PTSD and CBD Oil: Research, Dosage and Warnings

What does the research say?

In one study, 33 people with PTSD were randomly assigned to two groups. One group took a single dose of CBD, the other an inactive pill (placebo), and then they were asked to recall their trauma. There was no difference in anxiety. Only on a subscale measuring a feeling of mental fogginess was there an improvement in favor of CBD, lasting one week. A later analysis found that this effect appeared only in people whose trauma was not sexual.

In 42 people who had been through trauma and had high stress levels, taking CBD for one week did not change the anxiety, distress or heart rate response triggered by a reminder of the trauma. In people with both PTSD and alcohol use disorder, 6 weeks of CBD was also no better than placebo for PTSD symptoms, anxiety or alcohol use. A pooled review from 2025 also found no clear overall effect in anxiety-related disorders.

The only study reporting a positive result is a retrospective review of the records of 11 patients who used CBD in addition to their treatment: their symptom score fell by 28% over 8 weeks, but because there was no comparison group, it is not known how much of this was due to CBD. A laboratory study in healthy volunteers suggested that CBD may strengthen the learning that allows fear to fade; as of October 2026, this idea has not been confirmed in people with PTSD. In a study of 96 veterans, CBD or placebo was added for 16 weeks to a trauma-focused therapy (prolonged exposure). Its results were entered only in the trial registry in October 2026: symptom scores fell in both groups, and at the end of treatment the score on a 0–80 scale was 22.6 in the CBD group and 23.1 in the placebo group. The registry gives no statistics comparing the groups, and there is no peer-reviewed publication yet. Similar laboratory studies with THC in people with PTSD showed changes on brain imaging, but no difference in measures of the fear response.

In a US study of 80 military veterans with PTSD, participants used smoked cannabis for 3 weeks: high-THC, high-CBD, equal parts of both, or an inactive cannabis (placebo). Symptoms decreased in all groups, including placebo; no type of cannabis was better than placebo.

The strongest data for nightmares are new: in a study in Germany, 171 adults with PTSD and frequent nightmares were randomly assigned to groups and used synthetic THC (dronabinol) every night for 10 weeks. The frequency and intensity of nightmares decreased more than with an inactive pill (placebo); the share of people whose nightmares were cut at least in half was about twice as high. The difference in overall PTSD symptoms, however, was not clear on the clinician's assessment. Most people taking dronabinol could tell from the drug's effects which group they were in, and the drug's manufacturer supported the study; this single study's finding needs to be repeated in other studies. In an earlier small study of 10 soldiers, nabilone, a THC-like medicine, had also reduced nightmares. Cannabinoid research on insomnia, sleep apnea and restless legs is on the insomnia page; the nightmare studies are also listed in that page's table.

Real-world studies that follow people with PTSD who use cannabis paint two different pictures. In a US study following 2,276 veterans, those who started using cannabis after treatment had more severe symptoms over time; in another study of veterans, cannabis use was linked to more intense intrusive memories in the following months. On the other hand, records of patients using prescribed cannabis in the UK, and a 150-person study comparing users with non-users, reported improvement in symptoms; in these studies there was either no comparison group at all or one that was not formed by random assignment, and some of their authors had ties to cannabis companies. Because the groups were not randomly assigned, none of these studies shows that cannabis caused the outcome, in either direction. The 2023 guideline of the US Departments of Veterans Affairs and Defense makes a strong recommendation against cannabis and cannabis products in the treatment of PTSD.

Strength of evidence: ●●○○ Weak evidence. Finding: No effect found. What does this mean?

Ongoing research on this topic

Trials testing CBD for this topic on ClinicalTrials.gov: 6 under way (3 recruiting). A trial being under way does not show that CBD works.

NCT06381180: Phase 1/2, recruiting, target 500 participants
NCT06222268: Phase 1/2, not yet recruiting, target 280 participants
NCT05269459: Phase 2, recruiting, target 140 participants

All trials on this topic and their history

Possible side effects and drug interactions

The most common side effects are drowsiness, tiredness and a slight drop in attention. In the alcohol use disorder studies, 22.6% of people taking CBD had side effects that required limiting the dose; in one anxiety study, raised liver values and one severe drug rash were reported.

CBD can slow the breakdown of some medicines in the body; for example, blood levels of antidepressants such as citalopram and escitalopram may rise. In a small additional analysis of people taking certain antidepressants, fear faded again less well in those taking CBD than in those taking placebo.

Side effects are more common with medicines that contain THC: in the nightmare study, dizziness, dry mouth, diarrhea and increased appetite were more common with dronabinol than with placebo; serious side effects occurred in 7 people in the dronabinol group and in none in the placebo group (the researchers did not attribute any of them to the drug). Two people treated with nabilone who had a prior psychotic illness developed symptoms of psychosis.

Interactions with medicines used for this condition

Interactions reported for these medicines in peer-reviewed publications and official prescribing information. A medicine that is not listed can still interact with CBD; tell your doctor about every medicine you take.

PTSD: medicines used and their interactions with CBD
Medicine	Importance	Warning	Source
Citalopram (SSRI) (Cipram, Citol, Relaxol), Escitalopram (SSRI) (Cipralex, Citoles, Eslorex)	Caution	Citalopram, escitalopram: CBD raised citalopram blood levels: a significant rise in a series of 6 patients; in healthy volunteers, citalopram exposure (AUC) rose by 43% after 12 days of approved CBD (2.5 mg/kg twice a day). If you take these medicines, talk to your doctor.	PMID 34121064, PMID 41622700
Sertraline (SSRI) (Lustral, Selectra, Serdep)	Caution	Sertraline: In a 78-year-old patient who had taken sertraline for 20 years, blood sodium dropped and forgetfulness developed in the months after over-the-counter CBD was added; it resolved after CBD and then sertraline were stopped (a single case; CBD's role was rated "possible"). In two adverse event reports from Poland, people taking CBD oil together with sertraline had diarrhea, vomiting, fever and severe fatigue. Tell your doctor or pharmacist.	PMID 36627907, PMID 39502532, PMID 34121064
Fluoxetine (SSRI) (Prozac, Depreks, Fulsac)	Caution	SSRIs (fluoxetine, paroxetine): In a young person taking fluoxetine, worsening insomnia, restlessness and obsessive symptoms were reported after CBD oil was added (a single case). The reported case of a child taking paroxetine involved smoked marijuana, not CBD. Tell your doctor or pharmacist.	PMID 35652796, PMID 39719830
Alprazolam (Xanax, Ansiox, Stabina), Clonazepam (Rivotril, Rivoclon), Other benzodiazepines (lorazepam, etc.), Pregabalin (Lyrica), Hydroxyzine (Atarax)	Caution	Benzodiazepines, sleep medicines and sedatives: Drowsiness and sedation may increase together with CBD. Talk to your doctor before using them together.	DailyMed Epidiolex label 7.4, EMA Epidyolex SmPC 4.4
SNRIs (venlafaxine, duloxetine, etc.) (Efexor XR, Cymbalta)	Information	SNRIs and tricyclic antidepressants: There are no clinical data on interactions between CBD and SNRIs or tricyclics other than amitriptyline (see the separate warning for amitriptyline). Still tell your doctor about all the medicines you take.	PMID 38868665
Alprazolam (Xanax, Ansiox, Stabina), Buspirone (Buspon)	Information	Alprazolam, trazodone, buspirone: No study has looked at these medicines and CBD together. The data on the enzyme that breaks them all down (CYP3A4) conflict: approved CBD (750 mg twice a day) did not affect this enzyme, while a single dose of a CBD-dominant extract raised the blood level of midazolam, which is broken down the same way, by 56%. Drowsiness and sedation may increase when they are taken together. Tell your doctor.	FDA Epidiolex label §7.4 and §12.3 (05/2026), PMID 37313955, Alprazolam prescribing information §12.3, dailymed.nlm.nih.gov, Trazodone prescribing information §12.3, dailymed.nlm.nih.gov, Buspirone prescribing information (Drug Interactions), dailymed.nlm.nih.gov
Mirtazapine (Remeron, Zestat)	Information	Mirtazapine: In the lab, CBD had very little effect on the breakdown of mirtazapine; this has not been measured in patients. Still tell your doctor about the medicines you take.	PMID 34121064
Clonazepam (Rivotril, Rivoclon)	Information	Clonazepam: In 45 people with epilepsy taking approved CBD, clonazepam blood levels did not change as the CBD dose went up. Still, drowsiness may increase when they are taken together; tell your doctor.	PMID 37593907

To check interactions with the medicines you take: CBD drug interaction checker.

How was CBD used in studies?

In the studies, CBD was taken by mouth: in the trauma-recall studies, a single dose of 300 mg (study dose; not a personal recommendation) or 300 mg a day for one week (study dose; not a personal recommendation), and in the study of people with both alcohol use disorder and PTSD, 600 mg a day (study dose; not a personal recommendation). No clear benefit was seen at these doses.

Medicines containing THC were given by mouth in the studies: in the nightmare study, dronabinol was started at 2.5 mg a day before bed and raised by 2.5 mg every few days up to a maximum of 15 mg (study dose; not a personal recommendation); nabilone was started at 0.5 mg and raised to a maximum of 3 mg (study dose; not a personal recommendation). In these studies the medicines were used under a doctor's supervision; they are not the same thing as CBD oils.

