Introduction
In accordance with the European legislation on medicinal products, herbal medicinal
products – whether authorized (“well-established medicinal use”) or registered (“traditional
use”) – have to prove their quality, safety, and efficacy or tradition, respectively,
in case of traditional herbal medicinal products. This also includes the discussion
of potential interactions and respective information in the package leaflet of the
products, if justified by scientific data. For this reason, the relevance of potential
herb-drug interactions has to be evaluated, and a decision must be taken whether respective
information has to be included in the package leaflet.
Examples for Regulatory Actions
Some herb-drug interactions are well-known, e.g., between high-dose St. Johnʼs wort
(Hypericum perforatum) extract and several chemical substances, e.g., warfarin and cyclosporin, which have
been described in some of the first case reports of herb-drug interactions in the
literature and nowadays belong to the most well-known examples for herb-drug interactions.
Based on these case reports, an official risk assessment (“Stufenplanverfahren”) of
St. Johnʼs wort products was initiated by the German health authority in March 2000.
As a consequence, the package leaflets of products with a daily dose of more than
200 mg of the herbal drug or the equivalent amount of extract were obliged to include
specific information on potential interactions with a number of chemical substances
thus permitting the safe use of the respective products. The decision of the health
authority was explained by the ability of hyperforin and hypericum extracts to induce
CYP450 isoenzymes and P-glycoprotein. The limitation of the daily dose was of particular
relevance for traditional herbal medicinal products containing St. Johnʼs wort in
the German market, because according to the national legal situation existing at that
time, the “traditional use” status was not compatible with the labelling of any risks
including potential interactions.
In January 2004 the BfArM issued a draft guidance document on the assessment of potential
pharmacokinetic interactions with herbal medicinal products [1] for public consultation. This guidance document explained in detail that examinations,
on such interactions or, alternatively, information that the use of the respective
chemical substance is contraindicated, were generally required for herbal medicinal
products. Interested parties who commented on this draft argued that experimental
data should only be required in justified cases when valid case reports or scientific
literature data point to potential pharmacokinetic interactions. Later on, a list
of 12 (out of approximately 180 commonly used) plant species was made available for
which the health authority considered a risk of interactions with other medicinal
products possible, including some plants for which in fact detailed information on
interactions was requested, e.g., Eucalyptus, artichoke, milk thistle, St. Johnʼs wort, and Colchicum.
Between 2005 and 2008, a further official risk assessment was performed in Germany.
Based on a study published by Piscitelli et al. (2002) [2], the health authority announced its intention to implement measures for herbal medicinal
products containing garlic (Allium sativum) due to the suspicion of interactions between garlic and saquinavir. However, the
design and the results of the study were put into question because the preparation
was not exactly defined, and the decrease mean area under the curve (AUC) of saquinavir
was seen in six of the patients whereas in three patients an increase was observed.
Furthermore, a control group as well as information on the nutrition which may influence
the bioavailability of saquinavir were missing. For these reasons, the measures were
not considered justified by interested parties. The final decision of BfArM, however,
did not differ from the earlier draft.
Current Regulatory Basis
Authorized herbal medicinal products as well as registered traditional herbal medicinal
products are obliged to submit a summary of product characteristics (SmPC) as part
of their application for marketing authorization or registration, respectively. The
SmPC serves as a basis, e.g., for information given in the package leaflet. General
criteria on its content are laid down in the “Guideline on Summary of Products Characteristics
(SmPC)” published as a 2nd revision in September 2009 [3]. In the chapter on interactions, this document states that the respective section
4.5. should provide information on the potential for clinically relevant interactions based on the pharmacodynamic properties and in vivo pharmacokinetic studies of the medicinal product including in vivo interaction results. Interactions not studied in vivo but predicted from in vitro studies or deducible from other situations or studies should be described if they
result in a change in the use of the medicinal product.
In April 2010, a revised version of the “Guideline on the Investigation of Drug Interactions”
was published as a draft [4]. As compared to the existing document, the new draft includes an additional chapter
specifically addressing the issue of herbal medicinal products and specific food products.
