July 7, 2021

Goldenseal-Drug Interactions: A Concern?

Results from an open-label crossover study
Goldenseal increased the levels of the drug midazolam and decreased the levels of the drug metformin, but flaws in the study design call into question the practical value of the results.


Nguyen JT, Tian DD, Tanna RS, et al. Assessing transporter-mediated natural product-drug interactions via in vitro-in vivo extrapolation: clinical evaluation with a probe cocktail. Clin Pharmacol Ther. 2021;109(5):1342-1352.


Open-label, 2-arm, fixed-sequence crossover study


Researchers screened 29 subjects and enrolled 19. The final 16 subjects included were Caucasian (7 men, 6 women) and Asian (1 man, 2 women) whose ages ranged from 23 to 42 years.


This is a 3-part study involving both laboratory testing and clinical assessment. Part 1 tested Hydrastis canadensis, commonly known as goldenseal, with 3 known alkaloids as an inhibitor of multiple transporters using an in vitro system. Part 2 was a prediction of transporter-mediated interactions. Part 3 assessed goldenseal in the presence of 4 pharmaceuticals in healthy volunteers to predict herb-drug interactions in vivo.


Researchers chose the goldenseal product based on its popularity, using sales data from among 35 goldenseal products sold on Amazon. They compared the selected goldenseal product to others in an interactions database1 and standardized it to berberine. The pharmaceuticals obtained from a local pharmacy included furosemide (FU), metformin (ME), rosuvastatin (RO), and midazolam (MI). Subjects were given 1 g goldenseal 3 times daily, at 0800, 1200, and 1600 h, for 5 days. On day 6 they received the drug probe consisting of 1 mg FU, 50 mg ME, 10 mg RO, and 2.5 mg MI plus 1 g goldenseal, and then at 4-h intervals 2 additional 1-g goldenseal doses. Blood was collected from the subjects at 24, 48, 72, and 96 h after the drug cocktail, and urine was collected from 0–12 h post-drug cocktail.


Goldenseal or its alkaloids were found to have inhibitory effects on both influx and efflux transporters, namely OAT1 (organic anion transporter 1), OAT3, OATP1B1, OAT1B3, OCT1 (organic cation transporter 1), OCT2, BCRP (breast cancer resistance protein), MATE1 (multidrug and toxin extrusion 1), and MATE2-K at 50% inhibition or greater. The greatest effect was seen with the whole goldenseal extract, followed by singular-constituent berberine, beta-hydrastine, and hydrastinine alkaloids.

Key Findings

MI increased 40% during goldenseal exposure while FU was unaffected. ME AUC (area under the curve) decreased by 23%, and RO was not affected by goldenseal ingestion. Goldenseal was classified as a CYP450-3A inhibitor, and therefore, the authors classified it as similar to grapefruit juice. The active transport inhibitor could be berberine, synergy of the 3 characterized alkaloids, or other unknown constituents. Goldenseal likely reduced ME bioavailability by altering intestinal permeability, transport, and/or absorption. ME is itself taken up by transporters OCT1, OCT3, ThTr-2 (thiamine transporter 2), PMAT (plasma membrane monoamine transporter), and SERT (serotonin transporter) on the luminal membrane, so the interaction with goldenseal was assumed to be at OCT1.

The goldenseal extract taken for 5 days prior and 3 times during the day of the challenge increased the concentration of MI (CYP3A4), had no effect on RO (BCRP, OATP1B1, OATP1B3) or FU (OAT1, OAT3), and decreased ME (OCT1, OCT2, MATE1, MATE2-K) AUC but had no effect on the half-life of ME or its renal clearance. The mechanism could be altered intestinal permeability, transport across the membrane, or other unknown processes and may alter glucose control.

Practice Implications

This study uses doses that are not clinically relevant for goldenseal or the drugs, making it yet another study on interactions between natural health products and pharmaceuticals designed to limit the use of the former. Specifically, the subjects received a 4-drug cocktail that was well below the therapeutic dose for the drugs. MI (normally given at 2–2.5 mg initially, titrated at 1 mg/dose up to 3.5–7.5 mg in adults for surgical anesthesia) was kept at the low-end dose of 2.5 mg. FU (often given at 20–80 mg for prompt diuresis) was merely 1 mg/day in this study. RO (typically prescribed at 5–40 mg/d with a maintenance dose of usually 20 mg/day) was given at 10 mg. ME is usually prescribed at 500–1,000 mg twice daily, and in this study, it was given at 10 mg. The goldenseal dose was 1.0 g 3 times daily. The lower dose of the pharmaceutical and the higher dose of goldenseal could have easily skewed the results as goldenseal was the dominant medication throughout this trial. Typical doses of goldenseal are 100–500 mg dried root 3 times daily, supplying 8%–12% standardized alkaloids. To put this in perspective, all subjects received goldenseal plus a surgical anesthetic, a potent diuretic, a common statin, and a common antihyperglycemic all at the same time.