Study details (for health professionals)

A single 300 mg dose did not reduce anxiety during trauma recall; it reduced only the 'cognitive impairment' subscale, and in a secondary analysis the effect was seen only in non-sexual trauma. In trauma-exposed people, 300 mg/day for 1 week was ineffective; in alcohol use disorder + PTSD, 600 mg/day for 6 weeks was ineffective. In a 21-person pilot study in which 500 mg/day was added to prolonged exposure therapy for 18 days (registry results only), two scales differed in opposite directions, and no statistics were published. In a 96-person RCT in veterans in which CBD (600 mg/day according to the protocol) was added to the same therapy for 16 weeks (NCT03518801; registry results only, October 28, 2026), scores on the clinician-rated CAPS-5 scale fell in both groups (36.9 and 35.1 at baseline; 22.6 and 23.1 at the end of treatment; 21.2 and 21.9 at month 3); the registry gives no between-group statistics, and 23 of 50 people in the CBD group and 15 of 46 in the placebo group did not complete the study. The only clearly positive data for CBD are an open-label case series of 11 patients. In an RCT with smoked cannabis in veterans (80 people), no type of cannabis, including one containing 11% CBD, reduced PTSD symptoms more than placebo. For PTSD-related nightmares, in a 10-person crossover RCT with nabilone, a THC-like synthetic medicine, nightmares decreased more than with placebo. In a 171-patient multicenter RCT, dronabinol (synthetic THC, 2.5–15 mg at night, 10 weeks) lowered the nightmare score by 1.50 points more than placebo (0–8 scale; 95% CI 0.71–2.28); on the main secondary outcome measuring overall PTSD severity (clinician assessment), the difference was not significant (p = 0.08). Of the people taking dronabinol who answered the question, 81% correctly guessed which group they were in, reports of improvement were strongly linked to this guess, and the drug's manufacturer supported the study with an unrestricted grant; the nightmare finding rests on a single large study. In two uncontrolled case series with nabilone (47 and 104 people), nightmares decreased clearly. In two laboratory RCTs with a single dose of THC in people with PTSD, brain activity changed but the fear response did not. Real-world data point both ways: in a follow-up of 2,276 veterans, symptoms were more severe in those who started cannabis after treatment, and in a cohort of 361 veterans, cannabis use was linked to more intense intrusive memories in the following months; on the other hand, a 150-person cohort comparing cannabis users with non-users (5 of 7 authors with company ties) and two patient registries in the UK (no comparison group) reported improvement. In none of these studies were the groups formed by random assignment. In data from 410 people with PTSD who also had alcohol or substance problems, cannabis users benefited from trauma-focused therapy as much as non-users. The most recent meta-analysis, which finished its search before the nightmare study (2026, 54 RCTs), found no significant effect in PTSD; the 2023 guideline of the US Departments of Veterans Affairs and Defense makes a strong recommendation against cannabis, and the American Academy of Sleep Medicine (2018), relying only on the 10-person study, states that nabilone "may be used" for nightmares. October 2026 search (PubMed, Cochrane, ClinicalTrials.gov): 11 controlled studies with cannabinoids on this topic (6 treatment studies lasting weeks; 5 single-dose or short laboratory studies, one in healthy volunteers) and 4 systematic reviews were found; together with 2 guidelines and 9 observational studies, all are in the table below. Cochrane has no review dedicated to this question; 2 controlled studies that were completed but whose results have not been released, and studies left unfinished, are under "unverifiable".

Evidence assessment: Weak; controlled studies are mostly negative

CBD alone

In controlled studies in PTSD, pure CBD did not reduce anxiety or symptoms more than placebo. These studies are small (21–42 people) and short (from a single dose to 6 weeks). Results from a 96-person study of veterans in which CBD or placebo was added for 16 weeks to prolonged exposure therapy (NCT03518801) were released in October 2026, only in the trial registry: on a clinician-rated scale (0–80), scores fell in both groups, and at the end of treatment they were 22.6 in the CBD group and 23.1 in the placebo group; the groups were not compared statistically and there is no peer-reviewed publication.

Products containing THC

For overall PTSD symptoms: smoked cannabis containing THC or THC+CBD did not reduce symptoms more than placebo over 3 weeks (Bonn-Miller 2021); in the nightmare study, dronabinol was also not clearly different from placebo on the main secondary outcome measuring overall PTSD severity (clinician assessment); on the scale patients filled in themselves the difference favored dronabinol, but this was an additional analysis without correction for multiple comparisons (Roepke 2026). For nightmares: dronabinol, a synthetic THC, reduced the frequency and intensity of nightmares more than placebo in a multicenter, 10-week RCT with 171 patients (Roepke 2026); nabilone, a THC-like synthetic medicine, gave a similar result in a small crossover RCT with 10 soldiers (Jetly 2015). There is only one large study, most people taking dronabinol noticed which group they were in, and the drug's manufacturer supported the study; the finding needs to be repeated. Real-world data point both ways: some studies following cannabis users reported improvement, others a more severe course; the 2023 guideline of the US Departments of Veterans Affairs and Defense makes a strong recommendation against cannabis in the treatment of PTSD.

Evidence: Weak for overall PTSD symptoms (no clear difference from placebo in two RCTs); moderate for nightmares (a 171-person multicenter dronabinol RCT and a 10-person nabilone RCT were positive; the large study is the only one and is manufacturer-supported, and most people taking dronabinol noticed which group they were in)

Dose range in studies

300 mg single dose, 300 mg/day (1 week), 500 mg/day (18 days, added to therapy; NCT05132699), 600 mg/day (16 weeks, added to therapy; according to the protocol; NCT03518801), 600–1200 mg/day (6–8 weeks); 25–100 mg/day in the open-label case series; THC-containing: dronabinol 2.5–15 mg/day at night (10 weeks; single doses of 5–10 mg in laboratory studies), nabilone 0.5–3 mg/day (RCT) and 0.2–6 mg/day in case series, smoked cannabis up to 1.8 g a day

Starting dose and titration

Not reported in detail in the studies.

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.

