In brief
| What is shown | First on ranking probability in uncomplicated SIBO |
| What is not shown | A direct comparison with rifaximin |
| Studied dose | 400 mg twice daily, two weeks |
| Main risk | It raises the levels of other drugs |
| Who must not take it | Pregnancy, breastfeeding, newborns |
In our guide to SIBO berberine appears in a single line — as the intervention that topped a network meta-analysis. We were asked to open that line up, and it has to be opened honestly: the headline «berberine beat the antibiotic» is true right up to the first caveat, and becomes a lie without it.
So: what is actually shown, what is still missing, what the dose is in milligrams, and why safety matters more here than efficacy.
Where the «first place» comes from
The source is a 2025 network meta-analysis (PMID 41394885[1]): 30 randomised trials, 1552 participants, 12 different interventions, with the protocol registered in PROSPERO. It gathers trials that never overlapped with each other into a single network and arranges the interventions into a ranking.
Berberine received the highest SUCRA value in that ranking. The authors write that this positions it as a potentially favourable option for SIBO eradication.
And now the three caveats, without which this is a lie
First: first place is not a win in a direct comparison. Berberine never met rifaximin face to face in those trials. A network ranking and the result of an actual head-to-head are different things, and they diverge regularly.
Second: first place only in uncomplicated SIBO. As soon as a comorbidity appears the ranking rearranges itself — and berberine gives way.
| Who is in front of you | What topped the ranking | SUCRA |
|---|---|---|
| Uncomplicated SIBO | Berberine alone | highest |
| Plus a functional GI disorder | Rifaximin + prokinetic | 89% |
| Plus chronic liver disease | Prokinetic alone | 79.6% |
Third: the authors themselves ask you to read this as a hypothesis. Verbatim — the results should be interpreted as generating hypotheses for future validation in well-controlled, direct-comparison studies. That is not politeness: three of the included trials carried a high risk of bias, though meta-regression indicated this did not significantly influence the outcome.
What is still missing — and it is the most interesting part
The head-to-head is already under way. The BRIEF-SIBO trial (PMID 36873985[2]): 180 patients, berberine against rifaximin, 400 mg twice daily for two weeks, six weeks of follow-up, primary outcome a negative breath test. The stated hypothesis is that berberine is not inferior to rifaximin.
The protocol was published in 2023. As of September 2026 there are no results — checked in PubMed by the trial name and by the two drugs together. The trial is single-centre, and that is worth holding in mind in advance: a single-centre result opens a conversation rather than closing it.
The honest frame, then: there is a first place, there is no direct evidence yet, and the work that will supply it is already running. There is nothing shameful in that — it is what ordinary science looks like halfway through. What is shameful is selling the halfway point as the finish line.
What it is in milligrams
The most practical part hides here, and it is almost never printed large on the label.
The dose studied in SIBO is 800 mg of berberine a day. In metabolic research the usual range is 0.9–1.5 g a day for one to three months.
Now look at what a supplement actually contains. Very often it is not berberine but barberry extract standardised to 2% berberine. Which means a «500 mg» capsule holds roughly 10 mg of the active compound. And the label usually reads «up to 2%» — a ceiling rather than a promise: the real figure may be lower.
The practical rule in one line: look at the milligrams of berberine, not the milligrams of extract. If the package gives only the weight of the extract and the percentage of standardisation, multiply them.
What berberine does beyond the gut
The second reason to be interested is metabolism, and the evidence base there is considerably larger. A 2024 meta-analysis pooled 50 trials and 4150 participants (PMID 39640489[3]).
| What was measured | Berberine alone | With glucose-lowering drugs |
|---|---|---|
| Fasting glucose | −0.59 mmol/L (p = 0.048) | −0.99 mmol/L |
| 2-hour glucose | −1.57 mmol/L | −1.07 mmol/L |
| HbA1c | — | −0.69% |
| Triglycerides | −0.35 mmol/L | — |
Note the p = 0.048 for fasting glucose with berberine alone. That is significance which barely cleared the threshold: the effect exists, but it is modest and sits on the edge.
