What does the research say?
Research on these three syndromes paints a similar picture: almost all studies with cannabinoids are very small, many were stopped early, and some of their results were published not in a peer-reviewed journal but only on the international trial registry website (ClinicalTrials.gov). In our October 2026 search, we found four studies in these syndromes in which participants were assigned to groups at random; all four were stopped early, and none of them could run a statistical test. So the numbers below do not show whether an effect exists or not; they only describe what was tried.
Rett syndrome: Rett syndrome is a developmental disorder seen almost only in girls, usually caused by a change in a gene called MECP2. Development looks typical in the first year of life, and then speech and hand skills regress; breathing irregularities, sleep problems, anxiety and restlessness are common. Following positive findings in animal studies, several cannabinoids have been tried in this syndrome. The symptoms looked at in the research were behavior and mood (with the Rett behavior scale filled in by a caregiver), anxiety, sleep and movements.
The only study comparing purified CBD with placebo in Rett enrolled 29 girls aged 2–18; it was stopped early because of difficulty finding participants and the COVID-19 pandemic, and only 17 girls completed the 24 weeks. In the 9 girls on the high dose (15 mg per kilogram of body weight per day), the behavior score fell by an average of 12.1 points, and in the 10 girls on placebo by 6.1 points (in the placebo group, only 4 of the 10 girls completed the 24 weeks, and this group's starting score was slightly higher); in those on the low dose, the score barely changed. Doctors' overall impression and illness severity scores were similar in the three groups. Because the groups were so small, no statistical comparison was made, and the result was published only in the study registry record; these numbers cannot tell us whether CBD works. In the open-label extension that 21 girls from the main study moved into (everyone took CBD; there was no comparison group), the behavior score fell by an average of 7.4 points; this extension study was also ended early.
In Rett, a full-spectrum cannabis extract containing very little THC (0.08%) and made up mainly of CBDA, the acid form of CBD, was tried for 12 weeks in 11 girls aged 5–16. Doctors' overall assessment, the behavior scale, communication, anxiety, quality of life and caregiver burden improved. However, this study had no comparison group, the family and doctor knew which product was being used, the study was funded by the companies that developed the product, and 4 of the 8 authors work at these companies; so the evidence is very weak. A placebo-controlled study with the same product is registered, but as of October 2026 it had not started enrolling patients.
Cannabidivarin (CBDV), a cannabinoid close to but distinct from CBD, was tried in 5 girls with Rett who had drug-resistant epilepsy. In this small study, no clear change was seen in Rett's non-seizure symptoms. We describe the studies without a comparison group that used purified CBD for the epileptic seizures common in Rett syndrome (seizures fell clearly in some children; the evidence is very weak) on the Drug-Resistant Epilepsy page. A systematic review published in 2026 that looked at Rett together with autism and fragile X syndrome reported, for Rett, only a reduction in seizures (from studies without a comparison group); it gave no separate result for Rett's other symptoms and concluded that, in all three conditions, the current evidence is not enough to support routine use.
22q11.2 deletion syndrome: This syndrome, caused by a small missing piece of chromosome 22, occurs in about 1 in every 3,000–6,000 births. Besides heart conditions, cleft palate, low calcium and immune problems, anxiety, social withdrawal, attention problems and learning difficulties are common in children; about 4 in every 10 people have an anxiety disorder. In this syndrome, the risk of psychosis and schizophrenia in adolescence and young adulthood is also clearly higher. The symptoms looked at in the research were anxiety, mood and behavior problems.
In the only treatment study with cannabinoids in 22q11.2 deletion syndrome, lab-made, THC-free pure CBD was applied to the skin (upper arm or shoulder) as a gel. 20 children and teens aged 4–15 with marked anxiety used the gel for 14 weeks. The anxiety score fell on average from 14.7 to 8.5; mood, irritability and inappropriate speech scores also improved, and doctors rated three quarters of the participants as improved; on the other hand, the score doctors gave for overall illness severity did not change. This study had no placebo group, and it was funded by the company that developed the product. In a study without a comparison group, the placebo effect and improvement that happens on its own over time cannot be separated, so this result alone does not count as proof of an effect; the authors themselves write that a placebo-controlled study is needed. The same gel was tested against placebo in 212 children with fragile X syndrome, described on the site's Autism page; there, children on placebo improved similarly, and on the main measure the gel was not clearly different from placebo.
There are no treatment studies with THC-containing products in 22q11.2 deletion syndrome. Because the risk of psychosis is high in this syndrome, cannabis use is of particular interest: in a study of 434 people, regular cannabis use among people with this syndrome was clearly lower than among both healthy people and people with psychosis, and within the group with the syndrome no link was seen between substance use and psychosis. The difference in illegal drug use, including cannabis, was no longer clear once sex and IQ score were taken into account; in addition, information on people with the syndrome was collected from medical records, while information on the other groups was collected in a separate interview. This was a one-time measurement; it does not show whether cannabis raises the risk of psychosis in this syndrome.
Prader-Willi syndrome: In this syndrome, caused by a region of genes on chromosome 15 not working, the most prominent problems are a hard-to-stop urge to eat that begins in childhood (hyperphagia) and weight gain; learning difficulties, short stature, low reproductive hormones, and problems with sleep and body temperature regulation are also seen. The finding that people with Prader-Willi have higher blood levels of one of the body's own cannabinoids (2-AG) suggested that this appetite-regulating system might play a role in the syndrome. The symptoms looked at in the research were overeating, weight, irritability and rigid, repetitive behaviors.
In Prader-Willi, three studies were started to compare pure CBD or CBDV with placebo, and all three were stopped early: one with 7 people because the manufacturer went bankrupt, one with 4 people because company priorities changed (none of them was assigned to the placebo group), and one with 6 people because funding could not be found. At week 13, the overeating score had fallen by a median of 4.5 points in the 4 people in the CBD group and by 2 points in the 3 people in the placebo group; but only two people completed the 13 weeks, and no statistical test was done. In the CBDV study, the irritability score fell by a similar amount in both groups. Overeating and rigidity scores appeared to fall more in the CBDV group; but the groups were very different at the start (the CBDV group was 19 years old on average, the placebo group 10; the lethargy and withdrawal score was four times higher in the CBDV group at the start and stayed higher at week 12), and each group had 2–3 people. None of these results was published in a peer-reviewed journal, and with groups this small it is not known whether CBD affects overeating. There are no studies with THC-containing products in Prader-Willi.
A weight-loss drug (rimonabant) that is not a cannabis product but instead blocks the body's cannabinoid receptor had also been tried in adults with Prader-Willi; although there was a trend toward weight loss, half of those taking the drug left the study because of psychiatric side effects, and the study was stopped early. This drug is not CBD or cannabis; we mention it here only to explain why this system is being studied in Prader-Willi.