PTSD and CBD: studies reviewed
Study	Participants	Product	Dose and titration	Duration	Result	Source
Bolsoni 2022a Double-blind RCT	33 (17 CBD, 16 placebo) PTSD patients	Oral CBD (99.6% pure, BSPG-Pharm), capsule in corn oil (according to the 2022b full text on the same sample)	300 mg single dose None	Single dose; repeated without drug 1 week later	No difference in anxiety; only the cognitive impairment subscale decreased (lasted 1 week); no difference in physiological measures. There is no conflict-of-interest statement in this publication's PubMed record; according to the 2022b publication by the same authors with the same patients, 3 of 5 authors are inventors on CBD patents, and 1 is a consultant to pharmaceutical companies, including the company that supplied the CBD in the study (BSPG Pharm) Side effects: Not reported (in the abstract)	PMID 35029706, doi:10.1007/s00213-021-06043-y
Bolsoni 2022b Exploratory secondary analysis of the same sample (1:1 randomization by minimization; no correction for multiple comparisons)	33 (sexual trauma 14: 7 CBD/7 placebo; non-sexual 19: 10 CBD/9 placebo) PTSD patients	Oral CBD (99.6% pure, no other cannabinoids; BSPG-Pharm), capsule in corn oil	300 mg single dose (90 min before recall) None	Single dose	In non-sexual trauma, the rise in anxiety (difference −9.82; p=0.033) and cognitive impairment (−8.21; p=0.008) were lower than with placebo; no difference in sexual trauma; no difference in physiological measures. Funding: Brazilian public funds (FAPESP, CNPq); 3 of 5 authors are inventors on CBD patents, and 1 is a consultant to pharmaceutical companies, including the company that supplied the CBD in the study (BSPG Pharm) Side effects: No side effect reporting in the full text	PMID 35293520, doi:10.1590/1516-4446-2021-2317
Gournay 2024 Double-blind, placebo-controlled RCT	42 Trauma-exposed people with high stress levels	Oral CBD	300 mg/day None	Single dose + 1 week	No effect on the anxiety, distress or heart rate response to a trauma script. Funding: according to the PubMed record, the cannabis company Canopy Growth (the conflict-of-interest statement and the company's role could not be read) Side effects: Not reported (in the abstract)	PMID 38959155, doi:10.1002/jts.23072
Bogenschutz 2026 Two double-blind, placebo-controlled RCTs	S1: 27 (14 CBD, 13 placebo); S2: 30 (17 CBD, 13 placebo) Alcohol use disorder (S1); alcohol use disorder + full or subthreshold PTSD (S2; 66.7% full PTSD)	Oral CBD capsule (Tilray, 200 mg, oil-based; hemp-derived, THC <0.1%), twice daily	S1: 600 mg/day for 4 weeks (400 in the morning + 200 in the evening), then 1200 mg/day for 4 weeks (600 × 2); S2: 600 mg/day (same split) In S1, dose increase after 4 weeks; the dose could be reduced if not tolerated	8 / 6 weeks	No advantage over placebo for alcohol, craving, mood, anxiety or PTSD symptoms (in S2, the model-estimated PTSD symptom score, PCL-5, over treatment was 27.5 with CBD and 25.9 with placebo; difference 1.56, 95% CI −7.66 to 10.78; p = 0.74). Tilray provided the study drug and funding for part of the pharmacokinetic analysis (according to the publication, the funders had no role in design, conduct or analysis); 2 of 7 authors reported research support from Tilray Side effects: 7/31 (22.6%) of people taking CBD reduced or stopped the dose because of side effects; nausea 20% vs 0% (p=0.025), diarrhea 27% vs 12%; liver enzymes/bilirubin 1.5–6.5×ULN: 2 CBD, 1 placebo; 1 serious event (diabetic ketoacidosis, 'possibly related')	PMID 41312717, doi:10.1111/acer.70212, NCT03252756, results in the registry)
Elms 2019 Retrospective case series, flexible dose	21 started; 11 analyzed (those who attended the week 4 and week 8 follow-ups) PTSD (in addition to medication and psychotherapy)	Hemp-derived CBD capsule (22–28 mg/capsule) and/or liquid spray (CV Sciences); according to the authors, may contain trace THC	Starting mean 33 mg/day (capsule median 25 mg, 25–100); mean 48.6 mg/day at week 8 (2–100) Flexible dose; adjusted every 4 weeks, increased in most	8 weeks	Mean PCL-5 51.82 → 37.14 (−28%) (no control) Side effects: No one stopped because of side effects; among 21 people, 2 tiredness, 1 next-day grogginess/difficulty focusing, 2 digestive complaints, 1 increased reflux	PMID 30543451, doi:10.1089/acm.2018.0437
Straud (NCT05132699, registry results) Double-blind, placebo-controlled pilot RCT (phase 1/2); added to 10 sessions of intensive prolonged exposure therapy over 2 weeks; no peer-reviewed results publication	21 (11 CBD, 10 placebo); 3 people withdrew from the CBD group PTSD	Epidiolex oral solution (according to the registry); matching strawberry-flavored placebo	250 mg twice daily (500 mg/day) Not defined in the registry	18 days of drug + 2 weeks of therapy; 1-month follow-up	Clear decrease in both groups at month 1; clinician-rated CAPS-5 favored placebo (15.2 vs 10.1), self-report PCL-5 favored CBD (19.97 vs 27.3); no statistical comparison published Side effects: No serious side effects; other side effects CBD 10/11, placebo 7/10 (sleep disturbance 4 vs 0, digestive 4 vs 2)	NCT05132699, results October 10, 2024), protocol: Straud 2024, PMID 38404650
Das 2013 (fear extinction) Double-blind, placebo-controlled	48 Healthy	Vaporized (inhaled) CBD (according to the full texts of Kwee 2022 and Elms 2019; the Das full text was not read)	32 mg None	Before or after extinction	CBD given after extinction enhanced the consolidation of extinction learning (laboratory model) Side effects: Not reported (in the abstract)	PMID 23307069, doi:10.1007/s00213-012-2955-y
Bonn-Miller 2021 Double-blind, placebo-controlled, 2-stage crossover RCT	80 (stage 1), 74 (stage 2) US military veterans with PTSD	Smoked cannabis: high THC (12% THC), high CBD (11% CBD, 0.5% THC), THC+CBD (7.9% + 8.1%), placebo	Self-directed use, up to 1.8 g a day	3 weeks (each stage)	In stage 1, no type of cannabis reduced PTSD symptoms more than placebo; clear improvement in all groups, including placebo. Everyone in the THC-containing groups correctly guessed that they were receiving an active product. Funding: a Colorado state health department grant and MAPS, a nonprofit that designed the protocol and analyzed the data; 2 of 11 authors reported fees or shares from cannabis or pharmaceutical companies, 7 receive a salary from MAPS, and 1 from a private research company (Scottsdale Research Institute) Side effects: Generally well tolerated; most common cough 12.3%, throat irritation 11.7%, anxiety 10.4%; 8.4% stopped because of side effects	PMID 33730032, doi:10.1371/journal.pone.0246990, NCT02759185
Raminelli 2025 Systematic review and meta-analysis (21 placebo-controlled RCTs; SAD 5, GAD 1, PTSD 7, OCD 1, Tourette 7)	Anxiety-related disorders and Tourette syndrome	Pure or CBD-rich products			Pooled effect not significant (SMD −0.40; −0.84/0.03); only in the pure CBD subgroup, SMD −0.61 (−1.15/−0.07). Authors: for SAD and GAD, 'low-quality evidence, limited efficacy'	PMID 40956670, doi:10.1177/25785125251378242
Jetly 2015 Double-blind, placebo-controlled, randomized crossover pilot study	10 (mITT) Canadian male soldiers diagnosed with PTSD whose trauma-related nightmares continued despite standard treatment	Nabilone (THC-like synthetic cannabinoid, oral capsule)	Started at 0.5 mg and raised until nightmares were suppressed, up to a maximum of 3.0 mg Increased up to the effective dose or the 3.0 mg ceiling	2 × 7 weeks, with a 2-week washout in between	Reduction on the CAPS nightmare item −3.6 with nabilone, −1.0 with placebo (p = 0.03); CGI-C 1.9 vs 3.2 (p = 0.05); 5 of 10 people improved greatly with nabilone, 1 of 9 with placebo Side effects: Treatment-related side effects 50% with nabilone, 60% with placebo; no severe side effects or stopping because of side effects	PMID 25467221, doi:10.1016/j.psyneuen.2014.11.002, Health Canada registration
Roepke 2026 Multicenter (Germany), double-blind, placebo-controlled, parallel-group RCT (phase 2)	171 randomized (87 dronabinol, 84 placebo); efficacy analysis 169 (87/82); safety analysis 170 (87/83) Adults diagnosed with PTSD who have recurrent nightmares (mean age 37.9; 79.3% women; 46% were in psychotherapy at the same time, 62% were taking antidepressants)	Dronabinol (synthetic THC; BX-1, oral), before bed; matching placebo	Started at 2.5 mg once daily, up to a maximum of 15 mg/day According to the protocol, 2.5 mg for 4 days, then 2.5 mg increases on days 5, 8, 12, 15 and 19; the goal was the dose at which nightmares disappeared (maximum 15 mg)	10 weeks (4 weeks of dose adjustment + 6 weeks at a fixed dose); follow-up at week 12	On the primary outcome (nightmare frequency and intensity, CAPS-IV B2, 0–8 points), the change at week 10 was −3.66 points with dronabinol and −2.17 with placebo; difference −1.50 (95% CI −2.28 to −0.71; p < 0.001; effect size d = 0.65); similar in sensitivity analyses. At least a 50% reduction in nightmares 57.9% vs 29%, complete disappearance 36.8% vs 14.5% (no correction for multiple comparisons). On overall PTSD severity (clinician-rated CAPS-5, main secondary outcome), the difference was not significant (−3.43; 95% CI −7.33 to 0.47; p = 0.08); after correction for multiple comparisons, only the patient's global impression of improvement remained significant among the main secondary outcomes. On the PTSD scale patients filled in themselves (PCL-5), the week 10 difference was −6.03 points (95% CI −10.30 to −1.75; uncorrected additional analysis). At the end of the study, among those who answered the question, 80.8% of people taking dronabinol and 52.5% of those taking placebo correctly guessed which group they were in (most based on the drug's effects); reports of improvement were strongly linked to which group people believed they were in (of those on placebo, 80% of those who thought they were taking dronabinol and 33% of those who thought they were taking placebo said they had improved). Investigator-initiated; the manufacturer Bionorica provided an unrestricted grant and the study drug (according to the publication, it had no role in design or analysis; the 2023 protocol publication also names the cannabis company Canopy Growth as a funder); 1 of 20 authors (the lead author) had received personal fees from Bionorica before the contract Side effects: Any side effect 89.7% vs 77.1%; serious side effects 7 (8.0%) people vs 0 (the researchers did not attribute any to the drug); stopping because of side effects 5.7% vs 7.2%; dizziness 23% vs 14.5%, increased appetite 16.1% vs 8.4%, diarrhea 11.5% vs 3.6%, dry mouth 9.2% vs 1.2%	PMID 42557458, doi:10.1038/s41591-026-04546-9, NCT04448808, October 28, 2026), PMID 37147642
Ayers (NCT03518801, registry results) Double-blind (quadruple-masked), placebo-controlled, parallel-group RCT (phase 2); added to 12 sessions of prolonged exposure therapy; no peer-reviewed results publication	96 started the study (50 CBD, 46 placebo; the registry's 'actual enrollment' figure is 136, the difference is not explained); 27 and 31 completed; 93 people in the baseline analysis of the primary outcome (3 people were excluded because the validity of their symptom reports was found doubtful), 31 and 30 at the end of treatment, 24 and 28 at month 3 Military veterans diagnosed with PTSD (US, VA San Diego Healthcare System; mean age 42.8; 31% women); after March 2020, visits and therapy were remote	Synthetic, pharmaceutical-grade crystalline CBD, capsule (according to the protocol; the registry says only 'active drug'); placebo cellulose capsule	According to the drug section of the protocol, 300 mg capsule twice daily (600 mg/day); another sentence in the same section says 'target daily dose of 300 mg'. No dose in the registry; the dose patients actually took was not reported None; the dose could be reduced in case of side effects (in steps of 50 mg in the protocol's drug section, 0.10 mL in its risk section)	16 weeks of drug (12 weeks of therapy); 1- and 3-month follow-up	Primary outcome was total severity on the clinician-rated CAPS-5 scale (0–80; higher score is more severe): at baseline 36.9 in the CBD group and 35.1 with placebo; at the end of treatment (week 16) 22.6 vs 23.1; at the 1-month follow-up 22.9 vs 21.5; at month 3 21.2 vs 21.9. Scores fell from baseline in both groups; the registry gives no statistical comparison between groups. On the secondary outcome (number of therapy sessions until a 10-point drop on the patient-rated PCL-5 scale, in responders), median 5 vs 5 sessions (27 and 26 people). Non-completers: 23 of 50 people in the CBD group, 15 of 46 in the placebo group. Sponsor: US Department of Veterans Affairs Office of Research and Development (public); results were entered on ClinicalTrials.gov on October 28, 2026 Side effects: No deaths and no serious side effects (CBD 0/46, placebo 0/45); other side effects above the 5% threshold in 11/46 people with CBD and 15/45 with placebo: tiredness 6 vs 8, headache 3 vs 6, dry mouth 3 vs 3, nausea 0 vs 3, drowsiness 3 vs 2. The side effect scale (UKU) was dropped when follow-up switched to remote	NCT03518801, results October 28, 2026), protocol and statistical analysis plan in the registry (version 1.44, November 6, 2023), collaborating institution University of California San Diego
Zabik 2023 Double-blind, placebo-controlled, parallel-group RCT (single dose; laboratory fear extinction study with brain imaging)	71 analyzed (THC 34, placebo 37): 19 with PTSD, 26 trauma-exposed without PTSD, 26 healthy; according to the registry, 86 people were enrolled Adults with PTSD and trauma-exposed adults, plus a healthy comparison group (US, Detroit)	Dronabinol (synthetic THC, oral capsule) and placebo	7.5 mg single dose, about 2 hours before extinction learning None	Single dose; 3-day fear learning, extinction and recall experiment	On behavioral fear measures (skin conductance, expectancy ratings), no difference between THC and placebo in the extinction, recall and fear return phases; in people with PTSD who took THC, brain imaging showed higher activity in the frontal brain (vmPFC) and amygdala. Clinical symptoms were not measured. Funding: US National Institute of Mental Health (NIMH) Side effects: Not reported	PMID 37088409, doi:10.1016/j.nlm.2023.107758, NCT02069366, PMID 32162103, in response to threat images, THC reduced amygdala reactivity in people with PTSD) and Pacitto 2022, PMID 35981598, 51 people, THC reduced negative emotion during emotional reappraisal)
Zabik 2026 Double-blind, placebo-controlled, 3-arm RCT (single dose; laboratory fear extinction study with brain imaging)	44 randomized; 36 analyzed (placebo 11, 5 mg 11, 10 mg 14) Adults with PTSD (US, Detroit)	Dronabinol (synthetic THC, oral capsule) and placebo	5 mg or 10 mg single dose, before extinction learning None	Single dose; testing 24 hours and 1 week later	No difference between dose groups in fear expectancy ratings; on brain imaging, activity in some frontal brain and hippocampal regions was higher 24 hours later with 5 mg and 1 week later with 10 mg. Clinical symptoms were not measured; the sample was small. Funding: US National Institute of Mental Health (NIMH) Side effects: Not reported	PMID 39412674, doi:10.1007/s00213-024-06702-w, NCT03008005, PMID 42463794, 37 people with PTSD), done under the same grant (R61MH111935) and with the same dose groups: THC reduced frontal brain activity during emotional reappraisal, negative emotion ratings did not change (abstract only)
