Phyto Pharmaco Vigilance

See discussions, stats, and author profiles for this publication at:
http://www.researchgate.net/publication/5280710
Pharmacovigilance of Herbal
Medicines
ARTICLE in DRUG SAFETY FEBRUARY 2003
Impact Factor: 2.82 DOI: 10.2165/00002018-200326120-00001 Source: PubMed
CITATIONS
READS
79
261
1 AUTHOR:
Joanne Barnes
University of Auckland
88 PUBLICATIONS 2,088 CITATIONS
SEE PROFILE
Available from: Joanne Barnes Retrieved on: 29 December 2015
L
E
A
DI
N
G
Dru
g
Saf
ety
200
3;
26
(12)
:
829
851
011
4591
6/0
3/0
012
082
9/$
30.
00/
0
A
R
TI
C
L
E
Adis Data Information BV 2003. All rights reserved.
Pharmacovigilance of
Herbal Medicines
A UK Perspective
Joanne Barnes
Centre for Pharmacognosy and Phytotherapy, School of
Pharmacy, University of London, London, United Kingdom
Abstract
Thereisanincreasingawarenessatseveral
levelsoftheneedtodevelop
pharmacovigilancepracticesforherbal
medicines.Thecurrentmodelofpharma
covigilanceanditsassociatedtoolshavebeen
developedinrelationtosynthetic
drugs,andapplyingthesemethodsto
monitoringthesafetyofherbalmedicines
presentsuniquechallengesinadditiontothose
describedforconventionalmedi
cines.Severalproblemsrelatetothewaysin
whichherbalmedicinesarenamed,
perceived,sourced,andutilised.Other
importantchallengesarisefromthecurrent
regulatoryframeworkforherbalmedicinesin
theUK.
IntheUKatpresent,theCommitteeon
SafetyofMedicines/Medicinesand
HealthcareproductsRegulatoryAgencys
(CSM/MHRA)yellowcardscheme
foradversedrugreaction(ADR)reportingis
themainmethodofmonitoringthe
safetyofherbalmedicines.Despiterecent
initiativestostimulatereportingof
suspectedADRsassociatedwithherbal
medicines,suchasextendingthescheme
tounlicensedherbalproducts,andincluding
communitypharmacistsasrecog
nisedreporters,numbersofherbalADR
reportsreceivedbytheCSM/MHRA
remainrelativelylow.Underreporting,an
inevitableandimportantlimitationof
spontaneousreportingschemes,islikelytobe
significantforherbalmedicines,
sinceuserstypicallydonotseekprofessional
adviceabouttheiruseofsuch
products,orreportiftheyexperienceadverse
effects.TheherbalsectorintheUK
hasinitiatedvariousspontaneousreporting
schemes,basedontheyellowcard
scheme,buttargetedmainlyatherbal
medicinepractitioners.Itisimportantthat
theseschemeshavealinkwiththe
CSM/MHRAsothatpotentialsignalsarenot
missed.Severalothertoolsusedin
pharmacovigilanceofconventionalmedicines,
suchasprescriptioneventmonitoring,andthe
useofcomputerisedhealthrecord
databases,currentlyareofnousefor
evaluatingthesafetyofherbalandother
nonprescriptionmedicines.
ProposedEuropeanUnionlegislationfor
traditionalherbalmedicinalproducts
willrequiremanufacturersofproducts
registeredundernewnationalschemesto
complywithregulatoryprovisionson
pharmacovigilance.Inthelongerterm,
otherimprovementsinsafetymonitoringof
herbalmedicinesmayincludemodifi
cationstoexistingmethodology,patient
reportingandgreaterconsiderationof
830
Barnes
pharmacogeneticsandpharmacogenomicsinoptimisingthesafetyofherbal
medicines.
Pharmacovigilance has been defined as the

study of the safety of marketed drugs under the
practicalconditionsofclinicalusageinlargecom
[1]
munities. Itinvolvesmonitoringdrugsafetyand
identifying adverse drug reactions (ADRs) in
humans,assessingrisksandbenefits,andrespond
ing to and communicating drug safety concerns;
recently,ithasbeensuggestedthattherecouldbe
moreemphasisonextendingknowledgeofsafety
[2]
ratherthanfocusingondemonstratingharm.
The above definition makes no distinction be
tweenpharmacovigilanceofconventionalandherbal
1
medicines. Indeed, there is no need, nor is it

desirable, to separate the two; pharmacovigilance
should embrace all preparations used medicinally
regardless of their regulatory status, pharmaceutical
composition, cultural use and philosophical frame
work. Hence, the same aims and activities of
pharmacovigilanceapplytoherbalmedicines.How
ever,pharmacovigilanceactivitieslargelyhavebeen
focused on conventional medicines, and the current
model of pharmacovigilance and its science and
processes have developed in relation to synthetic
drugs. Applying the existing model and its tools to
monitoring the safety of herbal medicines presents
uniquechallengesinadditiontothosedescribedfor
conventionalmedicines,anditisimportantthatthese
areunderstoodbyallstakeholders.
Thereisanincreasingawarenessatseverallevels
of the need to develop pharmacovigilance practices
for herbal medicines; the WHO, for example, has
[3]
produced draft guidelines on this. Awareness has

arisennotonlybecauseoftheextensiveuseofherbal
medicines,butalsobecauseinrecentyearstherehave
beenseveralhighprofileherbalsafetyconcernswhich
havehadanimpactonthepublic
health.Againstthisbackground,thispaperaimsto
provideacriticaloverviewofthecurrentstateof
pharmacovigilanceofherbalmedicinesintheUK,
and to discuss the particular challenges that this
area presents. The paper is written from a UK
perspective, but also has wider international
relevance, particularly to those countries with
spontaneous reporting schemes and a healthcare
systemsimilartothoseoftheUK.
1. Herbal Medicines: Challenges
for Pharmacovigilance
Theuniquecharacteristicsofherbalmedicines,
andthewaysinwhichherbalmedicinesarenamed,
perceived,sourced,utilisedandregulatedraiseim
portant issues and challenges for pharmacovigi
lance.
Some issues arise because most herbal
medicinescanbeobtainedwithout aprescription
from various outlets, not only pharmacies. Thus,
the problems that apply to pharmacovigilance of
conventional nonprescription medicines, for
example,thatgenerallytheirusedoesnotinvolve
a prescriber and is not recorded or monitored
through the National Health Service (NHS), also
apply to herbal medicines. Other problems are
specifictoherbalmedicinesandpresentdifficulties
over and above those described for conventional
prescriptionandnonprescriptionmedicines.
1.1 Characteristics of Herbal Medicines
Herbal medicinal products (also known as

phytomedicines or phytotherapeutic preparations)
are "medicinalproductscontainingasactivesub
stances exclusively herbal drugs or herbal drug
[4]
preparations", i.e.theycontainasactiveingredi
Drug Safety 2003; 26 (12)
Pharmacovigilance of Herbal Medicines
831
Table I. Formal definitions and other terms used in relation to herbal medicines
Term
Definition
Herbal remedy
a medicinal product consisting of a substance produced by subjecting a plant or plants to drying, crushing
or any other process, or of a mixture whose sole ingredients are two or more substances so produced, or of
a mixture whose sole ingredients are one or more substances so produced and water or some other inert
[5]
Herbal substance
(herbal drug)
substances.
all mainly whole, fragmented or cut plants, plant parts, algae, fungi, lichen in an unprocessed, usually dried
form but sometimes fresh. Certain exudates that have not been subjected to a specific treatment are also
considered to be herbal substances. Herbal substances are precisely defined by the plant part used and the
[6]
Herbal preparation
(herbal drug preparation)
Herbal medicinal product
botanical name according to the binomial system (genus, species, variety and author).
preparations obtained by subjecting herbal substances to treatments such as extraction, distillation,
expression, fractionation, purification, concentration and fermentation. These include comminuted or
powdered herbal substances, tinctures, extracts, essential oils, expressed juices and processed exudates.
any medicinal product, containing as active ingredients one or more herbal substances or one or more
herbal preparations, or one or more such herbal substances in combination with one or more such herbal
preparations; in addition, the product may contain vitamins or minerals or other non-biological
substances for which there is well-documented evidence for its safety; the action of the non-herbal
substances must be ancillary to that of the herbal active ingredients.
[6]
[7]a
Herbal constituent
a specific chemical compound found in a herbal ingredient, e.g. hyperforin found in the aerial parts (herb) of
St Johns wort.
Herbal ingredient
a specific individual medicinal plant and the plant part, present in a herbal medicine, e.g. St Johns wort
herb present in St Johns wort tablets.
The text in bold is a recent amendment, the wording of which is still to be confirmed at a European level.
entsonlycrudeand/orprocessedplantsand/orplant
parts; an isolated chemical constituent which
originatesfromplantmaterialisnotaherbalmedi
cine.Thetermherbalmedicinesisalsousedgener
ally to describe both relatively crude preparations,
suchasherbaltinctures,usuallysuppliedbyherbal
medicine practitioners (medical herbalists), and
manufacturedorfinishedherbalmedicinalproducts,
usually formulated as tablets or capsules and avail
ableforpurchasewithoutaprescription.Legaldefi
nitionsrelatingtoherbalmedicinesandexplanations
ofothertermsusedaregivenintableI.
In contrast with conventional medicines, herbal
medicines are chemically rich complex mixtures
comprising several hundreds of constituents, often
more. For many herbal medicines, the chemical
constituents are unknown, and even for those with
welldocumented phytochemistry, there are few for
whichthespecificconstituentsresponsibleforphar
macologicalactivityarefullyunderstood.Theprofile
ofconstituentsisnotuniformthroughoutaplant,and
formanyplants,onlyaspecificplantpart,orparts,
such as roots or leaves, is (or should be) used
medicinally. Moreover, the precise profile of con
stituentsislikelytovarybothqualitativelyand
quantitativelybetweendifferentbatchesofherbal
starting materials because of one or more of the
followingfactors:
interorintraspeciesvariationinconstituents
environmental factors, such as climate, and growing

conditions
timeofharvestingtheprofileofconstituentscanvaryeven
overthecourseofaday
postharvestingfactors,suchasstorageconditionsand
drying.[8]
The method ofprocessing crude herbalmaterial,

forexample,thetypeofextraction,canalsoinfluence
the precise chemical composition of a herbal
preparation or product. Many herbal medicinal
products contain several herbal ingredients, and
medicalherbalistsusuallyprescribecombinationsof
herbaltincturesoftensuppliedasamixture,inboth
cases, furtheradding tothe chemicalcomplexity of
theherbalmedicinetakenbythepatient.Thechemi
cal complexity of herbal medicines creates difficul
ties in determining their clinical pharmacokinetics,
pharmacodynamics and toxicology and, equally,
whereasafetyconcernhasbeenidentifiedinassoci
ationwithaparticularherbalmedicine,establishing
whichconstituent(s),evenwhichherbalingredi
832
Barnes
ent(s) with combination herbal medicines, are

implicatedisproblematic.
Forthereasonsgivenabove,itislikelythatthere
will be variations in the chemical composition of
herbal medicines containing the same herbal ingre
dient but produced by different manufacturers; this
willapplytobothlicensedherbalmedicinalproducts
and unlicensed herbal medicines. Several studies
have found important differences in the pharma
ceutical quality of different products on the US
market. For example, variations in the content of
major constituents in St Johns wort (Hypericum
perforatum) products (which in several cases also
differedmarkedlyfromconcentrationsstatedonthe
[9]
label), and variations in and unacceptably high

concentrations (in several cases >25 000 parts per
million) of ginkgolic acids, which potentially are
[10]
allergenic,inginkgoproducts. Standardisationon
contentofcertainconstituentsisanapproachusedby
some manufacturers to achieve more consistent
pharmaceutical composition, but its usefulness is
limitedat present since thespecific active constitu
entsareknownonlyforafewherbalmedicines.
Because of the variations thatcan exist between
differentmanufacturersproductsandpreparationsof
the same herbal ingredient, evidence of safety (and
efficacy) should be considered in this light; strictly
speaking,evidenceisproductorextractspecific,and
should be extrapolated only to those products or
extracts which have been shown to be
[11]
pharmaceutically equivalent and bioequivalent.