Strength of evidence: ●○○○ Very weak evidence. Finding: Not studied enough. What does this mean?
Ongoing research on this topic
No ongoing CBD trial on this topic appears on ClinicalTrials.gov right now. All ongoing trials
Possible side effects and drug interactions
In the studies in these syndromes, the side effects most often reported with pure CBD were diarrhea, vomiting, decreased appetite, lethargy, fever and rash. In the controlled study in Rett, raised liver enzymes in the high-dose group were recorded as a serious side effect; in the long-term extension, one girl's liver enzyme also rose above three times the normal level. In the Rett extension study, serious problems such as respiratory failure, seizures and infections occurred in 5 of 21 girls, and 2 girls died; none of these was considered related to the treatment. Because Rett itself can involve breathing problems and seizures, it is hard to tell the cause of these events apart in an uncontrolled study. In 22q11.2 deletion syndrome, the only side effects of the skin gel considered treatment-related in the first 14 weeks were pain and itching where it was applied; according to the study registry record, counting the extension period, 3 of the 20 children had a serious health problem (low calcium, testicular torsion, viral lower respiratory tract infection). The record does not say whether these were related to the gel; low calcium is also common in the syndrome itself. With CBDV, excessive sleepiness and increased drooling were reported, and in one person a feeling of detachment (dissociative symptoms).
Many of the children in these studies were also taking other medicines for epilepsy, anxiety or behavior. CBD can slow the breakdown of some medicines in the body. The 22q11.2 study also did not enroll people taking certain epilepsy medicines such as clobazam, phenobarbital, carbamazepine and phenytoin, or medicines broken down in the liver by the same enzyme (CYP3A4). Details are on the site's epilepsy medicines, clobazam and valproate interaction pages.
To check interactions with the medicines you take: CBD drug interaction checker.
How was CBD used in studies?
In Rett syndrome, purified CBD was given by mouth as a solution at 5 or 15 mg per kilogram of body weight per day (study dose; not a personal recommendation); in the open-label extension, it was started at 5 mg per kilogram per day and increased by 5 mg each week up to at most 15, or 20 mg if needed (study dose; not a personal recommendation). In the Prader-Willi studies, lab-made CBD was taken by mouth with food at 10, 20 or 40 mg per kilogram of body weight per day (study dose; not a personal recommendation).
In 22q11.2 deletion syndrome, the pure CBD gel was applied to the skin: 250 mg a day for those weighing up to 35 kg and 500 mg a day for heavier participants (study dose; not a personal recommendation), split into two equal doses. CBDV was used in the Rett and Prader-Willi studies with a target of 10 mg per kilogram of body weight per day (study dose; not a personal recommendation). The dose of the low-THC cannabis extract tried in Rett is not given in the article's abstract.
In the studies, these doses were used under a doctor's supervision with regular blood tests and checkups; because most studies were stopped early, it is not known which dose, if any, works.
Study details (for health professionals)
Rett syndrome: the only RCT with purified CBD (NCT03848832; 29 girls, aged 2–18, 24 weeks, 5 or 15 mg/kg/day) was stopped early because of difficulty finding participants and COVID-19; 17 girls completed it and the planned statistics were not done; results are only in the ClinicalTrials.gov record: change in RSBQ total score −12.1 with 15 mg/kg/day (SD 13.6; n = 9), −6.1 with placebo (SD 7.2; n = 10), +0.4 with 5 mg/kg/day; CGI-I and CGI-S values similar between groups (no testing); in the placebo group, 4 of 10 girls completed the 24 weeks. In the open-label extension (NCT04252586; 21 girls, up to 20 mg/kg/day, none completed), RSBQ −7.4; serious side effects in 5 girls, 2 deaths, none considered treatment-related. With a full-spectrum, CBDA-dominant extract with 0.08% THC (Keating 2025; 11 girls, 12 weeks, open-label, manufacturer-funded), improvement from baseline in CGI-I (p = 0.028), RSBQ (p = 0.0005) and quality of life. In a phase 1 study with CBDV (Hurley 2022; 5 girls with drug-resistant epilepsy, 10 mg/kg/day), no clear change in non-seizure symptoms; seizures decreased (seizures are described on the Drug-Resistant Epilepsy page). A systematic review covering Rett, autism and fragile X together (Boeri 2026; 17 studies, 2 RCTs) reports, for Rett, only a descriptive reduction in seizures and finds the evidence insufficient for routine use. 22q11.2 deletion syndrome: in an open-label phase 2 study with a synthetic CBD gel applied to the skin (Heussler 2026, INSPIRE; 20 children and teens, aged 4–15, 250 or 500 mg a day, 14 weeks + 24-week extension; manufacturer-funded), the PARS-R anxiety score fell from 14.7 to 8.5 (−41.9%; p = 0.0001), ADAMS total −60.0% (p = 0.0014), ABC-C irritability −45.8% (p = 0.006); CGI-S change in the registry record 0.0; no THC in plasma; the only treatment-related side effects were pain and itching where applied; in the registry record, serious side effects in 3 people across the whole study. In a cross-sectional study of 434 people with the syndrome, regular cannabis use was clearly lower than in healthy controls and people with psychosis, and among people with the syndrome there was no link between any substance use and psychotic disorder (OR 0.80; 95% CI 0.54–1.20; cannabis not looked at separately); the group difference in illegal drug use was not significant after accounting for sex and IQ; in the syndrome group, data came from medical records (Vingerhoets 2019). Prader-Willi syndrome: all three placebo-controlled studies were stopped early, with no peer-reviewed publication: synthetic CBD 40 mg/kg/day in 7 young people (NCT02844933; HQ-CT median −4.5 vs −2; 2 people completed), synthetic CBD 10–40 mg/kg/day in 4 people (NCT05098509; all in the CBD groups, no efficacy data), CBDV 10 mg/kg/day in 6 people (NCT03848481; ABC irritability 22.3 vs 24.0 at baseline, 10.0 vs 10.3 at week 12; groups very different at baseline in age and some scores). The selective CB1 blocker rimonabant (not a cannabis product) had been stopped early with 10 people because of psychiatric side effects (Motaghedi 2011; background, not in the table). The 'Not studied enough' label is based on the evidence for pure CBD: placebo-controlled studies exist, but all were stopped early with few people (4–29), and their direction cannot be read. October 2026 search (PubMed, Cochrane, ClinicalTrials.gov, Europe PMC): in these three syndromes, it found 4 studies started to compare cannabinoids with placebo (1 in Rett, 3 in Prader-Willi; one with CBDV rather than CBD; all stopped early, one could not enroll anyone in the placebo group, none published in a peer-reviewed journal, results in the registry records), 4 uncontrolled studies (3 in Rett, 1 in 22q11.2), 1 systematic review covering Rett and 1 observational study looking at cannabis use in 22q11.2; all are in the table below. There is no controlled study in 22q11.2 deletion syndrome and no Cochrane review of cannabinoids in these syndromes. A 40-person placebo-controlled study with the low-THC extract in Rett is registered but has not started enrolling patients; a 14-person open-label study with the same extract ended in December 2025 and has no results.
