Berberine and your medications: what interacts, and what to avoid entirely.

It has a real, drug-like effect on blood sugar. That is the reason to take it seriously, and the reason it belongs on the list you hand your GP.

Berberine has spent the last few years being sold as nature's Ozempic. It is not that, and the comparison manages to flatter and insult it at the same time. What berberine actually is, on the evidence, is a supplement with a genuine, medicine-like effect on blood sugar. That is the reason to take it seriously. It is also the reason it collides with prescriptions unusually often for a supplement, and why the question people usually ask about it, whether berberine is safe, has no useful answer on its own. Safe alongside what?

Does berberine interact with medications?

Yes. Keep it away from two groups of medicines, transplant immunosuppressants and antiretroviral therapy for HIV. With the other medicines it interacts with, the answer is to monitor and tell your GP. Nearly all of it comes back to two mechanisms.

The first is the one it is sold for. Berberine lowers blood glucose in its own right, and a meta-analysis of 37 randomised trials found, across the 31 of them that measured it, that berberine lowered HbA1c, a measure of your average blood sugar over roughly the last three months, by 0.63 percentage points (PMID 36467075). So combining it with a medicine that also lowers glucose can take you lower than either was aimed at, which calls for close glucose monitoring in the first few weeks. The trials in that review most often used between 0.9 and 2.4 grams a day (PMID 36467075). Berberine is conventionally split across meals, and a typical shelf product sits at the lower end of that range.

The second is the one almost nobody mentions. Berberine appears to slow two of the body's clearance routes: CYP3A4, a liver enzyme, and P-glycoprotein, a pump in the gut wall that pushes drugs back out before they are fully absorbed. The enzyme half has been measured directly in people: two weeks of berberine raised exposure to a test drug cleared by CYP3A4 by about 40% in a randomised crossover trial in healthy men (PMID 21870106). The pump half has been shown in rats; the one test in people used goldenseal, a herb that contains berberine, at a much lower dose, and found little effect. Between them those two routes handle a long list of prescriptions, so slowing them tends to leave more of a drug in the blood than the dose intended. For most medicines that is a monitoring question, but for a few, where the gap between the working dose and the harmful one is narrow, it is not.

We have assessed berberine against every medicine our database held when we checked it. As of September 2026, twenty-four of those pairs carry a caution: four red, twenty amber. The rest came back clear.

That is more cautions than most supplements we assess collect, and it is not evidence that berberine is uniquely dangerous. It is what you would expect from a compound with a real pharmacological effect.

Can you take berberine with metformin?

Berberine and metformin both lower blood glucose, so taking them together adds one effect to the other and can push your glucose too low, which makes this a monitor-and-tell-your-GP combination rather than a flat no. The effect on blood sugar itself is well tested. A placebo-controlled trial in 116 people with type 2 diabetes and raised blood lipids found berberine lowered fasting glucose, blood sugar after a glucose drink, HbA1c and LDL cholesterol over three months (PMID 18397984).

How careful you need to be depends on which diabetes medicine you take, and the map is more useful here than a single warning. The steepest combinations are with the drugs that can cause low blood sugar on their own: gliclazide, glimepiride and the other sulfonylureas, and insulin. Berberine adds to a risk those already carry, so if you take one of them, check your glucose closely when you start berberine and expect your prescriber to want to review the dose.

The lower-risk combinations are with sitagliptin and the other gliptins, dapagliflozin and the SGLT2 drugs, and pioglitazone. These rarely tip into hypoglycaemia by themselves, so the combination tends to be well tolerated. Lower-risk is not no-risk, though, and checking your glucose at the start and telling your GP you have added berberine is still the sensible step.

One honest qualification, because it cuts against the warning above. The 37-trial review looked for hypoglycaemia across nine of its trials, 820 people in total, covering berberine on its own and berberine added to standard diabetes tablets. In seven of those nine, nobody in either group had a single episode. So the comparison rests on the two trials that recorded any, 186 people between them, and across those berberine showed no significant rise (PMID 36467075). The caution above is about where the risk concentrates, in the first weeks and with the drugs that can drop your sugar on their own, and it is not a claim that low blood sugar is the usual outcome.

One rule holds across all of them: do not change your own dose. If berberine is working, whoever manages your diabetes may need to reduce your prescription, and that is their decision, made on your glucose readings.

Which medications should you never take with berberine?

Transplant immunosuppressants. Berberine slows CYP3A4, the main enzyme the body uses to clear ciclosporin, tacrolimus, sirolimus and everolimus, so it can push those drug levels toward the range where they cause harm instead of protecting the graft, and here the answer is do not combine outside your transplant team's direct supervision.

The evidence here is stronger than most supplement warnings and thinner than it first looks, and both halves deserve saying. A randomised controlled trial in 104 kidney-transplant recipients gave berberine to half of them alongside ciclosporin for three months, and their ciclosporin trough levels, the lowest level between doses and the one clinics usually measure, finished about 29% higher than in the group who took ciclosporin alone. In a smaller group of six, measured before and after, total exposure rose by about 35% (PMID 16133554). That is a trial in patients, with numbers attached, and it is the strongest interaction evidence berberine has.

Ciclosporin is also the only one of the four tested in a trial. Tacrolimus has a single case report, in a young patient treated for a kidney condition rather than a transplant (Hou 2013, PMID 23765408). Sirolimus has one rat study (Yang 2025, PMID 40396352), and everolimus has no direct data. So those three are graded red largely by inference, because they travel through the same enzyme and the same pump. All four are drugs whose blood levels are monitored precisely because the margin between too little and too much is thin. We could wait for more trials. Given what is at stake for a transplant if the inference is right, we would rather grade them red now and tell you openly that for three of the four the case rests mostly on the mechanism.

