The Essential Guide to Herbal Safety by Simon Mills & Kerry Bone. St. Louis, MO: Elsevier Churchill Livingstone; 2005. 703 pages, hardcover. ISBN: 0-443-07171-3. $59.95. ABC catalog # B535.
Written by two esteemed clinicians, researchers, and authors, and in association with nine others, this book “contains comprehensive reviews of the published safety data for 125 common herbs.” These make up 420 pages in Part 2 and serve as referenced summaries on these herbs’ potential risks. The foundational portion of the book is in the 220 pages of Part 1 comprised of 12 chapters on safety issues. Half of these chapters are written entirely or in part by the other 9 contributors. In several early chapters, Mills addresses several aspects specific to individuals, including reactions due to nocebo (adverse placebo) and idiosyncratic (unexpected, unpredictable) responses.
The cytochrome P450 (CYP) enzyme system that metabolizes drugs is the focus of Mills’ chapter 5, “Human-Plant Interactions.” This highly variable set of CYP enzymes are responsible for the metabolism of many conventional drugs, as well as numerous other xenobiotic compounds, that may find their way into the human digestive system. As an overview of CYP isozymes and their roles, this chapter illustrates how different types of plants and their derivative preparations and components influence drug metabolism. Based on the criteria in Box 6-1 (“Rational Classification of Potential HDIs” [HDIs = Human-Drug Interactions]), many of the natural product influences listed in Table 5-1 must be considered either highly speculative, inaccurate, or misleading. The table provides short lists of drug substrates, inducers, and inhibitors for major isozymes, along with a few supposedly inducing and inhibitory herbs and several components.
There is a growing body of safety-related literature based on scientific and medical research on medicinal plants and plant-derived compounds. Reviewing and acknowledging the source and type of each report/study help in attempts to determine levels of clinical relevance. Unfortunately, Table 5-1 and the text of this chapter put excessive reliance on in vitro research that discredit these portions’ reliability for some of the herbs and phytochemicals listed. In so doing, the text is unclear about the types of data being utilized. For example, ginkgo (Ginkgo biloba L., Ginkgoaceae) and St. John’s wort (Hypericum perforatum L., Clusiaceae) are listed as inhibitors of the isozyme CYP2C19, based upon cited in vitro research on isolated gingkolic acids and hyperforin.1 Human studies show that 280 mg/day standardized ginkgo extract EGb 761 (W. Schwabe Pharmaceuticals, Karlsruhe, Germany) taken for 12 days2 and 900 mg/day St. John’s wort (SJW) extract with 4% hyperforin after 14 days likely induce CYP2C19.3 This table also lists diallyl sulphide and its herb source, garlic, as inducers of CYP2E1, while the cited studies show they are in vivo inhibitors, as can be extensively documented. There are other such errors. Isothiocyanates from herbs and vegetables in the family Brassicaceae—such as horseradish (Amoracia rusticana P. Gaertn. et al), mustard (Brassica juncea [L.] Czernov var. rugosa Bailey), broccoli (Brassica oleracea L., Cymosa group), and cabbage (Brassica oleracea L. var. capitata L.)—are also listed (with no citation) as CYP2E1 inducers. The two mentioned in the text are both CYP2E1 inhibitors (phenethyl isothiocyanate in vivo and sulforaphane in vitro),4,5 and the text citation referred to a CYP1A study.