Wilson 2026 Systematic review and meta-analysis of RCTs (1980–May 13, 2025; GRADE; PROSPERO CRD42023392718)	54 RCTs; 2,477 participants (all mental disorders and substance use disorders; the number of PTSD studies is not in the abstract) People whose primary diagnosis is a mental disorder or substance use disorder; PTSD was examined under a separate heading	CBD, THC, THC and CBD combinations, and other cannabinoids			No significant difference was found between cannabinoids and placebo for PTSD-related outcomes (the number of studies and effect size of the subanalysis are not in the abstract; full text could not be read). Significant benefit was seen only in cannabis use disorder, insomnia, tics and Tourette syndrome, and autism; the certainty of evidence is low for most outcomes, 44% of studies had a high risk of bias; the authors conclude that routine use of cannabinoids in mental disorders can rarely be justified for now. Because the search ended in May 2025, the 2026 dronabinol nightmare study (Roepke 2026) is not included. The correction published in the journal (June 16, 2026) corrected the axis of a figure and the study counts in the table; the results did not change Side effects: Across all studies, any side effect was more common with cannabinoids (OR 1.75; 95% CI 1.25–2.46; number needed to harm 7); no difference in serious side effects or dropping out of the study	PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged)
Aviram 2026 Scoping review (PRISMA-ScR; 5 databases and ClinicalTrials.gov, up to December 22, 2024; with quality assessment; not preregistered)	26 studies: 7 RCT publications (6 trials, because two publications of the same trial were counted separately), 18 observational studies, 1 RCT with unpublished results; 3,598 patients in total People diagnosed with PTSD	Medical cannabis, THC, CBD, dronabinol, nabilone			Of the 7 RCT publications, only one (nabilone, nightmares; Jetly 2015) showed a clear clinical advantage over placebo. The RCT with smoked cannabis was not better than placebo; a second study planned with vaporized cannabis stopped at 6 people and so could not be compared with placebo (the review also counts it as 'not better'). Two laboratory RCTs with THC found changes in brain activity but showed no improvement in symptoms; the two RCTs counted for single-dose CBD are actually two publications of the same 33-person trial (Bolsoni 2022a and 2022b), and the effect was small and temporary. Most observational studies reported improvement in nightmares, sleep and hyperarousal, but the risk of bias was high and most had no comparison group. Authors: high-quality RCT evidence is insufficient to support clinical use. Because the search ended in December 2024, Roepke 2026 is not included. 1 of 4 authors is tied to Syqe Medical, which develops a medical cannabis delivery device, and 1 is a former employee of the company; the same company paid the open-access fee Side effects: Side effects generally mild (dry mouth, dizziness)	PMID 42210342, doi:10.1186/s42238-026-00451-7
Rodas 2024 Systematic review (PubMed, PsycINFO, Embase; January 1990–February 2023; no meta-analysis)	14 studies (3 in people with PTSD together with cannabis use disorder) PTSD patients who use cannabis	Cannabis and cannabinoids			Of 10 studies examining overall PTSD symptoms in people without cannabis use disorder, 5 reported benefit and 5 reported no effect or worsening; some benefits were seen in re-experiencing and hyperarousal symptoms. All 3 studies in patients who also had cannabis use disorder reported a risk of worsening symptoms; a risk of worsening suicidal thoughts and violent behavior was also reported. The authors found no clear benefit for overall PTSD symptoms	PMID 38353645, doi:10.4088/JCP.23r14862, PMID 38353646
Schnurr 2024 (VA/DoD 2023 guideline) Clinical practice guideline synopsis (US Department of Veterans Affairs and Department of Defense; systematic evidence review and GRADE)	Adults with PTSD and acute stress disorder	Cannabis and cannabis derivatives; for the nightmare question, cannabis and nabilone were also assessed			Recommendation 20: strong recommendation against cannabis or cannabis derivatives in the treatment of PTSD (very low confidence in the quality of evidence). Rationale: the lack of large, well-designed RCTs and potentially serious side effects; because no new evidence was found in the 2023 review, the recommendation rests on the evidence assessment of the 2017 guideline. The 2021 RCT with smoked cannabis (Bonn-Miller 2021) was not included in the assessment because its follow-up was short; it had also found no difference. For nightmares, cannabis and nabilone were among the drugs assessed, but no recommendation was written about them; for nightmares there is only a weak recommendation for prazosin. The guideline predates the 2026 dronabinol nightmare study (Roepke 2026) Side effects: Based on the 2017 evidence assessment, the guideline mentions possible harms such as impaired attention, memory and driving ability, psychosis, suicide attempts, paranoia and agitation	PMID 38408360, doi:10.7326/M23-2757, Full guideline: VA/DoD Clinical Practice Guideline for the Management of Posttraumatic Stress Disorder and Acute Stress Disorder, June 2023 (healthquality.va.gov), Recommendations 20 and 32–33
Morgenthaler 2018 (AASM) Position paper (American Academy of Sleep Medicine; literature search March 2009–August 2017 and expert opinion; a position paper instead of a guideline because the evidence is limited)	Adults with nightmare disorder and PTSD-related nightmares	Nabilone (THC-like synthetic cannabinoid); for cannabis, it says only 'no evidence to make a recommendation'			For PTSD-related nightmares, nabilone is among the options described as "may be used"; in the paper this is a weaker level than "recommended," where the evidence or expert opinion is uncertain. The basis is a single 10-person RCT (Jetly 2015). It states there is no evidence to make a recommendation for cannabis. The only "recommended" treatment for nightmares is imagery rehearsal therapy (a psychotherapy method) Side effects: For Jetly 2015, dry mouth and headache are reported as most common	PMID 29852917, doi:10.5664/jcsm.7178
Fraser 2009 Retrospective chart review (the author calls it an "open-label study"; no control group)	47 Patients with PTSD whose nightmares had continued for at least 2 years despite antidepressants and sleep medicines (Canada, private psychiatric clinic; 27 women, 20 men)	Nabilone (THC-like synthetic cannabinoid, oral capsule), added to existing medicines	Started at 0.5 mg 1 hour before bed; mean effective dose 0.5 mg (0.2–4 mg) Weekly increase or decrease according to effect and side effects	Not reported (some patients used it for 4–12 months and longer)	In 34 of 47 patients (72%), nightmares stopped completely (28) or eased clearly (6); in most of those who tried stopping the drug, nightmares returned within a few nights. The assessment relies on patients' own nightly records; there is no comparison group, and the patients were those referred to a single specialist clinic Side effects: 13 patients (28%) stopped because of mild to moderate side effects: lightheadedness, forgetfulness, dizziness, headache	PMID 19228182, doi:10.1111/j.1755-5949.2008.00071.x
Cameron 2014 Retrospective chart review (no control group)	104 Male inmates with serious mental illness (Canada, prison treatment center); nabilone was most often given for insomnia (97%), nightmares (87%) and chronic pain (65%)	Nabilone (THC-like synthetic cannabinoid); in powder form with water to prevent misuse	Starting mean 1.4 mg/day (0.5–2), final dose mean 4.0 mg/day (0.5–6) Increased according to effect	Mean 11.2 weeks (1 day–36 weeks)	Nights with nightmares fell from 5.2 to 0.9 per week (90 people), sleep duration rose from 5.0 to 7.2 hours per night (101 people), and the PTSD symptom score (PCL-C) fell from 54.7 to 38.8 (58 people; all p ≤ 0.001). Before-and-after comparison; no control group, and measurements rely on patient reports and clinical records Side effects: Side effects in 31 people (29.8%), 10 (9.6%) stopped; the most serious was psychosis in 2 people with a prior psychotic illness; also drowsiness 12.5%, dry mouth 6.7%, feeling "high" 3.8%	PMID 24987795, doi:10.1097/JCP.0000000000000180
Wilkinson 2015 Observational, prospective follow-up (national evaluation data from US Veterans Health Administration intensive PTSD treatment programs, 1992–2011)	2,276 veterans Veterans admitted to intensive PTSD treatment programs (those with problematic alcohol and other substance use were excluded)	Cannabis (marijuana) use; not prescribed	Use in the last 30 days (yes/no)	Up to 4 months after discharge	PTSD symptoms were most severe in those who started cannabis after the program (effect size +0.34 vs never-users); symptoms were mildest in those who stopped and in never-users (effect size of stopping −0.18). Cannabis use was also linked to more violent behavior and alcohol and drug use. Because it is observational, it does not show cause and effect Side effects: Violent behavior was highest in those who started cannabis	PMID 26455669, doi:10.4088/JCP.14m09475, PMID 33035104, in 1, 413 veterans using nonprescribed opioids, cannabis use neither improved nor worsened PTSD symptoms over 4 months)
Metrik 2022 Observational, prospective cohort (baseline, month 6 and month 12; cross-lagged panel analysis)	361 veterans US veterans who served after September 11, 2001	Cannabis use (not prescribed)	Frequency of use	12 months	Cannabis use at baseline was linked to more severe symptoms of involuntary re-experiencing of traumatic memories 6 months later (β = 0.46; 95% CI 0.16–0.77; p = 0.002); the link in the opposite direction was also significant but weaker (from intrusive memories at baseline to cannabis use 6 months later, β = 0.22; p = 0.04). PTSD at baseline predicted cannabis use disorder 12 months later. Observational; does not show cause and effect	PMID 32546286, doi:10.1017/S003329172000197X
Bonn-Miller 2022 Observational, prospective controlled cohort (cannabis users and non-users; not randomized)	150 (75 cannabis users, 75 non-users) Adults diagnosed with PTSD living in Colorado (73% men; veterans and civilians)	Cannabis bought from licensed medical or recreational outlets; 91% of users mainly used THC-rich products (mostly smoked flower)	Self-directed use (at least once a week)	1 year (assessment every 3 months)	On the primary outcome (CAPS-5), PTSD severity decreased in both groups, faster in cannabis users (group × time interaction −0.32; 95% CI −0.59 to −0.05; p = 0.02); at the end of the year, users were 2.57 times as likely to no longer meet diagnostic criteria (95% CI 1.12–6.07). The difference came mainly from hyperarousal symptoms; no group difference in sleep, functioning or physical activity. The groups differed at baseline in age, proportion of veterans and type of trauma. Funding: Colorado state health department; 5 of 7 authors have ties to cannabis or pharmaceutical companies (two Canopy Growth employees, one Jazz Pharmaceuticals employee, two company consultants) Side effects: Not reported	PMID 33998874, doi:10.1089/can.2020.0056
Hill 2024 Individual patient data meta-analysis (4 RCTs from Project Harmony, a set of 36 RCTs on treating PTSD and substance use disorder; cannabis use was not randomized and was weighted using propensity scores)	410 people (33.2% reported using cannabis at baseline; 70% men) Adults with PTSD together with alcohol or substance use disorder	Cannabis use (in the month before treatment; not prescribed)	Use in the last 30 days (15 days or more in the sensitivity analysis)	Treatment duration (varies by study)	Trauma-focused therapy (COPE, which includes prolonged exposure) produced more improvement in PTSD than non-trauma-focused therapy, whether or not people used cannabis (d = 0.70 in users); session attendance and PTSD improvement did not differ between cannabis users and non-users, and the outcome was the same in frequent users too. It does not test whether cannabis treats PTSD; it shows that cannabis users can benefit from a proven therapy. 1 of 10 authors consulted for Canopy Growth in the last 2 years	PMID 38266511, doi:10.1016/j.janxdis.2024.102827
Lynskey 2024 Observational, prospective patient registry (Project Twenty21, UK; no control group)	238 (month 3 data for 116; 114 for PCL-C) Patients with PTSD starting prescribed cannabis treatment at a private clinic (77% screened positive for depression)	Prescribed cannabis-based products (oil and flower); 74.8% of patients were prescribed at least one THC-rich flower	Not reported (products were classified by form and THC/CBD ratio)	3 months	PTSD symptom score (PCL-C) fell from 58.0 to 47.0 over 3 months (p < 0.001); the drop was larger in those with depression (15.3 vs 7 points). Fewer than half of the patients at baseline have month 3 data; no comparison group. The organization running the registry (Drug Science) receives an unrestricted grant from a consortium of medical cannabis companies; 1 of 5 authors is a scientific advisor to two companies Side effects: Not reported	PMID 38468390, doi:10.1192/bjo.2024.13, PMID 39474239, 58 patients prescribed THC-rich flower, improvement in symptoms reported at months 3 and 6, all 3 authors work at the cannabis company Khiron or a clinic linked to it)
Datta 2025 Observational, prospective patient registry (UK Medical Cannabis Registry; no control group)	269 PTSD patients using prescribed cannabis products who had been followed in the registry for at least 18 months	Prescribed cannabis-based products (the product breakdown for this analysis is not in the abstract; full text could not be read)	Not reported (in the abstract); in the same registry's 2022 PTSD analysis (162 people), median daily dose 145 mg THC and 5 mg CBD	18 months (measurements at months 1, 3, 6, 12 and 18)	PTSD-specific symptoms (IES-R), anxiety, sleep quality and overall quality of life improved significantly from baseline at all follow-up points (p < 0.001); men were less likely to report improvement (OR 0.51). No comparison group; the authors state that cause and effect cannot be inferred. 14 of 18 authors work at the private cannabis clinic that runs the registry (Curaleaf Clinic); the conflict-of-interest statement is in the full text, which could not be read Side effects: Side effects in 70 patients (26%); most common insomnia (15.6%) and tiredness (14.9%)	PMID 40235073, doi:10.1080/14737175.2025.2490539, PMID 36503404, 162 patients, 6 months, 88.9% had used cannabis before, a total of 220 side effects in 33 people (20.4%), most common insomnia and tiredness)