Evidence assessment: Pure CBD: Very weak (the placebo-controlled studies that started with 29 people in Rett and with 7 and 4 people in Prader-Willi were all stopped early, no statistical tests were done, and the results are only in the registry records; in 22q11.2 deletion syndrome, the only data come from a manufacturer-funded skin gel study in 20 children without a control group). CBDV was not used as a basis for this level because it is a cannabinoid distinct from CBD; THC-containing products were assessed under the THC evidence
CBD alone
In Rett syndrome, the only placebo-controlled study with purified CBD (29 girls, 24 weeks; NCT03848832) was stopped early and published only in the registry record, without statistics: the behavior score fell by an average of 12.1 points with high-dose CBD and 6.1 points with placebo (9 and 10 girls); the doctors' overall impression scores were similar between groups. In its open-label extension (21 girls), the score fell by an average of 7.4 points. In Prader-Willi, two studies with synthetic CBD (7 and 4 people) were stopped early; in one, the overeating score fell by a median of 4.5 points with CBD and 2 points with placebo, and in the other no efficacy data were collected at all. In 22q11.2 deletion syndrome, in 20 children and teens using a pure CBD gel applied to the skin, without a comparison group, the anxiety score fell by about 42% (INSPIRE, manufacturer-funded). With CBDV, a cannabinoid close to but distinct from CBD, no clear change in non-seizure symptoms was seen in 5 girls with Rett; in a 6-person placebo-controlled study in Prader-Willi, irritability was similar. The evidence is very weak: no study is large enough to show or rule out an effect.
Products containing THC
With THC-containing products, there is only one study, in Rett syndrome: with a full-spectrum cannabis extract containing 0.08% THC and made up mainly of CBDA, the doctor's overall impression (p = 0.028), the behavior scale, anxiety and quality of life improved after 12 weeks in 11 girls aged 5–16; the study had no comparison group and was funded by the companies that developed the product. There are no treatment studies with THC-containing products in 22q11.2 deletion syndrome or Prader-Willi syndrome. The risk of psychosis is high in 22q11.2 deletion syndrome; regular cannabis use was uncommon among people with this syndrome, and within the group no link was found between substance use and psychosis (434 people, cross-sectional), but the safety of THC-containing products in this group has not been studied.
Evidence: Very weak (with THC-containing products, only a single 12-week, manufacturer-funded study without a control group in 11 girls with Rett syndrome, using an extract with 0.08% THC made up mainly of CBDA; no treatment studies with THC-containing products in 22q11.2 deletion and Prader-Willi syndromes)
Dose range in studies
Rett: purified CBD by mouth 5 or 15 mg/kg/day (24 weeks; up to 20 mg/kg/day in the extension); Prader-Willi: synthetic CBD by mouth 10–40 mg/kg/day; 22q11.2: synthetic CBD gel applied to the skin, 250 or 500 mg a day by body weight; CBDV 10 mg/kg/day (Rett and Prader-Willi); the dose of the low-THC extract is not in the abstract
Starting dose and titration
In the Rett extension, started at 5 mg/kg/day and increased by 5 mg/kg each week; with CBDV, started at 2.5 mg/kg/day and increased by 2.5 mg/kg each week; for the skin gel, no dose adjustment according to the article (fixed by body weight; the registry record states it could be raised to 500 or 750 mg a day if needed); doses twice a day
These are doses researchers gave under medical supervision; they are not a recommended dose for you. "mg/kg/day" means milligrams per kilogram of body weight per day.
| Study | Participants | Product | Dose and titration | Duration | Result | Source |
|---|---|---|---|---|---|---|
| NCT03848832 (Rett syndrome, registry results) Double-blind, placebo-controlled, 3-arm parallel RCT (phase 3). Stopped early because of difficulty finding participants and the COVID-19 pandemic; because the groups were small, the planned statistical comparisons were not done. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. Run by the manufacturer (Jazz Pharmaceuticals / GW) | 29 randomized (10 CBD 5 mg/kg/day, 9 CBD 15 mg/kg/day, 10 placebo); 17 completed (7 / 6 / 4) Rett syndrome, aged 2–18 (mean 9.7), all girls | Purified, plant-derived CBD oral solution (GWP42003-P, 100 mg/mL); placebo solution | CBD 5 or 15 mg/kg/day, divided into two doses a day No details in the registry record | 24 weeks | Primary outcome (change at week 24 in the total score of the caregiver-completed Rett Syndrome Behaviour Questionnaire, RSBQ, 15 mg/kg/day vs placebo): mean −12.1 (SD 13.6; 9 people) vs −6.1 (SD 7.2; 10 people); no statistical testing was done. +0.4 (SD 12.5) in the 5 mg/kg/day group. Baseline score 45.9 in the 15 mg/kg group, 50.0 on placebo, 39.8 in the 5 mg/kg group. In the placebo group, only 4 of the 10 girls completed the 24 weeks; the registry record still gives the week 24 change for 10 girls and does not say which measurement was used for those who left early. Doctor's overall impression (CGI-I, week 24; 15 mg/kg / 5 mg/kg / placebo) 2.9 / 3.2 / 3.1; change in illness severity (CGI-S) −0.1 / −0.1 / 0.0. On the sleep scale (CSHQ; same order) −5.0 / −1.6 / −3.7; on the anxiety-fear subscale −1.3 (15 mg/kg) vs placebo −1.7. Because the groups were very small and no statistics were done, no conclusion about effect can be drawn Side effects: Serious side effects 2/11 (5 mg/kg), 2/9 (15 mg/kg), 1/9 (placebo): one COVID-19 case in each group; lower respiratory tract infection in the 5 mg/kg group; raised liver enzymes (ALT and AST) and involuntary movements (dyskinesia) in the 15 mg/kg group. Diarrhea 2/11, 2/9, 0/9; lethargy 0, 3/9, 0 (sleepiness in 1 person in the 5 mg/kg group); vomiting and decreased appetite only in the 15 mg/kg group (2/9). 1 person stopped because of side effects. Because 1 person assigned to placebo was mistakenly given CBD, the groups in the side effect table have 11 / 9 / 9 people. No deaths | NCT03848832, results December 29, 2021, last updated September 2, NCT04252586, 2026) |
| NCT02844933 (Prader-Willi syndrome, registry results) Double-blind, placebo-controlled parallel RCT (phase 2). Terminated early with 7 people because the original manufacturer (Insys Therapeutics) went bankrupt and stopped all its studies. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record (the responsible manufacturer is now Radius Pharmaceuticals) | 7 randomized (4 CBD / 3 placebo); 2 completed (1 / 1) Prader-Willi syndrome, aged 8–17 (mean 12.7), overeating (hyperphagia) behavior | Oral CBD solution (according to the program's extension record NCT03458416, non-plant-derived, pharmaceutical-grade synthetic CBD); matched placebo | 40 mg/kg/day, divided into two doses a day, with food No details in the registry record | 13 weeks (planned) | Primary outcome (HQ-CT hyperphagia questionnaire, change from baseline at week 13): median −4.5 (lowest −14, highest −3; 4 people) vs placebo −2 (−12 to 2; 3 people); no statistical testing; the registry record gives week 13 values for all 7 people and does not say which measurement was used for those who left early. Median change in body weight −1.05 kg vs +1.3 kg (median baseline weight 88.4 kg vs 56.7 kg); responders 1/4 vs 1/3. Other measures (overall impression, eating questionnaire, quality of life, physical activity) were not collected because of early termination. With only 7 people and 2 completers, no conclusion can be drawn Side effects: No serious side effects. Diarrhea 2/4 (CBD) vs 1/3 (placebo); rash 2/4 vs 0/3; abdominal cramps, skin infections (impetigo, cellulitis) and ear infection in one person each in the CBD group; seizure in 1 person in the placebo group | NCT02844933, NCT03458416, no data reported, citing confidentiality). No peer-reviewed publication found in PubMed (October 30, 2026) |
