Williamson, EM. Drug Safety 2003. Herbal Drug Interactions

Drug Safety 2003; 26 (15): 1075-1092REVIEW ARTICLE 0114-5916/03/0015-1075/$30.00/0

Adis Data Information BV 2003. All rights reserved.

Drug Interactions Between Herbal andPrescription MedicinesElizabeth M. Williamson

The School of Pharmacy, Centre for Pharmacognosy and Phytotherapy, University of London,London, United Kingdom

ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10761. Mechanisms of Drug Interactions with Herbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1077

1.1 Pharmacokinetic Interactions Involving Herbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10771.2 Pharmacodynamic Interactions Involving Herbs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1078

2. Problems Monitoring Herb-Drug Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10782.1 Definition of Herbal Medicines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10782.2 Retrieval of Herb-Drug Interaction Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10782.3 Classification of Herb-Drug Interaction Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1079

3. Herb-Drug Interaction Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10793.1 The Cardiovascular System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1079

3.1.1 Warfarin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10793.1.2 Digoxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10813.1.3 Calcium Channel Antagonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10813.1.4 ACE Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10823.1.5 Antiplatelet Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1082

3.2 The Central Nervous System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10823.2.1 Antidepressants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10823.2.2 Antipsychotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10833.2.3 Lithium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10843.2.4 Levodopa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10843.2.5 Acetylcholine Receptor Antagonists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10843.2.6 Antiepileptic Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1085

3.3 Pain and Inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10863.3.1 Salicylates in Herbs and Prescription Medicines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10863.3.2 Other Unconfirmed Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1086

3.4 Sex Hormones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10863.4.1 Confirmed Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10863.4.2 Theoretical and Unlikely Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1086

3.5 Diabetes Mellitus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10873.5.1 Theoretical and Unlikely Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10873.5.2 Potential Additive Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1087

3.6 Antibacterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10873.7 Antiviral Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10883.8 Immunosuppressants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10883.9 Cancer Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10883.10 Miscellaneous Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10893.11 Herbs and Surgery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1089

4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1090

1076 Williamson

Until reports of interactions between St John’s wort and drugs such as digoxin,Abstractwarfarin, protease inhibitors and oral contraceptives began to appear, very fewherb-drug interactions were documented. These are now becoming more com-mon, although still rare compared with drug-drug interactions. In the absence ofhard data, potential interactions are being highlighted, and this review attempts todistinguish between the speculative and the proven. The subject is approachedfrom a therapeutic point of view since in most cases the patient is already takingone or more prescription drugs, and the question is whether or not it is safe for aparticular herb to be added to the regimen. Although many of the examples ofherb-drug interactions are minor or theoretical at present, the fact remains thatsome are serious and life threatening, and these almost exclusively concerncyclosporin, anticoagulants, digoxin, antidepressants and protease inhibitors, tak-en with the herb St John’s wort. Ginkgo and ginseng are implicated in a number ofreports, but many of these are unsubstantiated. To date, the cardiovascular, centralnervous and immune systems are the most common therapeutic categories cited inthe literature and other than those, examples are very limited. Although manyherbal drugs have good safety profiles, it must be borne in mind that herbalsupplements are intended to be taken over an extended period of time, whichprovides the opportunity for enzyme induction and other mechanisms of inter-action to take effect.

“The data on interactions are of widely varying manufacturers, experts and herbal organisations,quality and reliability. Sometimes they are no more covering the period 1990–2000. This is a smallthan speculative and theoretical scaremongering proportion of drug interaction reports, and of these,guesswork, hallowed by repeated quotation until they could only classify 13% as ‘well-documented’they become virtually set in stone” – Stockley.[1] and 18.5% as ‘possible’ interactions; the remainderThese statements have been applied to all types of were ‘unevaluable’. With the extent of the problemdrug interaction reports, but are probably even more still unknown, and in the absence of hard data,appropriate in the case of herbal medicines and potential interactions are being highlighted in-prescription drugs. These will be referred to as stead.[9-13] This approach has both benefits and‘herb-drug’ interactions for convenience, although drawbacks. It acts as an alert to a potential inter-herbs are of course, in this context, drugs. Very few action, and if the hypothesis is used properly theherb-drug interactions were documented until re- evidence can then be examined, and can exoneratecently, when reports of interactions between St as well as implicate the herb in an interaction. ForJohn’s Wort and drugs such as digoxin, warfarin, example, Miller and Murray[10] state that “echinaceaprotease inhibitors and oral contraceptives began to may interfere with immunosuppressive therapy”,appear. They are now becoming more common, Barnes et al.[11] suggest that saw palmetto “maywith the increasing use of complementary therapies interfere with existing hormone therapy” and Ang-and a greater awareness of the potential problem, Lee et al.[12] state that valerian “could increase thealthough they are still rare compared with drug-drug sedative effects of anaesthetics”. In contrast, wheninteractions. Fugh-Berman and Ernst[3] reviewed literature re-

ports with the aim of assessing the quality of theSeveral reviews on herb-drug interactions haveclinical evidence, they found no reports of interac-been published recently[2-8] but in all cases thesetions between conventional treatments and echina-have been hindered by the lack of reliable documen-cea, valerian or saw palmetto.tary evidence. For example, Fugh-Berman and

Ernst[3] identified 108 cases of suspected interaction The Mayo clinic website (www.MayoClinic.by searching the major databases and contacting com) provides a list of herbs and specifies drugs that

Adis Data Information BV 2003. All rights reserved. Drug Safety 2003; 26 (15)

Drug Interactions Between Herbal and Prescription Medicines 1077

should not be taken concurrently, and is a useful site, affected by the presence of another, at the site ofparticularly for consumers. However, to ensure that action (sometimes called pharmacological interac-patients are well protected, the information provided tions), which are less predictable and more difficulton the website includes some potential, unproven to classify.and theoretical interactions, and is not referenced to

1.1 Pharmacokinetic Interactionsthe scientific literature. It is widely recognised thatInvolving Herbsmore research is urgently needed, and it is also

extremely important that the identity and constitu- Most interactions affecting absorption usually re-tion of the herbal drug is known accurately, because sult in a reduction of the absorption of the drug,of the possibility of substitution, contamination or although increases in absorption can occur. Changesother forms of adulteration. in intestinal pH will affect absorption, as will com-

Most articles on the subject of herb-drug interac- plexing mechanisms and drugs affecting gastroin-tions approach it from the point of view of the herb, testinal motility. Herbal drugs are more likely toand those prescription drugs that may interact with inhibit absorption by forming complexes, for exam-the herb. However, in most cases the emphasis is on ple with metal cations such as calcium, tannins andthe prescription drugs already being taken by the polyphenols in water extracts. In practice, this maypatient, and whether it is safe for a particular herb to not be significant – after all, tea, which contains allbe added to the regimen. In this review, therefore, of these, is often used by patients to help swallowdrug interactions will be discussed according to drug tablets, and does not appear to render many drugstherapeutic type, and a distinction made between inactive. Drug displacement from protein-boundthose interactions for which evidence is available, forms, by concurrent drug administration, causes anand those that are unreliable or purely speculative. A increase in serum drug levels and therefore an in-very brief outline of mechanisms will be given so crease in the therapeutic effect. This is very impor-that a particular type of interaction can be put in tant in the case of antiepileptic agents for example,context. but no clinical cases have been reported for herbal

drugs. Stockley[1] considers that this mechanism of1. Mechanisms of Drug Interactionsinteraction “has been grossly over-emphasised” be-with Herbscause in vitro studies are not necessarily reflected bywhat happens in the body.Many herbal products are treated as ‘food supple-

ments’ for legal and regulatory purposes, although Drug metabolism is the most important mecha-strictly speaking they do not fall into that category, nism so far reported for herbs, and the best knownand here they will be considered as medicines. Food example is St John’s wort. If the cytochrome P450intake – time, quantity and content – can certainly (CYP) enzymes which metabolise a particular drugaffect drug treatment and this subject has been well are induced, they will obviously remove it from thereviewed recently by Schmidt and Dalhoff.[2] There circulation more rapidly, and if another drug isare some high profile food-drug interactions, partic- present which involves the same enzymes, that tooularly involving grapefruit juice, and drug absorp- will be metabolised faster. This is a very commontion (and therefore bioavailability of the drug) seems mechanism, and applies to the way in which Stto be the main parameter affected by the presence of John’s wort may reduce the efficacy of the oralfood in the stomach. Similar principles will apply to contraceptive pill or blood levels of warfarin, digox-herbs, which often contain similar compounds, al- in or theophylline.[14] The opposite may apply,beit in much higher concentrations. Mechanisms of where enzyme inhibition leads to increased druginteractions are necessarily complex, and more than levels, which can be an advantageous mechanismone may be involved. They are usually divided into unless drug serum levels reach toxic levels. Changespharmacokinetic and pharmacodynamic interac- in excretion will affect serum drug levels, and diure-tions, the former being processes involving absorp- tics are often cited as causing this effect. However,tion, distribution, metabolism and excretion, and the herbal diuretics are very weak (compared withlatter being those where the effects of one drug are furosemide for example), and unlikely to cause any