Reported side effects

Sedation/'mental sedation' (Crippa 2004, PMID 14583744); slight impairment of alertness, longer attention lapse time (Rudisill 2023, PMID 36625844)
Berger 2022 (PMID 35921510; 12 weeks, 200–800 mg/day): side effects in 80.6% (drug-related or possibly related in 61.3%), all mild to moderate — tiredness, low mood, hot flushes/chills
Crippa 2021 (PMID 34387679; 300 mg/day, 28 days): 5 serious side effects, all in the CBD group (4 liver enzyme elevations, 1 of them critical; 1 severe drug rash)
Bogenschutz 2026 (600–1200 mg/day): 22.6% of people taking CBD reduced or stopped the dose because of side effects; nausea (20% vs 0%) and diarrhea (27% vs 12%) were more common
Kwee 2022 (300 mg, 8 doses): no serious side effects; side effect frequency similar to placebo (10.3% vs 15.4%)
Roepke 2026 (dronabinol 2.5–15 mg at night, 10 weeks): any side effect 89.7% vs 77.1%; serious side effects 7 vs 0 people (none considered drug-related); dizziness 23% vs 14.5%, dry mouth 9.2% vs 1.2%, diarrhea 11.5% vs 3.6%, increased appetite 16.1% vs 8.4%; stopping because of side effects 5.7% vs 7.2%
Nabilone patient records: in Fraser 2009, 13 of 47 patients (28%) stopped because of lightheadedness, forgetfulness, dizziness and headache; in Cameron 2014, side effects in 29.8% of 104 patients, psychosis in 2 people with a prior psychotic illness

