involved a man taking berberine and dihydroberberine together (PMID 42684237).
NIH grants
Who the US government has paid to study berberine
2005-2026
NIH RePORTER lists 111 berberine-related projects (our pull covered 100 of them). The metabolic ones were 12 awards to Jingwen Liu at the Palo Alto VA research institute (about $3.78 million, 2005-2016, cholesterol mechanisms) and a VA project on fatty liver (2018-2021). We found no current NIH-funded human trial of berberine for blood sugar, cholesterol or weight; the largest recent berberine awards (about $0.9 million a year in 2023-2024) went to a company developing a berberine nanomedicine for food allergy.

Sources searched on 2026-10-28: ClinicalTrials.gov, Drug codes (HTD1801, BUDCA, dihydroberberine), Europe PMC preprints, Europe PMC patents, FDA CAERS (openFDA), NIH RePORTER. Where we look

Safety notes

The most common side effects in trials are stomach and bowel problems: constipation, diarrhoea, gas and stomach pain. In a placebo-controlled review they were reported by 2-23% of people on berberine against 2-15% on placebo. Source
Taking berberine with diabetes medicines can add to their effect and push blood sugar too low; low blood sugar has also been reported in a non-diabetic athlete taking it. Do not add it to, or swap it for, a prescription without your doctor. Source
Berberine slows liver enzymes that clear many drugs. In healthy volunteers, two weeks of 300 mg three times daily reduced CYP2D6, CYP2C9 and CYP3A4 activity, raising levels of a test drug by about 40%. Source
In kidney-transplant patients, berberine raised levels of the anti-rejection drug cyclosporine (trough 29.3% higher than without berberine). Anyone on cyclosporine, tacrolimus or other narrow-margin drugs should not start it without their specialist. Source
Three case reports from 2026 describe dangerous heart-rhythm problems (QT prolongation and torsades de pointes) in people taking berberine supplements, one aged 36, one 61 and one 92; the rhythm settled after stopping. Source
In lab and rat studies berberine pushes bilirubin off its carrier protein, which could worsen newborn jaundice; the 1993 authors advise avoiding berberine-rich herbs in pregnant women and jaundiced newborns. Source
Singapore banned two berberine-containing herbs for about three decades over fears of jaundice and brain damage in newborns with G6PD deficiency. Later human studies found little sign of harm, so the evidence conflicts; we found no human safety studies in pregnancy. Source
The drug form HTD1801 caused diarrhoea in 23.8% of people in a 24-week trial, against 1.1% on placebo. Source
'Natural GLP-1' skin patches that list berberine, a spin-off of the 'nature's Ozempic' trend, are illegal under US supplement law because supplements must be swallowed; a 2026 review of 25 such products found no certificates of analysis and frequent deceptive advertising. Source

Possible links

Where one finding points to others. These are leads for research, not conclusions.

Other conditions

Fatty liver (NAFLD / MASLD) 226 papers Several randomized trials and a 2024 meta-analysis of 10 trials (811 patients) found lower liver enzymes and lipids; the drug form HTD1801 is in trials for fatty liver. See on PubMed
Polycystic ovary syndrome (PCOS) 77 papers Not yet read by us; includes trials against metformin and in IVF. See on PubMed
Antipsychotic weight gain 14 papers Three randomized trials pooled in a 2026 meta-analysis; listed on the weight page. See on PubMed
H. pylori stomach infection 46 papers Not yet read by us; berberine is used in Chinese quadruple-therapy regimens. See on PubMed

Same mechanism

AMPK, the cell's energy switch (same pathway as metformin) 435 papers The most studied explanation for the blood-sugar effect, mostly from lab and animal work. Only 10 PubMed records on berberine and metformin are randomized trials or meta-analyses. See on PubMed
More LDL receptors in the liver (different from statins) 56 papers The 2004 study that launched berberine as a cholesterol treatment showed it stabilizes LDL-receptor messages; later work points to lower PCSK9. See on PubMed Source
Gut bacteria 397 papers The 409-person PREMOTE trial linked berberine's glucose effect to changes in gut bacteria and bile acids. See on PubMed Source
GLP-1 hormone release (the 'Ozempic' claim) Shown in lab and animal studies; the only registered human test (NCT05947370) has no published result. See on PubMed Source

Combinations

Berberine with red yeast rice (Armolipid Plus and similar) 88 papers Red yeast rice contains monacolin K, the same molecule as lovastatin, so these products' cholesterol results cannot be credited to berberine. See on PubMed
Berberine with silymarin (milk thistle) 57 papers A meta-analysis of 5 trials (497 people) found lower LDL and glucose; silymarin may increase berberine absorption. See on PubMed
Berberine with probiotics 106 papers Two large Chinese trials (409 and 297 people) tested the combination for blood sugar. See on PubMed

Who does the research

J.-D. Jiang and W.-J. Kong 13 papers Institute of Medicinal Biotechnology, Chinese Academy of Medical Sciences, Beijing (13 and 11 papers). Kong was first author of the 2004 cholesterol study; both are inventors on berberine patents for high cholesterol.
A. F. G. Cicero 13 papers University of Bologna; many trials and meta-analyses of berberine combination products.
F. Di Pierro 12 papers Velleja Research, Milan, a company; trials of berberine products and inventor on a berberine composition patent.
G. Derosa 8 papers University of Pavia; placebo-controlled lipid and glucose trials in Italy.
J. V. Zhao 4 papers University of Hong Kong; the placebo-only meta-analyses and the sex-specific trial in men.