1078 Williamson

problems. Most herbs do not change urinary pH to available without prescription, it is expected to beany large extent either; this is true for cranberry licensed soon for use in multiple sclerosis. Of coursejuice, which is used to treat cystitis. P-glycoprotein the illicit use of cannabis is very widespread indeed,is a pump mechanism, which has a considerable and similarly with tobacco, not as a medicine but aeffect on serum drug levels by blocking or facilitat- widely used drug.ing entry into cells. This is an important mechanism

2.2 Retrieval of Herb-Drugand the subject of an increasing number of reports,Interaction Reportsincluding one that implicates St John’s wort.[14]

Reports of herb-drug interactions are sparse (al-1.2 Pharmacodynamic Interactions

though they are increasing rapidly) for a number ofInvolving Herbs

reasons, and reliable figures of such interactions arenot available.[16] Patients may not tell their practi-Additive, synergistic or antagonistic effects oftioner that they have taken the herbal drug for fear ofcombinations of drugs are possible with any type ofdisapproval, or because they do not consider herbsdrug, including herbals. Therefore herbal sedatives,to be ‘drugs’. An interaction may not be recognisedanticoagulants, antihypertensives and others mayas such, and may only be reported if it is consideredpossibly increase the effect of a concurrent conven-‘serious’. It is difficult to assess the incidence oftional drug taken for the same purpose. Given thatherb-drug interactions, and even more difficult todrug combinations such as these are routinely givenassess the reliability of these reports, which oftenin medicine (e.g. antihypertensives, cancerinvolve only one patient or is a theoretical case withchemotherapies, and anti-HIV agents) and drugsno evidence, or an animal or in vitro experimenttend to be much more potent than herbs, the clinicalwhich is not relevant to a human clinical situation.importance of this type of interaction has not yetData are now accumulating, and there may soon bebeen properly evaluated. Reports at present appearmany such reports, although it is also possible thatto be confined to the serotonergic syndrome report-herb-drug interactions are either minor or tend toed for antidepressants taken with St John’s wort. Ininvolve drugs already notorious for their potential toorder to predict this type of interaction, a certaininteract.amount of basic pharmacological knowledge is

The following strategy was used initially to sur-needed about both herb and drug, but unfortunatelyvey the literature: the literature databases Embasethis is often lacking for the herb.and Medline were accessed from their inception

2. Problems Monitoring (1980 and 1966, respectively) until May 2003, usingHerb-Drug Interactions as key phrases ‘herb-drug interaction’, ‘herbal medi-

cine toxicity’, and ‘herbal pharmacokinetics’. Thesearch term ‘food-drug interaction’ was also includ-2.1 Definition of Herbal Medicinesed, to identify any interactions in which herbal med-

The definition of herbal medicines has been icines may have been classified as food supple-loosely interpreted, as a complex mixture of sub- ments. Standard references, particularly Stockley’sstances extracted from plant sources. Although caf- Drug Interactions 6th edition[1] and others[10,11,15]

feine, tobacco and cannabis are not normally classi- were checked, and the references cited in these, andfied as such, they have been included for the follow- also the reviews identified by the computer search-ing reasons: (i) caffeine is present in many stimulant es[2-9,12] were consulted, to attempt to rectify anyherbal products, in the form of cola and guarana, and omissions by the abstracting services. Report relia-are often sold as tonics; (ii) guarana may be included bility was difficult to assess; the scoring systemin herbal products for energy, and referred to as a described by Fugh-Berman and Ernst[3] which, al-natural stimulant from the Amazon containing though not validated, appeared to be the most useful.guaranine, to disguise the fact that guaranine is However, it proved to be impossible to apply thisactually identical to caffeine;[15] and (iii) cannabis is system to a large number of cases due to lack ofnot a herbal medicine, and although it will not be information given in them, a factor noted by those



authors, who classified such reports as unevaluable. are known for their many interactions due to anIt was also not possible to tabulate observed interac- inhibition of CYP enzymes. These drugs are potent,tions at this time without giving credence to too have a narrow therapeutic window and changes inmany unreliable reports. serum levels can have serious consequences. Other

interactions are known and will be outlined.2.3 Classification of Herb-Drug

3.1.1 WarfarinInteraction ReportsThe main reports concern an enhancement of the

The evidence available for each report has been effect of warfarin either by decreasing clearance andoutlined where possible, and an attempt made to therefore increasing bioavailability, leading to animplement a broad classification using the following elevation of the International Normalized Ratiodefinitions: (INR) and prothrombin time (PT) with concurrent

herbal use. At present no fatalities have been record-• Confirmed: interactions that are backed up byed, but this is probably due to the fact that thesedata on a number of reports, a rechallenge of thepatients are closely monitored in the clinic and theco-administered drug resulting in a similar effect,dose of warfarin adjusted routinely. Notable interac-or knowledge of the mechanism of action. Exam-tions occur between warfarin and several herbs usedple: St John’s wort (Hypericum perforatum) andin traditional Chinese medicine, such as donghormonal contraceptives (see section 3.4.1).quai,[17] dan shen (Salvia miltiorrhiza)[18] as well as• Potential: possible interactions with a sound the-garlic (Allium sativum) and ginkgo.[19]oretical basis and one or more reports already

received. Example: dong quai (Angelica sinensis Confirmed Interactionsroot) and warfarin (see section 3.1.1). St John’s wort is a popular herbal medicine taken

• Theoretical: with a more tenuous theoretical ba- for mild to moderate depression. It has been shownsis, and no clinical reports. This category is re- to interact with a number of drugs, including warfa-sponsible for most of the ‘theoretical scaremon- rin, mainly due to an induction of CYP2C9. Itgering guesswork’ described by Stockley.[1] Ex- caused a decrease in the anticoagulant effect of bothample: echinacea and immunosuppressants (see warfarin (seven cases) and phenprocoumon (onesection 3.8). case),[20] although paradoxically in another case, it

• Unlikely: a report of an adverse event occurring caused an increase in anticoagulant effect in onewhen two drugs are taken concurrently or in an patient taking phenprocoumon.[20] The UK Medi-animal experiment, but where theoretical consid- cines and Healthcare products Regulatory Agencyerations mean an interaction is unlikely or not (MHRA) [formerly the Medicines Control Agency]translatable into a human situation. Example: has recommended that patients on warfarin shouldcapsaicin cream with an ACE inhibitor (see sec- not take St John’s wort and that those already doingtion 3.1.4). so should stop, and adjust their dose of anticoagu-

• A single report: exactly that; unevaluable. Ex- lant.[21]

ample: ginkgo (Ginkgo biloba) and trazodonePotential Interactions(see section 3.2.1).Dong quai is used as a hormone regulator in

women for disorders such as painful or suppressed3. Herb-Drug Interaction Reportsmenstruation and symptoms associated with meno-pause, although it is considered to be devoid of3.1 The Cardiovascular Systemestrogenic activity itself.[15] Dong quai, together

Patients receiving anticoagulant therapy are usu- with warfarin, produced an increase in INR and PTally very cautious about concurrent drug use, al- in a single case report in one patient[17] and a phar-though they may not inform the prescriber of their macodynamic (but not pharmacokinetic) interactionherbal medicine consumption, for reasons already has also been documented for the same combinationmentioned. The majority of reports concern warfarin in rabbits.[22] Although dong quai is also used byand the cardiac glycosides (mainly digoxin) which traditional Chinese medicine practitioners for


1080 Williamson

cardiovascular indications, a study showed that it rin, leading to an increase in INR. The patient, aconferred no benefit to stroke patients.[4] 61-year-old woman, was also taking digoxin, ate-

nolol and fluvastatin.[25] Other single reports includeHaemorrhage has been reported in two patientsthe case of a woman on warfarin taking a herbaltaking a combination of warfarin with dan shen[18]

product containing boldo (Peumus boldo) together(which decreases warfarin clearance). Although inwith a separate product containing fenugreekthese instances the patients were also taking other(Trigonella foenum-graecum). The patient’s INRdrugs (furosemide and digoxin in both cases), thewas elevated modestly but no undesirable reactionsinteraction was ascribed to the dan shen component.such as bleeding or bruising were observed.[26] Simi-A Chinese herbal formulation, kangen-karyu, whichlarly, two patients taking the herbal remedyis a mixture containing dan shen, peony root, saf-curbicin, for prostate enlargement (containingflower, saussurea root, and cnidium and cyperuspumpkin seed, Cucurbita pepo, and saw palmetto,rhizomes, is traditionally used to treat hypertensionSerenoa repens) showed elevated INR values al-and atherosclerosis, with proven clinical efficacy.[18]

though only one was also taking warfarin.[27] Qui-A desire by some patients in Japan to take a combi-nine, derived from cinchona bark, is occasionallynation of kangen-karyu with warfarin led to an in-found in herbal products as well as carbonatedvestigation of the pharmacokinetic interactions indrinks. An interaction has been reported in twomice, although these have not been documented inwomen receiving warfarin, and a man taking phen-any clinical reports. This showed that kangen-karyuprocoumon, who were drinking large quantities ofsignificantly suppressed the metabolism and elimi-tonic water containing quinine. Stockley suggestsnation of warfarin, although serum binding and ab-that this is of little or no clinical significance.[1]sorption were not affected.[18] The authors conclud-

ed that the combined use of these drugs may be Many other items of food are used in concen-therapeutically advantageous provided that coagula- trated form as health supplements, and several au-tive status was checked regularly. thors suggest a theoretical interaction between gin-