And one more detail worth reading carefully in the table. HbA1c does not appear at all among the significant changes for berberine alone — it shows up only in combination with glucose-lowering drugs. And HbA1c is precisely the measure of what happens to blood sugar across months, unlike a single fasting reading. Hence the honest formulation: berberine on its own nudges the numbers of one day; the months are governed by prescribed treatment, which it may assist.
The detailed comparison with metformin is covered separately — berberine or metformin, on mechanisms, doses and how the side-effect profiles differ.
Safety: here it matters more than efficacy
Berberine is sold as a supplement, and that creates a sense that there is nothing serious to discuss. That is a mistake. It has at least three properties that in a pharmacy medicine would land in the contraindications section.
1. It raises the levels of other drugs
The weightiest evidence is not laboratory work but people. A randomised trial in renal transplant recipients (PMID 16133554[4]): 52 patients took cyclosporine together with berberine, 52 without it, for three months.
| Cyclosporine measure | Change with berberine |
|---|---|
| Area under the curve (AUC) | +34.5% |
| Trough concentration | +88.3% |
| Half-life | +2.7 hours |
| Clearance | −40.4% |
For a drug with a narrow therapeutic window that is a great deal. In animal work (PMID 19370549[5]) berberine likewise increased digoxin absorption — the area under the curve reached 170% of control, and only with oral dosing, meaning the effect is intestinal.
Hence the rule: if you take immunosuppressants, cardiac glycosides, anticoagulants or anticonvulsants, berberine is discussed with a physician, not added alongside them.
2. Pregnancy, breastfeeding and newborns
Berberine displaces bilirubin from its binding to albumin. In a laboratory study (PMID 8513024[6]) it proved, on a molar basis, about tenfold more potent than phenylbutazone — a known displacer of bilirubin.
For an adult that carries little weight. For a newborn it is a risk of free bilirubin reaching the brain. The US National Institutes of Health lactation database puts it plainly: most sources recommend avoiding newborn exposure to berberine, including through breast milk — and honestly adds that the extent of transfer from mother to infant is unknown.
3. The ordinary side effects
A 2023 umbrella review pooling eleven meta-analyses (PMID 36999891[7]) names constipation and diarrhoea as the common gastrointestinal effects. Overall it rates berberine as a safe ingredient — with the caveat that the methodological quality of the meta-analyses themselves needs improvement.
The bottom line
Berberine is a compound with real pharmacology, not a harmless herb. That is exactly why it deserves to be treated like a medicine: check the dose, check the interactions, check the contraindications.
On SIBO the honest summary is this: it topped the ranking, but a ranking is not yet a comparison. The direct comparison is running right now, and it is reasonable to wait for its results rather than buy the conclusion in advance.
And the thing worth taking away even if the rest is forgotten: with bloating, what decides the outcome is not the choice of agent but the order of steps — check first, then treat, and treat the person rather than the test.
This material is for information only. Berberine alters the blood levels of other medicines and is contraindicated in pregnancy and breastfeeding. Decisions about taking it, its dose and its combination with prescribed treatment are made by a physician.
References
Key facts
- In a 2025 network meta-analysis (30 RCTs, 1552 participants, 12 interventions) berberine had the highest SUCRA in uncomplicated SIBO.
- SUCRA is the probability of ranking higher within a network, not a win in a direct comparison. Berberine was never tested head-to-head against rifaximin in those trials.
- The subgroups diverge: with functional gastrointestinal disorders, rifaximin plus a prokinetic leads (89%); with chronic liver disease, a prokinetic alone (79.6%).
- A head-to-head trial is running: BRIEF-SIBO, 180 patients, berberine 400 mg twice daily against rifaximin. The protocol was published in 2023; as of September 2026 there are no results.
- The studied dose is 800 mg of berberine a day. Supplements often contain an extract standardised to 2% berberine: 500 mg of such an extract gives about 10 mg of the active compound.
- Berberine raises cyclosporine levels in humans: AUC +34.5%, trough concentration +88.3%. That is a randomised trial in 104 patients, not theory.