Failing to identify CYP isozyme inhibitory studies in Table 5-1 and elsewhere as performed in vitro can be misleading. In vitro research is cited to indicate that garlic is a CYP3A4 inhibitor. However, this is contradicted by a study that shows humans taking garlic twice daily for 3 weeks had 50% reduced plasma saquinavir content from induced CYP3A4 metabolism.6 In addition, alprazolam and midazolam metabolism by CYP3A4 in humans was unaffected by garlic standardized extract (Kwai®, Lichtwer Pharma, Berlin) or garlic oil, respectively.7,8 An egregious example of how in vitro inhibitory studies are misrepresentative of clinical effects is the citation used in Table 5-1 for “various herbs” as CYP3A4 inhibitors. This cited in vitro study found 18 of 21 common commercial tinctures were inhibitors. The two strongest were extracts of goldenseal (Hydrastis canadensis L. Ranunculaceae) and SJW.Yet Table 5-1 correctly indicates the well-known fact that SJW is a CYP3A4 inducer in humans. Goldenseal root has shown no significant effect on CYP3A4 metabolism in humans.9 The importance of not relying on in vitro projections can be further illustrated with another study by the same research group. In this case water extracts of 6 culinary spices, 8 medicinal herbs, 5 varieties of black tea (Camellia sinensis L., Theaceae), 7 herb combinations, and 7 soybean (Glycine max [L.] Merr., Fabaceae) varieties were studied in vitro. At least 3 isozymes among CYP 2C9, 2C19, 2D6, and 3A4 were inhibited by > (greater than) 65% for the herb formulas, > 70% for the herbs, and 80% to 99% for the spices and black teas, while soy extracts tested only on CYP 3A4 inhibited it by > 75%.10 Though the authors did not mention specific herbs from the former cited study by name or include this latter study in the book, concerns involving extrapolation of this type of in vitro research need to be expounded for consideration.
Use of in vitro metabolism and interaction models normally employed for isolated drugs appears inadequate for complex herbal preparations for which these models have not been validated. Inhibitor concentrations of some phytochemicals tested in vitro are often much higher than the same compound would be in the plasma of a human, assuming that it was able to move through the first-pass metabolism in the intestine and liver and survive without some form of chemical modification. Also, in vitro exposure to the entire complex phytochemistry simultaneously does not mimic in vivo effects due to variable absorption, metabolism, and distribution of herbal constituents. In this regard the tannin content of any plant preparation is crucial. Tannin-protein binding with isozymes in vitro can greatly reduce CYP isozyme activity.11 Furthermore, the reduction of a ferric ion to ferrous is necessary in the catalytic cycle of CYPs.12 Interference with this process in vitro could involve many herbs, as a human study found concurrent consumption of ferric chloride and black tea with tannins or herb teas with phenolic acids or flavonoids bind the ferric ions and reduce their bioavailability.13 The ubiquitous presence of polyphenolic phytochemicals make in vitro inhibition assays for herb extracts on CYP isozymes inappropriate for assessing drug metabolism effects in humans. This issue has not been widely recognized or acknowledged in scientific circles where research protocols typically apply the same methods for testing multicomponent herb preparations as they do to unimolecular drugs. Considering the growing demands for screening botanical products for potential interactions with these unrealistically simple cytochrome P450 inhibition assays, the uncritical acceptance of such questionable findings in this case is unfortunate, even if understandable.
The reviews in chapter 6 of adverse herb-drug interactions were written by both the principal authors and associates Michelle Morgan and Berris Burgoyne. The chapter states that misinformation on risks typically arise from speculative inaccuracies, often from the infamous and highly unreliable 1998 Archives of Internal Medicine article “Herbal Medicinals.”14 A glaring example is the contention that all phytochemical coumarin derivatives are potential anticoagulants, a view I also mistakenly propounded at that time for the prodrug coumarin (Mea culpa!). This myth is still being perpetuated though it is well established that, unlike synthetic pharmaceutical 4-hydroxycoumarins (with a dicoumarol structure like warfarin), neither natural coumarin itself nor its botanical derivatives are anticoagulants, although a few coumarins found in plants do have some antiplatelet activity. Tables 6-1 & 6-2 provide subjective recommendations in regard to documented and theoretical herb-drug and drug-herb interactions. These summary charts provide useful reference points from which to make clinical judgments. Chapter 6 includes short discussions of SJW, ginkgo, garlic, Asian ginseng (Panax ginseng C.A. Meyer, Araliaceae), and kava (Piper methysticum Forst., Piperaceae). The discussion on the pharmacokinetic influence of SJW is hindered by limited attention to the role of P-glycoprotein and its drug substrates, aside from cyclosporin. For example, critical information and its implications were omitted on how different SJW preparations affect P-glycoprotein and resultant digoxin levels.15 This failure is especially surprising, since I first learned of the variable effects of distinctive SJW preparations from a report by Bone in 2001 on a phytomedicine conference in Munich.16
Chapters that follow cover herb preparations in relation to pregnancy/lactation and allergic reactions. The discussion on pregnancy does not confront the issue of excessive dosage, a central safety issue in regard to relative effects. It also downplays the traditional cross-cultural use of herb extracts as emmenogogues and abortifacients. Instead, it emphasizes findings from a few limited studies and data (or lack thereof) on direct fetal damage, designating gradations of relative risks. Other chapters appropriately address products of poor quality due to substitution, contamination, and adulteration, especially Chinese herbal medicines, and the need and means for herbal pharmacovigilance in monitoring and reporting adverse events. A chapter by Bone, Morgan, Janice McMillan, and Mathias Schmidt, reviewing reports on kava hepatotoxicity in a case-by-case assessment, concludes Part 1 as an illustration of governmental mishandling of risk assessments. This kava chapter is especially thorough in considering different analyses of the factual data, describing possible rationales, and comparing relative risks of kava products and similar pharmaceuticals. However, specific cultivars and chemotypes with differing kavalactone makeup potentially bearing on this issue were barely mentioned.