Monitoring done in studies

Monitoring liver tests during long-term use is the doctor's decision

Who should be careful

Bipolar disorder: in mania, 600–1200 mg/day CBD was ineffective in 2 patients (Zuardi 2010, PMID 18801823); in bipolar depression, 150–300 mg/day was no better than placebo and no increase in mania was seen (Pinto 2024, PMID 37920963); there is no evidence that CBD works in treating mania
History of psychosis: no controlled data were found showing that CBD alone triggers psychosis, but in people with high paranoid tendencies, 600 mg showed a trend toward increased anxiety (Hundal 2018, PMID 29086614); the main risk is THC-containing products. A person with a history of psychosis should not use CBD without consulting a doctor (this is a precautionary recommendation, not a direct study finding)
Pregnancy and breastfeeding: no human studies show it is safe; in animals, signals of harm to prenatal development (Sarrafpour 2020, PMID 32524214; Argueta 2020, PMID 32425793)
Driving: 13.75 mg of vaporized CBD-dominant marijuana did not impair driving, but clinically important impairment could not be fully ruled out; CBD did not prevent THC's driving-impairing effect; after 300 mg oral CBD, attention lapse time lengthened (Arkell 2020, PMID 33258890; Rudisill 2023, PMID 36625844)
Liver: in healthy adults, ALT/AST >3×ULN in 5.6% with 5 mg/kg/day (Florian 2025, PMID 40622698); with 300 mg/day, 4 liver enzyme elevations, 1 of them critical (Crippa 2021, PMID 34387679)
Young people (ages 12–25): the only data are open-label, with side effects in 80.6% of participants (Berger 2022, PMID 35921510); an interaction has been shown with citalopram/escitalopram, which are commonly used at this age
Allergic skin reaction: 1 severe drug rash with 300 mg/day (Crippa 2021, PMID 34387679)
Athletes: in healthy volunteers who used 150 mg a day of broad-spectrum CBD oil for 10 weeks, CBG and CBDV, which are prohibited under WADA's 2026 list, were detected in urine (42% and 68% of pre-exercise samples, 74% and 84% of post-exercise samples); THC was not detected. CBDV was not listed on the product's certificate of analysis (Gillham 2026, PMID 40920736)
No study has shown that CBD is a first-line treatment for anxiety disorders
Cannabis (THC) use: in veterans with PTSD, those who started using cannabis after treatment had more severe symptoms and violent behavior (Wilkinson 2015, observational); the 2023 guideline of the US Departments of Veterans Affairs and Defense makes a strong recommendation against cannabis and its derivatives in the treatment of PTSD (Schnurr 2024). In records of people treated with nabilone, psychosis occurred in 2 people with a prior psychotic illness (Cameron 2014)

Drug interactions

Citalopram/escitalopram: when CBD (200–800 mg/day) was added, citalopram blood levels rose significantly in 6 patients; in the laboratory, breakdown via CYP3A4 and CYP2C19 was inhibited (Anderson 2021, PMID 34121064)
Sertraline, fluoxetine, mirtazapine: in the laboratory, CBD affected their breakdown very little; not measured in patients (Anderson 2021, PMID 34121064)
In people taking serotonergic antidepressants, CBD affected the re-extinction of subjective fear worse than placebo (Kwee 2024, PMID 38044339, exploratory subgroup)
Clobazam: in 13 children with epilepsy, the active metabolite N-desmethylclobazam rose by an average of 500%; the rise in clobazam itself (average 60%) was uncertain (confidence interval −2 to 91). In healthy volunteers, clobazam did not change and the metabolite rose ~3-fold (FDA/EMA). Side effects in 10 of the 13 patients, which resolved when the clobazam dose was reduced (Geffrey 2015, PMID 26114620)
Other benzodiazepines (alprazolam, clonazepam, etc.): no direct interaction study was found in the anxiety population
THC-containing products: CBD does not consistently reduce the effects of THC; full-spectrum products may contain THC (e.g., 0.23 mg/mL)
Antipsychotic + 1000 mg/day CBD (schizophrenia) was well tolerated (McGuire 2017, PMID 29241357)
No direct interaction study was found with venlafaxine, duloxetine, pregabalin, buspirone or hydroxyzine

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

CBD does not replace the therapies and medicines that have been shown to work for PTSD; do not stop your treatment on your own.
If you take an antidepressant, a sedative or another mental health medicine, talk to your doctor or pharmacist about CBD.
If you are struggling with alcohol or other substances, share this with your doctor too.
For long-term use, ask about liver tests; there are no human studies showing that CBD is safe during pregnancy and breastfeeding.
Check the label and the certificate of analysis (COA); 'full-spectrum' products may contain THC.

Choose your age group, medicines and situation to prepare a short summary you can take to your doctor. Nothing you enter is sent anywhere.

Open the personal warnings tool Convert drops ↔ mg

Bottom line

So far, controlled studies of CBD in PTSD have not shown a clear difference. In a large study of a medicine containing THC (dronabinol), PTSD-related nightmares decreased; but this is a single study (the evidence for nightmares is moderate) and the medicine is a THC drug, so the finding cannot be carried over to CBD products. Research is ongoing, and this picture may change in the future. For now, the most reliable support is ongoing treatment with a specialist experienced in PTSD.

Frequently asked questions

Does CBD oil work for PTSD?

Studies so far do not support this. In studies where patients were randomly assigned to groups, CBD did not clearly reduce anxiety or PTSD symptoms compared with an inactive pill (placebo). The only study reporting a positive result had 11 participants and no comparison group. In a study with smoked cannabis, no type of cannabis, including the high-CBD one, was better than placebo. A large study of a THC-containing medicine (dronabinol) for nightmares gave a positive result; this is a single study, and the finding does not apply to CBD oil.

Does CBD help if it is used together with therapy?

This idea is being studied. In a laboratory study in healthy volunteers, CBD strengthened the learning that allows fear to fade; however, as of October 2026, there is no controlled study showing that CBD added to therapy helps people with PTSD. In a 21-person pilot study whose results are only in the registry, both groups improved clearly and two scales pointed in opposite directions; results from a larger 96-person study of veterans (NCT03518801) were also released in October 2026, only in the trial registry: symptom scores fell in both groups (at the end of treatment, 22.6 in the CBD group and 23.1 in the placebo group on a 0–80 scale), but the groups were not compared statistically and there is no peer-reviewed publication. Laboratory studies with THC in people with PTSD also found no difference in the fear response; results of a study in which THC was added to therapy (NCT04080427) have not been released either.

Can I use CBD together with my PTSD medicine?

Make this decision with your doctor. CBD can slow the breakdown of some medicines in the body; for example, blood levels of citalopram and escitalopram may rise. CBD does not replace your medicine; do not stop your medicine on your own.

Does CBD help if someone has an alcohol problem along with PTSD?

In a 6-week study of people with alcohol use disorder and PTSD, CBD was no better than placebo for alcohol use, alcohol craving, anxiety or PTSD symptoms.

Can CBD oil for PTSD be obtained with a prescription in Turkey?

In Turkey, cannabis-derived products are governed by the Kenevirden Elde Edilen Ürünlere Dair Yönetmelik (Regulation on Products Derived from Hemp) dated January 31, 2026. These products, including CBD, can be sold only in pharmacies. Those classed as medicines and health products are dispensed with a prescription; for supplement products and cosmetics with a THC content of no more than 0.3%, the regulation does not state a prescription requirement. As of October 2026, there is no approved (licensed) CBD medicine in Turkey. Dronabinol (synthetic THC) and nabilone (a synthetic THC-like medicine), which were tested in PTSD studies, are not licensed in Turkey either: the list of licensed medicines of the Turkish Medicines and Medical Devices Agency (TİTCK) (October 16, 2026) contains no medicine containing cannabinoids, and dronabinol and nabilone are also not on the agency's list of active substances that can be brought in from abroad (September 28, 2026). If you are thinking of using it, talk with your doctor first.

Does using cannabis make PTSD worse?

There is no definite answer, because most studies on this are observational. In a US study following 2,276 veterans, those who started using cannabis after intensive treatment had more severe symptoms, while those who stopped had milder ones; in another study of 361 veterans, cannabis use was linked to more intense intrusive memories in the following months. There are also studies reporting that cannabis users improved, but these have no randomly assigned comparison group. In data from 410 people with PTSD who also had alcohol or substance problems, cannabis users benefited from trauma-focused therapies as much as non-users; in other words, using cannabis does not appear to prevent people from benefiting from these therapies. If you use it, it helps to talk about it openly with your treatment team.