Thisislargelyimpracticalatpresent,giventhelimited
dataavailableforherbalmedicines;nevertheless,the
differencesbetweendifferentpreparationsofaherbal
ingredient should not be ignored. In many cases,
becauseofthenatureofherbalmedicines,agroupof
relatedconstituents,ratherthanasingleconstituent,is
likelytoberesponsibleforanobservedadverseeffect.
Inthiscase,itmaybeappropriatetogrouptogether
different preparations and products containing the
same herbal ingredient or group of constituents in
ordertodetectsignals.Insomeways,thisissimilarto
investigating class effects with conventional
medicines.
The names that are usedfor herbal medicines

presentafurtherproblem(seealsosection2.1.1.,
ADRReportingForm).Often,commonorvernacu
lar names are used to describe herbal medicines,
butthesevarywidely,andmaybeusedtodescribe
morethanonespecies,andcannotbeprecise.
Contrarytopopularbelief,herbalmedicinesare
not safe because they originate from natural
sources; some plants are highly poisonous, and
manyothershaveinherentlytoxicconstituents.For
example,metabolitesofunsaturatedpyrrolizidine
alkaloids,suchassenecionine,arehepatotoxicin
humans, and carcinogenic and mutagenic in ani
[8]
mals. Senecionine is found inliferoot (Senecio
aureus) and in other Senecio species, such as
Senecioscandens,whichhasbeenreportedasan
ingredient in a traditional Chinese medicine
[12,13]
productQianbaiBiyanPianfoundintheUK.
Other known intrinsically toxic groups of
constituents, their effects and examples of plant
sourcesincludearistolochicacids(nephrotoxicand
carcinogenic), found in Aristolochia species
throughout the plant, sesquiterpene lactones
(allergenic), found in feverfew (Tanacetum
parthenium) and other species in the Asteraceae
family,andfuranocoumarins(phototoxic),foundin
angelica(Angelicaarchangelica)andotherspecies
[8]
belongingtotheApiaceaefamily.
1.2 Utilisation of Herbal Medicines and
Related Issues
Theuseofherbalmedicinesisapopularhealth
careapproachamongpatientsandconsumersinthe
UK,althoughtherearefewreliableestimatesofthe
prevalenceofuse.Estimatesofherbalmedicineuse
amongadultsinEnglandcomefromacrosssectional
survey(n=5010;responserate:59%)carriedoutin
1998whichfoundthat19.8%(95%CI18.321.3)had
purchased an overthecounter herbal medicinal
product and that 0.9% (95% CI 0.61.3) had con
[14]
sulted a medical herbalist in the previous year.

Therearenolongitudinaldataforprevalenceofuseof
herbal medicines for the UK at present, although
marketresearchdataindicateincreasingsalessales
oflicensedandunlicensedherbalmedicinal
833
productswereworth75millionin2002,an in
[15]
creaseof57%overtheprevious5years. Studies
carried out in other developed countries, such as
Australia and the US, also suggest increasing
prevalenceofuseofherbalmedicinesamongthe
[16,17]
general adult population.
Extrapolating
estimatesofherbalmedicineusefromsuchstudies
suggests that large numbers of people are being
exposed to herbal medicines; this in itself is of
concernforthepublichealth.
IntheUK,herbalmedicinesareusedbyawide
range of individuals for both acute and chronic
conditions.Manyherbalmedicinalproductsarepur
chased for maintenance of general health and well
being,andforuseinthepreventionandtreatmentof
minor,commonailments.Useisnotnecessarilybased
onevidence,norlimitedtosymptomsandconditions
suitableforselftreatment.Herbalmedicinalproducts
are also used by individuals with serious chronic
diseases,includingcancer,AIDS,multiplesclerosis,
and asthma, and many other conditions, by older
patients,andpregnantorbreastfeedingwomen,and
[18]
areadministeredbyparents/guardianstochildren.
Similarly,medicalherbalistsuseherbalmedicinesto
[19]
treatavarietyofconditions. Somepatientgroups,
suchaschildrenandolderpeople,areatincreasedrisk
ofadversedrugeffects,andthereisnoreasonwhy
this should not also apply where they use herbal
medicines. Other groups, for example, pregnant
women, may use herbal medicines in preference to
conventionalmedicinesbecausetheyareperceivedto
besafer,withoutrealisingthatlittleisknownabout
the effects of herbal medicines taken during
pregnancy.
Typically, users of herbal medicinal products do
not seek professional advice in selecting herbal
medicines, but rather rely on friends or relatives
recommendations, and information in the popular
[20,21]
media.
Herbal medicines are widely available
forpurchaseovertheinternetandfromretailoutlets
in which there is no trained healthcare professional
[22]
available (see section 1.3.1). Even where herbal

medicinalproductsarepurchasedfrompharmacies,a
consumerorpatientmaynothaveanyinteractionwith
apharmacistortrainedpharmacycounterassis
tant,orifaconsultationdoesoccur,pharmacystaff
maynothavesufficientknowledgetofeelconfident
about providing information and advice on herbal
[23]
medicines. Asmallproportionofusersofherbal
medicines seeks treatment from a herbalmedicine
practitioner,butatpresent,thereisnolegalrequire
mentforsuchpractitionerstohaveundertakentrain
inginherbalmedicineortobelongtoaprofessional
organisation for herbalmedicine practitioners, and
while many herbalmedicine practitioners will have
takenthesesteps,somewillnot.
A related issue is that some users of herbal

medicinesmaynotdisclosethisusetoahealthcare
[21]
professional; equally, healthcare professionals
donotasktheirpatientsroutinelywhethertheyare
using herbal medicines, even when receiving
reportsfrompatientsofsuspectedADRsassociated
with conventional medicines, and rarely record
information on herbalmedicine use on patient
[24,25]
records.
It is possible, therefore, that
undisclosed herbalmedicine use could be an
alternative explanation for reports of suspected
ADRsassociatedwithconventionalmedicines.
Disclosureofherbalmedicineusetohealthcare
professionals is particularly important where pa
tientsstart,stop,orarealreadyreceivingtreatment
with conventional medicines and, equally,
individuals consulting medical herbalists should
disclose their current use of conventional
medicines, because there may be a potential for
drugherbinteractions.Informationontheextentto
which concurrent use of herbal and conventional
medicinesoccursislimited,althoughpreliminary
datasuggestthatitmaybeextensive.Inacross
sectional survey of complementarytherapy use
among adults in the US (n = 2055 respondents;
60% weighted overall response rate), 44% were
regular users of prescription medicines and, of
these,18.4%wereusingconcurrentlyaherbalor
[16]
highdosevitaminpreparation. Inasmallstudy
conductedintheUK,59%ofherbalmedicineusers
identified in pharmacies and healthfood stores
claimed that they had used herbal medicines
concurrentlywithconventional medicines,mostly
prescriptionmedicines,inthepreviousyear.
[21]
834
Barnes
In summary, the ways in which herbal

medicines are described (named), perceived and
obtained,togetherwithusersbehaviourtowards
herbalmedicinesandissuesrelatingtohealthcare
professionals and herbal practitioners practice
present opportunities for herbal medicines to be
used inappropriately, even unsafely, and for
suspectedADRstogoundetectedandunreported.
1.3 Regulation of Herbal Medicines
1.3.1 Current Regulatory Framework
Aswithallmedicines,theoriginsofregulationand
pharmacovigilance for herbal medicines lie in the
thalidomidetragedyofthe1950sand1960s.Thiswas
the milestone which led, of course, to the
establishmentoftheCommitteeonSafetyofDrugs
(nowtheCommitteeonSafetyofMedicines,CSM),
anearlywarningsystem(theyellowcardscheme)
for doctors to report their suspicions on adverse
effects of drugs, and legislation in the form of the
Medicines Act 1968 requiring pharmaceutical
companiestosatisfythecompetentauthority(nowthe
Medicines and Healthcare products Regulatory
Agency,MHRA)ofthequality,safetyandefficacyof
theirnewmedicinesbeforemarketing.
Therearearound600licensedherbalmedicinal
productsontheUKmarket,althoughmostofthese
are not new marketing authorisations, but are
products which initially were granted product
licencesofright(PLRs)astheywereonthemarket
whenthemedicineslicensingsystemwassetupin
1971. When PLRs for herbal medicinal products
werereviewedbythecompetentauthority,manu
facturersofthoseintendedforuseinminorself
limiting conditions were permitted to rely on
bibliographic evidence to support efficacy and
safety,ratherthanbeingrequiredtocarryoutnew
[8]
testsandcontrolledclinicaltrials. So,although
many herbal medicinal products have product
licences, the products have not necessarily
undergonethestringenttestingrequiredtoobtaina
fullmarketingauthorisationtoday,butratherhave
reliedonevidencefromlongstandinguse.
Other herbal medicines available in the UK are
soldeitherasherbalremediesexemptfromlicensing
under section 12 of the Medicines Act 1968, or as

unlicensedfoodsupplementswithoutmakingmedical
claims and regulated under food, not medicines,
[8]
legislation. Herbalmedicinesmeetingthedefinition
ofaherbalremedy(seetableI)andcompoundedand
supplied by herbal practitioners on their own
recommendation currently are exempted from
[5]
licensing requirements under section 12 (1). This

exemption was initially intended to give herbal
practitioners the flexibility to prepare remedies for
theirpatients,althoughthetermisnotdefinedand,at
present,thereisnostatutoryregulationofherbalistsin
the UK. A Herbal Medicine Regulatory Working
Group has been established to consider appropriate
legislation and reform of section 12 (1) and is
expectedtoreportduring2003.
[26]
Manyherbalmedicinalproductsaresoldunderthe
section 12 (2) exemption, which exempts from
licensing requirements those herbal remedies con
sistingsolelyofdried,crushedorcomminuted(frag
mented)plantssoldundertheplantorbotanicalname
andwithnowrittenrecommendationsastotheiruse.
[5]
Inotherwords,suchproductsmustnotcontainany
nonherbal active ingredients, must not use
proprietarynamesandmustnotmakemedicalclaims.
Somemanufacturersareunawareof,orignore,these
conditionsandillegalunlicensedherbalproductscan
befoundforsale,ascurrentlythereisnorequirement
formanufacturerstoconsultthecompetentauthority
before placing an unlicensed herbal medicinal
productonthemarket.MHRAdoeshavethestatutory
power to decide whether a particular marketed
unlicensed product satisfies the definition of a
relevantmedicinalproductand,therefore,issubject
to the usual provisions of regulations relating to
MedicinesforHumanUse,unlessitmeetscriteriafor
exemption (i.e. as provided in section 12 of the
MedicinesAct).HerbalmedicinesavailableintheUK
include some traditionalChinese medicines (TCM)
and Ayurvedic medicines(bothof whichoften also
containnonherbalsubstances),andthesearesubject
to the same legislation as are Western herbal
[27]
medicines. Therearefurtherrestrictionsoncertain
toxicherbalingredients,suchasAristolochiaspecies,
foundin
835
someTCMproducts,andonotherherbalingredi
ents that may be confused with toxic herbal
[28]
ingredients. However, unlicensed herbal
productscontainingthesebannedspeciescontinue
[13]
tobefound.
IntheUK,licensedmedicinalproducts,including
licensedherbalmedicinalproducts,areclassifiedas
prescriptiononly medicines (POMs; generally may
onlybesoldorsuppliedfromaregisteredpharmacy
in accordance with a prescription given by an
appropriatepractitioner),pharmacymedicines(P;may
onlybesoldorsuppliedfromaregisteredpharmacy
andbyorunderthesupervisionofapharmacist),and
general sales list medicines (GSL; may be sold or
suppliedatregisteredpharmaciesorotherbusinesses
[5]
whichcanbeclosedtoexcludethepublic). Most
licensed herbal medicinal products are GSL
medicines. Potentially hazardous herbal substances
(e.g. Digitalis leaf)are controlledasPOMs;certain
other plants or plant parts (e.g. yohimbe bark;
Pausinystalia yohimbe) are controlled as P
medicines,butsomeofthese(e.g.belladonnaherb;
Atropabelladonna)canbesoldorsuppliedbyherbal
practitioners (see earlier in this section) if certain
conditions are met (e.g. limits on maximum dose,
maximum daily dose and/or strength). In practice,
theserestrictedherbalsubstancesarerarely,ifever,
sold or supplied by pharmacists, but some, such as
Ephedra species, where permitted, are utilised by
herbalpractitionersintheirpractice.
Thecurrentregulatoryframeworkpresentsseveral
major problems for pharmacovigilance of herbal
medicines.Whereasmanufacturersoflicensedherbal
medicinal products are required to comply with
regulatoryprovisionsonpharmacovigilanceassetout
[29]
in Directive 2001/83/EC, manufacturers of