| NCT03848481 (Prader-Willi syndrome, CBDV, registry results) Double-blind, placebo-controlled parallel RCT (phase 2; single center, Montefiore Medical Center, New York). Stopped early with 6 people because adequate funding could not be found; no robust statistical analysis was done (only an effect size for one secondary outcome). As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. Investigator-initiated; collaborators were a Prader-Willi research foundation and GW Pharmaceuticals, which supplied the product | 6 randomized (3 CBDV / 3 placebo); 5 completed (2 / 3) Prader-Willi syndrome, aged 5–30 (mean 14.7; 19.3 in the CBDV group, 10 in the placebo group) | Cannabidivarin (CBDV): a separate cannabinoid close to CBD, derived from the cannabis plant; THC below 0.2%; matched placebo | 10 mg/kg/day No details in the registry record | 12 weeks | Primary outcome (irritability subscale of the ABC behavior scale): mean 22.3 (CBDV, 3 people) and 24.0 (placebo, 3 people) at baseline, 10.0 (SD 5.7; 2 people) and 10.3 (SD 4.2; 3 people) at week 12; a similar decrease in both groups. Overeating score (HQ-CT) 17.3 and 15.7 at baseline, 11.0 and 14.3 at week 12; in repetitive behaviors (CY-BOCS), from 11.7 to 7.0 (1 person) and from 13.7 to 13.5. On the rigidity scale (MERS-R-PWS; with missing data filled in for the CBDV group), 28.3 and 28.0 at baseline, 13.7 and 28.3 at week 12 (effect size in the registry record d = 1.89). The lethargy and social withdrawal score was much higher in the CBDV group at baseline (16.3 vs 4.0) and stayed higher at week 12 (12.5 vs 3.7). No statistical testing was done; the groups had 2–3 people and differed greatly at baseline in age (mean 19.3 vs 10 years) and in some scores, so no conclusion can be drawn Side effects: No serious side effects. Dissociative symptoms (feeling detached) in 1/3 people in the CBDV group; no side effects in the placebo group. 1 person in the CBDV group withdrew from the study | NCT03848481, 2026). No peer-reviewed publication found in PubMed (October 30, 2026) |
| NCT05098509 (Prader-Willi syndrome, registry results) Planned as a double-blind, placebo-controlled, 4-arm RCT (phase 2/3). The manufacturer (Radius Pharmaceuticals) stopped it because company priorities changed, for reasons unrelated to safety; all 4 people enrolled were assigned to the CBD groups, no one was enrolled in the placebo group, and no efficacy analysis was done. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record | 4 (1 person 10 mg/kg, 1 person 20 mg/kg, 2 people 40 mg/kg; placebo 0); none completed Prader-Willi syndrome, aged 8–65 (age and sex not reported, citing confidentiality) | Oral solution containing synthetic CBD (RAD011); matched placebo | 10, 20 or 40 mg/kg/day, with food No details in the registry record | 34 weeks planned; longest treatment 68 days | No efficacy data: the overeating (HQ-CT), irritability and overall impression measures were not analyzed because of early termination. The row is in the table only for safety information and so that the result of an unpublished study is transparent Side effects: No serious side effects. Diarrhea in both of the 2 people in the 40 mg/kg group; one person each with skin inflammation (dermatitis; 10 mg/kg), lethargy (20 mg/kg), restlessness, headache, tiredness and ear discomfort (40 mg/kg) | NCT05098509, NCT05387798, 2026) |
| Heussler 2026 (INSPIRE) Open-label, multicenter phase 2 study (Australia and 1 US center; February 2020–November 2022); no control group. After 14 weeks of treatment, an optional 24-week extension for participants whose irritability improved by at least 35% or whom the physician considered suitable. Sample size based on feasibility, not on a power calculation. Funded by the manufacturer (Harmony Biosciences), which also paid for writing support; 2 of the 5 authors are manufacturer employees and 3 received research funding from the manufacturer (all five have ties to the manufacturer). Registered on ClinicalTrials.gov about 20 months after the study started (October 2021) | 20 started, 17 completed; 16 in the efficacy analysis; 13 moved into the extension, 11 analyzed Genetically confirmed 22q11.2 deletion syndrome, aged 4–15 (mean 9.9; inclusion criterion 4–17 years), 60% male; illness severity at least moderate (CGI-S ≥4) and high anxiety score (PARS-R ≥10); at most 2 psychiatric medicines, stable for 6 months | ZYN002: lab-made (synthetic) CBD gel applied to the skin (upper arm or shoulder); no THC detected in plasma | 125 mg every 12 hours (250 mg a day) for those weighing ≤35 kg, 250 mg every 12 hours (500 mg a day) for those >35 kg None according to the article (fixed dose by body weight, added to existing treatment); the product description in the registry record states it could be raised to 500 or 750 mg a day if needed | 14 weeks + optional 24 weeks (38 weeks in total) | The primary outcome was safety and tolerability (in the side effects column). Efficacy (secondary, no control group): the PARS-R anxiety scale fell on average from 14.7 to 8.5 (−41.9%; p = 0.0001; in the abstract −40.6%, p = 0.0005); ADAMS total score −60.0% (p = 0.0014; in the abstract −45.3%); ABC-C irritability −45.8% (p = 0.006; in the abstract −36.3%, p = 0.0055) and inappropriate speech −43.3% (p = 0.040); the reductions in withdrawal, stereotypy and hyperactivity were not significant. In the doctors' assessment, 75% of participants improved and 62.5% improved markedly; in contrast, the illness severity score (CGI-S) in the registry record did not change at week 14 (mean 0.0). In the extension (only those whose irritability had improved by at least 35% at week 14 or whom the physician considered suitable continued; 11 people analyzed), the reductions in anxiety and irritability were maintained. Because there was no placebo group, it is not known how much of the improvement is due to the gel; the authors write that a placebo-controlled study is needed Side effects: In the first 14 weeks, 13 side effects in 7 of 20 people (35%), all mild; 3 treatment-related side effects: pain (2) and itching (1) where applied; 1 person stopped because of side effects not related to treatment; no meaningful changes in lab tests, ECG or vital signs. In the registry record, across the whole study (about 306 days), side effects in 12/20 people and serious side effects in 3 people (low calcium, testicular torsion, blood collecting in the scrotum, viral lower respiratory tract infection); their relationship to treatment is not stated in the record | PMID 41888667, doi:10.1186/s11689-026-09687-z, NCT05149898, results March 9, 2026, collaborator Zynerba). The percentages in the abstract and the main text differ (e.g., PARS-R 40.6% and 41.9%), both are given in the row |