References

Show 27 references

Bolsoni 2022a. Double-blind RCT. PMID 35029706, doi:10.1007/s00213-021-06043-y
Bolsoni 2022b. Exploratory secondary analysis of the same sample (1:1 randomization by minimization; no correction for multiple comparisons). PMID 35293520, doi:10.1590/1516-4446-2021-2317
Gournay 2024. Double-blind, placebo-controlled RCT. PMID 38959155, doi:10.1002/jts.23072
Bogenschutz 2026. Two double-blind, placebo-controlled RCTs. PMID 41312717, doi:10.1111/acer.70212, NCT03252756, results in the registry)
Elms 2019. Retrospective case series, flexible dose. PMID 30543451, doi:10.1089/acm.2018.0437
Straud (NCT05132699, registry results). Double-blind, placebo-controlled pilot RCT (phase 1/2); added to 10 sessions of intensive prolonged exposure therapy over 2 weeks; no peer-reviewed results publication. NCT05132699, results October 10, 2024), protocol: Straud 2024, PMID 38404650
Das 2013 (fear extinction). Double-blind, placebo-controlled. PMID 23307069, doi:10.1007/s00213-012-2955-y
Bonn-Miller 2021. Double-blind, placebo-controlled, 2-stage crossover RCT. PMID 33730032, doi:10.1371/journal.pone.0246990, NCT02759185
Raminelli 2025. Systematic review and meta-analysis (21 placebo-controlled RCTs; SAD 5, GAD 1, PTSD 7, OCD 1, Tourette 7). PMID 40956670, doi:10.1177/25785125251378242
Jetly 2015. Double-blind, placebo-controlled, randomized crossover pilot study. PMID 25467221, doi:10.1016/j.psyneuen.2014.11.002, Health Canada registration
Roepke 2026. Multicenter (Germany), double-blind, placebo-controlled, parallel-group RCT (phase 2). PMID 42557458, doi:10.1038/s41591-026-04546-9, NCT04448808, October 28, 2026), PMID 37147642
Ayers (NCT03518801, registry results). Double-blind (quadruple-masked), placebo-controlled, parallel-group RCT (phase 2); added to 12 sessions of prolonged exposure therapy; no peer-reviewed results publication. NCT03518801, results October 28, 2026), protocol and statistical analysis plan in the registry (version 1.44, November 6, 2023), collaborating institution University of California San Diego
Zabik 2023. Double-blind, placebo-controlled, parallel-group RCT (single dose; laboratory fear extinction study with brain imaging). PMID 37088409, doi:10.1016/j.nlm.2023.107758, NCT02069366, PMID 32162103, in response to threat images, THC reduced amygdala reactivity in people with PTSD) and Pacitto 2022, PMID 35981598, 51 people, THC reduced negative emotion during emotional reappraisal)
Zabik 2026. Double-blind, placebo-controlled, 3-arm RCT (single dose; laboratory fear extinction study with brain imaging). PMID 39412674, doi:10.1007/s00213-024-06702-w, NCT03008005, PMID 42463794, 37 people with PTSD), done under the same grant (R61MH111935) and with the same dose groups: THC reduced frontal brain activity during emotional reappraisal, negative emotion ratings did not change (abstract only)
Wilson 2026. Systematic review and meta-analysis of RCTs (1980–May 13, 2025; GRADE; PROSPERO CRD42023392718). PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged)
Aviram 2026. Scoping review (PRISMA-ScR; 5 databases and ClinicalTrials.gov, up to December 22, 2024; with quality assessment; not preregistered). PMID 42210342, doi:10.1186/s42238-026-00451-7
Rodas 2024. Systematic review (PubMed, PsycINFO, Embase; January 1990–February 2023; no meta-analysis). PMID 38353645, doi:10.4088/JCP.23r14862, PMID 38353646
Schnurr 2024 (VA/DoD 2023 guideline). Clinical practice guideline synopsis (US Department of Veterans Affairs and Department of Defense; systematic evidence review and GRADE). PMID 38408360, doi:10.7326/M23-2757, Full guideline: VA/DoD Clinical Practice Guideline for the Management of Posttraumatic Stress Disorder and Acute Stress Disorder, June 2023 (healthquality.va.gov), Recommendations 20 and 32–33
Morgenthaler 2018 (AASM). Position paper (American Academy of Sleep Medicine; literature search March 2009–August 2017 and expert opinion; a position paper instead of a guideline because the evidence is limited). PMID 29852917, doi:10.5664/jcsm.7178
Fraser 2009. Retrospective chart review (the author calls it an "open-label study"; no control group). PMID 19228182, doi:10.1111/j.1755-5949.2008.00071.x
Cameron 2014. Retrospective chart review (no control group). PMID 24987795, doi:10.1097/JCP.0000000000000180
Wilkinson 2015. Observational, prospective follow-up (national evaluation data from US Veterans Health Administration intensive PTSD treatment programs, 1992–2011). PMID 26455669, doi:10.4088/JCP.14m09475, PMID 33035104, in 1, 413 veterans using nonprescribed opioids, cannabis use neither improved nor worsened PTSD symptoms over 4 months)
Metrik 2022. Observational, prospective cohort (baseline, month 6 and month 12; cross-lagged panel analysis). PMID 32546286, doi:10.1017/S003329172000197X
Bonn-Miller 2022. Observational, prospective controlled cohort (cannabis users and non-users; not randomized). PMID 33998874, doi:10.1089/can.2020.0056
Hill 2024. Individual patient data meta-analysis (4 RCTs from Project Harmony, a set of 36 RCTs on treating PTSD and substance use disorder; cannabis use was not randomized and was weighted using propensity scores). PMID 38266511, doi:10.1016/j.janxdis.2024.102827
Lynskey 2024. Observational, prospective patient registry (Project Twenty21, UK; no control group). PMID 38468390, doi:10.1192/bjo.2024.13, PMID 39474239, 58 patients prescribed THC-rich flower, improvement in symptoms reported at months 3 and 6, all 3 authors work at the cannabis company Khiron or a clinic linked to it)
Datta 2025. Observational, prospective patient registry (UK Medical Cannabis Registry; no control group). PMID 40235073, doi:10.1080/14737175.2025.2490539, PMID 36503404, 162 patients, 6 months, 88.9% had used cannabis before, a total of 220 side effects in 33 people (20.4%), most common insomnia and tiredness)

Points that could not be verified from the full text

Bolsoni 2022a, Gournay 2024, Das 2013 and Raminelli 2025 from the PubMed abstract only (full text not accessible; Gournay is open access, but the publisher's page blocks automated access); the other studies were checked against the full text (October 19, 2026)
Results of NCT05132699 and NCT03518801 are only on ClinicalTrials.gov (no peer-reviewed publication, no between-group statistics given). Results of NCT03518801 were entered in the registry on October 28, 2026; the registry states that 136 people were enrolled, while in the results tables 96 people start the study (the difference is not explained in the registry); the registry gives no dose, and the dose and product come from the protocol attached to the registry (synthetic CBD capsule, 300 mg twice daily). The protocol is not internally consistent: the same section has a sentence saying 'target daily dose of 300 mg', and other sections have sentences saying the drug would be given under the tongue and the dose would be reduced in 0.10 mL steps; the dose patients actually took was not reported. Ongoing or not yet started (October 28, 2026): NCT04550377 (CBD, 120 people; estimated completion of June 2026 has passed, still 'recruiting'), NCT05269459 (CBD; estimated 2028; CBD dose in the registry unclear), NCT06381180 (THC and CBD, 500 people, estimated 2030), NCT07224698 (THC-rich inhaled cannabis, 320 people, estimated 2028), NCT06222268 (CBD, THC and both together, added to prolonged exposure therapy; 280 people, estimated 2031)
Controlled studies completed but with results not released as of October 2026: NCT04080427 (dronabinol 7.5 mg or placebo before exposure sessions, 102 people; May 2025), NCT05226351 (fear extinction with a single dose of dronabinol 2.5 mg in people with PTSD and healthy people, 192 people; August 2024). A dronabinol study using a trauma film in healthy female volunteers (NCT04871269, 291 people) is "active, not recruiting". A 150-person study comparing CBD isolate with broad-spectrum CBD (NCT04197102) was suspended because it required approval from the US drug agency. The nabilone study NCT03251326 was stopped early with 4 people; its data are not in the registry. A manufacturer-supported crossover RCT planned with vaporized inhaled cannabis (NCT02517424) ended with 6 people instead of the targeted 42; because no comparison with placebo could be made, only change from baseline was published (Walsh 2023), and it was not included in the table. The status of Hadassah's randomized THC study (NCT00965809) is "unknown".
Full text could not be read (October 28, 2026): Roepke 2026 (paywalled; primary and secondary outcomes, side effects and group guessing were read from the publisher's publicly available table pages), Haggarty 2026 (closed on PMC until October 2027; abstract only), Wilson 2026 and Rodas 2024 (abstract only; the numbers of the PTSD subanalysis could not be read; the correction published for Wilson 2026 was read on October 29, 2026 and does not change the results), Datta 2025 (abstract only; product breakdown and conflict-of-interest statement could not be read). The 2019 PTSD recommendations of the International Society for Traumatic Stress Studies (ISTSS) could not be accessed (the site required human verification); in the drug meta-analysis commissioned by the society (Hoskins 2021), the only cannabinoid study included is Jetly 2015.
The status of dronabinol and nabilone in Turkey was read from two lists of TİTCK (Turkish Medicines and Medical Devices Agency) (List of Licensed Medicinal Products for Human Use, 16.10.2026; List of Active Substances from Abroad, 28.09.2026; downloaded on 29.10.2026). Whether a medicine that is not on the list can be brought in some other way was not researched in this review; the Turkish Pharmacists' Association's list of medicines from abroad was not searched because its current version could not be found.

