Evidence for adverse effects of cannabidiol (CBD) products and their non-conformity on the European food market – response to the European Industrial Hemp Association

Introduction

We actually agree with a main premise of the European Industrial Hemp Association (EIHA)’s comments; namely, that legal compliance and safety for both producers and consumers of cannabidiol (CBD) products must be ensured¹. If this can be achieved by their suggestion of a mandatory industry self-regulated approach¹ rather than by our suggestion of strict regulations^2,3 is another question. In light of the experience with industry self-regulations in other fields, this suggestion remains highly doubtful^4,5; it especially appears not well-thought-out how a self-regulation may be mandatory and how this demand can be enforced. Otherwise, we had previously suggested the necessity for a common regulatory approach regarding hemp food products on a European level, such as enforceable maximum levels for Δ⁹-tetrahydrocannabinol (Δ⁹-THC)⁶. To even increase the legal void by the assessment of hemp extract-based food products as non-foods within the international and national narcotic regulations⁷ is clearly not helpful.

Apart from these issues, which are political rather than scientific, the main finding presented in our study is that the levels of psychoactive Δ⁹-THC in many CBD products on the market exceed acceptable thresholds of toxicity. Furthermore, hemp extract-based CBD products were assessed as unapproved novel foods. For both reasons, the marketing of such products is illegal according to European Union (EU) food laws² (if we assume that the products are foods and not narcotics). The disputation provided by EIHA to refute our assessment is based on claims rather than facts and we therefore take this chance to corroborate our assessment by further data published since the last revision of our paper in January 2020.

Let us now consider each criticism of the EIHA¹ in detail:

Adverse effects of CBD products

The literature regarding adverse effects of CBD was considerably expanded since the writing of our article, so that besides the anecdotal reports and paediatric studies already mentioned, several case reports^8–10, a survey⁹, a meta-analysis of clinical trials¹¹ and a post marketing safety assessment of a full spectrum hemp extract¹² became available. Specifically the survey⁹ reports observations of adverse effects including “feeling high”, an effect likely to be explained by Δ⁹-THC contamination rather than by CBD. Similarly, effects of “a little high”, brief periods of mild intoxication, were described following ingestion of several brands of CBD products in Virginia, USA¹⁰. The post marketing safety assessment showed gastrointestinal effects as most commonly reported adverse effect¹². We have included this information and several more studies into the two newer versions (version 3; v3¹³ and version 4; v4¹⁴) of our article² to strengthen our arguments. Besides the mentioned human evidence, experimental research in vivo and in vitro resulted in concerns about hepatotoxicity¹⁵, teratogenicity¹⁶, and gut inflammation¹⁷.

The arguments of EIHA in refuting any adverse effects of CBD products are not convincing not only from a pharmacological standpoint, but it is also rather unscientific to refute adverse effects on the basis that the responsible authority in the UK has not been made aware of any safety incident till now¹. First, due to the very short time of public use of CBD, only acute toxic events would currently appear, while chronic toxic events, such as liver toxicity, may take years to develop. Second, there is currently no system of nutrivigilance implemented in the UK or most other EU member states, so that there is no formal registration of such cases.

“THC” definition and estimation of daily dose of products

Unfortunately, a misunderstanding regarding our definition of Δ⁹-THC has apparently occurred at EIHA. From the methods section and our definition of the abbreviation of THC as Δ⁹-tetrahydrocannabinol, it should be clear that we only include the psychotropic Δ⁹-THC and not the non-psychotropic Δ⁹-tetrahydrocannabinolic acid (THCA) in our risk assessment. Both compounds are baseline separated in our chromatographic conditions (Figure 1).

Figure 1. Typical liquid chromatography-tandem mass spectrometry chromatogram of a cannabinoid mixture showing the separation between CBD (a), CBN (b), Δ⁹-THC (c), Δ⁸-THC (d) and THCA (e) (concentration 200 µg/L; analytical conditions see Ref. 14).

In deviation of the suggested practice to implement the former German guidance values for total THC (i.e. the sum of Δ⁹-THC and THCA)¹⁸, we only have compared the psychotropic Δ⁹-THC with the guidance values. This practice is clearly in favour of the food business operator (FBO) because – as the EIHA correctly states – a major part of total THC may be comprised by THCA¹. We have updated our article to clarify this issue on several instances and to avoid future misunderstandings¹³.

