Which herbs should not be combined with medications? A guide
St. John's Wort, ginkgo, licorice, and cannabidiol alter drug effects. Check which combinations are confirmed by studies and which are myths without basis.
This guide primarily concerns combining herbs with prescription medications, and a separate section below answers questions about combining herbs with each other. Herbs are not divided into mild and strong but into studied and unstudied. The same plant that looks harmless in tea can accelerate the breakdown of an anticoagulant or slow the metabolism of an antidepressant, with effects sometimes visible only in blood tests. This guide collects what systematic reviews and clinical studies say about herb-drug combinations, separating documented risks from repeated myths. You will find a list of drugs sensitive to St. John’s Wort, the current evidence on ginkgo and bleeding, the mechanism of licorice on blood pressure, and a description of how cannabidiol alters liver enzyme activity. Each claim is sourced, and contradictions are explicitly noted.
KEY INFORMATION
• One in three people over 65 taking prescription drugs also takes herbal preparations or supplements, and one-third of this group faces interaction risks (Agbabiaka et al., 2018).
• St. John’s Wort lowers concentrations of warfarin, cyclosporine, digoxin, theophylline, and oral contraceptives by inducing CYP3A4, CYP2C9, CYP1A2, and P-glycoprotein (Henderson et al., 2002).
• Ginkgo has 15 published bleeding reports, including 8 intracranial (Bent et al., 2005), but a meta-analysis of 18 studies with 1985 participants found no significant effect of the standardized extract on coagulation parameters (Kellermann and Kloft, 2011).
• Cannabidiol inhibits CYP3A4 and CYP2C19; with clobazam, active metabolite concentration rose by 500% on average (Geffrey et al., 2015).
• Herbal products sold as dietary supplements do not have to have interaction studies. Adverse effects can be reported free to URPL, and in emergencies call 112.
Separately, the question arises about herbs in one mixture, e.g., in tea or smoking blend: the risk is usually lower than with drug combinations because doses of individual plants are small, but it does not disappear, and the mixture producer is not obliged to study interactions between ingredients. Which pairs of plants are truly problematic in such combinations is discussed in the section on herb-herb interactions.
Why do herbs interact with medications?
Herbs interact because they contain pharmacologically active compounds that compete with drugs for the same enzymes and receptors. There are two mechanisms. Pharmacokinetic changes alter drug concentration in blood. Pharmacodynamic effects sum effects without changing concentration. The first is visible in lab results; the second only in patient condition.
Most pharmacokinetic interactions involve cytochrome P450 family. Zanger and Schwab (2013) state that of 57 human CYP isoforms, about twelve from three families (CYP1, CYP2, CYP3) metabolize 70-80% of clinically used drugs. CYP3A4 and CYP2C9 are most abundant in liver, with smaller amounts of CYP2D6 and CYP2C19.
A plant can induce or inhibit this system. Induction means more enzyme, faster drug breakdown, and lower concentration. This is how St. John’s Wort works. Inhibition means enzyme is blocked, drug accumulates, and concentration rises. This is how cannabidiol and grapefruit juice act. The error direction differs with different consequences. Induction causes loss of protection as drug stops working. Inhibition causes effective overdose despite unchanged dose.
It is worth distinguishing three evidence levels because popular texts mix them. Theoretical interaction comes from cell studies and rarely translates clinically. Case reports concern specific patients with complications. Clinically confirmed interaction means measured concentration or effect change. The strongest evidence is for St. John’s Wort, which all European regulatory recommendations concern.
How often do patients combine herbs with prescription drugs?
More often than doctors expect. A UK cross-sectional study among people over 65 taking at least one prescription drug found 33.6% also used herbal or supplement products, and 32.6% of those were exposed to potentially harmful interactions (Agbabiaka et al., 2018).
The structure is instructive. Most (78%) used dietary supplements rather than herbs strictly: cod liver oil, glucosamine, multivitamins, and vitamin D. Herbal preparations were used by 20%, mostly evening primrose oil and valerian, less often ready mixtures of hops with gentian and passionflower. Women combined significantly more often than men (43.4% vs. 22.5%). The number of preparations per person reached eight.
A second study among oncology patients over 65 showed a narrower but sharper picture. Of 202 patients, 12% used herbs, 4% had potentially significant herb-drug interactions, half of which two pharmacists independently deemed clinically significant (Brokaar et al., 2025). Responsible herbs were red yeast rice, echinacea, turmeric, and cannabis. The scale is not massive but concerns patients with the smallest margin for error.
Which combinations have measurements, and which are based on mechanisms?