ger (Zingiber officinalis) and warfarin.[19,23] Againthis is unsupported by clinical evidence. Other un-Theoretical Interactions and Single Reportssubstantiated claims are that psyllium seedGarlic, a common supplement taken to improve(Plantago psyllium) and ispaghula husk (Plantagocardiovascular health has been reported (rarely) toovata), taken for the fibre content as a bulk laxative,cause an increase in INR when taken with warfa-may inhibit the absorption of warfarin, but a smallrin.[19] Ginkgo (taken to improve the circulation andstudy in six healthy individuals showed no effect onhence cognitive function) and ginseng (a generaleither the absorption or anticoagulant effect.[28]tonic and adaptogen) have both been subjects ofCases have been described where a high intake ofreports of possible interactions, although with nomango fruit together with warfarin caused an eleva-supporting data,[19] and Stockley concludes thattion in INR,[29] but the interaction is not importantthere is no real substance of such reports.[1] In theand patients need not avoid eating mangoes.case of ginseng, the combination resulted in a de-

crease in INR,[23] but the assessors of this report It is well known that dietary vitamin K, containedsuggest that causality is difficult to prove.[24] The in many green vegetables, can antagonise warfarinconsumption of ginseng is more widespread among (it is used as an antidote for warfarin overdose). Thisthe elderly population, who are most likely to be is suggested as a factor in the case of a patient ontaking warfarin, and these reports may be taken as warfarin who developed an elevated INR when con-an indication of a potential for interaction, rather suming large volumes of green tea (2–4L daily for athan confirmed, in view of the fact that there have week),[30] although this is an unlikely interaction, asbeen reports of spontaneous bleeding in patients it would be more logical to assume that the vitamintaking ginseng alone.[1] K would decrease the INR. A more likely scenario is

Recently, a single report has been documented that the polyphenols in the tea (given the largebetween a herb used in traditional Chinese medicine amount ingested) may have inhibited the absorption(Chinese wolfberry [Lycium barbarum]) and warfa- of the warfarin in some way. Possible interactions



for patients taking warfarin with, for example, fever- particular herb may be implicated in toxicity reportsfew (Tanacetum parthenium), for migraine prophy- with no foundation in truth.laxis, and ginkgo, as an aid to circulation, have been

3.1.2 Digoxinpostulated by several authors.[10,11] No clinical inter-The use of cardiac glycosides such as digoxin isactions have been reported to date but it would

decreasing because of the availability of safer drugsprobably be wise to avoid these combinations. Simi-for heart failure, but digoxin is still widely used andlarly, plants containing salicylates, such as meadow-poses a potential problem for patients wanting tosweet (Filipendula ulmaria) and willow bark (Salixtake herbal remedies.spp.)[15] should be avoided with anticoagulants.

It has been confirmed that the excessive use ofCloser monitoring of INR would probably be advis-laxatives, including the herbal drugs senna (leavesable for these patients also taking herbal medicines,and pods of Cassia senna and other Cassia spp.) andand this may also give information as to the extent ofcascara (Rhamnus purshiana), depletes blood elec-the problem.trolyte levels, particularly potassium, which contrib-

Potential Additive Effects utes to the toxicity of digoxin.[10,11]

Various herbs are used in both traditional Chi- There is clinical evidence that blood levels ofnese medicine and traditional herbal medicines for digoxin are reduced by the concurrent administra-cardiovascular problems. Natural anticoagulants in- tion of St John’s wort,[11,14] although, to date, noclude coumarin derivatives, contained, for example, therapeutic interactions between St John’s wort andin sweet melilot (Melilotus officinalis), tonka beans digoxin have been reported. If a patient insists on(Dipteryx odoratum) and sweet woodruff (Galium taking both he/she should be closely monitored.odoratum),[15] and obviously these may have addi- However, the UK Committee for the Safety of Med-tive effects with warfarin. A case study in a woman icines (CSM) recommends that St John’s wort andtaking a herbal tea containing all of these plants, digoxin should not be taken together.[11]

together with warfarin, showed an abnormal men- An elderly patient taking Siberian ginsengstrual bleeding and a reduction in clotting factors, (Eleutherococcus senticosus) with digoxin wasbut the PT returned to normal after she was given found to have grossly elevated digoxin levels, al-parenteral vitamin K.[1] Naturally-occurring couma- though no symptoms of toxicity were found.[33] Al-rins are only weakly anticoagulant, but if the plant though not everyone with high digoxin levels neces-material is not stored properly dicoumarol may be sarily has symptoms, another reason for this appar-formed by microbial transformation, and this com- ent contradiction might lie in the assay procedure forpound is much more potent. A list of herbs with blood digoxin levels, which is a serum immunoas-varying cardiovascular effects can be found in sev- say method. Both the active constituents have aeral texts,[4,7,10,11] but it must be borne in mind that steroidal chemical structure, which could interferemost of these have not yet been associated with any with the results of the assay.[1] However, it has beeninteraction reports. argued that the chemical structure of the compounds

present in Siberian ginseng are not very similar toA Case of Deliberate Adulteration that of digoxin, and a more likely explanation is thatA formulation called PC-SES, taken for prostate the ginseng had been adulterated with Periplocia

cancer, contains the Chinese herb, baical skullcap sepium, the constituents of which are closer in struc-(Scutellaria baicalensis), which is reputed to be an ture.[33] The substitution of Siberian ginseng hasanticoagulant. This product caused a massive in- been responsible for other adverse events ascribed tocrease in PT and associated clinical symptoms in a the herb, such as androgenicity.[34]

patient, although not in conjunction with any other3.1.3 Calcium Channel Antagonistsanticoagulant therapy.[31] On analysis, this product

was found to actually contain warfarin, and has now A theoretical interaction has been reported withbeen withdrawn from sale.[32] This example serves nicardipine and ginkgo. Ginkgo extract fed daily toas an illustration of a so-called herbal product, rats (0.5% w/w of the diet) was found to attenuatewhich is not what it appears to be, and how a the hypotensive response of nicardipine, by a mech-


1082 Williamson

anism involving the induction of CYP3A2 and other tivity, and has been associated with spontaneousliver metabolising enzymes.[35] The level of ginkgo bleeding when given alone without concurrentextract fed to the animals was much higher than medications. For example, a woman who had beenwould ever be when taken as a human supplement taking warfarin for several years also experienced(to put in perspective: daily consumption of 1kg intracerebral haemorrhage after initiating ginkgo.[39]

food would contain an intake of 5g extract, whereas Aspirin is also associated with similar reports.[40]

the recommended dose is 120mg). No clinical inter- Theoretically, feverfew, ginger, kava (kavain,actions have yet been reported, but caution should Piper methysticum - now largely withdrawn frombe taken with long-term, high dose ginkgo extract sale because of hepatotoxicity) and dong quai,supplementation, and those calcium channel ant- which have antiplatelet effects in vitro, may affectagonists which are metabolised by the same en- concurrent antiplatelet therapy,[10,11] but there are nozymes, i.e. nifedipine and diltiazem, as well as ni- reports published of clinical interactions. One fur-cardipine. ther example is that caffeine, present in cola and

guarana tonics (Cola nitida and Paullinia cupana,3.1.4 ACE Inhibitors respectively) as well as tea and coffee, antagonisesACE inhibitors are used for hypertension, and in the haemodynamic response to dipyridamole test-

the treatment of heart failure. In general, when taken ing, which may or may not be clinically relevant.[1]

with herbal medicines they appear to have a goodsafety profile, although a report has been made of 3.2 The Central Nervous Systeminteractions of garlic with lisinopril, by a medicalpractitioner. He described his own symptoms of 3.2.1 Antidepressantshypotension and fainting, which were restored to

Confirmed Interactionsnormal on cessation of the garlic.[36]

Mild depression may be self-treated by patientsA woman taking an unnamed ACE inhibitor ex-unwilling to consult a doctor for a condition per-perienced a cough after applying capsaicin creamceived to carry a stigma, and if so, a popular herbtopically as an analgesic.[37] Capsaicin is the pun-used to treat this condition is St John’s wort. If theregent ingredient in chilli peppers (Capsicum spp.),are no concurrent prescription drugs, St John’s wortand is included in several herbal products for coughshas a fairly good safety profile, but of course theand colds due to its expectorant and ‘warming’patient may be on other non-depression-relatedeffect. It is not known how much, if any, capsaicinmedication. It is also common for patients to addwas systemically absorbed. ACE inhibitors arethis herb to an existing antidepressant drug regimen,known to produce cough as an adverse effect and ifeither to boost the effect or with a view to replacingthis is indeed an interaction, which is not proven, itit with something natural on their own initiative, andis unlikely to be clinically significant.without consultation. This is where interactions can

3.1.5 Antiplatelet Drugs arise, and have been recorded. The most common isAntiplatelet drugs do not seem to interact with a potentiation of serotonergic effects with concomi-

most herbal medicines, although there is the poten- tant specific serotonin uptake inhibitors (SSRIs)tial for an additive effect with, for example, herbs such as sertraline, paroxetine and fluoxetine. In ad-containing salicylates, or ginseng or ginkgo, which dition to duplicating effects, it is also due to theare known to have antiplatelet activity.[11] inhibition of CYP2C9, CYP2D6 and CYP3A4 by

