Which herbs and supplements interact with common medications?

Most herb-drug problems come from the same few routes. Some herbs switch on gut and liver enzymes and transporters that clear drugs, so drug levels fall. St John's wort does this through CYP enzyme and P-glycoprotein induction¹. Others block those enzymes. Grapefruit furanocoumarins inhibit CYP3A4 in the intestine, so more of a drug gets into the blood⁹. A third group adds to a drug's own effect or side effects, such as licorice with corticosteroids¹⁰.

Herb by herb

• St John's wort (strongest evidence). It reduced exposure to contraceptive hormones by 13 to 15 percent, with more breakthrough bleeding and signs of ovulation³. It cut the HIV drug indinavir's exposure by 57 percent⁴ and warfarin's anticoagulant effect⁵. It also lowered cyclosporine, tacrolimus, digoxin and statin levels, and with SSRIs it has been linked to mild serotonin syndrome^1,10. See St John's wort.
• Grapefruit. It affects calcium-channel blockers, immunosuppressants, antihistamines and statins, but the size of the effect depends on the product, the amount, and the person^9,2. See grapefruit.
• Licorice. Regularly eating large amounts of licorice can cause high blood pressure and low potassium, sometimes with arrhythmias⁶. See licorice root.
• Kava. Rare liver injury has been reported, linked to overdose, long courses and co-medication⁷. See kava.
• Ginkgo and ginger. In a controlled study, neither changed warfarin's effect¹¹, and a meta-analysis found no clotting changes with ginkgo⁸. Case reports of bleeding exist but are hard to interpret¹. See ginkgo.
• Ginseng. It did not change warfarin levels or effect in a volunteer trial, though case reports suggest possible interactions^5,1.

How to protect yourself. Keep one up-to-date list of every prescription, over-the-counter drug and supplement, and show it to your pharmacist at each refill. Never stop a prescription yourself. Our Interaction Checker can flag obvious conflicts, but it is not a clinical screen.

The overall picture

Izzo and Ernst reviewed seven popular herbs (ginkgo, St John's wort, ginseng, garlic, echinacea, saw palmetto and kava) and found 128 case reports or series and 80 clinical trials. Trials showed that St John's wort reduced plasma levels of 27 drugs or drug classes, including warfarin, digoxin, simvastatin, ciclosporin, tacrolimus, oral contraceptives, indinavir and methadone. Ginkgo lowered omeprazole, ritonavir and tolbutamide levels, and no interactions were reported for saw palmetto. The authors concluded that the significance of many interactions is uncertain, but several, particularly with St John's wort, may have serious consequences¹. A review of herb-warfarin evidence rated nine of 38 herbs as major-severity interactions yet judged only four of the 38 as highly probable and 21 as doubtful, a reminder of how thin much of the evidence is¹². An earlier review warned that many reports are sketchy and lack analysis of the suspect product¹⁰.

St John's wort

Murphy and colleagues gave 16 healthy women a low-dose pill and placebo for two cycles, then added St John's wort 300 mg three times daily for two cycles. Contraceptive exposure fell 13 to 15 percent, breakthrough bleeding increased, and there was evidence of follicle growth and probable ovulation. The study was small and single-blind and did not measure pregnancy³. Piscitelli's letter reported that St John's wort cut indinavir's area under the curve by a mean of 57% and the trough level by 81% in healthy volunteers, a drop large enough to risk treatment failure and drug resistance⁴. Jiang and colleagues gave 12 healthy men a single warfarin dose alone or after 14 days of St John's wort in a randomized crossover. The herb raised S-warfarin clearance (ratio 1.29, 90% CI 1.16 to 1.46) and lowered the INR area under the curve (ratio 0.79, 0.70 to 0.95)⁵.

Grapefruit

In a randomized crossover, 10 volunteers took double-strength grapefruit juice or water for two days, then 60 mg of simvastatin. The juice raised simvastatin's peak concentration about 9-fold and its area under the curve 16-fold, probably by blocking first-pass CYP3A4 metabolism in the small intestine. The authors advised avoiding large amounts or reducing the dose². A review lists calcium-channel blockers, immunosuppressants and antihistamines among the affected drug classes and stresses the variability between products and people⁹. Note that this study used repeated glasses of double-strength juice, more than a typical breakfast glass.

Licorice and kava

A review of 104 case reports of licorice toxicity found pseudo-hyperaldosteronism with hypertension and low potassium as the typical picture, occasionally with arrhythmias, paralysis or rhabdomyolysis; toxicity often required hospital care but mortality was low⁶. Licorice has also been reported to potentiate corticosteroids¹⁰. Because low potassium is a recognized risk for digoxin toxicity and for some heart-rhythm problems, a hazard with digoxin or potassium-wasting diuretics is plausible, but we found no trial testing licorice with those drugs. For kava, a structured causality review of 14 patients with liver disease rated the link highly probable in one, probable in four and possible in nine. Risk factors included overdose, prolonged use and co-medication, and poor raw-material quality may contribute⁷.

Bleeding herbs: where the evidence is weak

Jiang and colleagues found that ginkgo and ginger, at recommended doses for a week before warfarin, did not change S-warfarin clearance (ginkgo ratio 1.05, 90% CI 0.98 to 1.21) or the INR area under the curve (ginkgo 0.93, 0.81 to 1.05; ginger 1.01, 0.93 to 1.15) in 12 healthy men¹¹. A meta-analysis of 18 randomized trials (1,985 adults) found no effect of standardized ginkgo extract on platelet aggregation or prothrombin time, so it indicated no higher bleeding risk⁸. These are surrogate measures in mostly short trials, not bleeding events in anticoagulated patients, and case reports persist¹. Reasonable caution is not the same as proof.

Limits of this evidence

Most data come from small volunteer studies with single drug doses, from case reports, and from products of unknown quality. Few studies follow patients on real regimens, and almost none compare brands.

Bottom line (confidence 4 of 5). St John's wort and grapefruit reliably alter drug levels and can matter clinically. Licorice and kava carry real but narrower risks. Ginkgo's bleeding risk is unproven. Always tell your pharmacist about everything you take.