unlicensed herbal products and those sold under
exemptionsfromlicensingarenotrequiredtodoso,
i.e.fortheseproducts,whichcomprisethemajorityof
herbalmedicinesontheUKmarket,manufacturers
havenoobligationtokeeprecordsofsuspectedADRs
associated with these products, nor to report these
suspectedADRstothecompetentauthority.Thisis
alsothecasewhereherbalmedicinesaresuppliedto
patientsbymedicalherbalists.
In addition, the range of possible regulatory ac

tions that the competent authority can take in res
ponsetoaherbalsafetyconcernislimitedforunli
censed herbal medicinal products and, for some
responses, requires the voluntary cooperation of
herbalmedicines manufacturers. For example, after
important interactions between St Johns wort and
certain prescription medicines emerged around
1999/2000,MHRAtookthedecisionthatprovisionof
warnings on St Johns wort products was an
appropriate partofthe regulatory response, but this
required the cooperation of manufacturers of unli
censed St Johns wort products. At the same time,
marketing authorisation holders of conventional
medicines believed to interact with St Johns wort
productswereobligedtomakevariationstoproduct
information for their relevant products. Similarly,
whenanassociationbetweenuseofkavakava(Pip
er methysticum) preparations and liver toxicity was
being investigated by the CSM, the herbal sector
agreed to withdraw kavakava products from sale.
Voluntary withdrawal worked reasonably well ini
tially,butastheperiodofevaluationdrewon,some
retailoutletsbegansellingkavakavaproductsagain.
Communitypharmacists,however,hadaprofessional
andethicalresponsibilitynottodoso.
[30,31]
Other issues relevant to pharmacovigilance arise

because manufacturers of unlicensed herbal medici
nal products are not required to demonstrate to
MHRA the quality, safety and efficacy of their
productsbeforemarketing.Theimportanceofphar
maceutical quality for the safety (and efficacy) of
[8,32,33]
herbal medicinal products is wellrecognised,

but manufacturers are required only to demonstrate
pharmaceutical quality standards for their licensed
herbal medicinal products. Some manufacturers of
unlicensed herbal medicinal products may have ap
propriate quality control and quality assurance pro
ceduresfortheirproducts,butothersdonot,andthe
pharmaceutical quality of many unlicensed herbal
medicinalproductsisofrealconcern.Inadditionto
difficultieswithassuringpharmaceuticalqualitydue
to the variation in chemical composition, quality
problems with unlicensed herbal products include
intentionaloraccidentalsubstitutionofspecies,con
836
Barnes
taminationwithrestrictedortoxicsubstances,in
cludingprescriptionmedicines,anddifferencesbe
[8,13]
tweenlabelledandactualcontents.
Itisessen
tial,therefore,whenassessingreportsofsuspected
ADRs associated with a particular unlicensed
herbal medicine to establish whether the herbal
ingredient(s) implicated are what the product
actuallycontains,andwhethertheproductcould
beadulteratedorcontaminated.Ideally,asampleof
thesuspectedherbal medicineshouldberetained
forpharmaceuticalanalysisifnecessary.
Thereisagenerallackofobjectiveinformationon
thesafetyofmanyherbalmedicines.Thishasarisen
in part because under the current regulatory
frameworkthereislittleincentiveformanufacturers
to carry out preclinical tests and clinical trials.
Postmarketing surveillance studies involving certain
herbal medicinal products have been conducted by
some manufacturers (usually those based in Ger
many),butthisistheexception.Generallyspeaking,
there is a lack of information on the types and
frequency of adverse effects, including interactions
withothermedicines,foods,alcohol,diseaseandso
forth,andotheraspectsrelevanttosafetyforherbal
medicines, such as their active constituents,
pharmacokinetics, pharmacology, use in special pa
tientgroups(e.g.children, olderpeople,individuals
with renal or hepatic disease, pregnant or breast
feedingwomen),effectsoflongtermuse,andsoon.
Itisoftenarguedthatherbalmedicineshavealong
historyoftraditionaluseandthatthisprovidesevi
dencefortheirsafety(andefficacy).However,while
thetestoftimemayhaveidentifiedinherentlytoxic
plants, it cannot, for example, identify delayed
adverse effects, effects that may arise from use in
patientswithmodernillnesses,suchasHIV/AIDS,
andsafetyissuesarisingfromhowherbalmedicines
are utilised today, for example, concurrently with
[34]
conventional medicines. Certainly, there are

examplesoftypeAreactions(thosethattypicallyare
dose dependent and related to the pharmacological
effects of the medicine) and type B reactions
(typicallyunrelatedtodose,idiosyncratic)andother
types of ADRs (e.g. delayed effects in the user or
offspringremotefrommedicineuseinthe
user) associated with the use of certain herbal

[35]
medicines.
Inaddition,theefficacyofmanyherbalmedi
cineshasnotbeenevaluatedinrandomisedclinical
trials.Evenforwelltestedherbalmedicines,such
as certain extracts of St Johns wort herb which
have been assessed in around 30 randomised
clinicaltrialsindepression,onlyasmallnumberof
clinicaltrial participants has been exposed to a
specificmanufacturersproduct.Furthermore,there
arefewlongtermclinicaltrialsofherbalmedicines
intended for longterm use. For comparison,
conventionalmedicineshavebeentestedinaround
1500 patients before they reach the market. The
lackofinformationonthesafetyandefficacyof
herbal medicines makes it difficult to carry out
benefitriskassessments.
Insummary,intheUK,thecurrentregulatory
framework allows unlicensed herbal medicines,
whichmaybeofinadequatepharmaceuticalquality
and for which there is a lack of information on
safety aspects, to be placed on the market and
obtainedbyconsumersandpatientsfromarangeof
retail outlets without a prescription or other
involvement of a healthcare professional.
Manufacturersareundernoobligationtocarryout
pharmacovigilanceofsuchproducts.Bycontrast,
conventionalPorGSLmedicinesarepermittedto
besoldorsuppliedwithoutaprescriptionbecause
theyhaveahistoryofrelativesafety.
1.3.2 Proposed New Regulatory Framework
The need for a new regulatory framework for

herbalmedicinalproductswasfirstdiscussedinthe
late1980sand,forseveralreasons,todayitisrecog
nisedwidelythat theexistingregulatoryframework
does not adequately protect the public health. In
particular,thecurrentsystemdoesnotgiveconsum
ers and patients adequate protection against poor
quality and unsafe unlicensed herbal medicinal
products.Italsodiscriminatesagainstmanufacturers
oflicensedherbalmedicinalproductsastheircostsare
likelytobehigherbecauseoftheneedtocomplywith
the principles of good manufacturing practice and
otherregulatoryprovisions.
837
Table II. Key features of proposed European Union (EU) Directive 2002/0008 (amends 2001/83/EC) on traditional herbal medicinal
products
[6,7]
Establishes a Committee on Herbal Medicinal Products which will be part of the EMEA and will take over the tasks of the CPMP with
regard to authorisations or registrations of herbal medicinal products by member states. Other tasks will include producing EC herbal
monographs, and establishing a positive list of herbal substances (to include indication, route of administration, strength and so on)
allowed under the directive.
Requires EU member states to set up a specified simplified national registration procedure for traditional herbal medicinal products that
could not fulfil medicines licensing criteria.
Main features of requirements for registration include:
products for oral, external or inhalation use only

minor indications only (suitable for self-diagnosis and self-treatment)
evidence that the herb has been used traditionally for at least 30 years, including at least 15 years within the EC;
period of traditional use can include the transition period
reliable identification of raw materials and use of appropriate quality herbal ingredients, i.e. compliance with Ph Eur
standards where they exist and with manufacturers own specification otherwise
systematic quality assurance and quality control throughout the manufacturing process: compliance with principles of GMP
and other relevant European guidelines; qualified person responsible for release of batches onto the market; manufacturers
licence or wholesale dealers licence where appropriate; inspection of premises
provision of bibliographic data on safety with an expert report
labelling, information and advertising requirements in accordance with 2001/83/EC and relevant national regulations
compliance with pharmacovigilance requirements in accordance with 2001/83/EC
transition period of at least 5 years, probably 7 years, once the directive comes into force
CPMP = Committee on Proprietary Medicinal Products; EC = European Community; EMEA = European Agency for the Evaluation of
Medicinal Products; GMP = good manufacturing practice; Ph Eur = European Pharmacopoeia.
Againstthisbackground,adraftEuropeanUnion
(EU) directive (2002/0008, which amends 2001/83/
EC) has been produced which aims to establish a
harmonisedlegislativeframeworkforauthorisingthe
[6]
marketingoftraditionalherbalmedicinalproducts.
ThedirectivewillrequireEUmemberstatestosetup
a simplified national registration scheme for
traditionalherbalmedicinalproductsmeetingdefined
criteria.Thekeyfeaturesofthedirectiveatthetimeof
writing are summarised in table II, although the
directive is still under discussion and amendments
mayyetbemade.Inessence,theproposeddirective
will require manufacturers wishing to obtain
registrations for their traditional herbal medicinal
productsunderanationalschemetodemonstratethe
quality and, to some extent, the safety of their
products, whereas the usual efficacy and, to some
extent, safety requirements will be replaced by
evidenceoftraditionaluse.Anothermajorchangeis
that manufacturers of products registered under the
directivewillberequiredtocomplywithinformation
and labelling requirements. Currently there is no
requirementformanufacturersofunlicensed
herbalmedicinalproductstoprovidesystematicin
formationwiththeirproducts.
The proposed directive will have an important
impact on pharmacovigilance of herbal medicines,
onceitcomesintoforce.Manufacturersoftraditional
herbal medicinal products registered under the UK
national scheme establishedunder thedirective will
be required to comply with relevant existing
pharmaceutical legislation, including the provisions
[29]
onpharmacovigilance. Severalofthesemaypose
problemsformanufacturerswithlittleornoexperi
ence in this area. For example, the requirement to

haveconstantaccesstoanappropriatelyqualifiedand
experienced person responsible for pharma
covigilance, implementation of the use of Medical
DictionaryforRegulatoryActivities(MedDRA)and
connectiontoandcompliancewithEudraVigilance.
Therehasbeensomediscussionaboutthepossibility
ofmakingsomeallowancesformanufacturerswith
regardtothequalifiedpersonresponsibleforproduct
quality;itisnotclearwhetheranysimilarallowances
will be made with respect to the qualified person
responsibleforpharmacovigilance.Accordingtothe
currenttimescale,thedirectiveis
838
Barnes
expectedtocomeintoforcearoundtheendof2004;
followingthis,therewillbeatransitionperiodofat
least5years(theprecisedurationiscurrentlyunder
discussionbutprobablywillbe7years).
[36]
Other developments at the European level are

concerned,atleastinpart,withpharmacovigilance
ofherbalmedicines.TheEuropeanAgencyforthe
EvaluationofMedicinalProductsHerbalMedici
nalProductsWorkingParty(previouslytheadhoc
HerbalMedicinalProductsWorkingGroup,setup
in1997)hasseveralpharmacovigilanceissueson
itsagenda.
[37]
2. Methods for Pharmacovigilance of

