A UK-based study to be published imminently in the peer reviewed journal BMC Medicine declares that 75% of the over-the-counter herbal products tested did not come with important safety information, including warning consumers about important interactions that can potentially occur with conventional medicines.

Nutraingredients reported on the new study yesterday, Monday 8th August, flagging the key finding that herbal products registered under the Traditional Herbal Medicinal Products Directive (THMPD) were better labeled than unregistered products and food supplements.

However, there may be very good reasons for this finding, reasons that the study authors failed even to mention. One of the most obvious is that products that have received ‘Traditional Herbal Registrations’ (THRs) under the EU directive are registered medicines, making it compulsory for them to contain detailed, relevant patient information, including possible adverse effects and contraindications.  Herbal products sold as food supplements, by contrast, are intended for healthy populations; as a result, indicating the product's risks, for example, to diabetics or other diseased populations might be sufficient for a regulator to deem the product as medicinal under the broad scope of EU medicinal law.  It is what might be described as a ‘Catch 22’ situation.

It appears that the paper’s findings may be compromised by a conflict of interest. The main author, David K Raynor, from the School of Healthcare, University of Leeds, is a director of Luto Research. One of the co-authors, Peter Knapp, was a co-founder of Luto, and Alison Blenkinsopp, who reviewed the manuscript, is a non-executive director of Luto. Luto Research "specialises in improving patient communication [by working] to enhance the clarity of information created for patients": in other words, they develop the type of patient information leaflets that this study finds are so sorely needed in OTC herbal products.  So by giving the paper its particular spin, they give themselves a high chance of earning lucrative consultancy income from a new industry.  Several of the authors have worked together in the past to evaluate patient information in the pharmaceutical sector – now it appears they have been led to new, fertile pastures.

But does this really matter?  Let’s make sure we’re not over-reacting just because these authors have generated a paper that will create a lot of mileage for the anti-herbals brigade. After all, as the paper points out, no research has previously been performed as to the extent and suitability of patient information leaflets in herbal products; and if these authors have the expertise for the job, then perhaps any hand-wringing over their commercial ties is unnecessary. In any case, there can be no complaints if the paper adds usefully to our existing knowledge, by objectively reporting a thorough analysis of patient information in the OTC herbal market.

Unfortunately, the paper contains some serious methodological flaws that, in the view of ANH-Intl, seriously weaken or even invalidate the authors' conclusions. In fact, the very hypothesis upon which the entire analysis rests, one that is inferred rather than stated, can be shown to be invalid.  Our reasons for this view are given in our critique below.

Our main concerns fall into three distinct categories.

1. Inadequate classification of herbal products under study

In the text of the paper, the authors distinguish between three categories of OTC herbal product on sale in the UK: licensed herbal medicines, which have a full product license in the same manner as conventional medicines; traditional herbal medicines, with a Traditional Herbal Registration (THR) granted under the THMPD; and unlicensed herbal products, which they took to include herbal food supplements. These three categories – THRs, licensed herbal medicines and unlicensed herbal medicines – are included in their data tables.

Unlicensed herbal medicines versus herbal food supplements

The authors only looked at a total of 68 products from eight retailers, representing just five single-herb products.  The breakdown was as follows: garlic, 21; St John’s wort, 15; echinacea, 13; ginkgo, 12; and Asian ginseng, 7.  Of these, the vast majority (93%) were found to be unlicensed, but here the authors made their first serious error. Stunningly, they did not seek in their analysis to separate out those products that were food supplements –legally intended only for healthy populations – from those that were selling under the now-defunct Section 12(2) exemption of the Medicines Act 1968, existing stock of which is still allowed to be ‘sold through’ in the UK.  This means that by far the largest group the authors studied was actually split between a category selling under a food regime, where no medicinal indications or contraindications can legally be made, and another category selling under medicinal law, albeit one exempted from licensing.

Even the products with full licenses have not been properly categorised, as some could have been granted licenses under the ‘well-established use’ provision in EU medicinal law, while others may have been granted national product licenses, many of which have been – or will be – grandfathered into the THR scheme.

No patient information needed for herbal food supplements

The authors' most significant mistake concerns herbal food supplements. These are an entirely separate category of products, governed not by the THMPD or any form of medicinal law, but by the Food Supplements Directive (FSD; Directive 2002/46/EC). Under the terms of the FSD, food supplements are legally obliged to state that they are food supplements on their packaging; these are the 48 products that the authors declared to be unlicensed herbal products, and which “stated that they were food supplements”.  (However, in the authors’ analysis, they did not distinguish between these and the 'medicines exempt from licensing' category!)

Since herbal food supplements are not medicines, but a sub-category of food, they are the responsibility of the Department of Health/Food Standards Agency in the UK, not the UK’s medicines regulator, the Medicines and Healthcare products Regulatory Agency (MHRA). Food supplements may be sold using brand names and can make particular health claims but, crucially, because EU law effectively states that food supplements are meant exclusively for use in healthy people, they do not need to include patient safety information. In fact, including patient information leaflets in this manner could easily bring them under the umbrella of EU medicines law, making them illegal unless they were in turn licensed!

It can be seen, therefore, that many of the products criticised by the study authors for not including patient safety information are not legally obliged to do so.