Conversely, an elderly man already taking aspi- this class of drugs. It is important to remember thatrin (acetylsalicylic acid) after coronary by-pass sur- the long half-life of fluoxetine, as well as the activitygery, developed spontaneous bleeding from the iris of its main metabolite norfluoxetine againstinto the anterior chamber of the eye and blurred CYP3A4, means the potential for interaction re-vision when he started taking ginkgo. The bleeding mains for several weeks after discontinuation.[41] Ststopped after cessation of ginkgo.[38] Ginkgo con- John’s wort also inhibits CYP3A4 (as well astains ginkgolides, which are unique terpene lactones CYP1CA2, CYP2CP,[42] and it also affects P-glyco-with specific platelet-activating factor antagonist ac- protein) which explains at least in part the mecha-



nism for this. A report of mania associated with An interaction between atropine and ami-triptyline resulting in mydriasis has been reported inconcomitant administration of sertraline and Sta neonate,[48] but this is probably a rare combinationJohn’s wort was considered to be serious.[43]

and of little clinical importance with regard to herbalMonoamine oxidase inhibitors (MAOIs) interactmedicines. Although atropine is a natural product,with reserpine (a rauwolfia alkaloid) in an interest-from Atropa belladonna, it is rarely used in itsing way, in that it depends on the order in which theherbal form unless by a qualified medical herbalist,drugs are given. Stockley[1] explains this interactionwho will be aware of the toxicity of the herb, andas follows: the rauwolfia alkaloids cause a releasemost unlikely to use it in babies. Tricyclic antide-and subsequent depletion of noradrenaline stores,pressants have been the subject of occasional inter-reducing the transmission of adrenergic nerve end-action reports but do not seem to pose a majorings, which leads to hypotension and depression. Inproblem with common herbal remedies.fact, an animal model of experimental depression

Trazodone is used in depression where sedationinduced by reserpine is used to test new antidepres-is required, particularly in the elderly. An interactionsant compounds. If the patient is already taking anresulting in coma was suspected in an Alzheimer’sMAOI, and rauwolfia is added, there may be apatient taking a low dose of trazodone with ginkgo,sudden release of accumulated noradrenaline (nore-but this is unproven.[49]

pinephrine) and serotonin, resulting in excessivePotential Additive Effectsstimulation of the receptors, manifesting as gross

central excitation and hypertension. It is referred to The root of rauwolfia, which contains reserpineand other alkaloids, was formerly used as a treat-as reserpine-reversal and has been reported forment for psychoses including schizophrenia, andphenelzine [44] and nialamide.[45] This reaction ishypertension. It is occasionally found in importedobviously undesirable, and if it is absolutely neces-medicines from Asia, where it is still used for thosesary to give both drugs, the MAOI should be givenpurposes and as a sedative. Rauwolfia itself canafter the reserpine or rauwolfia.cause depression so it is usually avoided in patientswith, or prone to, depression. However, it has beenTheoretical Interactions and Single Reportsshown that when taken in conjunction with the tricy-Two reports of theoretical interactions betweenclic antidepressants, imipramine and desipramine,an MAOI (phenelzine) and ginseng (Panax ginseng)there has been successful treatment of some formshave been noted.[46] In both cases, insomnia andof resistant depression in a few cases.[50] This hasheadache were seen. Ginseng is known to haveonly been attempted in a well controlled clinicalstimulant effects[11,15] but the mechanism of thissituation.interaction is not known, and MAOIs are known to

3.2.2 Antipsychoticsproduce insomnia and headache as adverse effects,even in the absence of other drugs. Potential Interactions

An isolated report of mania, considered to be due Several cases have documented lower bloodto fluoxetine interacting with cannabis has been levels of fluphenazine and chlorpromazine in pa-recorded.[47] This was attributed to the tetrahydro- tients who were heavy smokers (of both tobacco andcannabinol (THC) content, which is also an inhibitor cannabis) compared with that of non-smokers.[51-53]

of serotonin uptake. Although cannabis is not yet Generally, this is thought to be due to enzymeused as a medicine, it is widely used illicitly. Also induction as a result of smoking, which is wellTHC is already licensed for medical use in some known, and therefore leads to an increased metabol-countries (as dronabinol) and preparations of the ism of the antipsychotic drug. A comparative studyherb are likely to follow suit, so vigilance is re- of 403 patients on chlorpromazine found that drow-quired. This single report from 1991 has not been siness (an adverse effect and an indication of bloodrepeated since, but users of an illicit drug would not levels) was decreased according to the level ofnecessarily report such problems, which may ex- smoking,[51] and a patient who gave up smokingplain their absence in the literature. experienced increased sedation and higher blood


1084 Williamson

levels of the same drug.[52] When cannabis was evidence for this theory was obtained when it wasinvolved in addition to tobacco, the effect was even found that, in rats, tea abolished the cataleptic ef-greater, according to a study of 31 patients on chlor- fects of chlorpromazine, and was unrelated to thepromazine.[54] A similar effect was seen in patients caffeine content,[60] but a clinical study of 16 pa-taking fluphenazine, and clearance of the drug was tients with mental health problems showed that theirmuch higher in the smokers.[53] No change in beha- intake of tea and coffee had no effect on either bloodviour was seen. So, as far as clinical management is levels of antipsychotic drugs or their behaviour.[61]

concerned, it may be necessary to use larger doses in Since in the acid conditions of the stomach, thesmokers, and also to be alert to the possibility of tannin-drug complex would dissociate, it is likelyreducing the dose if the patient stops smoking. that this physico-chemical interaction is of no clin-

ical significance – and the original report remainsSeizures were reported in two patients takingunexplained.fluphenazine with evening primrose oil (which con-

tains gamolenic acid) and in one patient taking pla- 3.2.3 Lithiumcebo with evening primrose oil in a study of 23 Lithium is the most commonly used drug inpatients with schizophrenia.[55] Conversely, another, bipolar disorder and blood levels must be checkedcrossover study in 48 patients, the majority of whom regularly to avoid toxicity. Any substance which canhad schizophrenia, recorded no seizures when pa- affect this (and many do, as does food) have atients were given a combination of phenothiazines potential for interaction. Patients taking bulk fibreand evening primrose oil.[56] No reports have been preparations containing ispaghula or psyllium weremade of seizures in patients receiving evening prim- found to have lower blood levels of lithium. Therose oil who are not taking phenothiazines, (which mechanism of action is not known, but either ab-are known to be epileptogenic themselves), although sorption of lithium was reduced, or perhaps theanother study in hospitalised patients with schizo- (unnamed) preparations contained high levels ofphrenia found that evening primrose oil exacerbated sodium (in the form of bicarbonate) to aid dispersalthe disease, and also produced EEG evidence of in water before ingestion.[62]

temporal lobe epilepsy.[57] The situation is obviously A herbal diuretic was been implicated in theunclear and Stockley[1] considers the interaction to induction of lithium toxicity in a patient taking abe not well established, but suggests monitoring if proprietary mixture containing juniper, buchu,phenothiazines are taken together with evening horsetail, corn silk, bearberry, parsley, bromelainprimrose oil, as a precaution. and paprika.[63] Herbal diuretics are weak in compar-

ison to modern drugs and it is unclear whether thisTheoretical Interactions theoretical interaction was due to diuresis or someA study of the pharmacokinetics of clozapine in other effect such as enzyme inhibition. The patient

healthy volunteers concluded that the ingestion of had been actually taking it for slimming, and incaffeine 400–1000 mg/day inhibits the metabolism these cases people often use much higher doses thanof clozapine to an extent that may be clinically would normally be expected.significant.[58] No clinical reports have been docu-

3.2.4 Levodopamented and clozapine levels are monitored regularlyIt has been observed that the effect of levodopain all patients.

was reduced by the administration of reserpine (aMany herbal preparations contain polyphenolicsrauwolfia alkaloid).[1] Since reserpine depletesand tannins, and a single report of two cases ofneurotransmitter release (as explained in sectionschizophrenia being exacerbated by increased con-3.2.1), rauwolfia should not be taken by patientssumption of tea and coffee has been discussed.[1]

receiving levodopa.The original author of this case report considered it3.2.5 Acetylcholine Receptor Antagoniststo be an interaction with caffeine[59] but after studies

showed that many phenothiazines complex with tan- A confirmed interaction has been reported withnins and precipitate out in vitro, it was decided that betel (Areca) nut and procyclidine. Procyclidine isthis was probably the mechanism involved. Further often used to control the extrapyramidal (parkin-



sonian) adverse effects of antipsychotics. It has been may theoretically potentiate medication for epilep-reported that in an Indian patient on depot sy, since they have shown to possess anti-seizurefluphenazine, whose mild parkinsonian tremor was activity in animal experiments. Valerian containscontrolled by procyclidine, experienced severe ri- compounds that inhibit the breakdown of GABAgidity and jaw tremor when he started chewing betel and enhance benzodiazepine binding[67] which couldnut.[64] When he stopped chewing the nuts, these potentiate, for example, the effects of carbamaze-adverse effects disappeared. A similar effect was pine. Valerian also reduces the anxiogenic effects offound in a patient taking flupenthixol.[64] Betel nut is diazepam withdrawal,[68] which may prove a usefulsometimes used as a herbal medicine, but is much therapeutic effect. Passionflower contains chrysin, amore widely used as a social and recreational stimu- flavonoid, which acts as a partial agonist at thelant. This interaction seems to be established and benzodiazepine receptor, and in mice this compoundclinically significant, and patients on such drugs shows anti-seizure effects which are prevented byshould avoid chewing betel nut (found in pre-pre-

prior injection of a benzodiazepine antagonist.[69]pared ‘pan masala’ also). It is fairly easy to check

Many herbs (and foods) contain flavonoids, andwhether a patient has been chewing betel nut as ittheir activity is weak compared with the syntheticstains the mouth red. Betel nut contains arecolineantiepileptics, and so far there are no reports ofand arecaidine among other alkaloids, and high con-clinical interactions. Therefore the most likelysumption is associated with oral carcinoma, so it issource of problems is if patients attempt to replacebest avoided in any case.[15]

their current medication with an unproven herbal3.2.6 Antiepileptic Drugs product, thus precipitating seizures.