The reviews in the second part are where the rubber meets the road. These provide a clinical guide for practitioners, based on analysis of the literature data on the 125 herbs and their phytochemicals by Mills, Bone, Morgan, and McMillan. Herb monographs are arranged alphabetically by common names, though not necessarily those found in AHPA’s Herbs of Commerce 2nd ed (e.g., “Siberian ginseng” is used for Eleutherococcus senticosus, rather than the more appropriate term eleuthero, as required by law in the United States; since the authors are from the UK and Australia, respectively, one can understand this nomenclatural preference). The research evaluations provide realistic assessments and suggestions to minimize risks. The authors do a laudable job of specifying the form of the herb preparation being discussed, though occasionally lapsing into giving only the herb’s common name when referring to a derivative preparation, as seems to be the universal habit. Individual herb reviews’ content include safety and therapeutic summaries, actions and key constituents, adulteration, typical forms and doses, contraindications, warnings and precautions, machine operation, adverse reactions, interactions, use in children, overdosage, toxicology, regulatory status, and references citations.
The type of content for each of these categories varies, based on available data. Adverse reactions and interactions are gleaned from clinical studies and case reports. Emphasis in toxicology is on in vitro and animal studies. For example, toxicology tables typically include data from animal studies noting the lethal dose, often administered by injection, in 50% of the subjects sometimes for noncommercial extracts and frequently for fractions or components. This standard toxicological and mechanistic data, though fragmented and partial, can be useful for making relative comparisons. Often, subjective commentary is provided to help put the laboratory data into context.
In such a rapidly evolving body of knowledge, it can be easy to perceive shortcomings such as missing research or find fault with perspectives on controversial issues addressed in the introductory chapters or safety monographs. Given the inherent limitations of available research on this subject matter, this book presents an informed and balanced discussion of important issues that face every clinical practitioner in this age of uncertainties pertaining to pharmacological and toxicological activities of many herb products. An amazing amount of effort has obviously gone into gathering, organizing, and editing this remarkable work. Considering the wide scope of literature on botanicals use and research, a few factual inaccuracies or omissions can be forgiven. Aside from reservations about drug metabolism data interpretations and my own preferences on points of emphasis, the book on a whole is to be commended for its thorough coverage, in-depth content, intelligent discussions, and engaging descriptions of issues that can otherwise be perceived as staid. It stands as an important contribution on the subject of the safe use of herbal preparations as understood at this time.
—Francis Brinker, ND Clinical Assistant Professor, Program for Integrative Medicine, College of Medicine, University of Arizona, Tucson; Author of Complex Herbs—Complex Medicines; Herb Contraindications & Drug Interactions, 3rd ed; and Toxicology of Botanical Medicines, 3rd ed
1. Zou L, Harkey MR, Henderson GL. Effects of herbal components on cDNA-expressed cytochrome P450 enzyme catalytic activity. Life Sci. 2002;71:1579-1589.
2. Yin OO, Tomlinson B, Chow MS. Prediction and mechanism of herb-drug interaction: Effect of Ginkgo biloba on omeprazome in Chinese subjects. Clin Pharmacol Ther. 2003;73(2):P94.
