For the past 4 years, DevelRx and The Canna Consultants have worked in collaboration with the goal of achieving an objective and scientifically valid assessment the safety risks posed by the introduction of CBD as a novel food or food supplement.
Our motivation for responding to Dr Cristelle Santos’ recent article entitled “Does CBD really harm the liver?” is to correct the erroneous proposal that the way to address the safety of CBD is to generate yet more unnecessary and misrepresentative evidence in animal experiments.
When evaluating the safety risks posed novel compound for human use, the role of investigational toxicology is to identify a dose to initiate safety and tolerability testing in humans. This rationale applies irrespective of the intended use, e.g. as a novel food, nutraceutical or pharmaceutical. The explanation is safety pharmacology and toxicity studies in animals yield only predictions about safety in humans. Once an initial safe dose of the novel compound has been identified, the compound’s safety and tolerability can be definitively established in human subjects. As an example, safety pharmacology studies in animals predicted that CBD would not be abused by humans (Gray et al, J Pharmacol Exp Ther 382:54–65, July 2022), but definitive evidence of CBD’s lack of abuse potential came from a study in drug-experienced human volunteers (Schoedel et al, Epilepsy Behavior, 88:162-171, Nov. 2018).
Because Epidyolex (>98% purity CBD) was developed prevent seizures in children, the safety data generated in humans has been largely disregarded as irrelevant because CBD as Epidyolex is a medicine, it is not a novel food. What this argument fails to acknowledge is liver toxicity is the same adverse event irrespective of whether CBD is used as a drug to treat epilepsy or as a novel food. The key difference is what constitutes an acceptable level of benefit versus risk. No one would argue that the risk of liver toxicity when ingesting CBD as a novel food should not only be many times lower than when it is used as an anti-seizure drug, it should be negligible.
The key factor in the case of CBD is liver toxicity was observed as a drug-associated adverse event in clinical trials, and therefore, the threshold dose for this safety signal could be calculated based on data obtained from rigorous, placebo-controlled trials in humans.
Dr Santos states no effects of CBD occur in adult humans at doses that are multiples of the original FSA recommended daily dose of 70 mg (1 mg/kg for a 70 kg adult), although the FSA has now indicated it favours at a lower safe level of 10 mg/day. More recent data obtained in normal healthy subjects confirms no liver damage at doses that are multiples of the FSA recommended original daily dose of 70 mg/day and the author acknowledges that the co administration of CBD with other anticonvulsants that are known to cause liver damage is a confounder that unfairly points the finger of blame at CBD.
When we evaluated the safety of CBD as a novel food, we relied on results published in the European Medicines Agency (EMA) – Marketing Authorisation Application for Epidyolex (CBD) showing that daily administration of CBD as monotherapy, i.e. without the confounder of taking other co-prescribed anticonvulsant drugs, at a dose 10 mg/kg/day produced no liver toxicity signal in humans (Figure 34 [page 161], EMA Assessment report: Epidyolex). On this point, we in agree with Dr Santos’ interpretation in the human safety part of her article.
Where we disagree:
(i) If you already have information on liver damage from well controlled trials in human subjects (i.e. with Epidyolex), it is unnecessary and unethical to conduct scientific research in animal toxicity studies to obfuscate what is already known, i.e. CBD at doses that are multiples of the 70 mg daily intake, which the FSA initially recommended, do not produce liver damage in humans;
(ii) Citing misleading findings about the liver damage risk posed by CBD based on data obtained when it was co-administered with drugs known to produce liver toxicity. Table 44 (page 160 of the EMA Assessment report: Epidyolex) reports increased ALT (alanine aminotransferase) >3x ULN (upper limit of normal) occurred in 4.7% of subjects with normal liver function taking CBD 20 mg/kg/day without sodium valproate but in 34.9% when taken with sodium valproate. Similarly, Table 46 (page 161 of the EMA Assessment report: Epidyolex) reported 1.5% liver function serious adverse events (SAEs) and 2.2% drug discontinuations in patients taking CBD 20 mg/kg/day without clobazam, but 5.6% SAEs and 7.5% drug discontinuations when taken in combination with clobazam; and,
(iii) It is not appropriate to refer to REACH guidelines which are designed to evaluate the toxicity of chemicals:
“UK REACH is a regulation that applies to the majority of chemical substances that are manufactured in or imported into Great Britain. This can be:
a) A substance on its own;
b) A substance in a mixture, for example ink or paint; and,
c) A substance that makes up an ‘article’ – an object that is produced with a special shape, surface or design, for example a car, furniture or clothes.”
