Emerging therapy

Cannabidiol and cannabis-based products

CBD, purified cannabidiol (Epidiolex / Epidyolex), CBD oils, THC, nabiximols

The one product in this group that became a real medicine. Licensed for three named epilepsy syndromes, and supported beyond them by a large body of real-world evidence in other drug-resistant epilepsies. The oil sold in a shop and the licensed solution are still not the same thing.

Where it stands, condition by condition

The same therapy can be well supported for one problem and completely untested for another. This is the single most common place families are misled.

What it is

Three quite different things travel under this heading, and the differences decide everything.

Purified pharmaceutical cannabidiol is a licensed prescription medicine: a standardised oral solution, with a known concentration, a regulated manufacturing chain and approved indications. This is the one with randomised trial evidence.

CBD-rich plant extracts contain cannabidiol alongside other cannabinoids, often including some THC. Composition varies between batches and producers, and the “entourage effect” argument for whole-plant preparations is asserted far more often than it is demonstrated.

Over-the-counter CBD oils are sold as food supplements. They are not assessed for efficacy, the labelled content is frequently wrong, and no one is responsible for what is in the bottle.

THC is the psychoactive component and belongs in a separate conversation. It is not part of the licensed paediatric product, and concerns about effects on a developing brain apply to it, not to purified cannabidiol.

Legal status and availability differ substantially between countries, and in many places access runs through a named-patient or import route rather than an ordinary prescription. Check what applies where you live before planning anything.

How it is meant to work

Cannabidiol does not work the way people assume. It has little direct activity at the CB1 receptor — the one THC acts on — which is why it is not intoxicating.

Several mechanisms have been proposed instead: antagonism at GPR55, desensitisation of TRPV1 channels, and effects on adenosine reuptake, among others. None has been established as the main one.

This is an unusual and honest situation: the effect was demonstrated in trials before the mechanism was understood. That is the opposite of most treatments on this site, where the mechanism is attractive and the clinical effect is missing.

What has actually been tested

• The licensed indications rest on randomised, placebo-controlled trials published in major journals — in Dravet syndrome, in Lennox-Gastaut syndrome and in tuberous sclerosis complex. This is the standard of evidence that nothing else on this page comes close to.
• The size of the effect should be understood plainly. A median reduction of roughly 39% against 13% on placebo means fewer seizures for many children, not freedom from seizures for most.
• One genuine scientific argument deserves mentioning: cannabidiol raises blood levels of the active metabolite of clobazam, and part of the benefit in trials where most children were taking clobazam may come from that interaction. Analyses of children not on clobazam still suggest an independent effect, but the debate is not fully closed.
• Outside the three licensed syndromes the design of the evidence changes — open-label programmes, registries and multicentre retrospective cohorts rather than randomised trials — but the volume is considerable and the direction is consistent. Outside epilepsy it does drop sharply: for autism and for cerebral palsy spasticity, what exists is small, mixed and preliminary.

What we still do not know

• Whether benefit extends to drug-resistant epilepsies beyond the three licensed syndromes, and to which ones.
• Long-term effects of years of exposure on a developing brain — the licensed product is recent and the children treated first are still young.
• Whether whole-plant extracts add anything over purified cannabidiol, or only add variability.
• The right dose and the right point to stop in a child who has not responded.

Risks and costs

• Common and dose-related: sleepiness, reduced appetite, diarrhoea and tiredness. These are frequently why a dose has to be lowered.
• Liver enzymes. Raised transaminases occur, particularly in children also taking valproate. Liver function needs checking before starting and during treatment — this is a monitoring requirement, not a theoretical caution.
• Interactions. Beyond clobazam and valproate, cannabidiol affects several liver enzymes and can change the levels of other medicines. Every child's full medication list needs reviewing before starting.
• For unregulated oils, the risks are different and harder to manage: an unknown dose, possible THC content, contamination, and interactions that nobody has checked because nobody knows what is in the bottle.
• Cost and access. The licensed product is expensive, and in countries where it is not routinely reimbursed families sometimes substitute a shop-bought oil — which is not the same medicine at a lower price.

Questions to ask before you agree

1. Is this the licensed, standardised pharmaceutical product, or a plant extract or supplement?
2. Which indication is it being given for, and is that one of the licensed ones?
3. Have my child's liver tests been done, and when will they be repeated?
4. Has someone reviewed the interaction with every medicine my child takes, clobazam and valproate in particular?
5. What reduction are we aiming for, over what period, and at what point do we conclude it has not worked?

More in this section