Herbal Medicines
Some standard methods used in pharma
covigilance, particularly spontaneous reporting
schemes,areusedtomonitorthesafetyofherbal
medicines, although these methods are less well
establishedthanforconventionalmedicines.Other
methods, such as prescriptionevent monitoring,
havenotyetbeenappliedtoexploringthesafetyof
herbalmedicines.Allavailablepharmacovigilance
toolshaveimportantlimitationswithregardtotheir
useininvestigatingthesafetyofherbalmedicines,
in additiontothose already recognised, and it is
likelythat modified,evennovel,methodsarere
quired.Thissectiondiscussestheavailablemeth
ods, with a focus on spontaneous reporting
schemes,andtheparticularchallengesthatherbal
medicinespresentforeach.
2.1 Spontaneous Reporting Schemes
The future of spontaneous reporting schemes in

[2]
pharmacovigilancehasbeenquestioned, althoughit
is likely that this point was raised in relation to
conventional medicines for which other wellestab
lished tools, such as computerised healthrecord
databases, can be used for pharmacovigilance pur
poses.Bycontrast,spontaneousreportingforherbal
medicines is in the early stages of its development
and,atpresent,intheabsenceofothertoolsand/or
resources, is the main method of generating and
detecting signals of potential safety concerns asso
ciatedwithherbalmedicines.Spontaneousreporting
schemesappeartofunctionreasonablyeffectively
asapharmacovigilancetoolforherbalmedicinesin
countriessuchasGermanywhereherbalmedicinal
productsareregulatedasmedicines,frequentlypre
scribed by physicians and well known to other
healthcareprofessionals,particularlypharmacists.
[38]
However,spontaneousreportingislikelytobe
farlesseffectiveincountriessuchastheUKwhere
herbal medicines are marketed mainly as
unlicensed products with no obligation for
manufacturers to report suspected ADRs to the
competent authority, and where herbal medicines
are used mostly in selftreatment without any
supervisionfromahealthcareprofessional.
2.1.1 UK National Spontaneous Reporting Scheme
The CSM/MHRAs national spontaneous report

ingschemeforsuspectedADRreportingbyhealth
care professionals(alsoknownasthe yellowcard
schemebecauseoftheformusedtoreportsuspected
ADRs to the CSM/MHRA) has applied to licensed
medicines,includinglicensedherbalmedicines,since
its inception in 1964. However, the inclusion of
licensedherbalmedicinesintheschemewasnotwell
publicised until October 1996, over 30 years later,
whentheschemewasextendedtoincludereporting
[39]
for unlicensed herbal medicines. This move

followeda5yearstudyoftraditionalremediesand
food supplements, carried out by a UK Medical
[40]
Toxicology Unit, which identified suspected

ADRsassociatedwiththesetypesofproducts.The
extensionallowedthosewithofficialreporterstatus
atthetime,doctors,dentistsandcoronersonlyto
submit reports for unlicensed herbal medicines, but
didnot(andcouldnot)placeanystatutoryobligation
onmanufacturerstoreportsuspectedADRsassociated
withtheirunlicensedherbalproducts.
In April 1997 and November 1999, the scheme
underwent further extensions to allow reporting of
suspected ADRs by all hospital and community
[41]
pharmacists, respectively. Community pharma

cists were encouraged by the CSM and MHRA to
concentrateonareasoflimitedreportingbydoctors,
namelylicensedandunlicensedherbalproducts,and
othernonprescriptionmedicines.
[41]
Thisextension
839
140
120
Number of reports
100
80
60
40
20
0
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000 2002
Year received
Fig. 1. Numbers of reports of suspected adverse drug reactions associated with herbal medicines received by the UK Committee on Safety
of Medicines/Medicines and Healthcare products Regulatory Agencys yellow card scheme for the period 1964 to July 25th, 2003. Blue
bars (i.e. pre-1996) represent licensed herbal medicines; black bars (1996 onwards) represent reports for both licensed and unlicensed
herbal medicines following extension of the scheme to unlicensed herbal medicines in October 1996; the dotted line above the 2003 bar
represents an estimate of the total number of reports for the full year. Source: Adverse Drug Reactions On-line Information Tracking.
followed a 1year pilot scheme for community

pharmacistADRreporting,carriedoutinthefour
CSMregionsduring199798andinvolvingaround
3200 pharmacies, which showed that community
pharmacists,comparedwithgeneral practitioners
(GPs),submittedagreaterproportionofreportsof
suspectedADRsassociatedwithherbalmedicines
(thenumbersofherbalADRreportsasaproportion
of the total number of reports submitted by
pharmacistsandGPswere4/96(4.2%)and8/1975
[42]
(0.4%), respectively; p < 0.001). However,

numbersofherbalADRreportssubmittedbyboth
groupsofreporterswereverylowandrepresented
anaverageofonlyoneandtworeportsperCSM
regionforpharmacistsandGPs,respectively.
Despitetheseinitiativestostimulate reportingof
suspected ADRs associated with both licensed and
unlicensedherbalmedicines,numbersofherbalADR
reports submitted to the CSM/MHRA remain very
low. From 1964 until the end of 1995, 832 reports
werereceived.
[43]
Fortheperiod1996(when
[43]
theyellowcardschemewasextendedtounlicensed
herbal medicines and when its inclusion of herbal
medicines was first well publicised) to 2002 inclu
sive,467reportsofsuspectedADRsassociatedwith
herbalmedicineswerereceived(seefigure1).Most
frequently,thesereportsrelatedtoproductscontain
ingtheherbalingredientsStJohnswort(Hypericum
perforatum), ginkgo (Ginkgo biloba), peppermint
(Mentha piperita), Echinacea species, senna and
valerian (Valeriana officinalis). It is not known
whether the low numbers of reports of suspected
ADRs associated with herbal medicines simply re
flectalowfrequencyofadverseeffectswithherbal
medicines,orwhetherthereareotherexplanations,for
example,substantialunderreporting.
The number of herbal ADR reports received in
creasedovertheperiod19992002,withapeakinthe
year 2000 around the time that reports emerged of
suspected interactions between St Johns wort and
certain prescription medicines. In part, this simply
reflectedanincreaseinnumbersofreportsof
840
Barnes
suspectedADRsassociatedwithStJohnswort
60%(82/140)ofherbalreportsfortheyear2000
(42%for1999and13%for1998),and40%(138/
345) of all herbal reports received during this
periodrelated toSt Johns wort, witharound 40
reportsintotaldescribingdruginteractionswithSt
Johnswortbuttherewasalsoa(small)general
increase in numbers of herbal ADR reports
submittedduringthisperiod.
[43]
Itisnotyetclearwhetherthisjustreflectsyearto
year variation, or whether it has been sustained.
Typically,ayearonyearincreaseinthe numberof
herbal suspected ADR reports submitted could be
expected,withfurtherincreaseswiththeadditionof
new recognised reporters and initiatives aimed at
stimulating herbal ADR reporting. However, the
inclusionofpharmacistsdoesnotyetappeartohave
hadamarkedimpactinthisregard,sincetheaddi
tional reports of suspected ADRs associated with
herbal medicines submitted during 19992002 were
notsubmittedsolelybypharmacistsbutalsobyother
[43,44]
recognised reporters.
For comparison, in 1964
whentheyellowcardschemebegan,therewere1414
submittedreportsofsuspectedADRsassociatedwith
conventionalmedicines.Thisincreasedsteadilyuntil
1977 when there was a surge in reporting (10 921
reports for the year) around the time practolol was
associated with oculomucocutaneous syndrome and
withdrawnfromthemarket,andwheninitiativesto
encourage reporting of suspected ADRs were
introduced, for example, the introduction of the
newsletter Current Problems in Pharmacovigilance.
[45]
Annualreportingrateshavesinceincreased,with
somefluctuations,currentlytoaround20000reports
peryear(whichincludesthesmallnumberofherbal
ADRreports),givingatotalofover450000reports
totheendof2002.
Further extensions to the reporter base for the
yellowcardschemeoccurredinOctober2002when
all nurses, midwives and health visitors became
[46]
recognisedreporters. Atthesametime,electronic
reporting of suspected ADRs over the internet was
[47]
launchedinanattempttofacilitatereporting, and
inApril2003,apilotschemewasintroducedtoallow
patientreportingofsuspectedADRsviaone
oftheNHSs22NHSDirecttelephonecallcentres.
[48]
The impact that these new reporter groups and

initiatives will have on the reporting of suspected
ADRs associated with herbal medicines is not yet
known. The results of the pilot scheme for patient
ADRreportingviaNHSDirectwillberelevantfor
herbal medicines, given that many users of herbal
medicinesselectandselftreatwithsuchpreparations
without the advice or supervision of a healthcare
professional, and that some users may not report
suspectedADRsassociatedwithherbalmedicinesto
[20]
theirdoctororapharmacist. Evaluationsofpilot
schemeshaveindicatedthatthecompletenessofall
reports submitted by community pharmacists and
[42]
GPswassimilar, althoughinthepilotschemefor
ADR reporting by nurses there was some evidence
that completeness of doctors reports was slightly
betterthanthatofnurses.
[49]
ADR Reporting Form
Theminimuminformationrequiredforareportof
asuspectedADRisthesameforbothconventional
andherbalmedicines,andastandardyellowcardis
used to collect data, regardless of the type of
preparationimplicated.Intheyear2000,amodified
yellow card was introduced which included in the
sectionforOtherdrugstheprompt(includingself
[50]
medicationandherbalremedies). Apartfromthis,
the yellow card does not mention herbal medicines
specifically, and its current design has several
deficiencies with regard to prompting for and
collectinginformationonherbalmedicines.
The section Suspected drug presents several

problems.First,thereporterisaskedtoprovidethe
brand(proprietary)nameofthesuspecteddrug(s).
Whilelicensedherbalmedicinalproductsarelikely
to have brand names, unlicensed herbal products
legally are not permitted touse them only the
vernacularand/orbotanicalname,suchasStJohns
wortorHypericumperforatumshouldbeused,al
thoughthisisignoredbysomemanufacturers.For
unlicensedherbalmedicinesitwouldbemoreap
propriatetorequestthenameoftheherbalingredi
ent(s)andthenameofthemanufacturer/supplier.
Identifyingthemanufacturerisparticularlyimpor
tantforreasonsmentionedearlier,namelybecause
the composition of products containing the same

herbal ingredient can vary both qualitatively and
quantitatively between manufacturers. Also, there
may be other problems with the pharmaceutical
quality (e.g. contamination) of unlicensed herbal
productswhichshouldbeconsideredwhenassess
ingADRreports.Ideally,theformshouldalsoin
clude space to indicate whether a sample of the
suspectedproduct(s)isavailable.
Toidentifyspecificallytheherbalingredient(s)
implicated,thebinomialbotanicalname(genusand
species)shouldbegiven.Forexample,echinacea
isinsufficient,sincethreedifferentEchinaceaspe
cies(E.purpurea, E.pallida and E.angustifolia)
areusedmedicinallyandthesedifferintheirphyto
chemical composition. In addition, the specific
plantpartusedshouldalsobestated,sinceoneor
more plant parts may be used medicinally and,
again,thephytochemicalcompositioncanvary.For
example,boththerootandtheherb(aerialparts)of
E. purpurea and nettle (Urtica dioica) are used
medicinally. However,thereisnospecificrequest
forthesedetailsontheyellowcard.
841
existingreportingcardcouldbemadesothatimpor
tantdetailsonherbalmedicinescanberequested.
Signal Detection and Assessment
Atpresent,becauseoftherelativelysmallnumber
ofreportsofsuspectedADRsassociatedwithherbal
medicines held on the MHRAs ADROIT (Adverse
Drug Reaction Online Information Tracking)
database reports, signals are detected simply by
numbers of reports. It may be possible to obtain
proportional reporting ratios for some suspected
ADRsassociatedwithcertainherbalmedicines,such
asStJohnswortforwhicharound150reportsintotal
have been received since 1996. In this case, the
comparisonismadeagainsttherestofthedatabase,
rather than only against the subset of herbal ADR
reports. The assumptions made in proportional
analysis,andtheimportanceofconsideringtheeffect
of selected backgrounds, has been discussed in the
[51]
preparation, the strength of the preparation (e.g.