| NCT04252586 (Rett syndrome, open-label extension, registry results) Open-label extension study (phase 3; 21 girls who had taken part in the main study NCT03848832; the main study's flow chart appears to show 17 girls completing the 24 weeks, and the registry record does not explain this difference); no control group. Stopped early because of difficulties in the main study and COVID-19; no one completed it. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. Run by the manufacturer (Jazz Pharmaceuticals / GW) | 21 (moved over from the main study; 0 completed) Rett syndrome, aged 2–18 (mean 10.2), all girls | Purified, plant-derived CBD oral solution (GWP42003-P, 100 mg/mL) | Started at 5 mg/kg/day, up to at most 15 mg/kg/day; up to at most 20 mg/kg/day if needed After 1 week, increases of 5 mg/kg/day per week based on response and tolerability (at the investigator's discretion) | Up to 104 weeks planned; side effects were tracked from the start of the main study to day 767 | The primary outcomes were safety (in the side effects column). Secondary (uncontrolled): mean change in RSBQ total score from the start of the main study −7.4 (SD 14.9); doctor's overall impression (CGI-I) mean 3.3 (between 'minimally improved' and 'no change'), illness severity (CGI-S) 4.3; −3.0 on the sleep scale; +0.7 on the motor behavior scale. Because there was no placebo group and everyone left early, no conclusion about effect can be drawn Side effects: Side effects in 19 of 21 people; treatment-related in 7; 3 stopped because of side effects (1 treatment-related). Serious side effects in 5 people (respiratory failure 2, respiratory arrest 1, seizure and status epilepticus, lower respiratory tract infection, COVID-19, surgical wound infection, worsening of Rett syndrome); none considered treatment-related. The registry record lists 2 deaths. Most common: vomiting 7, diarrhea 4, fever 4, decreased appetite 3, restlessness 3, seizure 3. ALT above 3 times normal in 1 person. On ECG, QTcB >450 ms at some visits in at most 4 people, >480 ms in 1; none >500 ms | NCT04252586, NCT03848832, 2026) |
| Keating 2025 (NTI164, Rett syndrome) Open-label phase 1/2 study; no control group; baseline and week 12 compared with a paired t test. Funded by the manufacturer (Fenix Innovation Group and Neurotech International); 4 of the 8 authors work at these companies (3 at Fenix, 1 at Neurotech); no declaration of interests in the PubMed record | 11 Girls with Rett syndrome carrying a disease-causing change in the MECP2 gene, aged 5–16 (Australia) | NTI164: full-spectrum cannabis plant extract, 0.08% THC; according to the article on the autism study with the same product, the main component is CBDA, the acid form of CBD | Twice a day; dose not in the abstract (full text could not be read) Not reported (in the abstract; the full text could not be read) | 12 weeks | The primary outcome (doctor's impression of overall improvement, CGI-I) improved at week 12 (p = 0.028); improvement also in communication, mental alertness, eye contact, attention and anxiety. Illness severity (CGI-S, p = 0.008), RSBQ total score (p = 0.0005), the general mood, breathing problems, repetitive face movements and fear-anxiety subscales, quality of life (p = 0.0002) and caregiver burden (p = 0.006) also improved. Effect sizes are not in the abstract. Because there was no placebo group, the caregivers and doctors knew which product was used and the study was manufacturer-funded, it is not known how much of the improvement is due to the extract Side effects: The abstract only says 'safe'; the type and frequency of side effects are not in the abstract (full text could not be read) | PMID 40568811, doi:10.1111/jpc.70122, NCT06621043, aged 5–20, ended in December 2025, no results in the registry record, NCT07257978, not yet enrolling). That the product is CBDA-dominant: Keating 2026 (autism, PMID 42561512, a row on the site's Autism page) |
| Hurley 2022 (CBDV, Rett syndrome) Investigator-initiated; open-label, single-center phase 1 dose-escalation study (Sydney Children's Hospital, February 2019–August 2020); no control group. Funded by the NSW Ministry of Health; product supplied by GW Research; the authors declared no conflicts of interest | 6 found eligible, 5 started and were analyzed Girls with Rett syndrome with an MECP2 change and drug-resistant epilepsy, aged 6–13.7 (median 12.6) | Cannabidivarin (CBDV) oral solution, plant extract, 50 mg/mL; THC below 0.2% by weight | Target 10 mg/kg/day (average reached 9.9; 9.6–10.1), divided into two doses a day Started at 2.5 mg/kg/day and increased by 2.5 mg/kg every 7 days | 13.75–15.5 months (treatment duration varied by patient) | No clear change was seen in Rett's non-seizure symptoms, the question for this page: in the 3 girls with both a baseline and a final RSBQ score, median 34 to 33; slight improvement of 3% and 5% in two, 20% worsening in one; most domains of the Rett symptom severity scale did not change, with small improvements in mental alertness, attention and eye contact. Seizures (secondary outcome): monthly seizure counts fell in all 5 girls, median reduction 79% in the abstract and 82% in the main text; no clear change on EEG. Because the study is very small and uncontrolled, no conclusion can be drawn Side effects: 77 side effects, 77.9% mild; 3 definitely related events: excessive sleepiness (2) and increased drooling (1); 24 events possibly related; none of the 16 serious side effects was considered definitely related (the 5 possibly related ones were attributed by experts to Rett itself or other causes); no one stopped the drug. The dose was briefly reduced in 3 patients | PMID 35364618, doi:10.1111/epi.17247, PMC9544893. The seizure result is the subject of the Drug-Resistant Epilepsy page. On May 16, 2022, a correction was made that only added an open-access funding sentence |
| Boeri 2026 Systematic review (PRISMA 2020; GRADE and design-specific bias tools); randomized and non-randomized studies; descriptive pooled rates only | 17 studies (2 RCTs; the rest observational studies and case series); number of Rett studies and patients not in the abstract Autism, fragile X and Rett syndrome under 18 years of age | CBD, CBDV, THC or mixtures of these | For Rett syndrome, the abstract reports only a reduction in seizures (descriptive pooled rates); there is no separate result in the abstract for Rett's behavioral and other symptoms. Across all diagnoses, side effects were mostly mild to moderate and the dropout rate was low. The authors' conclusion: there is possible benefit for some behavioral symptoms and co-occurring epilepsy, but the current evidence is not enough to support routine use; well-designed RCTs are needed Side effects: Mild to moderate side effects, low dropout rate (all diagnoses together) | PMID 42339654, doi:10.1177/10445463261462382, The full text is not open access, which Rett studies it included and the GRADE levels could not be read. The authors declared no conflicts of interest. For autism and fragile X, there are more comprehensive reviews on the site's Autism page | ||