As a basic principle of this chapter, the potential for interactions should be investigated
for new herbal preparations. For traditional and well-established herbal preparations,
such a potential should be clarified if reports point to clinically relevant interactions
in humans. In vitro studies on the enzyme inhibitory potential of the constituents or the herbal preparations
are encouraged in case in vivo information may give rise to clinically relevant drug interactions. The guideline
also mentions the option of extrapolation from one herbal preparation to other ones.
An in vivo drug interaction study with a specific herbal product should be considered in case
product-specific labeling information is required. Finally, some advice is given on
how to proceed if there is a wish to investigate the effect of an herbal medicinal
product or a special kind of food on the pharmacokinetics of a medicinal product.
All in all, this guideline takes into account the specific issues of existing herbal
preparations with regard to interaction studies and restricts the need to perform
them to specific cases.
With a view to the existing regulatory framework and the recommendations of the respective
guidelines, it can be concluded that on a European level, herb-drug interactions have
so far not been a major issue, and recent guidance documents demonstrate that the
need to perform interaction studies should not be overestimated.
HMPC Monographs
Specific assessment criteria for herbal medicinal products which can be used by applicants
and health authorities are provided by the documents elaborated by the Committee on
Herbal Medicinal Products (HMPC) and published by the European Medicines Agency (EMA).
In this context, the community herbal monographs are of particular interest. They
consist of basic information on benefit and risk of herbal medicinal products and
traditional herbal medicinal products and can be used as reference documents for marketing
authorization/registration applications.
Out of more than 100 HMPC monographs published or drafted so far, only approximately
20 % include information on potential interactions, whereas the majority of the available
monographs state “none reported”.
As an example, the monograph on the well-established medicinal use of St. Johnʼs wort
(Hypericum perforatum, herba) [5] is mentioned which states under 4.5. ‘Interactions’: “Hypericum dry extract induces
the activity of CYP3A4, CYP2C9, CYP2C19 and P-glycoprotein. The concomitant use of
cyclosporine, tacrolimus for systemic use, amprenavir, indinavir and other protease
inhibitors, irinotecan and warfarin is contraindicated (see section 4.3. ‘Contraindicationsʼ).
Special care should be taken in case of concomitant use of all drug substances the
metabolism of which is influenced by CYP3A4, CYP2C9, CYP2C19 or P-glycoprotein (e.g.,
amitriptyline, fexofenadine, benzodiazepines, methadone, simvastatin, digoxin, finasteride),
because a reduction of plasma concentrations is possible. The reduction of plasma
concentrations of oral contraceptives may lead to increased intermenstrual bleeding
and reduced safety in birth control. Women using oral contraceptives should take additional
contraceptive measures. Prior to elective surgery possible interactions with products
used during general and regional anaesthesia should be identified. If necessary the
herbal medicinal product should be discontinued. The elevated enzyme activity returns
within 1 week after cessation to normal level.”
In contrast to this, the final monograph on the traditional use of low dose St. Johnʼs
wort (Hypericum perforatum, herba) [6] states for the internally used preparations: “In the case of a daily intake of hyperforin
less than 1 mg and of a duration of use not longer than 2 weeks (see section 4.2.
‘Posology and method of administration’), no clinically relevant interactions are
to be expected. Patients taking other medicines on prescription should consult a doctor
or pharmacist before taking Hypericum.”
However, as the draft monograph on the traditional use of low doses had included the
same information on interactions like the monograph on the clinically proven high-dose
preparations, interested parties had submitted detailed comments demonstrating that
traditionally used, low-dose St. Johnʼs wort products do not bear a risk of interactions
with other medicines. In the past few years several studies were performed in order
to elucidate the potential risk of various Hypericum preparations and to provide arguments for a balanced assessment with the German “Stufenplanverfahren”,
e.g., several studies in healthy volunteers using various marker substances such as
digoxin, cyclosporine A, or midazolam, respectively [7], [8], [9], [10], and one study in patients receiving ciclosporin A as a permanent medication after
kidney transplantation [11].
