FDA treatment of botanical drugs
Dr Yuan-Yuan Chiu (US Food and Drug Administration and chairman of the FDA botanicals working group) focused on dietary supplements and botanical drugs. In 1994, the US Congress had enacted the Dietary Supplement, Health and Education Act (DSHEA), which newly permitted “dietary supplements” and herbs other than tobacco to be marketed with a claim that the product was intended to “affect the structure or function of the human body”. In Europe, this is only possible for “medicinal products”.
She emphasised that clinical trials with FDA oversight are only needed if the label claims that a product will “affect the progress of disease”. Through this DSHEA, many foreign herbal drugs are marketed in the US as dietary supplements with a structure and function claim, but in practice are actually purchased and consumed by people to treat various diseases. Dr Chiu explained that, because dietary supplements are [also] considered to be “food”, these herbal products are legally required to comply with a set of quality standards different from, and much lower than, those required for drugs. However, this does not apply to Chinese traditional medicines.
She listed reasons for the differences, including present and past experience in the US and the presence of multiple components (congeners) of different quality for which the total efficacy and quality can be expressed as “combination drugs”, provided that the congeners are from the same part of the plant.
Dr Chiu indicated that new DSHEA regulations, effective since 7 February 2000, require labelling compliance within 18 months, and permit a structure/function statement but allow no specific reference to disease (the FDA definition of which includes ageing). If the product is “generally regarded as safe”, evidence should be supplied for free sale, whereas if not generally regarded as safe, then an Investigational New Drug (IND) submission, with FDA-overseen clinical trials, is required for the active substance.
In respect of a marketing application for botanical drug products (as distinct from the purified or synthesised active substance), there needs to be comprehensive quality standards, as well as prior experience of manufacture and control, any perceived risks, extent of plant variability and the nature of congeners, and non-identifiable active constituents. She recognised difficulties special to HMPs, such as quantifying potency, establishing reference standards and assessing safety. There are no specific clinical recommendations and an IND is not required for clinical research with dietary supplements (unless specifically claiming to treat disease).
In August 2000, her working group issued for public comment a draft “Guidance for industry: botanical drug products”. This provided a discussion on both regulatory and scientific issues unique to botanical drug products, and presented recommendations on how to develop herbal products as human drugs.
She said that “numerous comments were received and reviewed by the FDA” and that the final, revised version “should be published soon”.
Dr Chiu accepted that sometimes botanical identification is difficult but multiple fingerprints are expected, with some test of batch variation with relevant chemical markers, and an environmental assessment of wild plant harvest. Impurities include endogenous toxins and agrochemicals, and radioisotopes and some appreciation of potential adulteration (one third of listed HMPs might have some deliberate addition of synthetic chemical, eg, indometacin). She looked to a higher standard of potency for botanical “drugs” and ideally specification of all constituents.
In respect of international harmonisation of quality of HMPs, she believed that botanical raw material should have agreed common names, universal “Good Agricultural Practice”, harmonised identification, recognised marker compounds and set limits for contaminants. She accepted the prospect of multiple monographs if different extraction principles were used. She would also like “negative dispensing information published for toxic plants” and national quality monographs for dosage forms. She concluded that the “FDA goal was to ensure safety and efficacy as an integral part of the public health system.”
European Union perspective
Professor Konstantin Keller, of the German Medicines Control Institute and head of the European Medicines Evaluation Agency (EMEA) permanent working party on HMPs, was unable to attend and his paper was presented by Dr Friedrich Lang (see below). The working party recognised a common understanding of quality, safety and efficacy of HMPs and had supplied guidance available on the EMEA website. Its concept of Good Agricultural Practice was published in 2001. He was suspicious of HMPs being described as “harmless” and he quoted reports about hyperi*****, garlic, sassafras oil and methyleugenol, and noted the banning in Europe of kava and celandine — but not colchicine in ginkgo, as had been elsewhere suggested.
He referred to a pyramidal model of recommendations in four tiers, with the quality of information progressing from “truth” at the top to “gossip and lies” at the base:
• Grade A, in the top tier, signifies the fully characterised product supported by properly randomised trials
• Grade B, subsumes HMPs that are at least characterised but have had no full trials
• Grade C relates to traditional herbs adequately described by bibliographic data but with no direct studies
• Grade D HMPs are below the line of expert opinion, with only anecdotal or historical description
Professor Keller commented that manufacturers of grade D extracts tend to register their HMPs as dietary supplements because they cannot afford to undertake the clinical studies required for a place in the higher grades.
Dr Friedrich Lang, Schwabe GmbH, Karlsruhe, Germany, examined the scientific justification for substitution of HMPs from the same plant and whether clinical or bioavailabity/bioequivalence studies could be replaced by in vitro biopharmaceutical investigations. He noted that the existing tight network of rules concerning dissolution testing and the investigation of bioavailability and bioequivalence has not yet completely been transferred to HMPs.