3. Wang L-S, Zhou G, Zhu B, et al. St John’s wort induces both cytochrome P450 3A4-catalyzed sulfoxidation and 2C19-dependent hydroxylation of omeprazole. Clin Pharmacol Ther. 2004;75:191-197.
4. Li Y, Wang EJ, Chen L, et al. Effects of phenethyl isothiocyanate on acetaminophen metabolism and hepatotoxicity in mice. Toxicol Appl Pharmacol. 1997;144(2):306-314.
5. Barcelo S, Gardiner JM, Gescher A, et al. CYP2E1-mediated mechanism of anti-genotoxicity of the broccoli constituent sulforaphane. Carcinogen. 1996;17(2):277-282.
6. Piscitelli SC, Burstein AH, Welden N, et al. The Effect of Garlic Supplements on the Pharmacokinetics of Saquinavir. Clin Inf Dis. 2002;34(1):234-238.
7. Gurley BJ, Gardner SF, Hubbard MA, et al. Cytochrome P450 phenotypic ratios for predicting herb-drug interactions in humans. Clin Pharm Ther. 2002;72(3):276-287.
8. Markowitz JS, DeVane CL, Chavin KD, et al. Effects of garlic (Allium sativum L.) supplementation on cytochrome P450 2D6 and 3A4 activity in healthy volunteers. Clin Pharmacol Ther. 2003;74(2):170-177.
9. Sandhu RS, Prescilla RP, Simonelli TM, et al. Influence of Goldenseal Root on the Pharmacokinetics of Indinavir. J Clin Pharmacol. 2003;43:1283-1288.
10. Foster BC, Vandenhoek S, Hana J, et al. In vitro inhibition of human cytochrome P450-mediated metabolism of marker substrates by natural products. Phytomed. 2003:10:334-342.
11. Butterweck V, Derendorf H, Gaus W, et al. Pharmacokinetic Herb-Drug Interactions: Are Preventive Screenings Necessary and Appropriate? Planta Med. 2004;70:784-791.
12. Hollenberg PF. Characteristics and common properties of inhibitors, inducers, and activators of CYP enzymes. Drug Metab Rev. 2002;34(1&2):17-35.
13. Hurrell RF, Reddy M, Cook JD. Inhibition of non-haem iron absorption in man by polyphenolic-containing beverages. Br J Nutrition. 1999;81:289-295.
14. Miller LG. Herbal medicinals—selected clinical considerations focusing on known or potential drug-herb interactions. Arch Intern Med. 1998;158:2200-2211.
15. Mueller SC, Uehleke B, Woehling H, et al. Effect of St. John‘s wort dose and preparations on the pharmacokinetics of digoxin. Clin Pharmacol Ther. 2004;75(6):546-557.
16. Bone K. Conference report: Third International Congress on Phytomedicine. Br J Phytother. 2001;5(4):209-217.
Mills and Bone Respond
This review is clearly well informed and we are grateful for the detailed and generally positive critique. However there does seem to be an undue focus on chapter 5. We here draw the reader’s attention to the potential issue of interactions between natural and synthetic substances at the cytochrome P450 enzyme level. The reason we do this is our view that we do not want to be caught off guard by another St. John’s wort situation—indeed it leads us to suggest that interactions might happen between almost any ingested substances, and to the political point that therefore the problem is more with the most powerful drugs like digoxin and cyclosporin (with “narrow therapeutic windows”) than with the wide diversity of natural materials in foods and herbs. However we are also quite clear that these are potential interactions only; i.e., they are not necessarily documented case reports. Perhaps the reviewer overlooked our disclaimer (page 43):>
“It is important to note that in the majority of cases the evidence for these reactions follows laboratory work, in vitro studies, and investigations of the effects of agents in rats (notably Sprague-Dawley rats). Many of the molecules studied are unlikely to reach the body’s tissues after oral consumption and digestion and liver metabolism. Some may not be absorbed, whereas others may reach the liver in the first pass effect from the portal system. Therefore, such observations only raise the possibility of interactions and activities in real life. Nevertheless, based on such interactions, since 1998 the Food and Drug Administration has removed or restricted terfenadine, mibefradil, astemizole, grepafloxacin, and, notably, cisapride.”