drug : extract ratio), and the formulation of the
product (e.g. tablets, tincture). Also, many herbal
medicinal products contain several herbal ingredi
ents, some include nonherbal ingredients, such as
vitaminsandminerals,andherbalpractitionersoften
prescribeseveralherbaltincturestogethersuppliedas
amixture.Withrespecttothesepreparations,oneor
more herbal ingredient(s) may be the suspected
agent(s), yet there is limited space on the current
yellowcardtoprovidethislevelofdetail.
contextofconventionalmedicines. Sincethereare
additional biases and other issues in
pharmacovigilance of herbal medicines, what is an
appropriate comparator requires consideration. This
subjectis,however,beyondthescopeofthisarticle.
Following confirmation of a signal relating to a
herbalsafetyconcern,thenextstagesinits evalua
tion are also difficult with respect to herbal medi
cines. In the UK, quantifying the risk is probably
impossibleasthereisnoreliablewayofdetermining
the number of individuals exposed to the herbal
medicine of interest. Benefitrisk analysis is prob
lematicbecauseofthelimitedclinicaldataonsafety
andefficacyofherbalmedicines,andidentifyingat
riskgroupsisalsodifficultbecausetheuserprofilefor
herbal medicines is poorly defined. A particular
problemisthataspecificherbalmedicinecanhave
numeroususesandmaybetakenbyhealthyindivid
ualsforgeneralwellbeing,aswellasbypatients
with chronic disease. These problems are further
compoundedifthevariationindifferentpreparations
ofthesameherbalingredientisconsidered.
Althoughitisnotdesirabletointroducedifferent
reportingformsfordifferenttypesofpreparations,it
could be argued that herbal medicines present a
special case and that a more specialised reporting
formisrequired.Alternatively,modificationstothe
The concerns regarding kavakava (Piper

methysticum)andhepatotoxicityillustratethepro
cessofassessingandrespondingtosafetyissues
relatingtounlicensedherbalmedicinalproducts.A
signalconcerningkavakavaandlivertoxicitywas
Other relevant information not specifically re

questedincludesthemethodofprocessingthecrude
herbal material (e.g. typeofextract), since this can
alsoinfluencetheprecisechemicalcompositionand,
therefore, the potential toxicity of a herbal
[38]
842
Barnes
first raised in 2000 following a cluster of cases

reported in Switzerland, and was strengthened a
yearorsolaterfollowingfurtherspontaneousre
[52]
portsfromSwitzerlandandGermany. TheUK
CSM undertook an initial evaluation, including
causalityassessment,andfoundthattherisksof
kavakava appeared to outweigh its benefits. No
regulatoryactionwastakenatthattime,although
theherbalsectorinstigatedavoluntarywithdrawal
ofproductscontainingkavakavawhilethesafety
concernswereinvestigatedfurther.
Thenextstageinvolvedfurtherdatacollectionand
evaluation.TheCSMsetupaworkinggrouptoassess
theissueandrequestedadditionaldataonbenefitsand
risks of kavakava from the herbal sector and
regulatoryauthorities.WhentheCSMnextconsidered
theissueinJuly2002,atotalof68reportsoriginating
from several countries had been received, although
[52]
onlythreeoriginatedintheUK. Theseverityofthe
liver damage described in the reports varied from
abnormalliverfunctiontestresultstoliverfailureand
death; six patients received liver transplants.
Different preparations of kavakava were available
(e.g.differenttypesofextracts)andconsiderationwas
givenastowhetheronlycertaintypesofkavakava
preparation might be associated with liver toxicity.
However, there appeared to be no relationship
between the method of processing/type of extract,
strengthordose,andtheadversereactions.Thus,on
thebasisofthedataavailable,theCSMadvisedthat
thepossiblebenefitsofpreparationscontainingkava
kavadonotoutweightherisks,thatkavakavahadthe
potential to cause hepatotoxicity which could be
serious in nature, and that kavakava should be
prohibited in unlicensed medicines. On January 13,
2003,astatutoryordercameintoeffectintheUK
prohibitingthesale,supplyandimportofunlicensed
medicinescontainingkavakava.Productlicencesfor
licensedkavakavaproductswererevoked.
[53]
kavakava preparations by patients consulting

medical herbalists, were not known; reports in
volved different types of kavakava preparations;
onlyaverylownumberofreportswasreceivedin
theUK;thequalityandcompletenessofthereports
waspoor,andsomereportswereduplicated;there
arefewclinicaltrialsofkavakavaproductsanda
lack of clear evidence of efficacy; regulatory
optionsin responding tothe signal were limited,
and alternatives, such as including warning
informationwithproducts,wouldhaverequiredthe
voluntary cooperation of manufacturers of
unlicensedkavakavaproductsandMHRAwould
havehadnomeansofenforcement.
Strengths and Weaknesses
Spontaneousreportingschemeshaverecognised
advantages and limitations, and several of these
maybeevenmoreimportantwithregardtoherbal
medicines (see table III). In particular, under
reporting is a wellrecognised, important and
inevitablelimitationofanyspontaneousreporting
scheme,butforseveralreasonsitmaybeaneven
greaterproblemforherbalmedicines.
UnderreportingofsuspectedADRsassociated
withherbalmedicinescouldoccuratseverallevels.
Table III. Summary of advantages and limitations of spontaneous
adverse drug reaction (ADR) reporting schemes with respect to
herbal medicines
Advantages
Monitor all drugs, including all herbal medicines, all the time and
for all consumers and patients
Provide early warnings of undocumented drug safety concerns;
important for herbal medicines as information on safety is limited
Relatively cheap to run; important as the herbal sector may not
have the resources to conduct large-scale postmarketing
surveillance studies
Limitations
Under-reporting; likely to be greater for herbal medicines
Poor quality of data available to or provided by reporter; yellow
card does not cater specifically for recording information on
Someofthedifficultiesinassessingsafetycon
cernswithunlicensedherbalmedicineswereevi
denthere.Forexample,thenumberofunlicensed
herbalproductscontainingkavakavaavailablein
theUK,theirextentofuse,andtheextentofuseof
herbal medicines as suspected drugs

Biases in reporting
Cannot estimate frequency of an ADR as do not provide accurate
information on number of individuals exposed to the drug of
interest; probably not possible to obtain denominators for
unlicensed herbal medicinal products
Suspected ADRs may be identified/reported outside the formal
system (e.g. to herbalists, health-food stores)
843
First,becauseoftheperceptionthatherbalmedicines
are safe, users of these preparations may not
associateanadverseeventwiththeiruseofaherbal
medicine, particularly if they are taking other (con
ventional)medicines.Iftheuserdoesmakeanasso
ciation between use of a herbal medicine and an
adverse event, they may take steps to resolve the
problem themselves (for example, stop taking the
preparation) and/or may not inform a healthcare
[20]
professional. Underreportingcanalsooccuratthe
level of the healthcare professional, since doctors,
pharmacistsandotherrecognisedreporterscouldfilter
outreportsofsuspectedADRsdescribedbypatients.
[54]
Reasons for underreporting among healthcare

professionalsarewelldocumented,althoughstudies
exploring this area have been carried out in the
context of conventional medicines, and it is not
knownifthesesamereasonsapplytounderreporting
forherbalmedicines.
Severalstudiesinvolvingcommunitypharmacists
indicatethatmanypharmacistsareunawarethatthey
shouldreportsuspectedADRsassociatedwithherbal
medicines. A crosssectional survey carried out in
1998ofover1300communitypharmacists(response
rate: 67%) not involved in the CSM/ MHRA pilot
scheme for community pharmacist ADR reporting
foundthat47%ofrespondentswerenotawarethatthe
yellowcardschemeappliedtoherbalmedicinesatall,
37% were aware it applied to licensed herbal
medicines, and only 16% knew it applied to both
[25]
licensed and unlicensed herbal medicines. This

finding is not so surprising since these pharmacists
werenotrecognisedreportersatthetimeofthestudy
and would not have received training materials on
ADR reporting. Of more concern is that studies
conducted since all community pharmacists became
recognisedreportersandwereencouragedtofocuson
reportingsuspectedADRsassociatedwithherbaland
other nonprescription medicines have continued to
findthatmanycommunitypharmacistsareunaware
oftheneedtoreportsuspectedADRsassociatedwith
cinesinformationpharmacistsworkinginaMedi
cinesInformationCentreinWalesfoundthat al
though they encouraged only 41% of enquirers
aboutADRstocompleteyellowcards,theywere
morelikelytogiveencouragementwhereanalter
native medicine was involved rather than a
[57]
conventional medicine. In addition, all these
studies revealed deficiencies in community
pharmacistsknowledgeonotheraspectsofADR
reporting, such as the level of certainty required
regardingacausalrelationship.
Todate,thereareveryfewstudiesthatprovideany
information on the extent of underreporting of
suspectedADRsassociatedwithherbalmedicines.In
onecrosssectionalsurveyofcommunitypharmacists
who were not involved in the CSM/MHRA pilot
scheme for community pharmacist ADR reporting
(seeearlierinthissection),respondentswereaskedto
describe any reports of suspected ADRs associated
withcomplementarymedicinesthattheyhadreceived
[24,25]
or identified over the previous 12 months.

In
total,among818respondents,44reportsofsuspected
ADRs associated with herbal medicines were
described, an average of one report per 19
pharmacists. By contrast, the CSM/MHRA pilot
scheme, which ran over approximately the same
period covered by the survey, and involved around
[42,58]
3200 pharmacies, received only four reports.

Conclusions cannot be drawn from these crude
comparisons, since these studies used different
methodologies, involved pharmacists/pharmacies in
different regions of the UK, and so on. They do,
however,raisethehypothesisthatthereissignificant
underreporting by pharmacists of suspected ADRs
associatedwithherbalmedicines.
particularlyunlicensedherbal
It is recognised that pharmacists can make an

important contribution to ADR reporting for herbal
medicinalproducts,butitislikelythatgreatervigi
lanceonthepartofthepharmacistandinitiativesto
encourageherbalADRreportingbypharmacistsare
required. Against this background, there have been
severalrecentpapersinajournalreceivedbyallUK
medicines. There may also be biases favouring

ADR reporting for herbal medicines. An audit of
medi
pharmacists,
andafactsheetonADRreport
ingbypharmacistshasbeenproducedbytheScience
CommitteeoftheRoyalPharmaceutical
herbalmedicines,
[55]
[55,56]
[18,44,59]
844
Barnes
SocietyofGreatBritain(theprofessionalandregu
latorybodyforallpharmacistsintheUK)which
providesguidanceandremindspharmacistsoftheir
professionalandethicalresponsibilitiesinthisre
gard.
[60]
A further limitation of spontaneous reporting

schemesisthatherbalmedicinesarewidelyavailable
from a range of outlets without the need for
interactionwithahealthcareprofessionaland,there
fore, suspected ADRs associated with herbal medi
cinesmaybeidentifiedbyorreportedtoanindivid
ual(e.g.herbalist)whoisoutsidetheformalsystem
for ADR reporting. Healthfood stores are a major
outlet for herbal medicinal products, but it is not
known if staff in these outlets receive reports of
suspectedADRsassociatedwithsuchproducts,andif
theydo,whataction,ifany,theytake.
2.1.2 WHO/Uppsala Monitoring Centre Traditional
Medicines Project
ADR reports, including herbal ADR reports,

fromtheCSM/MHRAyellowcardschemeand(in
2003)thosefrom70othercountrieswithnational
ADRmonitoringschemesarefedintotheWHO/
Uppsala Monitoring Centre (UMC). The UMC
recognisestheproblemsinherentinADRreporting
forherbalmedicinesandhasestablishedaTradi
tional Medicinesproject tostimulatereportingin
thisareaandtostandardiseinformationonherbal
medicines, particularly with regard to nomencla
[61]
ture. For example, a special set of herbal

anatomicaltherapeuticchemical(ATC)codeshas
beendevelopedwhichisfullycompatiblewiththe
regularATCclassificationsystemforconventional
medicines.
[62]
The UMC database, established in 1968, held

over2millionreportsofsuspectedADRsby1999,
[62]
ofwhicharound0.5%involveherbalmedicines.
Fortheperiod19681997,almost9000reportsin
volving herbal medicines were received by the
UMC.TheUKisamongthetopfivecountriesin
terms of absolute numbers of herbal suspected
ADRreportssubmitted.
2.1.3 Herbal-Sector-Initiated Spontaneous