The table below shows for each plant the method supporting the warning and the interaction direction. Rows without human measurements are described as mechanism to distinguish from those with measured drug concentration or coagulation parameters.
| Plant | Measurement basis | Direction | Implication for patient |
|---|---|---|---|
| St. John’s Wort (Hypericum perforatum) | Clinical studies and case reports; EU monograph | Induces CYP3A4, CYP2B6, CYP2C9, CYP2C19, and P-glycoprotein; drug concentration decreases | Drug stops working: loss of anticoagulant protection, transplant rejection, pregnancy despite contraception |
| Ginkgo (Ginkgo biloba) | Case reports vs. meta-analysis of randomized studies | Disputed: reports indicate bleeding, meta-analysis found no effect on coagulation parameters | Caution in persons with existing bleeding risk; extract alone does not create risk |
| Garlic in supplements (Allium sativum) | 10 volunteers, saquinavir concentration measurement (Piscitelli et al., 2002); two separate warfarin studies (Mohammed Abdul et al., 2008; Macan et al., 2006) | Lowers saquinavir concentration; no effect on warfarin in both studies, though reviews list it among interactions | Measured risk concerns HIV treatment; bleeding reports and perioperative caution remain |
| Ginseng (Panax ginseng) | 12 healthy men (Jiang et al., 2004) and 25 post-stroke patients with INR measurement (Lee et al., 2008) | No effect on warfarin in both trials; reports and case descriptions indicate both directions | Unpredictability, not a fixed direction; possible hypoglycemia separately |
| Licorice (Glycyrrhiza glabra) | EU monograph and scientific committee opinion on glycyrrhizic acid | Weakens blood pressure-lowering drugs; with diuretics increases potassium loss | Increased blood pressure despite treatment, hypokalemia, arrhythmias |
| Valerian (Valeriana officinalis) | Mechanism; EU monograph reports no interactions | Additive sedative effect, mainly described with anesthetics | Drowsiness and impaired driving ability; report use to anesthesiologist before surgery |
Which drugs does St. John’s Wort weaken?
St. John’s Wort (Hypericum perforatum) weakens drugs whose concentration depends on CYP3A4, CYP2C9, CYP1A2, or P-glycoprotein. A systematic review of European reports showed clinically significant interactions with warfarin, phenprocoumon, cyclosporine, HIV protease inhibitors, theophylline, digoxin, and oral contraceptives, all leading to decreased drug concentration or effect.
A broader review by Izzo and Ernst (2009) expanded this list with clinical trial data. St. John’s Wort lowers plasma concentration or accelerates elimination of alprazolam, amitriptyline, atorvastatin, cyclosporine, digoxin, erythromycin, fexofenadine, gliclazide, imatinib, indinavir, irinotecan, midazolam, nifedipine, omeprazole, simvastatin, tacrolimus, voriconazole, and warfarin. Case reports add psychotropic drugs: paroxetine, sertraline, venlafaxine, nefazodone, buspirone, bupropion.
| Drug group | Effect of combination with St. John’s Wort | Direction of change | Evidence type |
|---|---|---|---|
| Anticoagulants (warfarin, phenprocoumon) | decreased concentration and effect, thrombosis risk | drug concentration down | review of reports and clinical studies |
| Hormonal contraception | decreased concentration, reported unplanned pregnancies | drug concentration down | review of reports and case descriptions |
| Immunosuppressants (cyclosporine, tacrolimus) | decreased concentration, transplant rejection risk | drug concentration down | clinical studies and case reports |
| HIV drugs (protease inhibitors) | decreased indinavir concentration | drug concentration down | clinical studies |
| Statins (simvastatin, atorvastatin) | decreased concentration, weaker cholesterol control | drug concentration down | clinical studies |
| Oncology drugs (imatinib, irinotecan) | decreased anticancer drug concentration | drug concentration down | clinical studies |
| SSRIs antidepressants | weakened effect and serotonergic symptoms | drug concentration down, serotonergic effect up | case reports |
Two practical notes. Induction strength is unpredictable because active compound content varies between preparations, which the European review explicitly cites as why a safe dose cannot be given. Also, St. John’s Wort acts serotonergically, so with antidepressants the risk goes both ways: either reduced efficacy or serotonin excess.
Does ginkgo really increase bleeding risk?
Evidence is contradictory, and the honest answer is: probably in some people yes, but not to a degree detectable in controlled studies. This is a rare case where case reports and meta-analysis differ, and knowing both sides changes patient decisions.