Stimulants containing caffeine (e.g. guarana,Potential Interactionscola) may exacerbate seizures by lowering a pa-Evening primrose oil and borage oil are taken fortient’s seizure threshold.[70] Caffeine also has alsoconditions such as premenstrual syndrome and ecze-been shown to exacerbate seizures in animals.[70]

ma. They contain gamolenic acid, which reportedlyHowever, since epileptic patients are unlikely to belowers the seizure threshold, and it is often recom-consuming tea and coffee in greater quantities as themended that patients avoid it.[10,11] There are norest of the population, any potential problem, if atsubstantiated reports of potential interactions withall, will probably only arise from products contain-concurrent antiepileptic drugs, although as men-ing high levels of caffeine.tioned in section 3.2.2, seizures have been reported

in two patients taking fluphenazine with evening Essential (volatile) oils are present in many herbsprimrose oil (and one taking the placebo), indicating and spices, and contain some known epileptogenicthe possible risks of inducing seizures.[56]

compounds (cineole and camphor in particular, andalso fenchone).[71] The levels of these found in food

Theoretical and Unlikely Interactionsand herbs are unlikely to cause major problems asSt John’s wort is known to interact with somethey are not particularly important as constituentsmedicines by nature of its inhibition of cytochromefrom a flavouring point of view. However, in theenzymes. However, it does not interfere with carba-case of aromatherapy, it is possible that rosemary,mazepine clearance[65] and no clinical reports aresage, hyssop and fennel oils could be absorbedavailable to suggest that is unsafe, although cautionthrough the skin and should be avoided in epilepticis, as always, recommended in epileptic patientspatients as a precaution.[72] A report of three patientsbecause maintaining therapeutic blood levels of an-who experienced tonic-clonic seizures after usingtiepileptic drugs is of critical importance.[40]

essential oils transdermally and orally, and the ab-Spinella[66] has drawn attention to the theoreticalsence of these after discontinuation of the oils, con-possibility for some herbal medicines to interactfirms this possibility.[71] It should be noted that thesewith antiepileptic drugs, although there are no re-patients did not have prior history of epilepsy, norcorded instances. Sedatives such as valerian (Valeri-were they taking any other medication.ana spp.), passionflower (Passiflora spp.) and kava


1086 Williamson

3.3 Pain and Inflammation 3.4 Sex Hormones

3.4.1 Confirmed Interactions3.3.1 Salicylates in Herbs and A number of reports have now been receivedPrescription Medicines concerning interactions between St John’s wort and Salicylates are found in several plant medicines, the contraceptive pill, which can lead to contracep-

such as willow bark (Salix spp.) and meadowsweet tive failure.[14] Breakthrough bleeding has been re-ported, although this is a fairly common occurrence(Filipendula ulmaria) but when they naturally occurassociated with many other factors, particularlyin plants they are chemically bound (for instance, inmissed doses.[14] While pregnancies have occurredthe form of glycosides) and appear to have a muchwith concurrent use of St John’s wort and the contra-lower propensity to cause gastrointestinal bleedingceptive pill, they have also occurred with otherthan, for example, aspirin.[15] Thrombolytics anddrugs administered concurrently with the pill, asantiplatelet agents are considered to interact withwell as the pill alone. Definite causality of St John’ssalicylates,[13] although currently no recommenda-wort affecting the efficacy of the contraceptive pill

tions regarding dose adjustment can be made,[73] and has not been established, although there are goodby extrapolation, salicylate-containing herbs should reasons to suspect that it is a true pharmacokineticprobably be avoided, although no adverse reports interaction due to enzyme induction by St John’shave been made. A similar situation arises with ACE wort and an increase in the expression of P-glyco-inhibitors and aspirin, but reports are conflicting and protein, leading to lower drug levels, including thatit is very common for patients to be taking both of the oral hormonal contraceptive pill.[78]

types of medicine concurrently,[74] so the interaction Two cases of failure of emergency hormonalcontraception have also been reported when admin-is probably not important in practice.istered with St John’s wort (although emergencyhormonal contraception is not 100% effective even3.3.2 Other Unconfirmed Reportswithout St John’s wort) and it is recommended thatTamarind, an Asian fruit used as a flavour ingre-if patients are already taking St John’s wort then an

dient in cooking, and also as an Ayurvedic medi- increase in the dose of emergency hormonal contra-cine, markedly increased the absorption of a single ception should be considered.[1,14]

dose of aspirin 600mg in six healthy individuals.[75]The number of cases of contraceptive failure is

The authors concluded that it could result in an still fairly small, and the incidence of concomitantincrease in toxicity if large doses of salicylates are use of the pill and St John’s wort is not known, soingested. This may or may not apply to other salicy- the actual risk cannot yet be assessed.[79] Currentlates and the significance of this interaction, if any, advice from the FDA in the US and the MHRA in

the UK, is that concurrent use should be avoided.[21]is not known with regard to herbal medicines.If patients are already taking these medications con-A report of excessive bleeding by a patient takingcurrently, then St John’s wort should be discontin-rofecoxib with ginkgo was reported,[76] but the inter-ued or additional contraceptive measures taken.

action was not confirmed and is probably not impor-3.4.2 Theoretical and Unlikely Interactionstant.Tamoxifen is an estrogen antagonist widely usedKakkonto is a Chinese herbal medicine contain-

in the treatment and prevention of breast cancer. Iting liquorice, peony root, ephedra, ginger and other was suggested in a letter to the Australian Medicalherbs, which in high doses in animal tests caused an Journal[80] that some herbal medicines that are usedincrease in paracetamol (acetaminophen) levels.[77]

as either ‘hormone regulators’ or as a form of naturalHowever, in two separate studies in humans,[77] no hormone replacement therapy (HRT) may directlyeffect was seen, even at doses of 5g. The interaction stimulate breast cancer growth or oppose the actionsthus remains unconfirmed but is probably not signif- of competitive estrogen antagonists, includingicant.[77] tamoxifen. This raises questions about the mechan-



isms of action of these herbs, which is often not drugs metabolised by these routes.[81] No clinicalknown, and in many instances, hormonal effects reports have yet been made.have not been conclusively demonstrated. Those Miller and Murray[10] have suggested that thementioned included dong quai (Angelica sinensis), potential for interactions exists between insulin andchasteberry (Vitex agnus castus) and black cohosh oral hypoglycaemics and stimulant herbs, such as(Cimicifuga racemosa); however, generally these ma huang (Ephedra sinica), and also caffeine-con-are not considered to have estrogenic properties.[15] taining products such as cola and guarana. This isOn the other hand, herbs such as red clover (Trifoli- because stimulants tend to elevate blood glucose andum pratense) and soya (Glycine max) contain es- may therefore interfere with diabetic control. It is atrogenic isoflavonoids; however, a high consump- theoretical possibility, and no cases have yet beention of these as part of the diet is associated with a recorded.lower incidence of breast cancer. This paradox is Aspirin and other salicylates have been knownthought to be a result of the isoflavones binding to for over 100 years to cause hypoglycaemia in largeestrogen receptors and acting as HRT agents, but by doses. However, the amount of salicylates found indoing so they prevent the binding of the more potent herbal drugs is unlikely to affect diabetic control.endogenous estrogens, thus also acting as anti-estro-

3.5.2 Potential Additive Effectsgens. This is largely speculative at present, but sinceThe fruit of karela (Momordica charantia, ortamoxifen is a proven and highly effective drug in

bitter gourd) is a vegetable eaten widely in tropicalthe management of breast cancer, it would poseregions of Asia, the West Indies and Africa. The leafunnecessary risks to combine its use with ‘hormon-is also available as a form of ‘bush tea’, known asally active’ herbal remedies in the absence of more‘cerassie’ in the Caribbean. All parts of the plant aredata.used as a treatment for type 2 diabetes mellitus in

A psychotic episode was reported in a young manAyurvedic and other forms of traditional medicine,

taking testosterone after an orchidectomy, who wasand consumption leads to hypoglycaemia. This may

also given sertraline for depression and continued toaffect diabetic control when taken alone, and all

take St John’s wort against medical advice.[43] Ahypoglycaemic agents tend to have additive effects

manic episode ensued, involving elation, overspend-when taken in combination. A recent report showed

ing on a car, arguments with his wife, and culminat-that a patient who was poorly controlled when tak-

ing in arrest by the police for stealing fuel for theing chlorpropamide alone, experienced much better

car. Grandiose delusions followed, (in that he expec-diabetic control after ingesting karela with chlor-

ted to be rescued by a team of soldiers and his fatherpropamide.[82]

would visit in a private aeroplane) and the patientwas sectioned under the 1983 Mental Health Act.