Accordingly, the HMPC assessment report [12] concluded that the extent of induction of CYP3A4 is well-documented and directly
correlated with the content of hyperforin in the herbal preparation. Products containing
only small amounts of hyperforin (< 1 %) have not been shown to produce clinically
relevant enzyme induction. For high-dose preparations attached to the well-established
medicinal use, the assessment report mentions pharmacokinetic interactions for several
drug substances metabolized via CYP3A4 having a narrow therapeutic range. Thus it
is concluded that Hypericum extracts should not be used concomitantly with these substances, or doses have to
be adjusted.
Further examples of HMPC monographs with relevant and justified information on interactions
are mentioned in the following:
The monograph on Agnus castus (Vitex agnus-castus, fructus) [13] states: “Because of the possible dopaminergic and oestrogenic effects of Vitex agnus-castus,
fructus, interactions with dopamine agonists, dopamine antagonists, oestrogens and
antioestrogens cannot be excluded.” The same wording is included in the assessment
report; a proof for this assumption by clinical data, however, is not given.
The monographs on bulk-forming laxatives, e.g., linseed (Linum usitatissimum, semen) [14] and ispaghula (Plantago ovata, semen; Plantago ovata, seminis tegumentum) [15], [16], explain that because of their pharmacodynamic properties the enteral absorption
of other concomitantly administered medicines may be delayed, and therefore the laxative
should not be taken ½ to 1 hour before or after intake of other medicinal products.
Further information is given with regard to concomitant use of medicinal products
which inhibit peristaltic movement as well as of thyroid hormones or insulin.
Furthermore, all of the six monographs on anthraquinone-containing herbal substances
(e.g., Cassia senna, fructus, folium [17], [18], Rhamnus frangula, cortex [19], etc.) include the following statement: “Hypokalaemia (resulting from long-term
laxative abuse) potentiates the action of cardiac glycosides and interacts with antiarrhythmic
medicinal products, with medicinal products, which induce reversion to sinus rhythm
(e.g. quinidine) and with medicinal products inducing QT-prolongation. Concomitant
use with other medicinal products inducing hypokalaemia (e.g. diuretics, adrenocorticosteroids
and liquorice root) may enhance electrolyte imbalance.” This corresponds to current
scientific knowledge and is in principle also part of the respective ESCOP monographs
[20].
The monograph on willow bark (Salix, cortex) [21] states: “Willow bark may increase the effects of anticoagulants such as coumarin
derivatives.” Detailed justification for this statement can be found in the assessment
report [22], though its clinical relevance may be low, as it is partly based on marginal differences
in a small numbers of patients [23], [24].
More recently, a draft on Glycyrrhiza glabra and/or Glycyrrhiza inflata and/or Glycyrrhiza uralensis, radix [25] was published for public consultation with the following statement: “Liquorice root
may counteract antihypertensive action of prescribed medications. Not to be used concomitantly
with thiazide diuretics, cardiac glycosides, corticosteroids, stimulant laxatives
or other medications which may aggravate electrolyte imbalance.” These interactions
are known for the intake of high doses and are also included, e.g., in the respective
ESCOP monograph [20].
Some other monographs, however, include statements on potential interactions with
other drugs which should be assessed in a more critical manner, e.g., two monographs
on tannin-containing herbal drugs, Tormentil rhizome (Potentilla erecta, rhizoma) [26] and oak bark (Quercus robur, Q. petraea, Q. pubescens, cortex) [27]. They both include the following statement: “Internal absorption of concomitantly
administered medicine may be delayed. For this reason the product should be taken
1 hour or more before or after intake of other medicinal products.” As there is no
evidence on potential interactions available from the assessment reports, except that
there are no concerns about interactions with oak bark preparations [28], [29], the statement of the monograph seems to be based on theoretical considerations,
and its clinical relevance may be further questioned in case the preparations are
taken for the treatment of diarrhea, a condition significantly influencing the absorption
of medicines.