Who pays for it

National Natural Science Foundation of China 70 papers
US National Institute of Diabetes and Digestive and Kidney Diseases (NIH) 21 papers
US National Center for Complementary and Integrative Health (NIH) 14 papers

Counts are papers across our three PubMed lists (2,005 records, mostly lab and animal work) that list the funder. Most human trials list no funder in PubMed. The HTD1801 trials were funded by the company developing the drug, and several European combination-product trials were run by product makers; we found no berberine-only supplement trial declaring supplement-industry funding, which does not mean none exists.

What people report

We read public first-person posts and counted every outcome, good or bad. 40 reports, read on 2026-10-28; 13 mention side effects.

18Helped
7No effect
6Got worse
4Mixed
5Too early

Why these counts prove nothing on their own: people who post are not a random sample, most take other treatments at the same time, and nobody checks the diagnosis. We show them because they tell researchers what to test next.

Read all 40 reports

Helped Within about a week, after-meal readings were near 131 instead of 180-210 and morning readings fell from the 140s to the 120s. Diabetes.co.uk forum, 2023-12-16 (Only one week in; started cinnamon at the same time; planned to check with doctor because of heart medicines.)
Worse After two weeks developed severe pain across the upper abdomen that stopped within 48 hours of quitting; a later report mentions stomach ache again. Side effects: severe upper abdominal pain. Diabetes.co.uk forum, 2023-12-16 (Poster attributes the pain to berberine because nothing else changed; no glucose result given.)
Mixed A first brand made little difference and an HbA1c check was only slightly better; a later refrigerated brand seemed to bring readings into the 4s. Diabetes.co.uk forum, 2024-11-23 (Already in non-diabetic range; poster calls the first try inconclusive and possibly natural progress.)
Helped After four months with bitter melon, halved diabetes medicine and was still having low blood sugar episodes. Side effects: low blood sugar episodes (hypos) while on gliclazide. Diabetes.co.uk forum, 2024-12-30 (Two supplements started together; the hypos happened alongside a sulfonylurea, the combination doctors warn about.)
No effect Took it with every meal but could not tell how much it helped and it did not fix morning highs; metformin later worked far better. Diabetes.co.uk forum, 2024-08-12 (Poster judges berberine at most a small help; no numbers for the berberine period.)
Worse Stopped after one week at 1,500 mg a day because of intestinal pain; fasting glucose was unchanged; later attempts caused the same pain. Side effects: intestinal pain. Diabetes.co.uk forum, 2023-10-18 (Only one week of glucose data; an earlier attempt was interrupted by diverticulitis.)
No effect Switched from metformin because of side effects; berberine had no side effects but, like metformin, barely moved fasting glucose. Diabetes.co.uk forum, 2024-08-13 (Poster says HbA1c is hard to attribute because they use several measures at once.)
Helped Takes it twice a day and thinks it helps; most recent HbA1c was 39 mmol/mol, in the normal range. Diabetes.co.uk forum, 2024-08-22 (No before value or duration given.)
Worse Fine for a day, then severe stomach pain and diarrhoea worse than on metformin, so stopped; now controls glucose with diet. Side effects: severe diarrhoea, stomach pain. Diabetes.co.uk forum, 2024-05-02 (Short trial; no glucose result.)
No effect Took two capsules a day for a while; any lowering of glucose was marginal and not worth it next to free metformin. Diabetes.co.uk forum, 2024-05-17 (Poster says it may have helped slightly.)
Mixed Berberine alone did not work; after adding black seed oil, walking and fewer carbs, glucose became responsive. Quit twice early on from gut discomfort. Side effects: intestinal discomfort at the start. Diabetes.co.uk forum, 2023-09-26 (Several changes at once; no after value for HbA1c.)
Worse After a couple of months it caused persistent diarrhoea, worse than standard metformin; will not try again. Side effects: chronic diarrhoea. Diabetes.co.uk forum, 2023-10-19 (No glucose result given.)
Helped Cannot tolerate metformin, so takes 500 mg twice daily instead; says cholesterol is better and has had no side effects. Diabetes.co.uk forum, 2023-10-19 (No numbers; glucose effect not described.)
Helped Used it for about a year and thinks it helped blood sugar; stopped over worries about long-term effects on gut bacteria. Diabetes Daily forum, 2025-07-18 (No numbers.)
Mixed In 2020 said cholesterol labs improved on twice-daily berberine; in 2025 said a year of it made little or no difference. Diabetes Daily forum, 2025-07-18 (Two posts five years apart; the 2025 post does not say which measure.)
No effect Took three different brands for months at a time and saw no improvement in blood glucose or anything else. Diabetes Daily forum, 2025-07-22 (Brands not tested for content.)