| Vingerhoets 2019 Cross-sectional comparative study. Substance use was collected in the 22q11.2 group from medical records (asked systematically in the psychiatric interview) and self-report, and in the comparison groups with a structured interview (Composite International Diagnostic Interview, CIDI); the groups were selected from different centers and in different ways. In the 22q11.2 group, psychiatric diagnosis, IQ and COMT genotype were also determined. Supported by the Netherlands health research organization (ZonMw); the authors declared no conflicts of interest | 833: 434 people with 22q11.2 deletion syndrome, 265 patients with psychosis without the deletion, 134 healthy controls People with 22q11.2 deletion syndrome over 15 years of age (4 centers in the Netherlands, Belgium and Canada; 165 of them, 38.1%, had a psychotic disorder); comparison groups from a psychosis study in Amsterdam (aged 15–60) | Cannabis and other substances (alcohol, nicotine, illegal drugs; self-reported) | Any substance use: 36.9% in 22q11.2 deletion syndrome; 83.6% in healthy controls, 87.9% in patients with psychosis (p < 0.001). Substance use disorder 1.2%. Regular (daily or weekly) cannabis use was clearly lower in the 22q11.2 group than in the other two groups (χ² = 48.2; p < 0.001); none of the substances other than cannabis was used weekly. The group difference in illegal drug use, including cannabis, was no longer significant after accounting for sex and IQ (p = 0.162). Within the 22q11.2 group, no link was found between any substance use and psychotic disorder (OR 0.80; 95% CI 0.54–1.20); this analysis does not look at cannabis separately. Cross-sectional study, and the samples were collected from clinics (according to the authors, not fully representative of the population): it does not show whether cannabis raises the risk of psychosis in this syndrome. According to an earlier study cited by the authors in the introduction, schizophrenia spectrum disorders occur in about 20–25% of people with this syndrome at ages 18–25 and in about 40% over age 25 | PMID 30604657, doi:10.1192/bjp.2018.258, The cannabis-specific percentages are in Table 2 and the supplementary tables of the article, these tables could not be read (no numbers in the text). A smaller US study points the same way: Amir 2023, PMID 37151981, alcohol and cannabis use lower than in controls in 89 deletion carriers) |
Reported side effects
- Purified CBD, Rett, 24 weeks: serious side effects 2/11 (5 mg/kg), 2/9 (15 mg/kg), 1/9 (placebo); raised ALT and AST and dyskinesia with 15 mg/kg; lethargy 3/9, vomiting and decreased appetite 2/9, diarrhea 2/11 and 2/9 (0 on placebo) (NCT03848832)
- Purified CBD, Rett, open-label extension: side effects in 19/21, 7 of them treatment-related; serious side effects in 5 people (respiratory failure, respiratory arrest, seizure and status epilepticus, infection), 2 deaths, none considered treatment-related; vomiting 7, diarrhea 4, decreased appetite 3; ALT >3×ULN 1; QTcB >480 ms 1 (NCT04252586)
- Synthetic CBD, Prader-Willi: no serious side effects; diarrhea 2/4 vs placebo 1/3, rash 2/4 vs 0/3 (NCT02844933); with 40 mg/kg/day, diarrhea in both of the 2 people (NCT05098509)
- Synthetic CBD gel applied to the skin, 22q11.2: mild side effects in 7/20 people over 14 weeks; the only treatment-related ones were pain (2) and itching (1) where applied; in the registry record, 4 serious side effects in 3 people across the whole study (low calcium, testicular torsion and blood collecting in the scrotum, viral lower respiratory tract infection; the record does not state whether they were related to the gel) (Heussler 2026; NCT05149898)
- CBDV: in Rett, definitely related excessive sleepiness (2 events) and increased drooling (1 event) (Hurley 2022); in Prader-Willi, dissociative symptoms in 1/3 people (NCT03848481)
- Low-THC, CBDA-dominant extract, Rett: the abstract only says 'safe'; the type and frequency of side effects could not be read (Keating 2025)
Monitoring done in studies
- In the studies, blood tests including liver enzymes, ECG and vital signs were monitored at baseline and at visits (NCT04252586; Heussler 2026; Hurley 2022)
- In the Rett extension study, weight, QTc interval, menstrual pattern, pubertal stage and IGF-1 were also checked, and participants were asked about suicidal thoughts or self-harm (NCT04252586)
- The skin gel was applied to clean, dry, intact skin (upper arm or shoulder); at each visit, side effects and signs of not tolerating the drug were assessed (Heussler 2026)
- With CBDV, while the dose was being increased: weekly examinations for the first 28 days and a phone call on days 2 and 4 after each examination; then monthly visits and weekly phone calls for 90 days, followed by a visit every 3 months (Hurley 2022)
- Ongoing studies awaiting results (October 2026): in Rett, a 14-person open-label study with the low-THC extract (NCT06621043; ended in December 2025, no results in the registry record) and a 40-person placebo-controlled crossover study (NCT07257978; not yet enrolling, estimated completion October 2028)
Who should be careful
- 22q11.2 deletion syndrome: in this syndrome, the risk of psychosis and schizophrenia in adolescence and young adulthood is clearly higher (about 20–25% at ages 18–25, according to an earlier study cited by Vingerhoets 2019); the safety of THC-containing products in this group has never been studied. If the person or their family has a history of psychosis, cannabis products should not be used without consulting a doctor (this is a precautionary recommendation, not a direct study finding)
- Use with epilepsy medicines: the 22q11.2 study excluded people taking clobazam, phenobarbital, ethosuximide, felbamate, carbamazepine, phenytoin or vigabatrin and those taking strong CYP3A4 inhibitors or inducers (Heussler 2026); most of the girls in the CBDV study in Rett were also taking valproate or clobazam (Hurley 2022)
- Liver: in the controlled study in Rett, a rise in ALT and AST in 1 girl in the high-dose (15 mg/kg/day) CBD group was recorded as a serious side effect (NCT03848832); in the long-term extension, ALT rose above 3 times normal in one of 21 girls (NCT04252586)
- Breathing in Rett syndrome: in the extension study, respiratory failure and respiratory arrest were recorded as serious side effects, and there were 2 deaths; they were not considered related to the treatment, but without a control group it is hard to tell (NCT04252586)
- People who had used cannabis or cannabinoid products in the past 3 months, or who tested positive for cannabinoids in a urine test, were not enrolled in the studies in these syndromes (Hurley 2022; Heussler 2026)
- Product type: the product tested in 22q11.2 and fragile X is a lab-made gel applied to the skin; the product in the positive open-label study in Rett is an extract that contains a little THC and is made up mainly of CBDA. These results cannot be directly applied to CBD oils on the market
Drug interactions
- Detailed information and monitoring recommendations are on the 'epilepsy medicines', 'clobazam' and 'valproate' pages in the site's drug interactions section
- The 22q11.2 study (Heussler 2026) excluded people taking the following medicines (a study safety limit, not evidence of an interaction): clobazam, phenobarbital, ethosuximide, felbamate, carbamazepine, phenytoin, vigabatrin; strong CYP3A4 inhibitors or inducers and medicines sensitive to CYP3A4; more than 2 stable psychiatric medicines
- In the CBDV study in Rett, 3 of the 5 girls were also taking valproate, 3 clobazam and 3 levetiracetam; whether side effects were related to the treatment was assessed in this multi-drug setting (Hurley 2022)
Abbreviations: RCT, randomized controlled trial (a study in which participants are assigned to groups by chance); placebo, an inactive substance; SMD, OR and RR, effect measures used in analyses that pool studies; CYP, liver enzymes that break down medicines.