Trusted sources from around the world

Show 2 references

VA/DoD 2023: Clinical practice guideline for PTSD and acute stress disorderUS Department of Veterans Affairs and Department of Defense (VA/DoD), English, June 2023
This guideline from the US veterans and defense departments makes a strong recommendation against cannabis and cannabis-derived products for the treatment of PTSD (very low confidence in the evidence).
CADTH: Nabilone for the treatment of PTSD, 2023 update (rapid review)CADTH (Canadian Agency for Drugs and Technologies in Health), English, February 2023
A 2023 rapid review by Canada's drug assessment agency on nabilone, a synthetic THC-like medicine, for the treatment of PTSD (an update of its earlier reports).

These sources sell no products. All trusted sources

Research history

Here we record what changed on this page, when and why, with dates. The newest change is at the top.

November 1, 2026
The evidence level text now uses the same level name as the top of the page; the level and the result did not change.
October 31, 2026
The last sentence of the page's short summary was rewritten in calmer, more guiding language: instead of "Talk to your doctor before you start.", it now says who you can talk with, and about what, if you are considering CBD. The result and the strength of evidence did not change.
October 31, 2026
The Turkey question now adds that dronabinol and nabilone, which were tested in PTSD studies, are not licensed in Turkey: the list of licensed medicines of the Turkish Medicines and Medical Devices Agency (October 16, 2026) contains no medicine containing cannabinoids, and neither drug is on the agency's list of active substances from abroad (September 28, 2026) either. The result and the strength of evidence did not change.

Older changes (9)

October 30, 2026
The correction to the Wilson 2026 meta-analysis published in the journal was read: only the axis of one figure and the study counts in a table were corrected; the results on this page did not change.
Related studies: Wilson 2026, published April 2026 (PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged))
October 28, 2026
October 2026 additional review: sources found for this page during the searches for other topics were read and incorporated into the page. Results of the study in which CBD or placebo was added to prolonged exposure therapy in veterans (NCT03518801) were entered on ClinicalTrials.gov on October 28, 2026; the study was added to the table with its registry results (96 people started; symptoms decreased in both groups, the groups were not compared statistically, no peer-reviewed publication), and the 'results not released' sentences were corrected. Because the nightmare studies also appear on the insomnia page, a reference to that page was added. The conclusion and the strength of evidence did not change.
Related studies: Ayers (NCT03518801, registry results) (NCT03518801, results October 28, 2026), protocol and statistical analysis plan in the registry (version 1.44, November 6, 2023), collaborating institution University of California San Diego)
October 28, 2026
Display fix: internal working notes left over from the previous cleanup were removed; content and evidence did not change.
October 28, 2026
In the October 2026 search (PubMed, Cochrane, ClinicalTrials.gov), 16 studies were added to this topic: 3 controlled studies (a 171-patient multicenter dronabinol study in PTSD-related nightmares, two laboratory studies with THC in people with PTSD), 3 systematic reviews, 2 guidelines (the 2023 guideline of the US Departments of Veterans Affairs and Defense, American Academy of Sleep Medicine 2018) and 8 observational studies (nabilone patient records, cannabis use and the course of PTSD in veterans, therapy outcomes in cannabis users, two patient registries in the UK). The conclusion and evidence text for THC-containing products were updated in light of the new nightmare study: two RCTs positive for nightmares ('moderate' in the evidence text), no clear difference in overall PTSD symptoms ('weak'). The two-way nature of the real-world data, the guideline recommendations, and registrations with unreleased results or left unfinished were added to the text. Registration numbers, PTSD outcome, funding, company ties and blinding information were added to five existing rows; the product type of the smoked cannabis study was corrected to 'THC-containing'. The conclusion and evidence labels ('Weak' for CBD, 'Weak' for THC-containing products) did not change.
Related studies: Roepke 2026, published September 2026 (online August 5, 2026) (PMID 42557458, doi:10.1038/s41591-026-04546-9, NCT04448808, October 28, 2026), PMID 37147642); Zabik 2023, published May 2023 (PMID 37088409, doi:10.1016/j.nlm.2023.107758, NCT02069366, PMID 32162103, in response to threat images, THC reduced amygdala reactivity in people with PTSD) and Pacitto 2022, PMID 35981598, 51 people, THC reduced negative emotion during emotional reappraisal)); Zabik 2026, published February 2026 (online October 16, 2024) (PMID 39412674, doi:10.1007/s00213-024-06702-w, NCT03008005, PMID 42463794, 37 people with PTSD), done under the same grant (R61MH111935) and with the same dose groups: THC reduced frontal brain activity during emotional reappraisal, negative emotion ratings did not change (abstract only)); Wilson 2026, published April 2026 (PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged)); Aviram 2026, published May 29, 2026 (PMID 42210342, doi:10.1186/s42238-026-00451-7); Rodas 2024, published February 14, 2024 (PMID 38353645, doi:10.4088/JCP.23r14862, PMID 38353646); Schnurr 2024 (VA/DoD 2023 guideline), published March 2024 (PMID 38408360, doi:10.7326/M23-2757, Full guideline: VA/DoD Clinical Practice Guideline for the Management of Posttraumatic Stress Disorder and Acute Stress Disorder, June 2023 (healthquality.va.gov), Recommendations 20 and 32–33); Morgenthaler 2018 (AASM), published June 15, 2018 (PMID 29852917, doi:10.5664/jcsm.7178); Fraser 2009, published 2009 (PMID 19228182, doi:10.1111/j.1755-5949.2008.00071.x); Cameron 2014, published October 2014 (PMID 24987795, doi:10.1097/JCP.0000000000000180); Wilkinson 2015, published September 2015 (PMID 26455669, doi:10.4088/JCP.14m09475, PMID 33035104, in 1, 413 veterans using nonprescribed opioids, cannabis use neither improved nor worsened PTSD symptoms over 4 months)); Metrik 2022, published February 2022 (online June 17, 2020) (PMID 32546286, doi:10.1017/S003329172000197X); Bonn-Miller 2022, published April 2022 (online December 9, 2020) (PMID 33998874, doi:10.1089/can.2020.0056); Hill 2024, published March 2024 (PMID 38266511, doi:10.1016/j.janxdis.2024.102827); Lynskey 2024, published March 12, 2024 (PMID 38468390, doi:10.1192/bjo.2024.13, PMID 39474239, 58 patients prescribed THC-rich flower, improvement in symptoms reported at months 3 and 6, all 3 authors work at the cannabis company Khiron or a clinic linked to it)); Datta 2025, published May 2025 (online April 18, 2025) (PMID 40235073, doi:10.1080/14737175.2025.2490539, PMID 36503404, 162 patients, 6 months, 88.9% had used cannabis before, a total of 220 side effects in 33 people (20.4%), most common insomnia and tiredness))
October 27, 2026
A small controlled study of nabilone, a THC-like synthetic medicine, in PTSD-related nightmares (Jetly 2015) was added; the conclusion for THC-containing products was split in two: overall symptoms and nightmares. The conclusion and the strength of evidence did not change.
October 22, 2026
The names of the finding labels were simplified: this page's "No benefit seen" label became "No effect found". The assessment did not change.
October 21, 2026
PubMed and ClinicalTrials.gov were searched for newly published and previously missed clinical studies. Study added: Bonn-Miller 2021. For pure CBD, the conclusion and the strength of evidence did not change; for THC-containing products, the strength of evidence was stated for the first time, as "Weak".
Related studies: Bonn-Miller 2021, published March 17, 2021 (PMID 33730032, doi:10.1371/journal.pone.0246990, NCT02759185)
October 21, 2026
The full texts of the studies were reread and the information on the page was checked against its sources. Corrected: product, number of participants, side effect and dose information in the study table; drug interaction notes; the dose range in the studies; the "Who should be careful" list; reported side effects; the evidence summary. Study added: Straud (NCT05132699, registry results). Points that could not be verified from the full text went from 3 items to 2. The year of 7 publications named on the page was corrected to the journal's print year instead of the online publication year. The conclusion and the strength of evidence did not change.
Related studies: Straud (NCT05132699, registry results) (NCT05132699, results October 10, 2024), protocol: Straud 2024, PMID 38404650)
October 19, 2026
Page published. This first review looked at 7 studies.

Evidence changes across all topics