The National Center for Complementary and Integrative Health created the Center of Excellence for Natural Product-Drug Interaction Research (NaPDI) (napdicenter.org) to provide pharmacokinetic study of natural health products.1 Six of the top 40 herbal medicines are implicated in causing significant drug interactions, but clinical data are lacking.2 The top herbs on their list of concern include Vaccinium macrocarpon (cranberry), Echinacea spp (echinacea), Camellia sinensis (green tea), Actaea racemosa (black cohosh), Garcinia spp, Linum usitatissimum (flaxseed), Zingiber officinale (ginger), Hedera helix (ivy), Curcuma longa (turmeric), Valeriana officinalis (valerian), Trigonella foenum-graecum (fenugreek), Pausinystalia yohimbe, Aloe spp, Serenoa repens (saw palmetto), Silybum marianum (milk thistle), Allium sativum (garlic), Cocos nucifera (coconut) oil, Sambucus spp (elderberry), Cinnamomum verum, Boswellia spp (frankincense), Ginkgo biloba, Senna alexandrina, Rhodiola rosea, all bioflavonoids, Panax spp (ginseng), Eleuthrococcus senticosus (Siberian ginseng), all beta glucans, Lepidium meyenii (maca), Hypericum perforatum (Saint-John’s-wort), barley and wheat grasses, Salvia hispanica (chia) seed and oil, and Lycium barbarum (goji berry plant), among many others.2

This study uses doses that are not clinically relevant for goldenseal or the drugs, making it yet another study on interactions between natural health products and pharmaceuticals designed to limit the use of the former.

In an era of HIV and coronavirus pandemic fears, there are also concerns about antiretroviral therapy interactions.3 The botanicals the NaDPI expressed highest concern about as affecting antiretrovirals included Hypericum perforatum (Saint-John’s-wort), Piper nigrum (black pepper), Piper spp (ie, kava kava), Citrus paradisi (grapefruit), Hypoxis hemerocallidea (African potato), Ginkgo biloba, Panax spp (ginseng), Eleutherococcus senticosus (Siberian ginseng), Allium sativum (garlic), Hydrastis canadensis (goldenseal), and others.3 In March 2020 the 4 highest-priority herbs were established as Cannabis sativa (marijuana), Hydrastis canadensis (goldenseal), Camellia sinensis (green tea), and Mitragyna speciosa (kratom), with the recommendation to add Glycyrrhiza spp (licorice) due to its sales value in the natural health products market.4 In the last quarter of a century, the market has grown from 4,000 to 80,000 products, or from $4.8 billion to $8.4 billion in the USA from 2008 to 2018, with cannabinoids classified as dangerous and leading the way by potentially causing hazardous interactions with antiseizure drugs.4

NaPDI has 3 core areas of expertise: analytic chemists and pharmacologists to characterize natural health products, pharmacologists and medicinal chemists to precipitate interactions with selected drugs, and informatics to generate information for release to the public, pharmacists, and medical doctors.1 Data are generated from enzyme reactions, tissue fractions such as human liver microsomes, cell cultures, drug transport experiments, and human subjects using oral, renal, plasma concentration, half-life, and area under the curve measurements.1 The NaPDI has adopted the acronym FAIR to indicate that they want findable, accessible, interoperable, and reusable data in this repository.1

Mitragyna speciosa (kratom) was selected as 1 of the first herbals to study because it has medicinal and recreational uses and can assist in withdrawal from opioids as well as for anxiety and depression, but it has been linked with hallucinations, seizures, and liver damage and is banned in some states. Its known active constituents include mitragynine, 7-hydroxymitragynine, and speciofoline.5 Cannabis, another high-priority plant for NaPDI, is also an early plant in their data repository based on 2 Cannabis sativa extracts, bulk plant material, and its known constituents of cannabidiol (CBD), tetrahydrocannabinol (THC), and other cannabinoids.6

Given the above concerns, let’s examine the data on berberine, one of the principal isoquinoline alkaloids in goldenseal. A pilot study of berberine (500 mg 3 times daily) or metformin (500 mg 3 times daily) for 3 months in 84 Chinese subjects with type 2 diabetes (T2D) led to a significant decrease in hemoglobin A1c (HbA1c), fasting glucose, postprandial glucose, and plasma triglycerides.7 In this trial 36 subjects were on only berberine, and 48 subjects were on combination therapy of berberine with sulfonylureas, metformin, acarbose, or insulin. If subjects in the latter group had significant gastrointestinal side effects, the berberine dose was reduced to 300 mg 3 times daily. Results were lack of efficacy in only 3 subjects, 1 from participation time and 1 from lack of compliance. In the combination group, HbA1c, fasting glucose, and postprandial glucose decreased significantly, and there were no changes in liver or kidney function.7 While this is not a trial with whole-plant goldenseal, it does suggest that the concerns raised in the current study under review here, which was 5-days duration, did not manifest in a 3-month study using a main alkaloid constituent, berberine. In fact, the opposite occurred as there was improvement in the trial subjects’ glucose control as well as in other parameters associated with T2D.

