Purified cannabidiol is an approved add-on medicine for three rare childhood epilepsy syndromes, but several important questions about it are still open. Researchers disagree on how much of its effect comes from raising the level of another anti-seizure drug. They also disagree on whether whole-plant extracts work better. Unregulated cannabidiol oils are a separate problem, because tests show that many do not contain what the label says.
This part also tells the Charlotte Figi story accurately and lists what is still unproven. It then explains the legal route and the Thai clinical data for families in Thailand and Southeast Asia.
A short recap for readers starting here. The approved product is a purified cannabidiol oral solution. It is sold as Epidiolex in the United States and as Epidyolex in Europe and Australia. It was made by GW Pharmaceuticals, which is now part of Jazz Pharmaceuticals. It is approved for Dravet syndrome, Lennox-Gastaut syndrome and tuberous sclerosis complex, always on top of existing medicines. In the main trials, seizures fell about 20 to 25 percentage points more than on placebo. Common side effects were sleepiness, poor appetite, diarrhoea and rises in liver enzymes.
This is Part 2. Part 1: Cannabis and Epilepsy: What Cannabidiol Actually Does, and What It Does Not
Translator’s note: A family may ask you, as a grower, for cannabis oil for a child with seizures. If so, read the sections on unregulated products and on Thailand first.
1. How Much of the Benefit Came From Clobazam?
Clobazam is a benzodiazepine, a calming and sedating medicine that is widely used in these syndromes. Cannabidiol slows the liver enzyme that clears the active form of clobazam. As a result, that active form rises roughly 2.5 to 3 times in the blood. The open question is whether part of the seizure benefit comes from this extra clobazam.
In 2019 Rogawski, writing in the journal Epilepsy and Behavior, pointed to open-label data. In open-label studies, patients and doctors know what is being given, and there is no placebo group. In those data, 51 per cent of patients taking clobazam were responders, meaning their seizures fell by at least half. Among patients not taking clobazam, 27 per cent were responders. Rogawski argued that controlled trials were needed to show whether cannabidiol has any anti-seizure effect of its own.
The European licence leaves this question open. For Dravet and Lennox-Gastaut syndromes, it approves cannabidiol only together with clobazam, which is the combination the trials actually tested.
On the other side, Devinsky and colleagues, writing in Acta Neurologica Scandinavica in 2020, pooled four randomised trials covering 714 patients. They found a significant fall in seizures against placebo, both with and without clobazam.
These data cannot settle the question either way. Both sides rely on subgroup analyses, which compare groups of patients inside trials that were not designed to make that comparison.
If the child already takes clobazam, ask the neurologist how they will tell whether sleepiness comes from cannabidiol or from the higher clobazam level.
2. Purified Cannabidiol Against Whole-Plant Extract
The claim that “the whole plant works better” is often linked to the entourage effect. It usually comes from one paper, by Pamplona, da Silva and Coan, in Frontiers in Neurology in 2018. The paper was a meta-analysis, which pools the results of earlier studies, and every study it pooled was observational.
| Measure | Cannabidiol-rich extracts | Purified cannabidiol |
|---|---|---|
| Any improvement at all (loosest measure) | 71% | 46% |
| Seizures down by at least half | 37% | 42% |
| Average dose (mg per kg body weight per day) | about 6 | about 25 |
On the loosest measure, extracts looked better. At the stricter measure of a 50 per cent fall in seizures, the difference disappeared.
Why the comparison is not settled
Every included study was unblinded, meaning patients and families knew what they were taking. None had a placebo group. The authors themselves wrote that it is impossible to measure how much of the response was a placebo effect.
The doses were also very different. Any difference could come from the type of product or from the dose, and the studies cannot separate the two. Extracts looked better only on the loose measure, which expectation affects most. They did not look better on the strict measure. No randomised trial has compared the two products directly.
When a seller or a forum post claims that a full-spectrum oil works better than the approved medicine, ask for a randomised trial that compared the two directly. No such trial exists yet.
3. Unregulated Cannabidiol Products Are a Different Category
In 2017 a team led by Bonn-Miller, publishing in the Journal of the American Medical Association, tested 84 cannabidiol products bought from 31 online companies.
- About 30 per cent contained an amount of cannabidiol within 10 per cent of the label.
- About 42 per cent contained more than the label claimed.
- About 26 per cent contained less.
- A meaningful minority of products contained tetrahydrocannabinol, the compound that makes users high. That is serious when the user is a child.
Contamination and fake products
Between October 2017 and January 2018, public health authorities in Utah, a state in the United States, investigated 52 cases of acute poisoning. The products were sold as cannabidiol but actually contained a synthetic cannabinoid. This is a man-made chemical designed to act like tetrahydrocannabinol. Around 60 per cent of the people affected were seen at a hospital emergency department.
Their symptoms included confusion or reduced awareness, vomiting, and seizures or shaking. A product marketed to control seizures had caused them.
Other risks are pesticide residues, solvent residues and heavy metals. Products can also be contaminated by microbes such as mould and bacteria.