If you are considering CBD
- You can talk openly with the doctor who follows your child (a pediatric neurologist, geneticist or child psychiatrist) about which symptom you want to track and what you expect; in these syndromes, CBD has not been shown to replace established treatments.
- Go over all the medicines your child takes together, especially epilepsy, anxiety and behavior medicines, with interactions in mind; do not stop current medicines or change their dose on your own.
- Because the risk of psychosis is high in 22q11.2 deletion syndrome, THC-containing products call for extra caution; be sure to mention any history of psychosis in your child or your family.
- If a product is being considered, review the amount of CBD and THC on its label and its independent certificate of analysis with the doctor; plan monitoring with liver tests, as was done in the studies.
- In Prader-Willi, CBD has not been shown to reduce overeating or weight; keep working with your care team on eating routines and established treatments.
Choose your age group, medicines and situation to prepare a short summary you can take to your doctor. Nothing you enter is sent anywhere.
Bottom line
Interest in CBD for Rett, 22q11.2 deletion and Prader-Willi syndromes is high, but research is still at an early stage: all the controlled studies were stopped early with few participants, and the encouraging results come from small studies without a comparison group. So it is not known whether CBD works in these syndromes. Making decisions together with a doctor who knows your child is the safest path.
Frequently asked questions
Does CBD oil work for Rett syndrome?
There is not yet enough research to know. The only study comparing purified CBD with placebo in Rett started with 29 girls, but it was stopped early and statistics could not be done: the behavior score fell by an average of 12.1 points in the 9 girls on high-dose CBD and by 6.1 points in the 10 girls on placebo; the doctors' overall impression scores were similar between groups. A manufacturer-funded study in 11 girls with a cannabis extract containing very little THC reported improvement, but there was no comparison group. The evidence is very weak.
Is CBD used for seizures in Rett syndrome?
Purified CBD is used off-label in some countries for the epileptic seizures common in Rett syndrome; in small studies without a comparison group, seizures fell clearly in some children (the evidence is very weak). We describe these studies and the strength of evidence on the Drug-Resistant Epilepsy page.
Does CBD help anxiety in 22q11.2 deletion syndrome?
There is a promising but early finding. In a 14-week study of 20 children and teens using a THC-free pure CBD gel applied to the skin, the anxiety score fell by about 42%, and mood and irritability scores also improved. However, there was no placebo group, and the study was funded by the company that developed the product; because the placebo effect cannot be separated out without a comparison group, this result alone does not count as proof of an effect. The same gel was tested in a large controlled study in fragile X syndrome; children on placebo improved similarly, and the gel was not different from placebo on the main measure.
Does using cannabis raise the risk of psychosis in someone with 22q11.2 deletion syndrome?
No study has looked at this directly. The risk of psychosis is already high in this syndrome. In a study of 434 people, regular cannabis use was uncommon among those with the syndrome, and within the group no link was found between substance use and psychosis; but this was a one-time measurement and does not show whether cannabis raises the risk. Because cannabis use in adolescence has been linked to psychosis in the general population (details on the Schizophrenia page), THC-containing products call for extra caution in this syndrome; make the decision together with your child's doctor.
Does CBD reduce overeating in Prader-Willi syndrome?
It is not known. In Prader-Willi, all three studies started to compare CBD or the closely related CBDV with placebo were stopped early with 4–7 people, and their results are only in the study registry records. At week 13, the overeating score had fallen by a median of 4.5 points in the 4 people in the CBD group and by 2 points in the 3 people in the placebo group; but only two people completed the 13 weeks, and no statistics were done. With these data, it cannot be said whether CBD affects overeating or weight.
Why were the CBD studies in these syndromes left unfinished?
None of the reasons given in the registry records is a safety problem: the Rett study was stopped because of difficulty finding participants and the COVID-19 pandemic; of the Prader-Willi studies, one was stopped because the manufacturer went bankrupt, one because the company's priorities changed and one because funding could not be found. Reaching enough participants is hard in rare diseases; that is why well-designed studies in these syndromes are still needed.
Is there an FDA-approved CBD medicine for Rett syndrome, 22q11.2 deletion syndrome or Prader-Willi syndrome?
No. As of November 2, 2026, the only FDA-approved CBD medicine is Epidiolex, a prescription oral solution for seizures linked to Lennox–Gastaut syndrome, Dravet syndrome and tuberous sclerosis complex in people 1 year and older. Doctors can legally prescribe an approved drug for another use when they judge it medically appropriate (this is called “off-label” use), but Epidiolex hasn’t been approved for Rett syndrome, 22q11.2 deletion syndrome or Prader-Willi syndrome. CBD oils sold in stores and online aren’t FDA-approved. The FDA hasn’t evaluated them for safety, dose or quality, and when it tested some products, many didn’t contain the amount of CBD the label claimed. If you’re thinking about CBD, talk with your doctor or pharmacist first, especially about the medicines you already take.