2. Misreading of advice provided by NCCAM herbal monographs

The authors of this study developed their lists of precautions, interactions and side effects for each herb by analysing herbal monographs published by the USA’s National Center for Complementary and Alternative Medicine (NCCAM). This US body is part of the National Institutes for Health (NIH), and was chosen for the analysis because its information is free to access and because its monographs covered all five herbs in the authors' remit.  The authors deemed that the Community Herbal Monographs published by the European Medicines Agency were inappropriate for use.

The authors have assumed that, where an NCCAM monograph mentions an adverse event related to a particular herb, it represents a blanket warning that should apply whenever that herb is intended to be used by a patient. For example, in the case of ginkgo, the section of the monograph entitled 'Side Effects and Cautions' states that, “There are some data to suggest that ginkgo can increase bleeding risk, so people who take anticoagulant drugs, have bleeding disorders, or have scheduled surgery or dental procedures should use caution and talk to a health care provider if using ginkgo”. In the authors' 'Key Safety Issues' in Table 1, this translates to three separate 'Precautions', all of which the authors felt should have been mentioned as warnings on patient safety information included with all ginkgo products. “The public should be able to expect full disclosure of key safety information when they purchase herbal medicines over the counter,” the paper declares.

However, the authors have failed to take into account the matter of dosage, to which the text in the above example strongly alludes in the phrases “some data” and “can increase”. The NCCAM monographs – and the authors used abridged, 'Herbs At A Glance' monographs for their analysis – do not include any discussion of the dosages of herbal active ingredients needed to produce the listed side effects. More comprehensive sources of information, such as the full monographs found on the Natural Standard database, give full details of recommended doses of active ingredients and the effects that may occur in excess of the safe dose.  We presume the reason this data source was ignored is because it requires a subscription fee.  There is perhaps something to be said for the adage: you get what you pay for.

For example, Natural Standard states that ginkgo is, “Generally safe when taken by otherwise healthy adults [as an herbal food supplement, these are the people for whom ginkgo is intended] orally in suggested doses for up to six months”. The general dosing recommendation for ginkgo, taken like all of Natural Standard's data from clinical trials, is as follows: “Traditional recommendations include Ginkgo products containing 24% flavoglycosides (also called flavone glycosides or flavones) and 6% terpenes: 80-240mg of a 50:1 standardized leaf extract daily, or 3-6mL of a 40mg/mL liquid extract in 2-3 divided doses, or 30-40mg of extract in a tea bag, prepared as a tea, for at least 4-6 weeks. There is a lack of evidence in support of the clinical benefit of small concentrations of Ginkgo found in fortified foods.”  

We suggest that the authors would have been better advised to use comprehensive data sources for their analysis, even at the expense of a subscription fee. Without sound dosage recommendations from their information source, it is impossible to know whether the products included in the authors' analysis contain enough active ingredient to cause the side effects and interactions with which the active ingredient is associated. Furthermore, the authors do not provide any details about the studied products beyond where they were bought, their legal classification and the level of patient information supplied, so the reader cannot ascertain which are higher-dose products and therefore which are more or less likely to cause side effects.

3. Study hypothesis not supported by the evidence

The authors do not provide a written hypothesis that their study has been designed to test. However, given the rationale and design of the study, and the authors' conclusions, the following hypothesis can be inferred: “The greater the congruence between patient safety information included with OTC herbal products in the UK and the full listing of patient-related safety information contained in the NCCAM monographs, the greater the quality of the patient information and the consequent safety of consumers”.

However, in light of the above discussion, this hypothesis can be seen to be fatally flawed. The information in the NCCAM monographs is abbreviated and incomplete, pointing only to potential adverse events that might occur if an herbal ingredient is taken at a level above an unspecified safe dose. Therefore, it follows that patient information leaflets in close accord with those monographs – i.e. those that include all of the precautions, interactions and adverse events mentioned in the monograph – present an inaccurate and exaggerated picture of the health risks posed by the herbal products, particularly those containing less than full therapeutic dosages of active ingredients. Not only is patient safety not increased by greatly overstating these health risks, but patients who are discouraged from using herbal medicinal products by inaccurate patient safety leaflets may miss out on highly beneficial health effects.

More importantly, products selling as food supplements are not even intended for patients; they are intended for healthy people wanting to remain healthy.  The declaration of a full list of adverse effects and possible contraindications on such food supplements would not only be, in many cases, unnecessary, it would probably cause a regulator to classify the product as a medicine!


This study appears to be the product of a particular research group straying into territory unfamiliar to them, and overlooking the most important principle associated with the scientific method: that is, developing and testing a hypothesis.  The study is flawed because of an invalid inferred hypothesis and the incorporation of what is referred to scientifically as ‘confounding’.  Worse than that, it appears to have been self-serving, concluding that the provision of more patient information, based on a US government database, would – seemingly miraculously – make products that are already among the safest that we ingest, even safer.  And the authors, it seems, are ideally set up to do this work as consultants.  How convenient.

But without having categorised the products appropriately, by looking at only five different kinds of single-herb product in the UK, and by ignoring the dosage of herbal active ingredients within the products, we believe that the study is not worth the paper it is printed on.


ANH Homepage

ANH Good Science campaign page