A wellness oil is not a cheaper version of the approved medicine. The epilepsy evidence applies to the regulated product, which is made to a fixed standard. It does not apply to every product that contains cannabidiol.
Translator’s note: A certificate of analysis is a laboratory report on one batch of a product. It shows the cannabinoid content and the contaminant test results. It is useful for checking a product, but it does not make a market oil equivalent to the medicine tested in the trials. Growers should also know that residues of any spray used on the plants can carry over into an oil made from them. Use only pesticides registered with the Department of Agriculture (กรมวิชาการเกษตร), and only as the label directs; see what you may legally spray. In the Thai rainy season, high humidity also raises the risk of mould toxins on flower.
Do not give a child any cannabidiol or cannabis product that a doctor has not prescribed. If the child already uses one, tell the neurologist and bring the bottle and its label to the appointment.
4. The Charlotte Figi Story, Told Accurately
Charlotte Figi was born in Colorado, a state in the United States, on 18 October 2006. She had her first seizure at around three months old and was later diagnosed with Dravet syndrome. By the age of five, many medicines had failed. She was reported to have up to 300 convulsive seizures a week.
In 2012 her family obtained a cannabis extract that was low in tetrahydrocannabinol and high in cannabidiol. It came from a cannabis variety later renamed Charlotte’s Web. The family reported a dramatic fall in her seizures. In 2013 her story appeared in Weed, a documentary by Sanjay Gupta on the US news channel CNN.
Charlotte died on 7 April 2020, aged 13. Pneumonia triggered seizures, and she died of respiratory failure and cardiac arrest.
What her case showed, and what it did not
Her case changed policy quickly. Families moved to Colorado to get the same extract. Several jurisdictions passed laws giving access to cannabidiol specifically. The public pressure that followed made it possible to fund the formal trials described in Part 1.
Her case showed two things. One child improved dramatically after starting a cannabidiol-rich extract. And a single case can change the law faster than ten years of committee work.
Her case did not show the average effect of cannabidiol, because one case cannot show that. It did not show that a full-spectrum extract works better than a purified one, because no comparison was made. It did not establish any dose or formulation. None of this criticises the Figi family, who never claimed otherwise. The criticism is of how others later used the story. For the general reasons, see why personal testimony is not medical evidence.
When a post shares one child’s recovery, take it as a reason to talk to a neurologist. Do not treat it as proof or as a dosing guide.
5. What Is Still Not Established
Adult focal epilepsy
In focal epilepsy, seizures start in one area of the brain. It is the most common form of drug-resistant epilepsy in adults. Cannabidiol has no comparable body of evidence for it.
A randomised trial published in JAMA Network Open in 2022 tested cannabidiol applied to the skin (transdermal) in adults with focal epilepsy. Patients used it on top of their usual medicines. The trial found no difference from placebo:
| Group | Seizures per 28 days | P value against placebo |
|---|---|---|
| Lower dose | 2.51 | 0.89 |
| Higher dose | 2.59 | 0.32 |
| Placebo | 2.49 | — |
The open-label extension of the same trial looked encouraging. By month six, 60.8 per cent of patients had their seizures cut by at least half. The extension had no placebo group, and it suggested a benefit that the controlled trial did not find. This shows why an extension cannot replace a placebo group.
Tetrahydrocannabinol
No randomised trial shows that tetrahydrocannabinol reduces seizures. It also carries its own risks for a developing brain.
Long-term outcomes
Open-label extensions give the longest view available, but they have no placebo group. In the three-year Dravet extension of 315 patients, the median time on treatment was 444 days. Nine per cent of patients stopped because of side effects.
Side effects were reported by 97 per cent of patients. The most common were diarrhoea (43 per cent) and fever (39 per cent), followed by decreased appetite. The effects on thinking, development and behaviour over a whole childhood are still unknown.
No placebo-controlled trial supports cannabidiol for an adult with focal epilepsy. None supports a product that contains tetrahydrocannabinol either. In these cases, discuss the standard treatment options with the neurologist first.
6. What Families in Thailand and Southeast Asia Can Actually Do
The legal route in Thailand
Thailand’s rules changed sharply in 2025. The Notification of the Ministry of Public Health Re: Controlled Herbs (Cannabis), B.E. 2568, was signed on 23 June 2025. It has been in force since 26 June 2025. It made cannabis flower a controlled herb under the Thai Traditional Medicine Protection Act B.E. 2542.
Flower can now be supplied only on prescription. This needs an official form, the PT 33 (ภ.ท.33) prescription, issued by an authorised practitioner. A prescription covers at most 30 days.
Advertising and online sale of flower are prohibited. That is why no legitimate Thai source markets a cannabis product as a treatment for a child’s epilepsy.
Selling or processing a controlled herb without a licence carries up to one year in prison, a fine of up to 20,000 baht, or both. This penalty is in Section 78 of the Thai Traditional Medicine Protection Act B.E. 2542. For a family, the route is therefore the ordinary medical one: a paediatric neurologist, a documented diagnosis, and a prescription inside the licensed system.