Besides the issue about THCA and its potential degradation into Δ⁹-THC, the uncertainties and limitations of the exposure assessment include a lack of knowledge about the typical consumption amounts of hemp foods, which are not covered by the currently available representative nutritional surveys as a separate food category. This problem is less problematic for food supplements, which must be labelled with a recommended daily intake, but may lead to uncertainties with tea, beverages and other derivative products containing CBD or hemp extracts. The typical approach in such situations is to use survey data for the most similar food category (such as herbal tea for hemp tea, see detailed discussion in the examples below). Otherwise, to ensure a high level of consumer protection, worst-case consumption amounts, considering also sensitive consumer groups such as children, are typically considered in these situations.

Regarding the exposure assessment, we confirm to have conducted an estimation of the daily intake in those instances, where a maximum recommended daily dose was missing in the labelling. This is in accordance with the requirements of Art. 14 of Regulation (EC) No. 178/2002 laying down the general principles and requirements of food law¹⁹, which specifies that the information provided to the consumer must be considered in determining whether any food is unsafe. As we will elaborate further in the following, our exposure estimations are both practical and realistic from the standpoint of consumer health protection.

Mitigation of THC effects by interaction with CBD?

The allegation of EIHA that we have dismissed the interaction between THC and CBD¹, in the sense that CBD would mitigate the effects of THC, can be clearly rebutted. First and foremost, the underlying risk assessment in our expert opinions is based on the opinion of the Panel on Contaminants in the Food Chain of the European Food Safety Authority (EFSA)²⁷, which has considered interaction effects. However, EFSA concluded the information is controversial and not consistently antagonistic²⁷. This is consistent with more recent research of Solowij et al.²⁸ that the effects of Δ⁹-THC may even be enhanced by low-dose CBD (e.g., as found in food supplements) and may be particularly prominent in infrequent cannabis users. Positive findings regarding antagonistic effects (e.g. Pisanti et al.²⁹ cited by EIHA) were typically found for much higher dosing regimens, i.e. aiming to mitigate the adverse effects of THC in hashish and marihuana, while another study with smoked cannabis did not detect such an effect³⁰.

We strongly believe, in line with EFSA, that the current scientific evidence does not allow for considering cumulative effects in low dose CBD oils and hemp extracts. The applicability and appropriateness of the acute reference dose (ARfD) of 1 μg Δ⁹-THC per kg body weight – without considering interactions by CBD – was recently re-confirmed by EFSA³¹.

As the EIHA mentioned this argument, we have decided to include a short rationale into the v3 of our article¹³ for reasons of completeness. Otherwise, our article is not a basic toxicological research article about the rationale for risk assessment but an applied research article, which has based the risk assessment on the guidelines of the responsible risk assessment authorities BfR²⁰ and EFSA^27,31. Therefore, we would invite EIHA to correspond directly with these institutions, when they believe there is scientific evidence or new data that might change the available assessments. Currently, we see no such data. It should be noted that the EIHA has unsuccessfully tried lobbying the risk assessment bodies into providing more “reasonable” guidance values for THC (e.g., see Banas et al.³²), and we believe that a comment on our scientific article is not the right place to continue this effort.

Illegality of all hemp products containing isolated CBD or hemp extracts

While the regulatory status is not part of our chemical and toxicological research, we thank the EIHA for pointing out this issue, as there is a potential misunderstanding of the lobbyist regarding the most up-to-date regulations and decisions of EU and national legislators as well as of the courts, which is evidenced by the outdated references cited by EIHA¹.

We also thank the EIHA¹ for the re-iteration of our conclusion that “basically all available CBD products based on hemp extract marketed as food or food supplement within the EU are therefore illegally sold”. We still stand by this conclusion.

It is certainly true that case-by-case decisions have to be conducted in official food control, and of course we have exactly done this for each product, which was submitted to our laboratory for evaluation. However, the situation of hemp-extracts is a particular one, because of its regulatory status as unapproved novel food. This status allows for such a blanket statement, that each single product that contains hemp extract as ingredient can be judged as illegally placed on the market. It should be noted that this assessment is independent of the amount of hemp extract or its concentration of CBD. Regarding the THC levels found, which are widely variable, a case by case decision has to be made in any case, which spans from unsuspicious levels below the ARfD up to exceedance of the LOAEL dose, which we judge as a serious risk in consideration of Art. 14 of Regulation (EC) No. 178/2002¹⁹.

The EIHA¹ is also correct in considering the EU Novel Food Catalogue, which leads to this “blanket” assessment of hemp extracts as being novel, as legally not binding and that it is only an indicator for court decisions. What the EIHA, however, fails to mention is the fact that there are a number of court decisions that have actually endorsed the suggestions of the novel food catalogue and have confirmed the actions of the authorities in prohibiting the placing of the respective CBD product on the market^{25,26,33–39}. To our knowledge, there currently is no court ruling, that might have endorsed the EIHA opinion.