Helped About a week in, cut metformin to 500 mg and readings stayed in the 80s and low 90s; later said it beats metformin for them. Diabetes Daily forum, 2020-07-25 (Short time frame; another member notes the poster had lost weight and was mildly diabetic.)
Helped Could not take metformin; has taken 1,500 mg a day for years with no side effects and sees higher readings when a dose is missed. Diabetes Daily forum, 2019-11-26 (Posted across several threads 2016-2020; no lab values.)
Mixed Took it from October 2019; it worked somewhat but readings never fell below 195, and it stopped working after a few months. Diabetes Daily forum, 2020-07-29 (No baseline given.)
Helped Average fasting glucose fell from 144 to 106 within about two weeks; a year later, on metformin too, still felt it helped. Diabetes Daily forum, 2020-07-30 (Cinnamon started at the same time; poster says they cannot prove it.)
Helped With 500 mg three times daily and a diet change, HbA1c fell from 8.1% to 6.3% in three months, then 5.6%; lost 25 lb. Diabetes Daily forum, 2015-09-11 (Diet changed at the same time; another member questioned whether this new account was promoting a brand.)
No effect Took a generic brand for three months and saw no change in blood glucose. Diabetes Daily forum, 2016-01-04 (Brand not tested; later planned to try another brand.)
Helped Has gone on and off it and says it clearly cuts morning glucose release and improves after-meal readings after a few weeks. Diabetes Daily forum, 2016-01-16 (No numbers; long-time advocate in that forum.)
Too early A few days in, a fasting reading was 86 and after-meal numbers looked lower than before. Diabetes Daily forum, 2016-01-19 (Only days of use; vinegar taken too.)
Helped Replaced metformin with berberine plus a ketogenic diet; reports good glucose numbers and visible weight loss. Diabetes Daily forum, 2018-04-07 (Diet change is a large confounder; no lab values.)
Helped A multi-ingredient product containing berberine lowered readings by about 30-40 points on days taken; weight did not change. Diabetes Daily forum, 2019-11-26 (Product is a blend, so berberine's share is unknown.)
No effect Lost 4 lb in week one, then nothing; about eighteen months later said it had not worked much and she gave up. Side effects: hunger and rumbling stomach in first days. Mumsnet, 2025-01-18 (Early loss may be normal fluctuation, as she noted.)
Helped Taking 500 mg a day and lost 16 lb over about two months. Mumsnet, 2023-06-08 (Poster stresses a strict low-carb diet as well.)
Too early About two weeks in, losing weight on a health drive and feels berberine is curbing appetite. Side effects: mild constipation (poster blames diet). Mumsnet, 2023-06-22 (No amounts; had read it can take three months.)
Too early Ten days in, lost 4 lb and found appetite lower, while also fasting and eating fewer carbs. Side effects: more frequent bowel movements. Mumsnet, 2023-06-23 (Poster says diet may explain the appetite change.)
Helped Took it with myo-inositol for over a year with diet and exercise; lost a stone and a half with no side effects. Mumsnet, 2023-06-22 (Several changes together; poster credits the combination.)
Helped Lost 7.6 lb in three weeks and felt appetite suppressed, while also taking a fibre supplement and fasting 18 hours a day. Side effects: gas and bloating in week 3. Mumsnet, 2024-04-20 (Only three weeks; two other weight-loss measures at once.)
Too early Took it for one day, was very gassy, and did not take it again. Side effects: gas. Mumsnet, 2022-10-25 (One day only.)
Too early Nearly a month in with no ill effects once taken with food; had not weighed yet. Side effects: gas when taken before meals. Mumsnet, 2022-10-30 (No outcome measured.)
Worse Over a 3.5-month trial, triglycerides rose 15%, LDL rose 2 points and HDL fell 10%. QVC Community, 2024-09 (Diet changed at the same time; small LDL change may be noise; thread posts dated September 2024.)
No effect Took berberine with red yeast rice and other supplements for a long stretch on a cardiologist's advice with no improvement. QVC Community, 2024-09 (Familial high cholesterol; now on a prescription injection that works.)
Worse At 1,100 mg a day, normal blood pressure dropped too far; two weeks in had spinning vertigo that took two weeks to clear after stopping. Side effects: blood pressure dropped too low, vertigo lasting two weeks. City-Data forum, 2024-11-11 (Poster still believes it works; no glucose or weight result.)
Helped Takes a low dose with other herbs; says it boosts metformin so less is needed, and manages diabetes with herbs, diet and exercise. City-Data forum, 2019-07-07 (Several herbs at once; no numbers.)
Helped Replaced metformin with 1,200 mg a day plus diet; HbA1c 5.6%, the same as on metformin; also stopped a blood-pressure pill. City-Data forum, 2021-11-01 (Single HbA1c value; stopping a blood-pressure medicine was the poster's own choice.)
Helped With diet, exercise, stress and sleep changes plus berberine, HbA1c fell from 14 to 8 in a year; metformin and glipizide had not suited. City-Data forum, 2022-01-28 (Many changes at once; berberine's share unknown.)

Checked against PubMed on 2026-10-28.