Antiretroviral therapy for HIV is the other group to keep berberine away from, for a slightly different reason. Much of it is cleared by CYP3A4 too, and unlike transplant drug levels, antiretroviral levels are not routinely measured, so a change in them could go unnoticed. That is why our report leaves berberine out for anyone on HIV treatment, and why we would leave the decision with your HIV team. Our checker answers a narrower question, what the evidence shows about each pair, so it grades these pairs differently, and it is worth saying why. It grades the HIV protease inhibitors, such as darunavir, amber. What berberine does to liver enzymes in people points to their levels running higher than intended, which risks side effects rather than the virus becoming resistant, and protease inhibitors are usually taken with a booster that already blocks the same enzyme. Efavirenz is amber too, on less evidence. Newer medicines, including dolutegravir and bictegravir, show as not yet assessed. Do not stop or change your HIV medicine yourself.

What about statins, blood pressure tablets, digoxin and warfarin?

Four combinations people rarely think to check. None calls for stopping on the spot, and each is worth raising with whoever prescribes the medicine.

Statins. In the 32-patient study that first identified it as a cholesterol-lowering drug, berberine lowered LDL cholesterol by about 25% over three months, by a route those researchers called distinct from a statin's (PMID 15531889): it increases the number of receptors that pull cholesterol out of the blood, where a statin slows the making of it. Taken together the two effects add up, which can take your lipids lower than your GP was aiming for, so ask for a lipid panel sooner than your usual check, and do not stop either on your own. We grade this pair amber on limited evidence: the extra lowering from combining the two has only been seen in small trials (Lan 2015, PMID 25498346).

Blood pressure tablets. Berberine appears to lower blood pressure modestly. A meta-analysis that pooled the two trials adding it to a blood pressure medicine found the top number fell by about 5 points more than on the medicine alone (PMID 25498346). That is thinner than it sounds: blood pressure was one of three conditions the review covered, and only four of its twenty-seven trials were in people with high blood pressure. Alongside amlodipine, ramipril, losartan or their relatives, berberine can occasionally leave you light-headed or running lower than intended, so check your blood pressure in the first few weeks and tell your GP you have started. The effect is small and the underlying trials are not strong, which is why this is a monitoring note and not a warning.

Digoxin. Digoxin is a drug where a small shift in blood level matters. Berberine's effect on the gut pump that limits how much of it gets absorbed may raise that level. If you take both, it is worth knowing the signs of too much digoxin, which are nausea, disturbed vision and an irregular pulse, and worth asking your GP whether a level check is warranted. The evidence behind that one deserves saying plainly, because it is thinner than the others. In rats, berberine raised blood levels of swallowed digoxin by 23 to 70% and left injected digoxin untouched, which is what pump inhibition in the gut wall looks like (PMID 19370549). No study has measured berberine itself with digoxin in people. The closest, in healthy volunteers taking goldenseal, found only a small rise in peak digoxin level and no significant change in overall exposure, but that product supplied well under a tenth of a typical berberine dose, so it cannot settle the question (Gurley 2007, PMID 17079360). Because the evidence is this indirect, we grade the pair amber rather than red.

Warfarin. Nobody has tested berberine and warfarin together in people. The caution comes from the same healthy-volunteer study as above, where two weeks of berberine also slowed CYP2C9, the enzyme that clears the more powerful half of warfarin (Guo 2012, PMID 21870106). That could raise your INR, the clotting test your warfarin dose is set by, and with it your risk of bleeding. A systematic review that included eight studies of berberine and goldenseal described their effects on liver enzymes as weak and did not list CYP2C9 among the enzymes affected (Hermann 2012, PMID 22855269). That leaves the human evidence at one study in healthy volunteers, which is why we grade the pair amber on theoretical evidence. If you take warfarin, tell whoever manages it before you start berberine, keep to your usual INR checks, and do not change your warfarin dose yourself.

Is berberine safe with everything else?

Mostly yes, and the word mostly matters: the pairs that carry a caution are the ones set out above, and we can say the rest with more confidence for berberine than for most compounds because it is one we have taken all the way through. It has been assessed against all 296 medications our database held when we assessed it, the obscure ones included; a medicine added since then shows on our checker as not yet assessed until it has been checked. The pairs that produced nothing show green on our checker because somebody looked and found nothing, rather than because nobody looked.

That distinction matters more than it looks. "No known interaction" and "nobody has checked" read identically on almost every website, and they are not the same sentence. If your prescription is not on the list above, the useful thing is not our reassurance in an essay; it is checking the specific pair. One exception is worth stating plainly: if you take antiretroviral therapy for HIV, treat berberine as off the table unless your HIV team tells you otherwise.

One last thing, which is not an interaction but belongs in an honest piece about berberine. You may see it sold for longevity, on the grounds that it activates AMPK, an energy-sensing pathway that comes up often in ageing research. We removed longevity from berberine's entry in our own database, because a pathway is a mechanism and not a result, and not one of the studies we cite measures ageing, healthspan or how long anybody lived. The blood-sugar evidence is good. The ageing claim is borrowed from a diagram.

If you want to check berberine against your own prescriptions, our free interactions checker gives you the graded answer with the evidence behind it, and the berberine page lists every pair that carries a caution, each one graded and cited. Neither costs anything. We do not sell supplements and we earn no commission on anything you buy, so we have no stake in which way the answer comes out.

Berberine behaves like a medicine, which is the reason it works and the reason it collides. Tell your GP you are taking it.