Furthermore, the court rulings have also disproved the claims of the EIHA about the burden of proof for determining the novelty of a food. The opinion of EIHA¹ in this regard is based on outdated, incomplete evidence. In their decision about the marketability of a CBD product, the administrative court of the German Federal State Baden-Württemberg ruled that the food business operator has the burden of proof²⁶. This is in accordance with Article 4(1) of Regulation (EU) No 2015/2283, which states that the food business operator shall verify that foods which he or she has placed on the market in the EU, fall within the scope of this Regulation or not⁴⁰. Also outside the CBD field, the burden of proof has been imposed on the FBOs in several court rulings confirming Art. 4(1) of Regulation (EU) No 2015/2283 (see review of court rulings in Meyer et al.⁴¹). For a more detailed assessment of CBD court rulings see our recent review⁴².

Finally, we cannot follow the arguments of the EIHA¹ that European Court of Justice decisions regarding pharmacological effects might be relevant or that the novelty of a product is connected with an associated abstract health risk. The novelty of a product purely depends on the fact that it was not used for human consumption to a significant degree within the EU before 15 May 1997⁴⁰. The novelty does not depend on potential pharmacological effects or health risks of the product.

The German Federal Office of Consumer Protection and Food Safety (BVL) recently published a statement that the classification of food containing CBD in the press release of EIHA of March 3, 2020, is not correct⁴³. The BVL states that for extracts of Cannabis sativa L. and derived products containing cannabinoids (e.g. CBD) a significant history of consumption in the EU has still not been demonstrated by the economic operators, nor by the EIHA or any other association⁴³. For this reason, they are still considered EU-wide as novel foods⁴³.

In conclusion, we believe that the responsible authority can currently make conclusions on the non-marketability of CBD products based on a lack of novel food approval, and additionally based on the lack of safety when THC thresholds are exceeded. We must stress here that the responsible local authority’s tasks clearly include the enforcement of the Novel Food Regulation⁴⁰ as well of the food safety rules¹⁹. This is practiced all over Europe and evidenced by the numerous alerts found in the EU’s Rapid Alert System for Food and Feed (RASFF)⁴⁴.

Judgement about food producers of CBD products

Since the publication of our article², a number of studies have confirmed our analytical results. Food control authorities in Europe have reported various offences of FBOs selling CBD products against the European food law. More than 160 notifications regarding CBD as unauthorised novel food ingredient and/or unauthorised THC in CBD products were shared in the RASFF. In Belgium, about half of 213 products seized from CBD shops exceeded a threshold of 0.2% THC+THC-A and large discrepancies were observed between labelled and measured CBD concentration⁴⁵. The Food Safety Authority of Ireland (FSAI) reported that from 38 tested CBD products, 37% exceeded the safe limit of THC dosage set by EFSA (1 µg/kg body weight/day), 34% were classified as novel food lacking approval, 36% were food supplements lacking the necessary notification of the competent authority, 92% were tested to contain differences between analytical and declared CBD content of more than 10%, and finally 50% contained misleading claims such as unauthorised health claims or medicinal claims⁴⁶. An analysis of over-the-counter CBD products from the UK found that only 38% of 29 products were within 10% of advertised CBD content and 55% had measurable levels of THC or cannabinol⁴⁷. Similarly, only 3 out of 25 CBD products from the State of Mississippi (USA) were within 20% of label claim, and 3 exceeded 0.3% THC⁴⁸. Similar studies from Italy⁴⁹, the Netherlands⁵⁰, and the USA⁵¹ are available.

In consideration of these consistent reports worldwide, we actually cannot find a better wording than our original statement: “In our opinion the systematically high Δ⁹-THC content of CBD products is clearly a “scandal” on the food market. Obviously, the manufacturers have – deliberately or in complete ignorance of the legal situation – placed unsafe and unapproved products on the market and thus exposed the consumer to an actually avoidable risk.”

We fully stand by this conclusion and have even expanded our judgement of the CBD industry in a recent editorial, which concluded that the illegal market of CBD products may provide a strong rationale for the necessity of a paradigm shift towards pre-marketing approval in regulating food supplements³.

The following arguments of EIHA¹, starting with obsolete letters of the EU commission (written at a time when hemp extracts were not available on the market, highlighting their irrelevance to the current situation) and some disconnected information about novel food status, without providing any evidence at all besides unsubstantiated claims, cannot plausibly refute our conclusions. Instead, we have provided ample evidence – based on EFSA criteria²⁷ – that a substantial number of CBD products on the market is not safe (69% of samples above ARfD of EFSA) and all samples (100%) were judged on a case-by-case basis as unapproved novel foods. Additionally, all samples (100%) were non-compliant with mandatory labelling rules and/or used unapproved health claims². We feel that this is ample proof for our statement above, which is based on facts.