References
Show 10 references
- NCT03848832 (Rett syndrome, registry results). Double-blind, placebo-controlled, 3-arm parallel RCT (phase 3). Stopped early because of difficulty finding participants and the COVID-19 pandemic; because the groups were small, the planned statistical comparisons were not done. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. Run by the manufacturer (Jazz Pharmaceuticals / GW). NCT03848832, results December 29, 2021, last updated September 2, NCT04252586, 2026)
- NCT02844933 (Prader-Willi syndrome, registry results). Double-blind, placebo-controlled parallel RCT (phase 2). Terminated early with 7 people because the original manufacturer (Insys Therapeutics) went bankrupt and stopped all its studies. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record (the responsible manufacturer is now Radius Pharmaceuticals). NCT02844933, NCT03458416, no data reported, citing confidentiality). No peer-reviewed publication found in PubMed (October 30, 2026)
- NCT03848481 (Prader-Willi syndrome, CBDV, registry results). Double-blind, placebo-controlled parallel RCT (phase 2; single center, Montefiore Medical Center, New York). Stopped early with 6 people because adequate funding could not be found; no robust statistical analysis was done (only an effect size for one secondary outcome). As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. Investigator-initiated; collaborators were a Prader-Willi research foundation and GW Pharmaceuticals, which supplied the product. NCT03848481, 2026). No peer-reviewed publication found in PubMed (October 30, 2026)
- NCT05098509 (Prader-Willi syndrome, registry results). Planned as a double-blind, placebo-controlled, 4-arm RCT (phase 2/3). The manufacturer (Radius Pharmaceuticals) stopped it because company priorities changed, for reasons unrelated to safety; all 4 people enrolled were assigned to the CBD groups, no one was enrolled in the placebo group, and no efficacy analysis was done. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. NCT05098509, NCT05387798, 2026)
- Heussler 2026 (INSPIRE). Open-label, multicenter phase 2 study (Australia and 1 US center; February 2020–November 2022); no control group. After 14 weeks of treatment, an optional 24-week extension for participants whose irritability improved by at least 35% or whom the physician considered suitable. Sample size based on feasibility, not on a power calculation. Funded by the manufacturer (Harmony Biosciences), which also paid for writing support; 2 of the 5 authors are manufacturer employees and 3 received research funding from the manufacturer (all five have ties to the manufacturer). Registered on ClinicalTrials.gov about 20 months after the study started (October 2021). PMID 41888667, doi:10.1186/s11689-026-09687-z, NCT05149898, results March 9, 2026, collaborator Zynerba). The percentages in the abstract and the main text differ (e.g., PARS-R 40.6% and 41.9%), both are given in the row
- NCT04252586 (Rett syndrome, open-label extension, registry results). Open-label extension study (phase 3; 21 girls who had taken part in the main study NCT03848832; the main study's flow chart appears to show 17 girls completing the 24 weeks, and the registry record does not explain this difference); no control group. Stopped early because of difficulties in the main study and COVID-19; no one completed it. As of October 2026, no peer-reviewed publication; results from the ClinicalTrials.gov record. Run by the manufacturer (Jazz Pharmaceuticals / GW). NCT04252586, NCT03848832, 2026)
- Keating 2025 (NTI164, Rett syndrome). Open-label phase 1/2 study; no control group; baseline and week 12 compared with a paired t test. Funded by the manufacturer (Fenix Innovation Group and Neurotech International); 4 of the 8 authors work at these companies (3 at Fenix, 1 at Neurotech); no declaration of interests in the PubMed record. PMID 40568811, doi:10.1111/jpc.70122, NCT06621043, aged 5–20, ended in December 2025, no results in the registry record, NCT07257978, not yet enrolling). That the product is CBDA-dominant: Keating 2026 (autism, PMID 42561512, a row on the site's Autism page)
- Hurley 2022 (CBDV, Rett syndrome). Investigator-initiated; open-label, single-center phase 1 dose-escalation study (Sydney Children's Hospital, February 2019–August 2020); no control group. Funded by the NSW Ministry of Health; product supplied by GW Research; the authors declared no conflicts of interest. PMID 35364618, doi:10.1111/epi.17247, PMC9544893. The seizure result is the subject of the Drug-Resistant Epilepsy page. On May 16, 2022, a correction was made that only added an open-access funding sentence
- Boeri 2026. Systematic review (PRISMA 2020; GRADE and design-specific bias tools); randomized and non-randomized studies; descriptive pooled rates only. PMID 42339654, doi:10.1177/10445463261462382, The full text is not open access, which Rett studies it included and the GRADE levels could not be read. The authors declared no conflicts of interest. For autism and fragile X, there are more comprehensive reviews on the site's Autism page
- Vingerhoets 2019. Cross-sectional comparative study. Substance use was collected in the 22q11.2 group from medical records (asked systematically in the psychiatric interview) and self-report, and in the comparison groups with a structured interview (Composite International Diagnostic Interview, CIDI); the groups were selected from different centers and in different ways. In the 22q11.2 group, psychiatric diagnosis, IQ and COMT genotype were also determined. Supported by the Netherlands health research organization (ZonMw); the authors declared no conflicts of interest. PMID 30604657, doi:10.1192/bjp.2018.258, The cannabis-specific percentages are in Table 2 and the supplementary tables of the article, these tables could not be read (no numbers in the text). A smaller US study points the same way: Amir 2023, PMID 37151981, alcohol and cannabis use lower than in controls in 89 deletion carriers)
Points that could not be verified from the full text
- The full text of the study with the low-THC extract in Rett (Keating 2025) could not be read (publisher access blocked): the dose, side effects and effect sizes are not in the abstract. It could not be verified whether the open-label registry record with the same product, which enrolled 14 girls and ended in December 2025 (NCT06621043; according to the record, the dose started at 5 mg/kg/day and rose to at most 20 mg/kg/day over 4 weeks), is the same study as this article.
- The full text of the systematic review that also covers Rett (Boeri 2026) could not be read: which Rett studies it included, how many patients with Rett there were and the certainty-of-evidence levels are not known.
- The cannabis-specific percentages of the cannabis use study in 22q11.2 deletion syndrome (Vingerhoets 2019) are in the article's tables; the tables could not be read, so only the comparison given in the text is used on this page.
- The percentages in the abstract of the 22q11.2 study (Heussler 2026) differ from those in the main text (e.g., 40.6% and 41.9% for the anxiety score; 36.3% and 45.8% for irritability); this page uses the numbers in the main text. The article states that there were no serious side effects in the first 14 weeks, while the registry record states that 3 people had serious side effects across the whole study; whether they were related to the treatment is not given in the record.
- The open-label extension of the synthetic CBD program in Prader-Willi (NCT03458416) was ended with 2 people, and no data were reported in the registry record, citing confidentiality. In the CBDV study in Rett, the monthly seizure reduction is given as 79% in the abstract and 82% in the main text.
Research history
Here we record what changed on this page, when and why, with dates. The newest change is at the top.
Page published. This first assessment reviewed 10 studies.