Warning side: a systematic literature review collected 15 published bleeding reports temporally linked to ginkgo use, including 8 intracranial bleeds (Bent et al., 2005). Authors weakened their conclusion: in 13 of 15 cases other bleeding risk factors were present, and only 6 reports clearly stated bleeding did not recur after stopping ginkgo. In three cases bleeding time was measured and prolonged.
Reassuring side: a meta-analysis of 18 randomized studies with 1985 participants found no significant effect of standardized ginkgo extract on ADP-induced platelet aggregation, fibrinogen concentration, prothrombin time, or kaolin-cephalin time (Kellermann and Kloft, 2011). The only statistically significant effect in the subgroup taking 240 mg daily or more was deemed clinically insignificant.
The practical conclusion is conditional. Ginkgo alone in a healthy person does not appear to be a bleeding risk factor. Risk concerns those who already have it: taking warfarin, aspirin, or NSAIDs, post-surgery, or with coagulation disorders. Izzo and Ernst also note interactions with antiepileptics, risperidone, and trazodone, and decreased concentrations of omeprazole, ritonavir, and tolbutamide.
Which kitchen herbs affect coagulation?
Garlic and ginseng enter food in amounts no one treats as drug doses, yet they appear in bleeding warnings. A perioperative review lists garlic, ginkgo, and ginseng as three plants directly increasing bleeding risk during surgery. In warfarin trials in the table above, neither garlic nor ginseng changed INR.
| Plant | Interaction listed in review | Source |
|---|---|---|
| Garlic (Allium sativum) | warfarin, chlorpropamide, fluindione, ritonavir; decreased chlorzoxazone concentration | Izzo and Ernst 2009; warfarin measurements above did not confirm effect |
| Ginseng (Panax ginseng) | warfarin and phenelzine | Izzo and Ernst 2009; INR measurements above did not confirm effect |
| Ginkgo (Ginkgo biloba) | warfarin, aspirin, ibuprofen, antiepileptics | Izzo and Ernst 2009 |
| Garlic, ginkgo, ginseng | direct perioperative bleeding risk | Ang-Lee et al. 2001 |
| Ginseng | hypoglycemia in diabetic patients | Ang-Lee et al. 2001 |
Note what is missing in this table. There is no number indicating how much bleeding risk increases when combining several such plants. Such a number circulates in guides but does not come from any measured study. Summation of effects is mechanistically probable and why surgeons ask about herbs before surgery, but giving percentages here would be guessing.
Why do licorice and ephedrine burden the heart?
Both plants act directly on the cardiovascular system, not via drug metabolism, so they harm even if you take no prescription drugs. Mechanisms and manifestations differ.
Licorice (Glycyrrhiza glabra) contains glycyrrhizic acid, whose metabolite blocks 11-beta-hydroxysteroid dehydrogenase type 2 and also binds mineralocorticoid receptor as an agonist (Sabbadin et al., 2019). The result is pseudo-hyperaldosteronism: hypertension, metabolic alkalosis, and hypokalemia. A 2021 case report shows extreme consequences: sinus pause and syncope after years of daily root consumption. Authors emphasize chronic uncontrolled use can cause ventricular arrhythmias, including ventricular tachycardia and torsade de pointes. For patients on diuretics or digoxin, potassium loss is the most dangerous factor.
Ephedrine has a different story. An independent analysis of 140 reports to the US FDA found 31% of events definitely or probably related to ephedrine alkaloid supplements, with 47% of related events involving the cardiovascular system (Haller and Benowitz, 2000). Most common symptoms were hypertension, palpitations, tachycardia, strokes, and seizures. Ten events ended in death, thirteen in permanent disability. More on licorice and dosing is in our licorice properties guide.
What happens when sedative herbs meet benzodiazepines?
Sedation sums without changing drug concentration. This is a pharmacodynamic interaction: the patient takes a usual benzodiazepine dose, blood test is normal, yet in the morning cannot concentrate and falls on stairs. For an octogenarian, this scenario ends with a fracture, not discomfort.
The perioperative review describes this mechanism directly: kava and valerian enhance anesthetic sedation, a reason anesthesiologists ask about herbal use before anesthesia. The same applies outside hospital when valerian tea and a glass of wine accompany evening benzodiazepines. Three substances act on the same inhibitory system.
The practical problem is sedative herbs rarely appear singly. In the cited UK study, one of the most common herbal preparations was a mixture of hops, gentian, and passionflower, sold OTC as a sleep aid. A person taking it alongside a prescription hypnotic usually does not report it to the doctor, not treating herbs as medicine.
The safety rule is simple and needs no table: if a preparation helps you sleep, do not combine it with anything else that also helps sleep, including alcohol. For a gentler path, we described differences among popular evening herbs in a sleep tea comparison.