Two meta-analyses have concluded that berberine is effective in T2D, hyperlipidemia, and hypertension.8,9 Both berberine and metformin are from plants, the latter from Galega officinalis. Both have different structures, are tolerable, and may cause gastrointestinal discomfort in some. They appear to work together. When used in combination, they reduce the side effects of each other, they lower the dosage required for clinical effect, and they better overcome the issue of oral bioavailability since berberine plants are 80% metabolized by CYP450-2D6.9 A six-month randomized, controlled trial in 60 T2D patients found better results when berberine and metformin were combined compared when used separately, likely due to synergism even though they favor different transporters and metabolism.10,11

The various transporters used in this trial are found in diverse locations throughout the body. BCRP (breast cancer resistance protein) is an ATP efflux transporter identified in 1998 because it led to drug resistance of mitoxantrone and topotecan in breast cancer cells.12 Hundreds of BCRP inhibitors are now known, and BCRP is found in the placenta, intestinal epithelium, hepatocytes, endothelial cells of brain micro vessels, and renal proximal tubular cells. BCRP has been investigated for its effects on chemotherapy medications, statins, tacrolimus, olmesartan, and other medications. Study results are mixed regarding whether intracellular levels of drugs and steroids go up or down with BCRP transporter expression. Lastly, BCRP appears to have a role in gout development.12

OTC (organic cation transporter), also known at HOCT, is on the blood side of the sinusoidal membrane of hepatocytes, often works with MATE transporters, and is important for metformin pharmacokinetics, as well as neurotransmitters and tyrosine kinase inhibitors.13 It is found in the small intestine, renal proximal tubular cells, and basophilic granulocytes with significant differences between rodents and humans and among humans. Thus, a clinical trial could bias results through their screening and selection of human subjects or use of rodent data.13,14

MATE (multidrug and toxin extrusion) is a proton antiporter that mediates efflux of organic cations in the kidney and liver, especially metformin but also cisplatin and fluoroquinolones.15,16 It is also expressed in skeletal muscle, adrenal glands, testes, and heart and can be both an uptake transporter as well as an efflux agent in the kidney. It has multiple genetic polymorphisms and may play a role in mental retardation and behavioral issues, especially Smith-Magenis syndrome, which involves chromosome 17.17,18 As with the above 2 groups of transporters, the United States Food and Drug Administration (FDA) and European Medicines Agency (EMA) guidance policy do not strictly recommend them for consideration in drug interaction studies, but this may be required in the future.

Where I practice in Canada, patients are routinely ordered by hospital pharmacists and medical doctors to stop all supplements 14 days before any intervention or surgical procedure. The intense intimidation with which this message is delivered is seldom appropriate. Based on the current study under review, for example, such a warning would be warranted only for the goldenseal-MI interaction alongside the patient having a surgical emergency. The rarity of this combination is not a rational reason to avoid a clinically effective botanical medicine like goldenseal.

The diverse genetic polymorphisms characteristic of these transporters mean that patient response to the various drugs assessed in this trial can vary considerably from patient to patient. The transporters could also be a contributing factor explaining why each patient’s response to the same drug or herb is different, but our knowledge is in its infancy at best, and making blanket statements about interactions or proposals to ban selected foods and herbs based on petri dish or small human trials is likely overgeneralized. For example, we are all aware that metformin as a single drug works in some patients and not others, and that some have significant nausea and cannot tolerate it while others have no issues even at higher doses.19 The focus on transporters also ignores the role of the patient’s microbiome, which we have only just begun to research.20


In this trial, 16 poorly identified subjects were given 1 g Hydrastis canadensis (goldenseal) 3 times daily for 5 days and then were co-administered a drug cocktail of nonclinically relevant furosemide (1 mg), metformin (50 mg), rosuvastatin (10 mg), and midazolam (2.5 mg) followed by 2 more doses of 1 g each goldenseal with blood and urine collection. Drug levels were assessed to show that goldenseal increased the levels of midazolam and decreased the levels of metformin, and thus, the authors suggested goldenseal should not be ingested. Going forward, we should continue to take a critical look at the design of these trials from the NaDPI group on some of the most popular botanical medicines.