Dr Lang referred to an FIP working-group’s new concept [see Pharmazeutische Industrie, 2001, volume 63, no 10] of biopharmaceutical characterisation of HMPs based on classification of plant extracts, pharmaceutical equivalence of the active pharmaceutical ingredient and the formulation as well as the solubility of the extracts. He commented that some HMPs are complex mixtures, with up to 2,000 natural components. In this FIP concept, a Type A extract contains known constituents solely responsible for well documented bioactivity; there are only half-a-dozen standardised examples, such as senna and kava. Type B1 extracts have chemically defined constituents possessing relevant pharmacological properties (active markers likely to contribute to clinical activity), eg, valerian and stinging nettle.
The FIP working group does not consider dissolution testing generally to be adequate for HMPs, since in most cases only selected constituents can be traced but not the whole extract as the active pharmaceutical ingredient. HMPs containing extracts of categories A or B1 which are intended to be supplied in substitution should be compared with respect to their pharmaceutical and clinical equivalence. HMPs with the same active ingredient should be pharmaceutically equivalent, whereas if there are different excipients, or if the manufacturer has changed the solvent or extraction process, the products would no longer be bioequivalent. Alternatively, he suggested, they should comply with the EU/CPMP “Note for guidance on the investigation of bioavailability and bioequivalence”. This recognises the great importance of solubility in release and subsequent absorption of the HMP.
For the large group B2, HMPs containing the less well defined extracts, these need not comply with that guidance note. Analysis might not be helpful because of variation between batches. In the case of high solubility of the extracts — or equivalent solubility behaviour — bioavailability, bioequivalence or clinical studies can be waived. Nevertheless, he commented, if there are extreme differences in solubility of the extracts, clinical equivalence would still have to be demonstrated.
Dr Lang’s paper was directly complemented by a strategy to develop dissolution test protocols and acceptance criteria, described by the symposium chairman, Professor Helga Oeser. She emphasised that “dissolution testing is an important tool to characterise the biopharmaceutical quality of oral solid and semi-solid dosage forms”, as well as for comparing various formulations, assessing potential improvement during pharmaceutical development and optimising the design of bioavailability and bioeqivalence studies.
She quoted comprehensive “Guidelines for dissolution testing of solid oral products” that had been established by an FIP working group which require:
• Description, composition and manufacture of the product
• Physicochemical characteristics of drug substance
• In vivo performance relating pharmacokinetics and therapeutic concept
• In vivo/in vitro comparison
• Validated dissolution test conditions
• Setting specifications to be used worldwide
Professor Oeser listed factors that she considered important in qualification of the dissolution apparatus, such as dimensions, position and control of stirrer or flow rate (according to type of equipment), temperature volume and pH of medium, and defined sampling and tolerance allowed for test time and results. She quoted from the EMEA “Guidelines for dissolution and disintegration tests for synthetic drugs” and illustrated the development of dissolution procedures, based on current pharmacopoeial equipment, for immediate release HMPs containing extracts with known active compounds. Using the classification described by Dr Lang, she took milk thistle extract as an example from category A and St John’s wort and ivy leaf from category B1.
Thus, milk thistle dry extract contained between 40 and 80 per cent silymarin (calculated as silybin) and comprised at least six individual compounds of pH-dependent solubility. Professor Oeser reported bioavailability studies with 18 volunteers using three marketed forms of milk thistle extract (film coated, and soft and hard gelatine capsules). In in vivo studies, plasma analysis has shown that the rate of absorption (tmax in 60min) was similar for all three products but the innovator (hard gelatine) brand provided over twice the dose at tmax. In vitro tests, using a paddle (100rpm) system with 4L of pH7.5 buffer medium, demonstrated markedly different dissolution performance with the innovator brand giving 85 per cent release in 30 to 60min, compared with only 50 per cent for film coated product (confirming in vivo comparison), and virtually no release from the soft gelatine presentation. She decided that the test conditions were unsuitable and for these presentations, a faster paddle, or a flow-through system, with addition of 1 per cent sodium lauryl sulphate to the medium, was needed.
Her second example, St John’s wort products, contained lipophilic and water-insoluble active compounds. Investigations were undertaken using a paddle method with biorelevant simulated intestinal fluid (SIF) dissolution media. It was found that hyperforin release was poor when using “fasted state” SIF but was nearly comparable with in vivo performance of different HMPs if using “fed state” SIF . She assumed that the higher quantities of taurocholate and lecithin accelerated dissolution — but considered these “too expensive for routine use”.
Her third extract was common ivy leaf (French Pharm X): the principal component (2 to 5 per cent hederasaponin C) is water soluble and the product is available as tablets and oral solutions. She reported dissolution testing using the paddle method (75rpm) was satisfactory for all biorelevant pH (1.0 to 7.5), but the presence of the much less soluble congeneric component a-hederin required buffer adjustment to pH 7.5.
From these three studies with products containing known active constituents of milk thistle, St John’s wort and ivy leaf, she generalised that herbal medicinal products can be biopharmaceutically characterised by dissolution tests, provided that there was: identification of the active pharmaceutical ingredient and all constituents responsible for therapeutic activity; characterisation of solubility and permeability of the constituents; use of standard pharmacopoeial apparatus (paddle or flow-through system) for a variety of different dosage forms, with appropriate choice of dissolution medium; use of surfactant agents such as sodium lauryl sulphate, sodium taurocholate and lecithin for low solubility active compounds; and verification of the dissolution profile by in vivo performance.