Reporting Schemes
At present, herbal medicine practitioners are not

recognised as reporters by the CSM/MHRA yellow
cardscheme.Severalherbalmedicinepractitionerand
other herbalsector organisations have initiated their
own ADR reporting schemes for herbal medicines
based on the CSM/MHRA scheme. While this is a
responsibleandpotentiallyusefulstepforwardwhere
these schemes have developed a link with the
CSM/MHRAorWHO/UMC,adhocschemesarenot
encouragedbecausethereisariskthatreportswillbe
dispersedandsignalsmaybenotbedetectedasearly
as possible, or may be missed. As with any
spontaneousreportingscheme,herbalsectorinitiated
schemesarealsolikelytobepronetolimitationssuch
asunderreporting.Itisnotknownwhetherreasons
forunderreportingofsuspectedherbalADRsbythe
herbal sector are different to those for herbal ADR
reportingbyconventionalhealthcareprofessionals.It
is possible that there may be concerns among the
herbalsectorthattheavailabilityofherbalmedicines
andtheirfreedomtopractiseherbalmedicinemaybe
threatened if significant numbers of herbal ADR
reportsaresubmitted.
The National Institute of Medical Herbalists
(NIMH),themajororganisationformedicalherbal
istsintheUK,requestsfromitsmembersreportsof
suspected ADRs associated with herbal treatments.
Reports are submitted on a modified yellow card
form,whichhassomeadditionaldatafieldsrelevant
to herbalists prescriptions. The NIMH sends an
annualsummaryofreportsreceivedtoMHRA.For
theperiodJanuary1994November2001,23reports
[63]
were received by the NIMH. Most reports

describedreactionsexperiencedbypatientswhohad
received a combination of several herbs, which is
typicalofmedicalherbaliststreatmentapproach.A
similar scheme has been set up by the Register of
ChineseHerbalMedicine(RCHM)whichalsousesa
modified yellow card form to collect data from its
practitionersofChineseherbalmedicine.TheRCHM
schemealsohasalinkwithMHRA.Atthetimeof
writing,ninereports(threein2000)hadbeenreceived
bytheRCHMfromitsmembers.
[64]
Otherschemeshavebeenestablishedwhicharenot
restrictedtoherbalmedicinepractitioners.Phytonetis
a passwordprotected, internetbased system for
gatheringreportsofsuspectedADRsassociatedwith
herbalmedicinesthatwassetupbyaUKuniversity
onbehalfoftheEuropeanScientificCoOperativeon
[65]
Phytotherapyin1996. Phytonetusesanelectronic
form based on the CSM/MHRA yellow card, but
differs from the schemes described above in that it
acceptsreportsfromhealthcareprofessionals,herbal
practitioners, patients and the public. Submitted
reports are assessed by an expert panel and, where
appropriate, fed into the WHO/ UMC. Few reports
havebeenreceived,however,andsupportisneededto
revive the system. As there is no obligation for
manufacturerstoreportsuspectedADRsassociated
with their unlicensed herbal products, the British
Herbal Medicine Association (BHMA), whose
members include many herbalmedicines
manufacturers,hasaddressedthisinitsvoluntarycode
[66]
ofpracticeforitsmembers. Thecodeincludesthe
requirement that manufacturers send reports of
suspected ADRs associated with their unlicensed
herbal products to the BHMA, which may, at its
discretion, forward such reports to MHRA. At the
timeofwriting,theBHMAhadnotreceivedfromits
membersanyreportsof suspectedADRsassociated
with unlicensed herbal medicinal products. The
University of Westminster, London, UK, in
conjunctionwithasupermarketretailer,isdeveloping
anADRreportingschemeforconsumers,nutritional
advisers and other health practitioners to report
suspected ADRs associated with herbal and
[67]
complementarymedicines. Thereisatleastone
otherschemeintheUK,setupbyaherbalsupplierin
Leicester, UK, which is inviting purchasers of its
productstosubmitreportsofsuspectedADRs.Itis
notclearwhataretheaimsofthisscheme,giventhat
itisstatedthattheinformationwillbeusedonlyby
theherbalsupplier.Thisisof
845
concernif,forexample,aherbalistsubmitsareport
ofasuspectedADRassociatedwithaherbalmedi
cinetothisschemealone.
2.1.4 Intensive Monitoring Schemes
Extensions to the CSM/MHRA yellow card

schemewerelaunchedin1997and1998tostimulate
reporting of suspected ADRs associated with
medicinesusedinthetreatmentofHIVinfection,and
[45]
ofADRsoccurringinchildren. TheHIVreporting
scheme, launched in 1997, targets specialist
healthcareprofessionalsworkingwithHIVinfected
individuals and encourages reports of suspected
ADRsonamodifiedyellowcardformwhichdoesnot
request the patients name. The paediatric reporting
schemealsotargetedhealthcareprofessionalsbutdid
not involve a modified yellow card. Numbers of
reportsincreasedfollowingtheintroductionofthese
schemes, although the underlying reporting rate
increased only with the HIV scheme; ongoing
promotion of the scheme, for example, through a
[45]
regular newsletter is stated to be important. To

date,CSM/MHRAhavenotintroducedanintensive
monitoringschemeforherbalmedicinesand,given
thattheimpactofinitiativestostimulatereportingof
herbalADRshassofarbeenlimited,itmaybean
appropriatetimetosetupsuchaschemetogetherwith
a programme of ongoing promotion aimed at
maintainingitseffectiveness.
Theherbalsectorhasnotyet setupanysuch
schemesforpharmacovigilancepurposes,although
apilotstudyrunbytheMedicinalPlantResearch
Groupcollecteddetaileddata,includingoutcomes,
fromaround30herbalistsonalltheirpatientstreat
edforirritablebowelsyndrome.
[68]
2.2 Prescription Event Monitoring
The methodology of prescription event monitor

ing (PEM) in monitoring the safety of newly mar
ketedprescriptiondrugsiswellestablished.
[69]2
The
2 PEMisahypothesisgenerating,noninterventional,observationalformofmonitoringfornewlymarketedmedi
cinescarriedoutbytheDrugSafetyResearchUnit,Southampton,UK.CurrentPEMmethodologyinvolvessendinga
greenformtoGPswhohaveprescribedthemedicinebeingstudied;thesedataareobtainedfromtheUKPrescription
PricingAuthority.Thegreenformcomprisesasimplequestionnaire,whichrequestsdataonallhealtheventsthe
patientwhowasprescribedthedrugexperiencedduringtreatment.TheseformsareusuallysenttotheGPsaround6
monthsafterthepatientwasfirstprescribedthemedicineunderstudy.
846
Barnes
valuable contribution that PEM has made to

pharmacovigilance of conventional medicines is
clear, but the existing method is of no use at
presentforpharmacovigilanceofherbalmedicines
becausetheyarerarelyprescribed.
A protocol for modified PEM methodology has
been developed by the Drug Safety Research Unit,
Southampton,UK,incollaborationwiththeNIMH,
the University of Southampton, the Medical Toxi
cologyUnitatGuysandStThomasHospitalTrust,
London,andtheSchoolofPharmacy,Universityof
London.Thisapproachinvolvesusingherbaliststo
provideadverseeventdataongreenformsforpatients
treated with a specific herbal medicine. Where
patients give permission, a green form requesting
adverse event data would also be sent to their GP.
Therearelimitationstothismethod,suchaswhether
sufficient patient numbers could be achieved and,
particularly, that the herb of interest is not newly
marketedsotheremaybepreconceptionsaboutits
safetyprofile.Nevertheless,theprotocolrepresentsa
step forward in attempting to develop methods for
pharmacovigilance of herbal medicines. Funding is
beingsoughttocarryoutapilotstudyofthemodified
PEMmethodology.
Anotherpotentialapproach,basedonPEMcon
cepts,istousecommunitypharmaciststorecruita
cohortofpurchasers(whereconsentisgiven)ofa
specificherbalmedicinalproductwhowouldthen
befollowedupovertimeandadverseeventdata
collected.Thefeasibilityofthisapproachhasbeen
demonstratedinapilotstudyusingaconventional
[70,71]
nonprescription medicine,
but needs to be
evaluated as a method for pharmacovigilance of
herbalmedicinalproducts.
2.3 Other Pharmacoepidemiological
Study Designs
The methodology for casecontrol and cohort

studiesiswellestablishedandthesestudydesigns
can be used to investigate safety concerns with
herbalmedicines,althoughfewstudieshavebeen
carriedouttodate.Onestudyexploredtherelation
shipbetweencolorectalcanceranduseofprepara
tionscontaininganthranoidlaxatives.
[38]
The
strengths and limitations of casecontrol and cohort

[72]
studies are well documented, but as with other

studydesigns,someoftheproblemsarecompounded
when these study designs are applied to herbal
medicines.Forexample,toestablishandverifyboth
cases and controls exposure to the herbal
medicine(s) of interest is particularly problematic
since herbal medicines are rarely prescribed; even
whereherbalmedicinalproductsarepurchasedfrom
pharmacies,pharmacistsdonotroutinelyrecorduse
of herbal and other nonprescription medicines on
[24,25]
computerised patient medication records.

In
addition, for reasons explained earlier, there are
likely to be variations in different manufacturers
productsand,therefore,definingexposureprecisely
willbedifficultatbest.
Casecontrolandcohortstudiesinvolvingcon
ventionalprescribedmedicinescanbecarriedout
using UK computerised healthrecord databases
such as the General Practice Research Database
andtheMedicinesMonitoringUnit database,but
such tools currently are of no use for studies
involvingherbalmedicinessinceherbalmedicines
are rarely prescribed and information on non
prescriptionmedicines,includingherbalmedicines,
isnotrecordedonGPspatientrecords.
As with casecontrol and cohort studies, experi
mental studies can be applied to investigating the
safety of herbal medicines. At present, notwith
standingrecognisedlimitations,suchassamplesize
and ethical considerations, welldesigned and well
conducted randomised clinical trials (RCTs) over
come some ofthe difficulties thatherbal medicines
present for other pharmacoepidemiological studies.
Forexample,preciselyestablishingexposureissim
plersincecompliancecheckscanbecarriedout,and
RCTsareunlikelytouseherbalmedicinalproducts
(containingthe sameherbal ingredient)fromdiffer
entmanufacturers,soproductvariationand,usually,
batchtobatch variation in products, is eliminated.
Thereisalwaysthepossibilityofcoursethatclinical
trial participants could take purchased herbal
medicinesinadditiontothestudymedication.
Systematicreviewsandmetaanalysesofadverse
eventdatafromRCTsofspecificherbalmedicines
have been carried out, but this introduces other