The CVUA Karlsruhe as part of the food control system in the EU also clearly wishes to reject the allegation of EIHA¹ that the institute conducts “discrimination”, “undifferentiated action” and “arbitrariness”. We have assessed all products sent to our institute for evaluation in a transparent and consistent fashion (the criteria for evaluation were published in 2019⁵²), conducted our toxicological and regulatory assessment on a case-by-case basis², and even allowed public scrutiny by publishing our full dataset²³.

Regarding the concerns of EIHA to defend the reputable hemp industry against “free riders”, “black sheep” or “cowboys”, we can ensure them that food control includes this segment of the market as well, e.g. by conducting sampling of online stores. Otherwise, the EIHA has the possibility to take their own steps against such practices on the basis of the national laws against unfair commercial practices (e.g. in Germany “Gesetz gegen den unlauteren Wettbewerb (UWG)”).

The closing remarks of EIHA¹ in this section appear ill-considered. First, it is commendable that EIHA wants to ensure compliance with the law and consumer safety. But how can this solely be achieved by an industry standard? And how can an industry standard be made mandatory for all FBOs? Perhaps on a voluntary basis for the members of EIHA, but clearly not for the whole industry, and not for the “free riders”, “black sheep” or “cowboys”. As stated before, we would certainly agree with an improved legal basis for hemp food products similar to other vertical regulations in the food sector, such as the EU spirit drinks regulation. However, we fail to see how this can be achieved as an industry standard.

Regarding the lack of communication between EIHA and public authorities, we recall a technical discussion at our institute at the end of 2018 and are also aware that the EIHA was invited to present their evidence at the “Working Group Novel Food” in Brussels⁵³.

Finally, we congratulate the EIHA for the decision to facilitate novel food applications by conducting extensive toxicology studies.

Judgement of the hemp industry in the food sector

The quote "Currently CBD users must be aware that they may be ‘participating in one of the largest uncontrolled clinical trials in history’” of Pál Pacher included in a Newsweek article⁵⁴ is in our opinion very fitting to the reality of the market. First, Pál Pacher is clearly an authority regarding cannabis research (e.g., Refs.^55–58). Second, the comment is regarding CBD and not regarding THC, and we currently cannot see a substantial difference between CBD content of food or nutritional supplements on the markets in the USA and Europe. Along with the lack of labelling detected in our study and the suggestions of many manufacturers to “gradually increase the dosage”, pharmacologically active CBD dosages similar to prescription medications may be easily reached by commercial over-the-counter CBD products on the market in Europe. As noted above, no nutrivigilance is typically conducted and no safety assessment has been conducted for the products, because the manufacturers put them on the market before achieving novel food approval. Nevertheless, according to the peer reviewer’s suggestion, we have removed the statement from article version 4; v4¹⁴).

Proposal of a legal ban on hemp extracts

We would like to note that we actually have suggested a regulated legalization of CBD products. Therefore, we question how or why the EIHA is interpreting this as the proposal of a “ban”. We also wonder why our statement “For cannabis-derived products, such as CBD, the problem is aggravated by conflicting regulations in the narcotic, medicinal, and food law areas. For example, hemp extract based products of similar composition could be treated as illegal narcotics, prescription-based medicinal products, or novel foods” is criticized by the EIHA¹, when they actually provide supporting evidence with their examples of melatonin or garlic that certain substances could fall into either legal realm depending on labelling sometimes even when the concentration is similar (e.g., also compare sage tea⁵⁹ or Ginkgo biloba extract⁶⁰).

Our statement also has been validated by an assessment of the European Commission (according to press information⁷) to either consider hemp extracts as narcotics or as foods. As detailed elsewhere^3,42, we believe that it would be disproportional to regulate CBD products as narcotic drug according to the principle of "ultima ratio" in criminal law. Considering a decision by the Court of Justice of the European Union, the European Commission recently clarified its position to not further consider CBD as narcotic, but to advance the novel food approval procedure¹⁴.

Conclusions

We hope our response informs the F1000Research readership about the most recent evidence regarding the toxicological and regulatory evaluation of CBD products. We believe that the Correspondence article of the EIHA¹ has made many unsubstantiated claims and is unable to discredit our scientific work that was based on a validated and externally accredited analytical method² with fully transparent criteria for risk assessment based on BfR²⁰ and EFSA²⁷.

We hope that the promised extensive toxicological studies and quality standards of EIHA will include the following research questions:

Data availability

All data underlying the results are available as part of the article and no additional source data are required.