Research and evaluation of herbal medicines without a long history of use or which have not been previously researched, should follow WHO’s Research guidelines for evaluating the safety and efficacy of herbal medicines3.
For herbal medicines with a well-documented history of traditional use, the following procedures for conducting research and evaluating safety and efficacy may be followed.
In assessing the safety and/or efficacy of a herbal medicine, whether derived from a single plant or from a defined mixture of plants, the first step involves the evaluation of literature reports. The literature search should include reference books, review articles, systematic surveillance of primary sources, and/or database searches. It should be kept in mind, however, that reference books and review articles might contain inaccurate information. Nevertheless, these sources will cite primary references that can be consulted for in-depth analysis. The search profile used should be recorded, as should details of any references cited, whether or not they are available. The literature search should then be extended to gather information on closely related plant species for chemotaxonomic correlation.
If several investigators publish similar safety and/or efficacy data, they should be accepted as useful indicators. In vitro (biochemical or cellular) safety data should be viewed as indicators of potential toxicity, but not as absolute markers. In vivo data from animal studies are more indicative of toxicity and may be considered to be safety markers.
For both safety and efficacy, a pharmacological effect observed in vitro or in animal models is not necessarily applicable to humans. In vitro data usually serve to verify the reported mechanism of action in animals or humans. Such data have to be confirmed by clinical studies. Well-documented reports of pharmacological activity in animals or humans may be viewed as having scientific rationale.
Theories and concepts of systems of traditional medicine
The theories and concepts of prevention, diagnosis, improvement and treatment of illness in traditional medicine historically rely on a holistic approach towards the sick individual, and disturbances are treated on the physical, emotional, mental, spiritual and environmental levels simultaneously. As a result, most systems of traditional medicine may use herbal medicines or traditional procedure-based therapies along with certain behavioural rules promoting healthy diets and habits. Holism is a key element of all systems of traditional medicine. Therefore, when reviewing the literature on traditional medicine (both herbal medicines and traditional procedure-based therapies), the theories and concepts of the individual practice of traditional medicine, as well as the cultural background of those involved, must be taken into account.
Review of safety and efficacy literature
A review of the literature should identify the current level of evidence for the safe and effective use of a herbal medicine. The study design should be evaluated, taking note of, for example, the number of patients, specific diagnosis, dosage, duration of administration, criteria for evaluation (such as improvement of symptoms), absence of simultaneous therapy, and valid statistical analysis.
In cases where traditional use and experience of a herbal medicine in humans have not established its safety and efficacy, new clinical studies will be necessary. If well-known herbal medicines are formulated into a new mixture, however, the requirements for proof of safety and efficacy should take into account the well-established uses of each herbal medicine. Such information may appear in authoritative national documents (such as pharmacopoeias or official guidelines of national authorities) or in highly respected scientific publications. However, it should not be forgotten that new preparative methods may alter the chemical, toxicological and even pharmacological profiles of traditionally used herbal medicines.
Issues related to reviewing literature on clinical trials are presented in Part 3, and should also be consulted.
Reported and documented side-effects (recorded according to established principles of pharmacovigilance) of a herb or herb mixture, its closely related species, constituents of the herb and its preparations/finished herbal products should be taken into account when decisions are made about the need for new pharmacological or toxicological studies.
The absence of any reported or documented side-effects is not an absolute assurance of safety for herbal medicines. However, a full range of toxicological tests may not be necessary. Tests which examine effects that are difficult or even impossible to detect clinically should be encouraged. Suggested tests include immunotoxicity (e.g. tests for allergic reactions), genotoxicity, carcinogenicity and reproductive toxicity. The discussion presented in Annex III may be used for reference. WHO’s Research guidelines for evaluating the safety and efficacy of herbal medicines can also be consulted for these as well as for other appropriate toxicity tests (see Annex II).
Only when there is no documentation of long historical use of a herbal medicine, or when doubts exist about its safety, should additional toxicity studies be performed. Where possible, such studies should be carried out in vitro. Using in vitro tests can reduce the number of in vivo experiments. If in vivo studies are needed, they are to be conducted humanely, with respect for the animals’ welfare and rights. Toxicity studies should be conducted in accordance with generally accepted principles, such as those described in WHO’s Research guidelines for evaluating the safety and efficacy of herbal medicines3.
It is important for herbal medicines, and particularly for those made from mixture herbal products, that the requirements for proof of efficacy, including the documentation required to support the indicated claims, should depend on the nature and level of the indications. For the treatment of minor disorders, for non-specific indications, or for prophylactic uses, less stringent requirements (e.g. observational studies) may be adequate to prove efficacy, especially when the extent of traditional use and the experience with a particular herbal medicine and supportive pharmacological data are taken into account. The level of the evidence and the grading of recommendations must correspond to the nature of the illness to be treated or the nature of the physical or mental function to be influenced and regulated. Definitions of levels of evidence and the grading of recommendations from the USA Agency for Health Care Policy and Research may be used for guidance (see Annex IV). Many other national documents, such as the Australian Guidelines for levels and kinds of evidence to support claims for therapeutic goods (see Annex V), could also be used for reference.
The therapeutic alternatives available within the community and the risks of the herbal medicine have to be taken into account. It should be noted that in the case of herbal medicines made from herb mixtures, a therapeutic or scientific rationale must exist for the presence of each herb in the mixture. Research on possible therapeutic effects of herbal medicines made from herb mixtures or specific combinations of herbs, however, needs to be carried out.
The scope and design of such studies should be based on information on traditional use obtained from official national compendia and relevant literature, or by consultation with traditional medical practitioners.
In the case of a new herbal medicine, a new indication for an existing herbal medicine, or a significantly different dosage form or route of administration, the general principles and requirements for a clinical trial should be very similar to those which apply to conventional drugs (see, for example, WHO’s good clinical practice protocols, which are described in Annex VI). In some cases, however, the design of such studies must be adapted to deal with the particularities of herbal medicines.
Well-established, randomized controlled clinical trials provide the highest level of evidence for efficacy. Such studies facilitate the acceptance of herbal medicines in different regions and in people with different cultural traditions. However, methods such as randomization and use of a placebo may not always be possible as they may involve ethical issues as well as technical problems. For example, it may be not possible to have a placebo control if the herbal medicine has a strong or prominent smell or taste, as is the case for products containing certain essential oils. In addition, patients who have been treated previously with the herbal medicine under investigation that has a characteristic organoleptic property, cannot be randomized into control groups. In the case of herbal medicines with a strong flavour, placebo substances with the same flavour may have a similar function. In such cases, it may be advisable to use a low dosage of the same herbal medicine as a control. Alternatively, a positive control, such as well-established treatment, can be used. Other examples of control groups are presented in Part 3.
Observational studies involving large numbers of patients may also be a very valuable tool for the evaluation of herbal medicines. According to the theories and concepts of traditional medicine, as mentioned in Part 1 (section 1.3), the prevention, diagnosis, improvement and treatment of illness is often based on the specific needs of the individual patient. Therefore, single-case studies for the evaluation of efficacy of a herbal medicine should not be ignored. Due to the potential contribution of single-case studies to traditional medicine, a more detailed description of this and other study designs is given in Part 3.
Regulatory requirements of national authorities for evaluating herbal medicines differ from country to country. Many governments have recently developed their own national regulations for traditional medicine. For an extensive review of the regulatory situation in various countries, consult WHO’s Regulatory situation of herbal medicines: a worldwide review4.