problems. Many existing RCTs of herbal medicinal
productsareofpoororlimitedmethodologicalqual
ity,and/orpublishedreportsofstudiesdonotfollow
Consolidated Standards of Reporting Trials (CON
SORT) guidelines. In addition, clinical trials of a
particular herbal ingredient usually will have been
carried out using several different manufacturers
products, but systematic reviews and metaanalyses
oftenignorevariationsbetweenproducts.
3. Communication of Herbal
Safety Concerns
The importance of the timing, content and
method of delivery of messages regarding safety
concerns hasbeen discussedextensively,andthe
requirementsforsuccessfulcommunicationofsafe
tyconcernsshouldapplyequallytoherbalmedi
cines. However, communicating information on
herbalsafetyconcernspresentsadditionaldifficul
tiesforseveralreasons.DearDoctor/Pharmacist
letterscanbesent,buthealthcareprofessionalsare
unlikelytoknowwhichoftheirpatientsareusing
herbalmedicinesand,therefore,willbeunableto
pass on safety messages to specific individuals.
Medicalherbalistsmaykeepsomerecordsoftheir
patients treatment, but as there is no statutory
regulation for herbalmedicine practitioners, lists
ofallindividualspractisingherbalmedicinearenot
available.
Moreover,mostusersofherbalmedicinesobtain
these medicines from outlets where there is no
healthcare professional present and without seeking
professional advice. Methods aimed at reaching the
public directly (e.g. the internet) and the popular
media are often the only ways of communicating
herbalsafetyinformationtosuchindividuals.Thereis
a lack of research on how herbalmedicine users
interpretinformationonrisksassociatedwithherbal
medicines. It should not be assumed that users
understanding of risk associated with herbal medi
cinesisthesameasthatforprescriptionmedicinesor
conventionalnonprescriptionmedicines.Ithasbeen
shownthatindividualsmayoverestimatethe
847
risksofadverseeffectsassociatedwithprescription
medicinesandconventionalnonprescriptionmedi
[73,74]
cines,
but given that herbal medicines are
widely perceived to be safe, the hypothesis that
usersofherbalmedicinesmayunderestimaterisks
needstobetested.Furthermore,oncetheproposed
directive on traditional herbal medicinal products
comes into force, manufacturers of products
registeredunderthenewnationalschemewillbe
required to provide systematic information with
their products, including information on adverse
eventsandspecialwarnings.Theimpactofthison
users perceptions of the risks associated with
herbalmedicineswillalsorequireevaluation.
The action taken by MHRA to communicate
information on interactions between St Johns wort
and certain prescription medicines after this issue
emergedintheyear2000providesanexampleofthe
process of communicating information on herbal
safety concerns. Following its decision that manu
facturers should include warning information on
product packaging, MHRA used various ways of
communicating the message. Dear Doctor/Pharma
cistlettersweresent, andpharmacistsinparticular
were asked to provide advice to consumers and
patientsoninteractionsbetweenStJohnswortand
conventionalmedicines.Atelephonehelplinewasset
up, and information for patients was posted on the
MHRAwebsite.However,itisdifficulttoassessthe
effectivenessofthesemeasures.SinceFebruary2000
when the information was made public, the
CSM/MHRA yellow card scheme has continued to
receive reports ofsuspectedinteractionsbetweenSt
Johns wort and conventional medicines (>30 from
February 2000April2003), for example,reports of
breakthrough bleeding and unintended pregnancy in
womentakingStJohns wortproducts concurrently
withoralcontraceptives.
[43]
It is likely that there is scope for improving

communication with the public on herbal safety
issues. Recognising this, an area on the MHRA
websitehasbeensetupwhichisdedicatedtopro
viding early information on herbal safety con
cerns.
[13]
848
Barnes
4. The Future for Pharmacovigilance of

Herbal Medicines
The potential for herbal medicines to have a
significantnegativeimpactonthepublichealthneeds
tobekeptinperspective.Nevertheless,aparallelcan
be drawn between the lack of a formal medicines
regulatorysystembeforethethalidomidedisasterand
the current situation in the UK as regards herbal
medicinal products in that the sector largely is
unregulated. Most herbal medicinal products,
includingherbsfromChina,SouthAmericaandmany
othercountries,whichare newtothe UK,aresold
without any requirement to demonstrate to the
licensing authority evidence of quality, safety and
efficacy. Postthalidomide, new initiatives in drug
safety monitoring initially followed further high
[75]
profile drug safety problems. Likewise, several

recent highprofile herbal safety concerns, such as
renal failure and urothelial cancer associated with
exposuretoAristolochiaspecies,
withStJohnswort,
[76]
druginteractions
[77]
andhepatotoxicityassociated
[78]
withkavakava, havecontributedtotheincreasing
awarenessoftheneedtomonitorthesafetyofherbal
medicines.Againstabackgroundofincreasinguseof
herbal medicines, particularly by patients using
conventional drugs concurrently and those with
serious chronic illness, it is likely that new safety
concernswillcontinuetoemerge.
However, improvements in the safety and

pharmacovigilanceofherbalmedicinescanbeex
pected,iftheproposedEUdirectivefortraditional
herbal medicinal products is implemented as
[6]
planned. Manufacturersoftraditionalherbalme
dicinalproductsregisteredundernationalschemes
establishedunderthedirectivewillberequiredto
adhere to quality standards, to provide
bibliographic evidence of the safety of their
products,andtocomplywithregulatoryprovisions
onpharmacovigilance.Theseimprovementsmay
nothappenimmediatelyacrossallmanufacturers,
since some may take advantage of the transition
period(whichwillbeatleast5years,probably7
years, from the date the legislation comes into
force)beforesubmittingtheirdossierstoMHRA
inordertoapplyforproductregistrations.
Anothereffectofthedirectivemaybetoshiftthe
emphasisofresearchinvolvingherbalmedicines.At
present,mostresearchintheherbalmedicinesareais
aimed at discoveringthepharmacologicalactivities
ofmedicinalplantsandprovidingevidenceofclinical
efficacy;ratherlesseffortisfocussedoninvestigating
safety.However,sincetheproposedtraditionalherbal
medicinal products directive does not require
manufacturerstodemonstrateefficacy(otherthanby
way of traditional use), there may be more interest
among manufacturers and researchers in extending
knowledgeofthesafetyofherbalmedicines.While
researchintothesafetyofherbalmedicinesistobe
welcomed, research into efficacy is also needed in
order to develop herbal medicinal products with
favourablebenefitriskprofiles.
Statutory regulation of herbalmedicine practi

tioners,asrecommendedbyTheHouseofLords
SelectCommitteeonScienceandTechnologysre
[79]
port on complementary/alternative medicine is
alsoexpectedtobeimplementedoverthenextfew
years. Once this has been achieved, it seems
reasonabletoexpectthattheyellowcardscheme
would be extended to include stateregistered
herbalmedicine practitioners as recognised
reporters who would be encouraged to report
suspectedADRsassociatedwithherbalmedicines.
Inthelongerterm,modified,evennoveltoolsfor
monitoring the safety of herbal medicines may be
developed. Pharmacyrecord linkage is used in The
Netherlandsforpharmacovigilancepurposes,butno
suchtoolexistscurrentlyintheUK.AnewDepart
[80]
ment of Health report, however, discusses the

possibility of community pharmacists being able to
accessacommonelectronichealthrecordwhichwill
be created for all patients and, presumably, to add
communitypharmacydatatoit.Whilesuchasystem
probablywouldapplyonlytoprescriptionmedicines
initially,withtechnologicaladvancesitmightalsobe
developed into a computerised recordlinkage
databasethatcouldbeusedtomonitorthesafetyof
herbal and other nonprescription medicines.
Consideration should also be given as to whether
consumersandpatientscouldhaveagreaterrolein
pharmacovigilanceofherbalmedicines,
849
possiblybytheirinclusionasrecognisedreporters
inspontaneousreportingschemesandbycollecting
data directly from patients in studies based on
modifiedPEMmethodology.
8.
Thefutureforensuringthesafetyofherbalmedi
cinesmaylie,atleastinpart,withpharmacogenetics
and pharmacogenomics. The importance of genetic
factorsindetermininganindividualssusceptibilityto
10.
[81]
ADRs is well documented, and this applies to

herbal medicines as well as to conventional drugs.
However, optimising treatment, including reducing
the potential for ADRs, on the basis of a patients
genotypehasbarelybeendiscussedinthecontextof
herbalmedicines.
Acknowledgements
TheauthorthanksMrsLeighHenderson,Medicinesand
Healthcare products Regulatory Agency (MHRA) for
providing data from the ADROIT system, and Dr Linda
Anderson,MHRA,andtherefereesfortheircomments.
Nosourcesoffundingwereusedtoassistintheprepara
tion of this manuscript. The author has no conflicts of
interestdirectlyrelevanttothecontentsofthisreview.The
viewsexpressedarethoseoftheauthoraloneanddo not
necessarily represent the views of the MHRA or the
individualsmentionedabove.
References
1.
2.
3.
4.
5.
6.
7.
Mann RD, Andrews EB, editors. Pharmacovigilance.

Chicester:Wiley,2002
Waller PC, Evans SJW. A model for the future conduct of
pharmacovigilance.PharmacoepidemiolDrugSaf2003;12:
1729
WorldHealthOrganization.DraftWHOguidelinesonsafety
monitoring and pharmacovigilance of herbal medicines.
Geneva:WorldHealthOrganization,2003
European Agency for the Evaluation of Medicinal Products
(EMEA) Committee for Proprietary Medicinal Products
(CPMP).Noteforguidance onqualityofherbal medicinal
products. CPMP/QWP/2819/00 (EMEA/CVMP/814/00).
London:EMEA,July26,2001[online].AvailablefromURL:
http://www.emea.eu.int/pdfs/human/qwp/281900en.pdf [Ac
cessed2003Aug13]
TheMedicinesAct.London:TheStationeryOffice,1968
Commission of the European Communities. Proposal for
amending the directive 2001/83/EC as regards traditional
herbal medicinal products. 2002/0008 (COD). Brussels:
EuropeanCommission,2002
CommissionoftheEuropeanCommunities.Amendedproposal
foraDirectiveoftheEuropeanParliamentandoftheCouncil
amending the directive 2001/83/EC as regards traditional
herbal medicinal products. 2002/0008 (COD). Brussels:
EuropeanCommission,2003
9.
11.
12.
13.
14.
15.
Barnes J, Anderson LA, Phillipson JD. Herbal medicines: a