How does cannabidiol change drug concentrations?
Cannabidiol inhibits drug-metabolizing enzymes, raising their concentrations instead of lowering. A registration data review (Brown and Winterstein, 2019) identifies CYP3A4 and CYP2C19 as main targets and P-glycoprotein as a second mechanism via excretion. Authors note adverse effects occurred in nearly half of cannabidiol users and were dose-dependent.
The best-measured interaction concerns clobazam. In a study of 13 children with refractory epilepsy taking both drugs, clobazam concentration rose 60%, and its active metabolite norclobazam by 500% (Geffrey et al., 2015). Adverse effects occurred in 10 participants and resolved after clobazam dose reduction, not cannabidiol discontinuation.
| Drug or group | Known facts |
|---|---|
| Clobazam | 60% increase in drug and 500% in metabolite concentration; monitoring needed (Geffrey et al., 2015) |
| Warfarin | Reported case of INR increase requiring dose adjustment |
| CYP3A4 and CYP2C19 substrates | Risk of concentration increase; dose reduction recommended |
| Liver enzymes | Increased aminotransferase activity as common adverse effect |
Separate issue is grapefruit. Grapefruit juice inhibits intestinal CYP3A4, the same enzyme as cannabidiol, so both act in the same direction. A simple practical rule: if the drug leaflet warns against grapefruit, the same caution applies to cannabidiol. A full list of sensitive drugs is in a separate text on cannabidiol-drug interactions.
Which herbs should not be combined with each other?
The answer is shorter than demand suggests. Herb-drug interactions have clinical studies and monographs, but herb-herb interactions practically do not, so any number describing their strength would be guessing. What can be said honestly is based on summing the same effects.
The first pair to avoid is sedative plants combined: valerian, hops, passionflower, and lemon balm act in the same direction, and ready mixtures contain them together. The EU monograph on valerian states “None reported” for interactions and separately warns the preparation may impair driving. The perioperative review lists anesthetic enhancement as a real operating room problem (Ang-Lee et al., 2001). Sedation summation is mechanistically credible but not quantified.
The second pair concerns plants affecting potassium balance and blood pressure: licorice combined with diuretic herbs deepens potassium loss, which licorice alone causes. The third, often overlooked, is St. John’s Wort in mixtures acting on the same enzymes, potentially lowering concentrations of other herbs as it does drugs. Separately, note the nettle leaf monograph states “None reported” for interactions, so frequent warnings about combining it with diuretics lack basis.
Who is most at risk of herb-drug interactions?
Risk is unevenly distributed. It rises where three conditions meet: many drugs simultaneously, slower metabolism, and narrow therapeutic window of at least one preparation. Three groups meet all three simultaneously.
Seniors are in this group for arithmetic reasons. Polypharmacy increases possible herb-drug pairs faster than linearly, and the UK study showed one in three over 65 adds a self-purchased preparation to prescriptions. Also, slower liver and kidney function prolongs drug presence regardless of number of preparations.
Oncology patients are at risk for another reason. Anticancer drugs have a narrow margin between effective and toxic doses, so any concentration change has consequences. A 2025 study found 4% potentially significant herb-drug interactions, half clinically significant. Transplant recipients taking cyclosporine or tacrolimus are another high-risk group, where St. John’s Wort-induced concentration drop risks organ rejection.
The third group is pregnant and breastfeeding women. Here the problem is not documented interaction but lack of data: herbal preparations rarely enter studies with pregnant women, so safety remains unknown, not confirmed. The rule is one: no preparation without the attending physician’s consent.
How to check interactions and report adverse effects?
Checking takes a minute, reporting a quarter-hour, and is free. Both are done in different places and for different purposes: the first before taking a preparation, the second after an unexpected symptom.
Before starting, list everything you take: prescription drugs, OTC products, vitamins, herbs, cannabis products. The list should go to doctor and pharmacist because neither can guess you drink valerian tea at night. Professionals use Lexicomp, UpToDate, and Micromedex databases; patients can check a few herb-drug pairs for free at Drugs.com. Entering warfarin and ginkgo returns a warning with mechanism description.
The second habit is introducing one plant at a time and maintaining it for weeks before adding another. With five preparations started the same day, even a doctor cannot tell which causes rash or drowsiness. Symptoms warranting discontinuation and doctor contact include gum or nose bleeding, easy bruising, dizziness, excessive drowsiness, palpitations, and yellowing skin.