(iv) This approach is compounded by applying the REACH derived no-effect level (DNEL) from liver toxicity data in rats to calculate what is the safe dose in humans. The reason why multiple correction factors are applied (i.e. allometric scaling, interspecies variability and exposure duration) is because observing what happens in rats and applying it to predict what might happen in humans is fraught with uncertainty. Hence, numerous safety factors are built into the calculation to ensure that the DNEL in humans is actually safe. If enough safety factors are built when calculating the DNEL, it could result in a situation where homeopathic amounts of CBD are achieved in the formulation, resulting in no safety risk (but no potential benefit either).
Our overall view:
(i) Humans are the intended species irrespective of whether CBD is being used as a novel food, nutraceutical or as a prescription pharmaceutical;
(ii) After the safety signal of CBD had been identified and assessed in clinical trials in humans, one should not be proposing a return to more experiments in rodents to predict the magnitude of this risk, or to set safe amounts of CBD for daily intake. If there are safety concerns, the appropriate action is to answer those questions with studies in humans;
(iii) It is not relevant or appropriate to consult the REACH protocol that was designed to look at the toxicity of chemicals and environmental toxins to evaluate the toxicity of CBD, which has undergone a full safety analysis in animals and humans as part of its registration as a medication (Epidyolex) to treat epilepsy in children. The relevant approach is to identify a tolerated and safe daily dose of CBD in clinical studies and allow a fraction of that dose for use as a food supplement. For example: if 10 mg/kg/day CBD produced no liver toxicity in clinical trials (Epidyolex EMA dossier); one tenth of that dose should be safe as a food supplement;
(iv) It is unacceptable to make predictions about CBD safety when there are known confounders, e.g. don’t cast doubt on the safety of CBD based on findings when it was co-administered in humans with another drug that is a known liver toxin, e.g. sodium valproate; and,
(v) It is inappropriate and unethical to suggest more toxicity studies should be conducted using higher doses of CBD in rats (or other animal species) to generate more unreliable predictions about safe daily doses to be consumed by humans.
In summary, the toxicity risks of CBD were thoroughly investigated in animals as an intrinsic part of its safety evaluation for approval a prescription pharmaceutical. Additional rodent toxicity investigations have been demanded by the FSA and conducted by sponsors as part of the evaluation of CBD as a novel food. In our opinion, far too many animals have already been sacrificed to confirm what is already known if one is prepared to search for the published evidence. The last thing that is needed is an exhortation to conduct more animal research to answer this particular question.
Dr Santos quotes Paracelsus: “The dose makes the poison”, which is an apposite observation in this case. To this we would add the rider: “Don’t look to animals to offer predictions about the safety of CBD as a novel food when the definitive answers have already been generated by studies in human subjects .”
EMA Assessment report: Epidyolex [International non-proprietary name: cannabidiol]
Professor David J Heal PhD, DSc, FRSC, HonFBPhS (DevelRx)
Dr Sharon L Smith PhD (DevelRx)
Mr Stephen Oliver LLB (The Canna Consultants)
Mr Matthew Lawson LLB (The Canna Consultants)
DevelRx’s involvement with CBD is based on several years researching into its safety to support CBD’s approval as a novel drug to treat two rare forms of childhood epilepsy (see Gray et al, J Pharmacol Exp Ther 382:54–65, July 2022). Our careers for more than two decades have been in pharmaceutical R&D supporting the discovery, development and approval of safer and more effective drugs to address unmet clinical needs in psychiatry, neurology and metabolic disorders. As researchers in this field, we understand that experiments in animals are essential to achieve these objectives, but we are also committed to the 3Rs in animal research, viz Refinement, Reduction, and Replacement.
