What the research shows about CBD

Seizures: the proven use

• FDA approved the CBD prescription drug Epidiolex in 2018 for seizures in Lennox-Gastaut syndrome and Dravet syndrome, after randomized, placebo-controlled trials.
• In the Lennox-Gastaut trials, drop seizures fell by about 37% to 44% with CBD, compared with about 17% to 22% with placebo.
• Sleepiness and raised liver enzymes were among the common side effects.

Bottom line: this is what a proven CBD use looks like, and it shows CBD is biologically active. It is a pharmaceutical product at prescribed doses, and it does not show that other CBD products work for other conditions.

Sources: NeurologyLive: FDA approves cannabidiol for Lennox-Gastaut and Dravet syndromes; FDA: What You Need to Know (and What We're Working to Find Out) About Products Containing Cannabis or Cannabis-Derived Compounds, Including CBD.

Anxiety: the most consistent signal

• Randomized trial, 24 people with social anxiety (2011). A single 600 mg dose of CBD before a simulated public-speaking test reduced anxiety, cognitive impairment and discomfort compared with placebo. The CBD group responded much like healthy volunteers.
• Systematic review, 15 randomized trials, 673 people (2024). The authors concluded CBD "shows promise" for anxiety with a favorable safety profile. In the review's results table, 6 of the 15 trials found a benefit over placebo and 9 did not. They called the evidence conflicting and asked for larger, better-designed trials.
• Randomized trial, 63 people with high worry (2023). 300 mg of CBD daily for two weeks reduced anxiety symptoms compared with placebo. It did not reduce worry itself, and a single dose had no effect. The authors suggest CBD's effects may be modest and tied to physical symptoms of anxiety.
• Chart review, 72 adults at one psychiatry clinic (2019). Anxiety scores fell within the first month in 79.2% of patients, and CBD was well tolerated in all but 3. With no placebo group, this can suggest an effect but cannot prove one.

Bottom line: the strongest case for CBD in mental health. Several well-run trials found less anxiety, others found none, and the benefits that appear are modest.

Sources: Bergamaschi et al., Neuropsychopharmacology, 2011 (randomized, double-blind); Coelho et al., Life, 2024 (systematic review of 15 randomized trials); Gournay et al., Psychopharmacology, 2023 (randomized, placebo-controlled); Shannon et al., The Permanente Journal, 2019 (retrospective case series).

PTSD: early and promising, not proven

• Case series, 11 adults (2019). Patients added oral CBD to their usual psychiatric care for 8 weeks. Ten of 11 had lower PTSD symptom scores, and the average score fell 28% (from 51.8 to 37.1 on the PCL-5). CBD was generally well tolerated, and some patients with frequent nightmares reported relief. There was no placebo group.
• Randomized trial, 33 people with PTSD (2022). A 300 mg dose of CBD before recalling a traumatic event reduced the cognitive impairment people felt during recall, and the effect lasted a week. It did not significantly change anxiety, alertness or discomfort compared with placebo.
• VA's National Center for PTSD says CBD treatments are being tested for PTSD and that early studies suggest oral CBD may reduce anxiety.

Bottom line: the studies we found are small, one without a control group. They are a reason for larger trials, not proof that CBD treats PTSD. Proven treatments exist, and CBD is not a replacement for them. See Get help.

Sources: Elms et al., Journal of Alternative and Complementary Medicine, 2019 (case series); Bolsoni et al., Psychopharmacology, 2022 (randomized, double-blind); VA National Center for PTSD: Cannabis Use and PTSD Among Veterans (updated Sept. 3, 2025).

Sleep: mixed

• Chart review (2019). Sleep scores improved within the first month in 66.7% of the 25 patients treated for poor sleep, but the improvement fluctuated over time.
• Randomized pilot, 30 adults with moderate to severe insomnia (2024). 150 mg of CBD nightly for two weeks did not beat placebo on the main measures, including the Insomnia Severity Index. Wrist-sensor data showed higher sleep efficiency (about 6.7 percentage points) and people reported better well-being. The authors note the CBD group started with higher well-being and that the study was small and short.

Bottom line: some signals, no clear result. Sleep that stays poor is worth raising with a provider.

Sources: Shannon et al., The Permanente Journal, 2019 (retrospective case series); Narayan et al., Journal of Clinical Sleep Medicine, 2024 (randomized pilot).

Pain: where the evidence is weakest

• A 2024 review in The Journal of Pain looked at 16 randomized trials of verified CBD for pain, covering 12 pain conditions and doses from 6 mg to 1,600 mg. Fifteen of the 16 showed no benefit over placebo.
• The authors also found that CBD products on the market often vary in how much CBD they contain and may contain other chemicals.

Bottom line: for pain, the trials do not currently support CBD on its own, so we don't suggest it. Research on CBD combined with other cannabinoids is a separate question that this review did not cover.

Source: Moore et al., The Journal of Pain, 2024 (review of CBD pain trials).

Why policy is slower than the research

• Clinical guidelines are reviews that stop at a point in time. The VA/DoD PTSD guideline is dated 2023, and the guideline does not name CBD. Studies published around and after that date may not be reflected.
• VA clinicians can prescribe only FDA-approved medicines, and FDA has approved one CBD drug. So what VA clinicians can prescribe depends on FDA approval, not on how many small studies exist.
• FDA has said a new regulatory pathway for CBD is needed, and hemp rules are still changing.

Bottom line: caution in policy is not proof that CBD doesn't work, and a few promising small studies are not proof that it does. Read the policy page for the VA and federal view.

How to read this research

• Randomized, placebo-controlled trials carry the most weight. Case series and chart reviews have no comparison group, so improvement may come from other causes.
• Most studies here used 150 to 600 mg of pharmaceutical-grade CBD. Products sold online vary, and FDA has found many contain less CBD than their labels say.
• Most CBD studies are short, from one dose to a few weeks. Long-term effects are not well known.
• Side effects and drug interactions are real. Read CBD safety and talk with your provider before you start.