3.6 AntibacterialsThe testosterone is probably unconnected to theepisode and it is more likely to be an interaction An unlikely interaction between ciprofloxacinbetween St John’s wort with sertraline. and fennel (Foeniculum vulgare) has been postulat-

ed after experiments in rats showed that lower bloodconcentrations of the antibacterial were found with3.5 Diabetes Mellitusconcurrent administration of fennel extract.[83] Theauthors suggest that if this combination is taken

3.5.1 Theoretical and Unlikely Interactions clinically, then an adequate dose of ciprofloxacin beAn experiment in rats showed that an extract of taken. However, the dose of fennel used (2 g/kg) in

Angelica dahurica (which contains fura- the animal experiment would translate to be overnocoumarins and is widely used in traditional Chi- 100g of herb in humans, which is unlikely. Addi-nese medicine), delayed the elimination of tolbuta- tionally, the study found that none of the organicmide in vivo.[81] It was attributed to the inhibition of components of fennel seemed to interact, and theCYP2C enzyme, but the extract also inhibits most likely explanation was chelation in the stom-CYP3A and CYP2D1 and this would apply to other ach by metal cations found in the extract,[83] conse-


1088 Williamson

quently, this interaction (if such it is) is probably of cyclosporin with concomitant administrationunimportant. were found.[87-89] In some cases blood levels of

cyclosporin were well below therapeutic levels, andKhat (Catha edulis) is a mildly stimulant herbhad dropped by an average of 49%. Both drugs arechewed as a social drug (in the manner of tea andmetabolised via CYP3A4, and as this can havecoffee) in Africa and the Middle East, particularly inserious consequences St John’s wort should not beYemen, and by the same ethnic populations in othertaken at the same time as cyclosporin.parts of the world. It is occasionally used as a

medicine, and has been shown to reduce the absorp- A 54-year-old kidney transplant patient takingtion of ampicillin, and to a lesser extent, amoxicillin, cyclosporin, in combination with azathioprine, pred-in a study of eight healthy Yemeni volunteers.[84] It nisone, diltiazem and nifedipine, experienced ais thought that this theoretical interaction was poss- large increase (up to 8-fold) in cyclosporin bloodibly due to formation of a tannin-antibiotic complex levels. After investigation, it transpired that 2 weeksin the stomach. previously he had started drinking a tea containing

A small decrease in the absorption of penicillin V Geum chiloense, a herb reputed to enhance virility.was observed with concurrent oral administration Blood levels returned to normal on cessation of thiswith guar gum, in a study of ten individuals.[85] herb, but the mechanism of action was not deter-Since guar is taken to deliberately reduce absorption mined.[90]

from the stomach (in diabetic patients, to avoid Glycyrrhizin, found in liquorice (Glycyrrhizasudden peaks in glucose levels) this is not surprising glabra), was found to reduce the clearance of pred-and probably not clinically significant, especially in nisolone in healthy individuals.[91] It was consideredview of the fact that oral penicillin V should always to be due to inhibition of the metabolism of prednis-be taken on an empty stomach. olone, and the authors concluded that this inter-

action may have clinical benefits provided that toxic3.7 Antiviral Drugs serum levels are not reached.[91]

Echinacea is taken as an immune stimulant, soIndinavir, a protease inhibitor used to treat HIVthere would be little point in taking it with aninfection, is metabolised by CYP3A4, as is manyimmunosuppressant. Nevertheless, a potential forother drugs – and the herb St John’s wort. The FDAinteraction exists and it should probably be avoided.in the US and the CSM in the UK have warned aboutNo clinical interactions have been reported.the possibility of an interaction between indinavir

and St John’s wort leading to lower blood levels ofthe antiviral.[21] This has been confirmed in a study 3.9 Cancer Therapiesusing healthy individuals.[86] Although no clinicalreports have yet been found, it is certainly possible

There is a potential for interactions betweenthat HIV-positive people would (or already do) takeherbal remedies and cancer therapies, which hasSt John’s wort because of its antidepressant proper-been discussed by Block and Gyllenhaal,[92] becauseties. Lowered blood levels of indinavir could resultof the effects of some herbs (particularly St John’sin serious treatment failure, therefore St John’s wortwort) on drug metabolising enzymes. However, noshould be avoided by all patients on indinavir, otheractual herb-drug interactions have yet been reportedanti-retrovirals and also any drugs metabolised byclinically during cancer treatment. On the otherthe same enzyme.hand, a potentially useful interaction has been de-scribed by Meneendez et al. who found that3.8 Immunosuppressantsgamolenic acid, found in evening primrose and star-

Cyclosporin is an important immunosuppressant flower (borage) seed oils, potentiated the in vitroused after organ transplant. Confirmed, serious in- cytotoxicity of paclitaxel and vinorelbine in humanteractions with St John’s wort have been document- breast cancer cell lines. The authors suggest a roleed, including graft rejection in kidney, liver and for some unsaturated fatty acids as modulators ofheart transplant patients, and lowered blood levels tumour cell chemosensitivity.[93]



3.10 Miscellaneous Interactions 3.11 Herbs and Surgery

Diminished blood levels of theophylline (used in Patients undergoing surgery may be taking herbalmedicines without informing the anaesthetist, sur-respiratory conditions) were found in a patient alsogeon or physician, and this practice may have antaking St John’s wort.[13] Although this patient wasimpact on the peri-operative care of these patients. Aalso taking many other drugs, her serum levels ofhospital study of patients attending pre-anaesthetictheophylline returned to normal on cessation of Stevaluations in the US found that over half of theJohn’s wort, and it was considered to be an inter-patients were taking some form of supplement,action caused by induction of CYP enzymes.mainly vitamins but also herbal medicines.[97] The

Piperine is a constituent of black pepper and most common of these was garlic, followed byother species (e.g. Piper nigrum, Piper longum), ginkgo, St John’s wort, ma huang, echinacea, an-which are a component of many Ayurvedic formula- thraquinone laxative drugs and liquorice. Some oftions, including trikatu. Piperine increases the bio- these may well have the potential for interactionavailability of many drugs and a number of cases with drugs used during surgery, and it was recom-have been documented.[94] The significance is not mended that a history of herbal supplements by theknown, and no clinical reports have yet been made, patient be taken. This would be helpful in determin-but a regular, high dietary level of pepper (or inges- ing whether or not interactions occur, if they are oftion of an Ayurvedic product containing it) may clinical significance, and enable reliable advice toincrease blood levels of other drugs such as pheny- be given rather than suggesting that all supplementstoin, propranolol and theophylline, and caution be avoided.should be taken. The effect of herbal medicines on the response to

A case of myocardial infarction was reported in a anaesthetics has not been evaluated, and conse-quently the American Society of Anaesthesiologistspatient taking sildenafil for erectile dysfunction andhas recommended that all herbal medications bethen smoking cannabis.[95] Although this was con-stopped 2–3 weeks prior to an elective surgicalsidered to be an interaction, serious cardiovascularprocedure.[98] Stringent recommendations wereevents are listed as adverse effects in the generalmade recently regarding the precautions that shouldprescribing data for sildenafil.[96]

be taken, such as discontinuing ginkgo, ginseng andSome herbal medicines, including ma huang, garlic because of the possibility that they may in-

yohimbe (Pausinystalia yohimbe) and liquorice can crease bleeding.[99] Since low-dose aspirin taken asraise blood pressure even when taken alone.[15] Ma an antiplatelet drug to prevent myocardial infarctionhuang is taken as a slimming aid, decongestant and is not contraindicated before surgery, it is not clearfor asthma, and it has been abused as a stimulant in whether herbs would actually pose any risk. Sthigh doses. It contains the sympathomimetic alka- John’s wort is quite reasonably contraindicated forloid ephedrine, which can increase heart rate and its potential to interact with other drugs, as is maperipheral vascular resistance. Liquorice is an ingre- huang because it can increase blood pressure anddient of many traditional Chinese medicine prepara- heart rate. Furthermore, it has even been suggestedtions for a multitude of indications, especially stom- that echinacea should be discontinued as soon asach disorders, and excessive consumption is asso- possible prior to surgery, because it may cause poorciated with pseudo-aldosteronism.[15] Yohimbe wound healing.[12] The rationale for this is not clearbark, which contains yohimbine, a presynaptic and there is no evidence that it inhibits healing.α2-adrenoceptor antagonist (and possibly a MAOI), Traditional Chinese medicines have also been iden-is sometimes used to treat erectile dysfunction. The tified as potentially interacting with anaesthetics[100]

problems posed by these herbal medicines in hyper- and it was suggested that the anaesthetist be in-tensive patients have been reviewed by Valli and formed of any concurrent usage. Unfortunately thisGiardiana[4] and Mansoor.[7] information is rarely available and an informed deci-


1090 Williamson

2. Schmidt LE, Dalhoff K. Food-drug interactions. Drugs 2002; 62sion about the safety of such combinations cannot be(10): 1481-502

made. 3. Fugh-Berman A, Ernst E. Herb-drug interactions: review andassessment of report reliability. Br J Clin Pharm 2001; 52:587-954. Conclusions