According to the monograph on valerian (Valeriana officinalis, radix) [30], “only limited data on pharmacological interactions with other medicinal products
are available. Clinically relevant interaction with drugs metabolised by the CYP 2D6,
CYP 3A4/5, CYP 1A2 or CYP 2E1 pathway has not been observed. Combination with synthetic
sedatives requires medical diagnosis and supervision.” As the mentioned interactions
with various cytochrome P450 enzymes have not been observed, it is questionable whether such statements should
be made under “interactions” in the monograph and therefore in the SPC and package
leaflet of the products or should better be only part of the HMPCʼs assessment report.
From a scientific point of view, the statement on interactions in the monograph of
sage leaf (Salvia officinalis, folium) [31] is not quite understandable: “The intake of Salviae folium preparations might influence
the effect of medicinal products acting via GABA receptor (e.g. barbiturates, benzodiazepines),
even if not seen clinically. Therefore the concomitant use with such medicinal products
is not recommended.” In accordance with the “Guideline on Summary of Product Characteristics”
[3], information on interactions should be provided when they are clinically relevant. The assessment report [32] explains that no drug interactions are documented clinically, but assumes that preparations
of sage might interact with other medicines due to the effect of α-thujone on the γ-aminobutyric acid (GABA) type A receptor. However, in order to exclude any neurotoxic
effects, the presence of thujone in sage leaf preparations is restricted to a daily
intake of 5.0 mg/person for a maximum duration of 2 weeks in the HMPC monograph. Bearing
in mind that the content of thujone is restricted and the suspect of potential interactions
are solely deduced from pharmacological data without any clinical evidence, the statement
on interactions of sage leaf preparations should be discussed critically.
As could be shown by the examples mentioned above, only a small number of the existing
HMPC monographs include information on potential interactions. Most of them are scientifically
justified and based on sound data showing clinical relevance. Some of them, however,
must be put into question because they seem to be deduced from theoretical considerations
only. For most of the herbs which are of relevance in the European market and for
which HMPC monographs exist, there is no evidence of interactions with chemical drugs.
Discussion
With some exemptions, the discussion of herb-drug interactions does not seem to be
of high relevance on a European level. Information on potential interactions, e.g.,
of high-dose St. Johnʼs wort products with chemical substances such as cyclosporine,
indinavir, tacrolimus, warfarin, etc., is mandatory and must be included in SPCs and
the package leaflet in order to guarantee the safe use of herbal medicinal products
and thus contribute to the protection of patients and consumers. However, potential
interactions which are described in the literature and which are deduced from in vitro or animal data should be assessed critically for their clinical relevance.
Butterweck et al. [33] had already stated earlier that systematic in vitro screenings led to contradictory results and showed low predictability for clinical
effects. As drug interactions can also be mediated by food, beverages, etc., a preventive
risk assessment should at first be done for synthetic drugs with a narrow therapeutic
range. Thus it would be more important to provide information on the interaction to
the persons using the synthetic drugs instead of informing the wide population using
herbal medicinal products or food which contains herbal preparations [33].
Furthermore, the European regulatory situation of herbal and traditional herbal medicinal
products is not comparable to the situation in the United States where herbal preparations
are classified as dietary supplements without the need of being authorized or registered
before gaining access to the market. Presumed increased risks of interactions as described
for unauthorized products with a potential high variability in their phytochemical
contents [34] can therefore not be transferred uncritically to medicinal products which are obliged
to prove their quality, safety, and efficacy or traditional use, respectively.
Conclusion and Perspectives
Information on clinically relevant interactions is an important part of the package
leaflet of herbal medicinal products and traditional herbal medicinal products and
allows for the safe use of these products. However, published reports on in vitro interactions should be checked case-by-case for their clinical and regulatory relevance.
This is in line with regulatory guidelines stating that the potential of interactions
should be clarified if reports point to clinically relevant interactions in humans.
It does not seem justified to demand the conduct of clinical studies demonstrating
the lack of clinical relevance based on theoretical considerations only.
Therefore a balanced assessment of potential interactions is absolutely required,
and information on the interactions in the package leaflet of synthetic medicinal
products, in particular those with a narrow therapeutic range, is required as well.
All in all, a sense of proportion should be kept when regulatory documents such as
monographs are established or regulatory actions are undertaken.