Adverse effect reports are accepted by the Office for Registration of Medicinal Products, Medical Devices and Biocidal Products. Anyone can report, including patients, free and without certainty of causality. In emergencies, call 112.
What does Polish law say about herbs, supplements, and cannabis products?
Most plant preparations in Polish trade are marketed as dietary supplements, i.e., food supervised by the Chief Sanitary Inspectorate, not medicinal products. Supplement manufacturers do not have to prove clinical efficacy or conduct interaction studies, directly translating into the risks described above.
The practical effect concerns quality. Active substance content in plant preparations can differ from label declarations, and for St. John’s Wort the European review cited this variability as a reason for unpredictable enzyme induction strength. Choose producers who provide extract standardization and independent lab analysis results.
A separate regulation concerns hemp. According to Art. 4 point 5 of the Act of July 29, 2005 on counteracting drug addiction (consolidated text Journal of Laws 2023 item 1939), as amended by the Act of March 24, 2022 (Journal of Laws 2022 item 763), fiber hemp are plants with a sum of delta-9-THC and tetrahydrocannabinolic acid not exceeding 0.3% dry weight, rounded to one decimal place. The threshold concerns the sum of both compounds, not delta-9-THC alone, which changes lab test results. Until May 6, 2022, this value was 0.20%.
The national threshold corresponds to the EU threshold but does not derive from it. The 0.3% value for hemp varieties eligible for support under the Common Agricultural Policy was introduced by Regulation (EU) 2021/2115, effective January 1, 2023; the earlier Regulation 1307/2013 set 0.2% and was repealed. These are two separate regulations with the same number. Cannabidiol itself is not listed in any controlled substances list, but HHC is controlled in Poland.
Frequently Asked Questions
Does St. John’s Wort reduce the effectiveness of contraceptive pills?
Yes. A systematic review of European reports lists oral contraceptives among drugs whose concentration or effect decreases under the influence of St. John’s Wort, with the mechanism being induction of CYP3A4 and P-glycoprotein. Women using hormonal contraception should avoid St. John’s Wort preparations or additionally use barrier methods.
How long before surgery should herbs be discontinued?
There is no single timeframe for all plants, and recommendations depend on the preparation. More important than the date is providing the anesthesiologist with a full list of taken preparations. A perioperative review lists garlic, ginkgo, and ginseng among plants increasing bleeding risk during surgery, although for ginkgo itself a meta-analysis with a control group did not show a significant effect on coagulation parameters. Valerian enhances anesthetic effects, and St. John’s Wort accelerates the metabolism of perioperative drugs.
Can cannabidiol be combined with antidepressants?
It can, but only under medical supervision and with symptom monitoring. Cannabidiol inhibits CYP2C19, the enzyme responsible for metabolizing some antidepressants, so their concentration may increase. Never discontinue a psychiatric medication on your own to make room for a supplement.
Can grapefruit juice and cannabidiol be consumed together?
Better not. Both inhibit CYP3A4, so their effects add up and concentrations of drugs metabolized this way may increase more than from each agent alone. A practical rule: if the drug leaflet warns against grapefruit, the same caution applies to cannabidiol.
Do mild herbs in infusions also interact?
Usually not to a clinically significant degree, because an infusion contains many times less active substance than a standardized extract. The problem arises with supplements taken daily for weeks. If you take any chronic medication, discuss the planned herbal preparation with a doctor or pharmacist before the first dose.
Can herbs replace medications for hypertension or diabetes?
No. Discontinuing a chronic medication in favor of a herbal preparation on your own is dangerous and, in hypertension, diabetes, epilepsy, and heart diseases, can lead to complications within days. Herbs can play a complementary role, but any change in regimen should be decided by the attending physician.
Does the manufacturer of an herbal supplement have to study interactions?
No. Dietary supplements in Poland are supervised as food by the sanitary inspectorate and do not have to undergo interaction studies or prove efficacy. Responsibility for safe use thus falls on the patient and doctor, so it is worth choosing preparations with standardized extracts and available analysis results.
Where to report adverse effects after an herbal preparation?
Reports are accepted by the Office for Registration of Medicinal Products, Medical Devices and Biocidal Products. Anyone can report, including patients without a doctor’s mediation, free of charge and without certainty about causality. Data feed the European safety monitoring database. In emergencies, call 112.
If you are looking for plant raw materials with known origin and described composition, you will find them in our herbs category.
This article is informational and educational and does not replace consultation with a doctor. If you are pregnant, breastfeeding, taking medications, or have chronic diseases, consult a specialist before using supplements or herbs.
Author: Michał Waluk · Published: 2026-05-11 · Updated: 2026-09-22