4. Valli G, Giardiana EG. Benefits, adverse effects and drug inter-actions of herbal therapies with cardiovascular effects. J Am

It is apparent that herb-drug interactions may be Coll Cardiol 2002; 38: 1083-955. Sorensen JM. Herb-drug, food-drug, nutrient-drug, and drug-grossly under-reported for various reasons. Al-

drug interactions: mechanisms involved and their medicalthough many of the examples cited in this paper areimplications. J Altern Complement Med 2002; 8: 293-308

unproven or purely speculative, the fact remains that 6. Izzo AA, Ernst E. Interactions between herbal medicines andprescribed drugs: a systematic review. Drugs 2001; 61 (15):some are serious and indeed life-threatening. These2163-75almost exclusively concern cyclosporin, anticoagu-

7. Mansoor G. Herbs and alternative therapies in the hypertensionlants, digoxin, antidepressants and protease inhibi- clinic. Am J Hypertens 2002; 14: 971-5

8. Fugh-Berman A. Herb-drug interactions. Lancet 2000; 355:tors, with the herb St John’s wort. Ginkgo and134-8ginseng are also implicated in a number or reports,

9. Izzo AA, Borrelli F, Capasso R. Herbal medicine: the dangers ofbut many of these are unsubstantiated or may be drug interaction. Trends Pharm Sci 2002; 23: 358-9idiosyncratic. The most common drugs that have 10. Miller LG, Murray WJ, editors. Herbal medicinals: a clinician’s

guide. New York: Pharmaceutical Products Press, 1998been reported to interact with herbs are those involv-11. Barnes J, Phillipson JD, Anderson LA. Herbal medicines. 2nding the cardiovascular, central nervous and immune ed. London: The Pharmaceutical Press, 2002

systems. Additional drugs treating other systems 12. Ang-Lee MK, Moss J, Yaun C. Herbal medicines and peri-operative care. JAMA 2001; 286: 208-16have also been implicated, however, examples are

13. Abebe W. Herbal medication: potential for adverse interactionsvery limited. It is reasonable to assume that adding with analgesic drugs. J Clin Pharm Ther 2002, 401herbal products to a regimen of prescription drugs 14. Henderson L, Yue QY, Bergquist C, et al. St John’s wort

(Hypericum perforatum): drug interactions and clinical out-taken for the same purpose would lead to additivecomes. Br J Clin Pharm 2002; 54 (4): 349-56

effects, and the answer to that lies in patient educa- 15. Williamson EM. Potter’s herbal cyclopaedia. Saffron Walden,tion. Although many herbal drugs are safe, it must UK: CW Daniel & Co., 2003

16. WHO Drug Monitoring Centre Uppsala UR22 Apr 2003 [on-be borne in mind that herbal supplements are intend-line]. Available from URL: www.who-umc.org [Accesseded to be taken over an extended period of time, 2003 Nov 3]

which provides the opportunity for enzyme induc- 17. Page RL, Lawrence JD. Potentiation of warfarin by dong quai.Pharmacotherapy 1999; 19: 870-6tion and other mechanisms of interaction to take

18. Makina K, Wakushima H, Okamoto T, et al. Pharmacokineticeffect. As far as advising patients is concerned, it is interactions between warfarin and kangen-karyu, a traditionalChinese medicine, and their synergistic action. Jexpedient not to be seen as scaremongering, other-Ethnopharmacol 2002; 82: 35-40wise patients may feel that the medical establish-

19. Vaes LP, Chyka PA. Interactions of warfarin with garlic, ginger,ment is over-reacting and that drug companies are ginkgo or ginseng: nature of the evidence. Ann Pharmacother

2000; 34: 1478-82being self-serving, and if that happens even sound20. Maurer A, Johne A, Bauer S, et al. Interaction of St John’s wortadvice will be ignored. Articles exaggerating the

with phenprocoumon [abstract]. Eur J Clin Pharmacol 1999;problem, by treating hypothetical situations as fact, 55: A22

21. Monitoring the Safety and Quality of Medicines. Herbal safetydo not serve the users of herbs or the medical andnews. St John’s wort [online]. Available from URL: http://pharmaceutical professions well, although they are medicines.mhra.gov.uk [Accessed 2003 Nov 12]

often given undue prominence in the press. 22. Lo ACT, Chan K, Yeung JHK. Danggui (Angelica sinensis)affects the pharmacodynamics but not the pharmacokinetics ofwarfarin in rabbits. Eur J Drug Metab Pharmacokinet 1995; 20:Acknowledgements 55-60

23. Janetsky K, Morreale JP. Probable interaction between warfarinNo sources of funding were used to assist in the prepara- and ginger. Am J Health Syst Pharm 1997; 54: 692-3

tion of this manuscript. The author has no potential conflicts 24. Thomson Coon J, Ernst E. Panax ginseng: a systematic reviewof interest that are directly relevant to the content of this of adverse effects and drug interactions. Drug Saf 2002; 25:

323-44manuscript.25. Lam AY, Elmer GW, Mohutsky MA. Possible interaction be-

tween warfarin and Lycium barbarum L. Ann Pharmacother2001; 35: 1195-9References

1. Stockley IH, editor. Stockley’s drug interactions. 6th ed. 26. Lambert JP, Cormier A. Potential interaction between warfarinLondon: The Pharmaceutical Press, 2002 and boldo-fenugreek. Pharmacotherapy 2001; 21: 509-12



27. Yue Q-Y, Jansson K. Herbal drug Curbicin and anticoagulant 51. Swett C. Drowsiness due to chlorpromazine in relation to ciga-effect with and without warfarin: possibly related to the vita- rette smoking: a report from the Boston Collaborative Drugmin E component [letter]. J Am Geriatr Soc 2001; 49: 838 Surveillance Program. Arch Gen Psychiatry 1974; 31: 211-3

52. Stimmel GL, Falloon IRH. Chlorpromazine plasma levels, ad-28. Robinson DS, Benjamin DM, McCormack JJ. Interaction ofverse effects and tobacco smoking. J Clin Psychiatry 1983; 44:warfarin and non-systemic gastrointestinal drugs. Clin420-2Pharmacol Ther 1971; 12: 491-5

53. Ereshefsky L, Jann MW, Saklad SR, et al. Effects of smoking29. Monterrey-Rodrigues J, Feliu JF, Rivera-Miranda GC. Inter-on fluphenazine clearance in psychiatric inpatients. Biol Psy-action between warfarin and mango fruit. Ann Pharmacotherchiatry 1985; 20: 329-322002; 36: 940-1

54. Chetty M, Miller R, Moodley SV. Smoking and body weight30. Taylor JR, Wilt VM. Probable antagonism of warfarin by greeninfluence the clearance of chlorpromazine. Eur J Clin Psycho-tea. Ann Pharmacother 1999; 33: 426-8pharmacol 1994; 46: 523-631. Weinrobe MC, Montgomery B. Acquired bleeding diathesis in a

55. Holman CP, Bell AFJ. A trial of evening primrose oil in thepatient taking PC-SPES. N Engl J Med 2001; 345: 1213-4treatment of chronic schizophrenia. J Orthomol Psychiatry32. Shaw D. National Poisons Unit, Guy’s and St Thomas’s NHS1983; 12: 302-4Trust, London UK, 2003

56. Vaddadi KS, Courtney P, Gilleard CJ, et al. A double blind trial33. Awang DV. Siberian ginseng toxicity may be a case of mistakenof essential fatty acid supplementation in patients with tardiveidentity [letter]. CMAJ 1996; 155 (9): 1237dyskinesia. Psychiatr Res 1989; 27: 313-23

34. Waller DP, Farnsworth NR, Awang DV. Lack of androgenicity57. Vaddadi KS. The use of gamma-linolenic acid and linoleic acidof Siberian ginseng [letter]. JAMA 1992; 267 (17): 2329

to differentiate between temporal lobe epilepsy and schizo-35. Shinozuka K, Umegaki K, Kubota Y, et al. Feeding of Ginkgo phrenia. Prostaglandins Med 1981; 6: 375-9

biloba extract (GBE) enhances gene expression hepatic cyto-58. Hagg S, Spigset O, Mjorndal T, et al. Effect of caffeine onchrome P-450 and attenuates the hypotensive effects of ni-

clozapine pharmacokinetics in healthy volunteers. Br J Clincardipine in rats. Life Sci 2002; 70 (23): 2783-92Psychopharmacol 2000; 49: 59-63

36. McCoubrie M. Doctors as patients: lisinopril and garlic [letter].59. Mikkelson EJ. Caffeine and schizophrenia. J Clin PsychiatryBr J Gen Pract 1996; 46: 107

1978; 39: 732-337. Hakas JF. Topical capsaicin induces cough in patient receiving 60. Cheeseman HJ, Neal MJ. Interaction of chlorpromazine with teaACE inhibitor. Ann Allergy 1990; 65: 322-3 and coffee. Br J Clin Psychopharmacol 1981; 12: 165-938. Rosenblatt M, Mindel J. Spontaneous hyphema associated with 61. Bowen S, Taylor KM, Gibb IAM. Effect of coffee and tea on

ingestion of Ginkgo biloba extract. N Engl J Med 1997; 338: blood levels and efficacy of antipsychotic drugs. Lancet 1981;1108 I: 1217-8