Translator’s note: Cannabis oils are extracts, and a separate rule now covers them. The Ministerial Regulation on Cannabis or Hemp Extracts B.E. 2569 has been in force since 26 April 2026. Under it, only three kinds of body can be licensed for extracts with more than 0.2% tetrahydrocannabinol:
- Thai juristic persons that are not “foreign” under the Foreign Business Act B.E. 2542
- state agencies
- the Thai Red Cross Society
See the two 2026 ministerial regulations. A grower who makes oil from flower for a family without a licence is processing and selling a controlled herb without a licence.
What the Thai clinical data show
Thailand’s Department of Medical Services runs two centres that treated 14 children with drug-resistant epilepsy: the Neurological Institute of Thailand and the Queen Sirikit National Institute of Child Health. They used a product with at least 20 parts cannabidiol to one part tetrahydrocannabinol. The median dose, meaning the middle value, was 5.6 mg/kg/day.
Nine children continued for a median of 18 months. Among them, 50 per cent had their convulsive seizures cut by at least half. For all seizure types combined, the figure was 43 per cent.
Side effects were common:
- sleepiness: 64 per cent
- worsening seizures: 43 per cent
- irritability: 42 per cent
- reduced appetite: 28 per cent
Five of the 14 children (35 per cent) stopped treatment, most of them in the first three months because their seizures got worse. One child had hepatitis (liver inflammation). The study was uncontrolled and had no placebo group. It therefore cannot show how much of the benefit came from the product.
If your child is treated in Thailand, ask the neurologist whether this kind of product is an option. Also ask which side effects to watch for and which blood tests the child will need.
Cost, supply and crossing borders
The approved medicine is expensive and is not registered in every country in the region. Ask the neurologist whether it can be obtained where you live.
Where the approved medicine is not available, families may be offered locally made cannabidiol-enriched preparations. These are not the product studied in the main trials, even when they contain the same molecule. Their concentration, contaminants, batch-to-batch consistency and tetrahydrocannabinol content all differ. The trial evidence therefore does not apply to them.
Do not carry cannabis across borders in this region. Several Southeast Asian countries impose severe criminal penalties for possessing or importing it. For details, see cannabis laws across Asia.
Translator’s note: Families can take these practical steps:
- Keep a seizure diary. Record the date, time, type and length of each seizure for at least four weeks, as the trials did for their baseline. The neurologist can then judge whether a treatment change is working.
- Keep one medicine list. Include every prescription, supplement and oil, so the neurologist can check for interactions.
- Keep copies of liver blood tests. Rising results are easier to spot when they sit side by side.
- Store medicine as its label says. Thai homes often reach 30–35 °C, so do not leave medicine in a parked car or in direct sun.
If you are a grower and a family asks you for oil, explain the route above and refer them to a paediatric neurologist. Do not supply flower or oil yourself.
Share your experience
Your family may have been through this: the diagnosis, the waiting lists, the seizure diaries, or the decision about whether to try one more medicine. Your experience helps other families here more than any summary of trial data. Please reply below and describe the route to treatment where you live and what monitoring it involved. Share what you lived through, and do not recommend products. Every decision about starting, stopping, adjusting or replacing an anti-seizure medicine belongs to the child’s treating neurologist. A forum thread cannot make that decision.
References for Part 2
- Rogawski (2019). On cannabidiol, clobazam and the open-label responder data. Epilepsy and Behavior.
- Devinsky et al. (2020). Pooled analysis of four randomised trials (714 patients), with and without clobazam. Acta Neurologica Scandinavica.
- Pamplona, da Silva and Coan (2018). Meta-analysis of observational data, cannabidiol-rich extracts versus purified cannabidiol. Frontiers in Neurology.
- Bonn-Miller et al. (2017). Labelling accuracy of 84 cannabidiol products from 31 online companies. Journal of the American Medical Association.
- Utah public health investigation (October 2017 – January 2018): 52 acute poisonings from products sold as cannabidiol that contained a synthetic cannabinoid.
- O’Brien et al. (2022). Randomised trial of transdermal cannabidiol in adults with focal epilepsy, and its open-label extension. JAMA Network Open.
- Scheffer et al. (2021). Three-year open-label extension in Dravet syndrome (315 patients). Epilepsia.
- Lusawat et al. (2021), Journal of the Department of Medical Services 46(3). Neurological Institute of Thailand and Queen Sirikit National Institute of Child Health (Department of Medical Services): 14 children with drug-resistant epilepsy treated with a product of at least 20:1 cannabidiol to tetrahydrocannabinol.
- Sanjay Gupta, Weed (CNN documentary, 2013).
- Notification of the Ministry of Public Health Re: Controlled Herbs (Cannabis), B.E. 2568 (2025).
- Thai Traditional Medicine Protection Act B.E. 2542 (1999), Section 78.
- Ministerial Regulation on Cannabis or Hemp Extracts B.E. 2569 (2026).
This article is part of the Asiannabis Community educational series on medical cannabis. It is general information, not medical advice. Treatment decisions belong with a licensed practitioner. Content is for educational purposes within jurisdictions where medical cannabis is legally permitted.