guide for healthcare professionals. London: Pharmaceutical
Press,2002
De Los Reyes GC, Koda RT. Determining hyperforin and
hypericin contentin eightbrands ofSt Johns wort. Am J
HealthSystPharm2002;59:5457
KressmanS,MullerWE,BlumeHH.Pharmaceuticalqualityof
differentGinkgobilobabrands.JPharmPharmacol2002;54:
6619
LoewD,KaszkinM.Approachingtheproblemofbioequiva
lenceofherbalmedicinalproducts.PhytotherRes2002;16:
70511
WoodfieldR.Seneciospeciesinunlicensedherbalremedies
[letter].London:MedicinesControlAgency,2002
MedicinesandhealthcareproductsRegulatoryAgency.Herbal
safety news [online]. Available from URL: http://
medicines.mhra.gov.uk[Accessed2003Jul29]
ThomasKJ,NichollJP,ColemanP.Useandexpenditureon
complementarymedicineinEngland:apopulationbasedsur
vey.ComplementTherMed2001;9:211
Mintel International Group Ltd. Complementary medicines,
UK. April 2003. London: Mintel International Group
Limited,2003
16. EisenbergDM,DavisRB,EttnerSL,etal.Trendsinalternative
medicineuseintheUnitedStates,19901997:resultsofanational
followupsurvey.JAMA1998;280(18):156975
17. MacLennanAH,WilsonDH,TaylorAW.Theescalatingcost
andprevalenceofalternativemedicine.PrevMed2002;35:
16673
18. Barnes J. Herbal therapeutics (1): an introduction to herbal
medicinalproducts.PharmJ2002;268:8046
19. BarnesJ,ErnstE.Traditionalherbalistsprescriptionsforcom
mon clinical conditions: a survey of members of the UK
NationalInstituteofMedicalHerbalists.PhytotherRes1999;
12:36971
20. BarnesJ,MillsSY,AbbotNC,etal.Differentstandardsfor
reporting ADRs to herbal remedies and conventional OTC
medicines:facetofaceinterviewswith515usersofherbal
remedies.BrJClinPharmacol1998;45:496500
21. GulianC,BarnesJ,FrancisSA.Typesandpreferredsources
ofinformationconcerningherbalmedicinalproducts:faceto
faceinterviewswithusersofherbalmedicinalproducts[ab
stract].IntJPharmPrac2002;10Suppl.:R33
22. VickersAJ,ReesRW,RobinA.Advicegivenbyhealthfood
stores:isitclinicallysafe?JRCollPhysLond1998;32(5):
4268
23. Quinn CG, Waterman P. A comparison of the teaching of
herbal medicine in the United Kingdom and Europe.
UniversityofStrathclyde,1997
24. BarnesJ,AbbotNC.Experienceswithcomplementarymedi
cines:asurveyofcommunitypharmacists[abstract].PharmJ
1999;263:R3743
25. BarnesJ.Anexaminationoftheroleofthepharmacistinthe
safe,effectiveandappropriateuseofcomplementarymedi
cines[PhDthesis].UniversityofLondon,2001
26. DepartmentofHealth.HerbalMedicineRegulatoryWorking
Group[online].AvailablefromURL:http://www.doh.gov.uk/
herbalmedicinerwg[Accessed2003Jul29]
27. MedicinesControlAgency.Traditionalethnicmedicines:pub
lic health and compliance with medicines law. November
2001.London:MedicinesControlAgency,2001
28. The Medicines (Aristolochia and Mu Tong) (Prohibition) Order
2001(SI2001/1841).London:TheStationeryOffice,2001
850
29. CommissionoftheEuropeanCommunities.Directive2001/83/
EC.Brussels:EuropeanCommission,2001
30. AdcockH.MedicinesControlAgencyproposesbanforkava
containingproducts.PharmJ2002;269:128
31. RoyalPharmaceuticalSocietyofGreatBritain.Medicines,eth
ics and practice: a guide for pharmacists. London: Royal
PharmaceuticalSocietyofGreatBritain,2003Jul:92
32. De Smet PAGM, Hansel R, Keller K, et al., editors.
Toxicological outlook on quality assurance of herbal
remedies.In:Adverseeffectsofherbaldrugs.Vol1.Berlin:
SpringerVerlag,1992
33. BusseW.Thesignificanceofqualityforefficacyandsafetyof
herbalmedicinalproducts.DrugInfJ2000;34:1523
34. ErnstE,DeSmetPA,ShawD,etal.Traditionalremediesand
thetestoftime.EurJClinPharmacol1998;54(2):99100
35. De Smet PAGM. Health risks of herbal remedies. Drug Saf
1995;13(2):8193
36. Woodfield R. Proposed directive on traditional herbal medicinal
products: progress in European negotiations [letter]. London:
MedicinesandHealthcareproductsRegulatoryAgency,2003
37. Pressrelease(revised).MeetingoftheWorkingPartyonHerbal
MedicinalProducts.February2425,2003,EuropeanAgency
fortheEvaluationofMedicinalProducts,London.Document
reference:EMEA/CPMP/HMPWP/1090/03Rev.1
38. DeSmetPAGM.Anintroductiontoherbalpharmacovigilance.
In: De Smet PAGM, Keller K, Hansel R, et al., editors.
Adverse effects of herbal drugs. Vol. 3. Berlin: Springer
Verlag,1997
39. Extensionoftheyellowcardschemetounlicensedherbalreme
dies.CurrProbPharmacovigilance1996;22:10
40. ShawD,LeonC,KolevS,etal.Traditionalremediesandfood
supplements:a5yeartoxicologicalstudy(19911995).Drug
Saf1997;17(5):34256
41. Extensionoftheyellowcardschemetopharmacists.CurrProb
Pharmacovigilance1997;23:3
42. DavisS,CoulsonR.Communitypharmacistreportingofsus
pectedADRs:(1)thefirstyearoftheyellowcarddemonstra
tionscheme.PharmJ1999;263:7868
43. Medicines and Healthcare Products Regulatory Agency. Ad
verse drug reaction online information tracking (ADROIT)
system,Jul25,2003
44. MajorE.Theyellowcardschemeandtheroleofpharmacistsas
reporters.PharmJ2002;269:256
45. DavisS,RaineJM.Spontaneousreporting:UK.In:MannRD,
Andrews E, editors. Pharmacovigilance. Chicester: Wiley,
2002:195207
46. CommitteeonSafetyofMedicinesandMedicinesandHealth
careProductsRegulatoryAgency.Theyellowcardscheme:
extensionoftheyellowcardschemetonursereporters[on
line]. Available from URL: http://medicines.mhra.gov.uk/
aboutagency/regframework/csm/csmhome.htm [Accessed
2003Jul25]
47. MCAlauncheswebversionofyellowcardscheme.PharmJ
2002;269:631
48. PatientsabletoreportADRsviaNHSDirect.PharmJ2003;
270:608
49. MorrisonGriffithsS,WalleyTJ,ParkBK,etal.Reportingof
adversedrugreactionsbynurses.Lancet2003;361:13478
50. Updatedyellowcardlaunched.PharmJ2000;265:387
51. GogolakVV.Theeffectofbackgroundsinsafetyanalysis:the
impact of comparison cases on what you see.
PharmacoepidemiolDrugSaf2003;12:24952
Barnes
52. MedicinesControlAgency.ConsultationMLX286:proposals
to prohibit the herbal ingredient kavakava (Piper methys
ticum)inunlicensedmedicines.London:MedicinesControl
Agency,2002Jul19
53. The MedicinesforHumanUse (KavaKava)(Prohibition)Order
2002(SI2002/3170).London:TheStationeryOffice,2003
54. VanGrootheestK,deGraafL,deJongvandenGergLTW.
Consumer adverse drug reaction reporting: a new step in
pharmacovigilance?DrugSaf2003;26(4):2117
55. WingfieldJ,WalmsleyJ,NormanC.WhatdoBootspharma
cistsknowaboutyellowcardreportingofadversedrugreac
tions?PharmJ2002;269:10910
56. GreenCF,MottramDR,RavalD,etal.Communitypharma
cistsattitudestoadversedrugreactionreporting.IntJPharm
Prac1999;7:929
57. BiscoeR,HoughtonJE,WoodsFJ.Anauditofthelevelof
encouragementgivenbymedicinesinformationpharmaciststo
enquirersofsuspectedadversedrugreactionstocompletea
yellow card report: perspectives in patient safety [abstract].
28th UK Medicines Information Conference Proceedings;
2002Sep1921;Chester,UK
58. PharmacistsadversedrugreactionreportingtostartonApril1.
PharmJ1997;258:3301
59. CoxA.EmbracingADRreportingcouldimprovepharmacists
standing[letter].PharmJ2002;269:14
60. MoffatT.Adversedrugreaction(ADR)reportingbypharma
cists.London:RoyalPharmaceuticalSocietyofGreatBritain,
2003.Inpress
61. FarahMH,EdwardsR,LindquistM,etal.Internationalmoni
toringofadversehealtheffectsassociatedwithherbalmedi
cines.PharmacoepidemiolDrugSaf2000;9:10512
62. World Health Organisation Collaborating Centre for Interna
tionalDrugMonitoring.DraftguidelinesforherbalATCclas
sification.Uppsala:TheUppsalaMonitoringCentre,2002
63. BroughtonA.Yellowcardreportingscheme.EurJHerbMed
2001;Dec:36
64. WardT.RegisterforChineseMedicalHerbalists.2002Sep18
65. European Scientific Cooperative on Phytotherapy (ESCOP).
PhytoNET [online]. Available from URL: http://www.es
cop.com[Accessed2003Jul29]
66. British Herbal Medicine Association.Code of goodpractice:
unlicensed herbal remedies. Bournemouth: British Herbal
MedicineAssociation,1997Mar
67. Nutraceuticals adverse events database. Executive summary
[online].AvailablefromURL:http://www.wmin.ac.uk/sih/re
search/proj_nutraceuticals.html[Accessed2003Jul25]
68. LogueM,PendryB,WatersE,etal.Irritablebowelsyndrome:
herbalresearchupdate[abstract].PhytotherapyResearchCon
ference;2002Apr,Glasgow
69. Shakir SAW. PEM in the UK. In: Mann RD, Andrews EB,
editors.Pharmacovigilance.Chicester:Wiley,2002:33344
70. LaytonD,SinclairHK,BondCM,etal.Pharmacovigilanceof
overthecounter products based in community pharmacy:
methodologicalissues frompilotworkconductedinHamp
shireandGrampian,UK.PharmacoepidemiolDrugSaf2002;
11:50313
71. SinclairHK,BondCM,HannafordPC.Pharmacovigilanceofover
thecounter products based in community pharmacy: a feasible
option?PharmacoepidemiolDrugSaf1999;8:47991
72. Strom BL. How should one perform pharmacoepidemiology

studies?Choosingamongtheavailablealternatives.In:Strom
BL,editor.Pharmacoepidemiology.3rded.Chicester:Wiley,
2000:40113
73. Berry DC, Knapp PR, Raynor DK. Is 15% very common:
informingpeopleabouttherisksofmedicationsideeffects.Int
JPharmPrac2002;10:14551
74. BerryDC,RaynorDK,KnappP,etal.Patientsunderstanding
ofriskassociatedwithmedicationuse:impactofEuropean
Commissionguidelinesandotherriskscales.DrugSaf2003;
26(1):111
75. EdwardsIR,OlssonS.WHOprogramme:globalmonitoring.
In: Mann RD, Andrews EB, editors. Pharmacovigilance.
Chicester:Wiley,2002:16982
76. CosynsJP.AristolochicacidandChineseherbsnephropathy:a
reviewoftheevidencetodate.DrugSaf2003;26(1):3348
77. Henderson L, Yue QY, Bergquist C, et al. St Johns wort

(Hypericum perforatum): drug interactions and clinical out
comes.BrJClinPharmacol2002;54:34956
78. Kavakavaandhepatotoxicity[newsitem].CurrProbPharma
covigilance2002;28:6
851
79. HouseofLordsSelectCommitteeonScienceandTechnology,
Session19992000,6threport.Complementaryandalternative
medicine.London:TheStationeryOffice,2000
80. DepartmentofHealth.AvisionforpharmacyinthenewNHS.
London:DepartmentofHealth,2003Jul
81. Pirmohamed M, Park BK. Genetic susceptibility to ADRs.
TrendsPharmacolSci2001;22(6):298305
Correspondence and offprints: Dr Joanne Barnes, Centre for

Pharmacognosy and Phytotherapy, School of Pharmacy,
University of London, 29-39 Brunswick Square, London, WC1N
1AX, UK.
E-mail: joanne.barnes@ulsop.ac.uk

Phyto Pharmaco Vigilance

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Phyto Pharmaco Vigilance

Hochgeladen von

Copyright:

Verfügbare Formate

See discussions, stats, and author profiles for this publication at:

Available from: Joanne Barnes Retrieved on: 29 December 2015

Adis Data Information BV 2003. All rights reserved.

Pharmacovigilance has been defined as the

medicines. Indeed, there is no need, nor is it

produced draft guidelines on this. Awareness has

Herbal medicinal products (also known as

Pharmacovigilance of Herbal Medicines

environmental factors, such as climate, and growing

The method ofprocessing crude herbalmaterial,

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

ent(s) with combination herbal medicines, are

label), and variations in and unacceptably high

pharmaceutically equivalent and bioequivalent.

The names that are usedfor herbal medicines

sulted a medical herbalist in the previous year.

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

Pharmacovigilance of Herbal Medicines

available (see section 1.3.1). Even where herbal

A related issue is that some users of herbal

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

In summary, the ways in which herbal

under section 12 of the Medicines Act 1968, or as

licensing requirements under section 12 (1). This

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

Pharmacovigilance of Herbal Medicines

in Directive 2001/83/EC, manufacturers of

In addition, the range of possible regulatory ac

Other issues relevant to pharmacovigilance arise

herbal medicinal products is wellrecognised,

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

conventional medicines. Certainly, there are

user) associated with the use of certain herbal

The need for a new regulatory framework for

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

Pharmacovigilance of Herbal Medicines

products for oral, external or inhalation use only

provision of bibliographic data on safety with an expert report

compliance with pharmacovigilance requirements in accordance with 2001/83/EC

ence in this area. For example, the requirement to

Other developments at the European level are

2. Methods for Pharmacovigilance of

The future of spontaneous reporting schemes in

The CSM/MHRAs national spontaneous report

for unlicensed herbal medicines. This move

Toxicology Unit, which identified suspected

pharmacists, respectively. Community pharma

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

Pharmacovigilance of Herbal Medicines

followed a 1year pilot scheme for community

(0.4%), respectively; p < 0.001). However,

Adis Data Information BV 2003. All rights reserved.

Drug Safety 2003; 26 (12)

The impact that these new reporter groups and

ADR Reporting Form

The section Suspected drug presents several