Categorized Under


  1. Birer-Williams C, Gufford BT, Chou E, et al. A new data repository for pharmacokinetic natural product-drug interactions: from chemical characterization to clinical studies. Drug Metab Dispos. 2020;48:1104-1112.
  2. Spanakis M, Sfakianakis S, Sakkalis V, Spanakis EG. PharmActa: Empowering patients to avoid clinical significant drug-herb interactions. Medicines. 2019;6(1):26.
  3. Fasinu PS, Gurley BJ, Walker LA. Clinically relevant pharmacokinetic herb-drug interactions in antiretroviral therapy. Curr Drug Metab. 2015;17(1):52-64.
  4. Gaston TE, Mendrick DL, Paine MF, Roe AL, Yeung CK. “Natural” is not synonymous with “safe”: toxicity of natural products alone and in combination with pharmaceutical agents. Regul Toxicol Pharmacol RTP. 2020;113:104642.
  5. Todd DA, Kellogg JJ, Wallace ED, et al. Chemical composition and biological effects of kratom (Mitragyna speciosa): in vitro studies and implications for efficacy and drug interactions. Sci Reports. 2020;10(1):19158.
  6. Qian Y, Want X, Markowitz JS. In vitro inhibition of carboxylesterase-1 by major cannabinoids and selected metabolites. Drug Metab Dispos Biol Fate Chem. 2901;47:465-472.
  7. Yin J, Xing H, Ye J. Efficacy of berberine in patients with type 2 diabetes. Metabol. 2008;57(5):712-717.
  8. Lan J, Zhao Y, Dong F, et al. Meta-analysis of the effect and safety of berberine in the treatment of type 2 diabetes mellitus, hyperlipidemia, and hypertension. J Ethnopharmacol. 2015;161:69-81.
  9. Koppen LM, Whitaker A, Rosene A, Beckett RD. Efficacy of berberine alone and in combination for the treatment of hyperlipidemia: a systematic review. J Evid Based Comp Alt Med. 2017;22(4):956-968.
  10. Pang B, Zhao LH, Zhou Q, Zhou TY, Want H, Gu CJ, Tong XL. Application of berberine on treating type 2 diabetes mellitus. Int J Endocrinol. 2015;2015:905749.
  11. Wang H, Zhu C, Ying Y, Luo L, Huang D, Zhijun L. Metformin and berberine, two versatile drugs in treatment of common metabolic diseases. Oncotarget. 2018;9(11):10135-10146.
  12. Mao Q, Unadkat JD. Role of the breast cancer resistance protein (BCRP/ABCG2) in drug transport: an update. AAPS Journal. 2015;17(1):65-82.
  13. Jonker JW, Schinkel AH. Pharmacological and physiological functions of the specific organic cations transporters: OCT1, 2, 3 (SLC22A1-3). J Pharmacol Exp. 2004;308(1):2-9.
  14. Jonker JW, Wagenaar E, Mol CA, et al. Reduced hepatic uptake and intestinal excretion of organic cations in mice with targeted disruption of the organic cation transporter 1(OCT1, SLC22A1) gene. Mol Cell Biol. 2001;21(16):5471-5477.
  15. Otsuka M, Matsumoto T, Morimoto R, Arioka S, Omote H, Moriyama Y. A human transporter protein that mediates the final excretion step for toxic organic cations. Proc Natl Acad Scid USA. 2005;102(50):17923-17928.
  16. Terada T, Inui K. Physiological and pharmacokinetic roles of H+/organic cation antiporters (MATE/SLC47A). Biochem Pharmacol. 2008;75(9):1689-1696.
  17. Slager RE, Newton TL, Vlangos CN, Finucane B, Elsea SH. Mutation in RAI1 associated with Smith Magenis syndrome. Nat Genet. 2003;33(4):466-4668.
  18. Crain CA. An assessment of obesity and hyperphagia in individuals with Smith-Magenis syndrome. The University of Texas MD Anderson Cancer Center UTHealth Graduate School of Biomedical Sciences Dissertations and Theses (Open Access) website. https://digitalcommons.library.tmc.edu/utgsbs_dissertations/39. Accessed June 27, 2021.
  19. Li CL, Pan CY, Lu JM, et al. Effect of metformin on impaired glucose tolerance. Diabet Med. 199;16(6):477-481.
  20. Rodrigues RR, Gurung M, Li Z, et al. Transkingdom interactions between Lactobacilli and hepatic mitochondria attenuate western diet-induced diabetes. Nat Comm. 2021;12(1):101.