39. Matthews MK. Association of Ginkgo biloba with intracerebral 62. Toutong M, Schultz P, Widmer J, et al. Probable interactionhaemorrhage. Neurology 1998; 50: 1933-4 entre le psyllium et le lithium. Therapie 1990; 45: 357-60

40. British Medical Association, Royal Pharmaceutical Society of 63. Pyevich D, Bogenschutz MP. Herbal diuretics and lithium toxic-Great Britain. In: Mehta DK, editor. British National Form- ity [short communication]. Am J Psychiatry 2001; 158: 1329ulary 2003; 46: 118 64. Deahl M. Betel nut induced extrapyramidal syndrome: an unu-

41. Hemeryck A, Belpaire FM. Selective serotonin reuptake inhibi- sual drug interaction. Mov Disord 1989; 4: 330-3tors and cytochrome P-450 mediated drug interactions: an 65. Burstein AH, Horton RL, Dunn T, et al. Lack of effect of Stupdate. Curr Drug Metab 2002; 3: 13-37 John’s wort on carbamazepine pharmacokinetics in healthy

42. Markowitz JS, DeVane CL, Boulton DW, et al. Effect of St volunteers. Clin Pharmacol Ther 2000; 68: 605-12John’s wort (Hypericum perforatum) on cytochrome P-450 66. Spinella M. Herbal medicines and epilepsy: the potential for2D6 and 3A4 activity in healthy volunteers. Life Sci 2000; 66: benefit and adverse effects. Epilepsy Behav 2001; 2: 524-32133-9 67. Riedel E, Hansel R, Ehrke G. Inhibition of gamma-ami-

43. Barbenel DM, Yusufi B, O’Shea D, et al. Mania in a patient nobutyric acid catabolism by valerenic acid derivatives. Plantareceiving testosterone replacement post-orchidectomy taking Med 1982; 46: 219-20St John’s wort and sertraline. J Psychopharmacol 2000; 14: 68. Leuchner J, Muller J, Rudmann M. Characterisation of the84-6 central nervous system depressant activity of a commercially

44. Essor AH. Clinical observations on reserpine reversal after available valerian root extract. Arznemittel Forschung 1993;prolonged MAO inhibition. Psychiatr Neurol Neurochir 1967; 43: 44-770: 59-63 69. Medina JH, Viola H, Wolfman C, et al. Overview: flavonoids: a

45. Gradwell BG. Psychotic reactions and phenelzine [letter]. BMJ new family of benzodiazepine receptor ligands. Neurochem1960; 2: 1018 Res 1997; 22: 419-25

46. Jones BD, Runikis AM. Interaction of ginseng with phenelzine. 70. Ault B, Olney MA, Joyner JL, et al. Proconvulsant actions ofJ Clin Psychopharmacol 1987; 7: 201-2 theophylline and caffeine in the hippocampus: implications for

47. Stoll Al, Cole JO, Lukas SE. A case of mania as a result of the management of temporal lobe epilepsy. Brain Res 1987;fluoxetine-marijuana interaction. J Clin Psychiatry 1991; 52: 426: 93-102280-1 71. Burkhard PR, Burkhardt K, Haenggeli CA, et al. Plant-induced

seizures: reappearance of an old problem. J Neurol 1999; 246:48. Yung M, Herrema I. Atropine interaction with amitriptyline667-70leading to mydriasis in a neonate: case report [abstract]. Reac-

tions 2000; 826: 7 72. Tisserand R, Balacs T. Essential oil safety. Edinburgh: Churc-hill Livingstone, 199549. Galluzi S, Nanetti O, Binetti G, et al. Coma in a patient with

Alzheimer’s disease taking low-dose trazodone and Ginkgo 73. Harder S, Klinkhardt U. Thrombolytics: drug interactions ofbiloba. J Neurol Neurosurg Psychiatry 2000; 68: 679-83 clinical significance. Drug Saf 2000; 23 (5): 391-9

50. Amsterdam JD, Berwish N. Treatment of refractory depression 74. Mahe I, Meune C, Diemer M, et al. Interactions between aspirinwith combination reserpine and tricyclic antidepressant ther- and ACE inhibitors in patients with heart disease. Drug Safapy. J Clin Psychopharmacol 1987; 7: 238-42 2001; 24 (3): 167-82


1092 Williamson

75. Mustapha A, Yakasai IA, Aguye IA. Effect of Tamarindus 90. Duclos J, Goecke H. ‘Heirba del clavo’ (Geum chiloense)indica L.on the bioavailability of aspirin in healthy human interfiere niveles de ciclosporin: potencial riesgo para trans-volunteers. Eur J Drug Metab Pharmacokinet 1996; 21: 223-6 plantados. Rev Med Chil 2001; 129: 789-90

76. Hoffmann T. Ginkgo, Vioxx and excessive bleeding: possible 91. Chen M-F, Shimada F, Kato H, et al. Effect of oral administra-drug interactions: case report [letter]. Hawaii Med J 2001; 60: tion of glycyrrhizin on the pharmacokinetics of prednisolone.290 Endocrinol Jpn 1991; 38: 167-74

77. Qi J, Toyoshama A, Honda Y, et al. Pharmacokinetic study on 92. Block KI, Gyllenhaal C. Clinical corner: herb-drug interactionsacetaminophen: interaction with a Chinese medicine. J Med in cancer chemotherapy: theoretical concerns regarding drugDent Sci 1997; 44: 31-5 metabolizing enzymes. Integrative Cancer Ther 2002; 1 (1):

83-978. Hennessy M, Kelleher D, Spiers JP, et al. St John’s wortincreases expression of P-glycoprotein: implications for drug 93. Meneendez JA, Ropero S, Barbacid M, et al. Synergistic actioninteractions. Br J Clin Pharmacol 2002; 53 (1): 75-82 between vinorelbine and gamma-linolenic acid in human

79. Murphy PA. St John’s wort and oral contraceptives: reasons for breast cancer cell lines. Breast Cancer Res Treat 2002; 72:concern? J Midwifery Womens Health 2002; 47: 447-50 1-17

80. Boyle FM. Adverse interaction of herbal medicine with breast 94. Williamson E. Synergy and other interactions incancer treatment [letter]. Med J Aust 1997; 167: 286 phytomedicines. Phytomedicine 2001; 8: 401-9

81. Ishihara K, Kushida H, Yazurihara M, et al. Interaction of drugs 95. McLeod AL, McKenna CJ, Northridge DB. Sildenafil: firstand Chinese herbs: pharmacokinetic changes of tolbutamide report of an interaction with concomitant marijuana leading toand diazepam caused by extract of Angelica dahurica. J Pharm myocardial infarction: case report [short communication]. ViaPharmacol 2000; 52: 1023-9 Reactions 2002; 900: 11

82. Aslam M, Stockley IH. Interaction between curry ingredient 96. Mehta D, editor. British Medical Association and the Royal(karela) and drug (chlorpropamide) [letter]. Lancet 1979; I: Pharmaceutical Society of Great Britain. London; British Na-607 tional Formulary 2003; 46: 406

83. Zhou M, Wong PYK, Li RC. Effect of oral administration of97. Kaye AD, Clarke RC, Sabar R, et al. Herbal medicines: currentfennel (Foeniculum vulgare) in ciprofloxacin absorption and

trends in anesthesiology practice: a hospital survey. J Clindisposition in the rat. J Pharm Pharmacol 1999; 51: 1391-6Anesth 2000 Sep; 12 (6): 468-71

84. Abdel Ghani YM, Etman MA, Nada AH. Effect of khat chewing98. Sabar R, Kaye AD, Frost EA. Peri-operative considerations foron the absorption of orally administered amoxicillin. Acta

the patient on herbal medicines. Middle East J AnesthesiolPharm 1999; 49: 43-502001; 16: 287-314

85. Huupponen R, Seppala P, Iisalo E. Effect of guar gum, a fibre99. Monitoring the Safety and Quality of Medicines. Herbal safetypreparation, on digoxin and penicillin absorption in man. Eur J

news. MHRA’s report on the safety of herbal medicinal prod-Clin Pharmacol 1984; 26: 279-81ucts [online]. Available from URL: http://medicines.m-86. Piscitelli SC, Burstein AH, Chaitt D, et al. Indinavir concentra-hra.gov.uk [Accessed 2003 Nov 12]tions and St John’s wort. Lancet 2000; 355: 547-8

100. Kam PC, Liew S. Traditional Chinese herbal medicine and87. Breidenbach T, Hoffman MW, Becker T, et al. Drug interactionanaesthesia. Anaesthesia 2002; 57: 1083-9of St John’s wort with ciclosporin [letter]. Lancet 2000; 355:

1912

88. Barone GW, Gurley BJ, Ketel BL, et al. Herbal supplements: a Correspondence and offprints: Dr Elizabeth M. Williamson,potential for drug interactions in transplant recipients. Trans-Centre for Pharmacognosy and Phytotherapy, The Schoolplantation 2001; 71: 239-41of Pharmacy, University of London, Brunswick Square,89. Ahmed SM, Banner NR, Dubrey SW. Low cyclosporin-A levelLondon WC1N 1AX, UK.due to St John’s wort in heart transplant patients [letter]. J

Heart Lung Transplant 2001; 20: 795 E-mail: [email protected]


Documents

Williamson, EM. Drug Safety 2003. Herbal Drug Interactions