Dronabinol Eliminates PTSD Nightmares in Over One-Third of Patients in Phase III Trial
Key Insights
A ten-week randomized trial found dronabinol, a cannabis-derived THC medication, reduced PTSD nightmare scores by 3.66 points versus 2.17 for placebo, a statistically significant difference.
Complete nightmare remission was achieved in 36.8% of patients on dronabinol compared to 14.5% on placebo, with 57.9% seeing scores at least halved.
Sleep quality improved significantly, though clinician-rated overall PTSD severity and depression scores did not shift, suggesting a nightmare-specific benefit.
A nightly dose of dronabinol—the medical form of THC—eliminated trauma-related nightmares in more than one-third of patients with post-traumatic stress disorder (search), according to results from a randomized, placebo-controlled trial conducted across four German university medical centers and published in Nature Medicine.
The ten-week study found that 36.8% of patients receiving the cannabis-derived drug achieved complete nightmare remission by the end of treatment, compared with 14.5% in the placebo group. The finding addresses a long-standing treatment gap: nightmares affect 50% to 70% of PTSD patients and are linked to substance abuse, suicidal thoughts, and suicide, yet no drug is currently approved specifically for this symptom.
“What surprised me most was the magnitude of the treatment effect,” said Stefan Röpke, a psychiatrist at Charité Universitätsmedizin Berlin who designed and supervised the trial. “Not only did we observe a substantial reduction in trauma-related nightmares, but more than 85% of participants receiving THC also reported an overall improvement in their health by the end of the trial.”
Trial Design and Primary Endpoint
The THC-PTSD trial enrolled 171 adults randomly assigned to dronabinol or placebo, running from October 2020 to January 2025 at Charité Universitätsmedizin Berlin, University Medical Center Hamburg-Eppendorf, and the Central Institute of Mental Health in Mannheim. Participants averaged 37.9 years old, and 79.3% were women. Most were civilians whose trauma stemmed from physical or sexual violence. All had at least two nightmares per week at baseline, were on stable medication, and had no lifetime cannabis-use disorder.
Patients took the medication as oil drops approximately one hour before bed. Doses were titrated from 2.5 to 15 milligrams daily over the first four weeks, then held steady for six additional weeks. The placebo consisted of the same oil flavored with cannabis to match the active drug in appearance, smell, and taste.
The primary outcome measure was the nightmare item of the Clinician-Administered PTSD Scale, scored from 0 to 8 based on nightmare frequency and severity. Scores dropped by 3.66 points in the dronabinol group and by 2.17 points in the placebo group—a 1.50-point difference that was statistically significant with a moderate effect size.
Remission and Sleep Outcomes
Among patients assessed at week 10, 57.9% of those on dronabinol saw their nightmare score reduced by at least half, compared with 29% on placebo. Complete remission—defined as no nightmares at all—was reached by 36.8% of the dronabinol group and 14.5% of the placebo group. Every sensitivity analysis the research team conducted pointed in the same direction.
Sleep improvements extended beyond nightmare metrics. Both a PTSD-specific sleep questionnaire and a general sleep quality index favored the drug. Approximately 86% of the dronabinol group reported improvement on a global change scale at week 10, against 57% on placebo.
Clinician-rated PTSD severity, however, did not shift significantly, and neither did clinician-rated depression after adjusting for multiple comparisons. The benefit therefore appears specific to PTSD nightmares and sleep rather than to the disorder as a whole.
Safety Profile
Adverse events were common in both arms, occurring in 89.7% of the dronabinol group and 77.1% of the placebo group. Most were mild or moderate. Dizziness, increased appetite, diarrhea, and dry mouth were reported more frequently with the drug. Discontinuation due to side effects was similar between groups—5.7% on dronabinol versus 7.2% on placebo.
Seven serious adverse events occurred in the dronabinol arm and none in the placebo arm. Investigators judged all seven unrelated to the medication. Two were planned hospital admissions, and one involved surgery for a benign brain tumor. No deaths occurred, and no meaningful withdrawal symptoms followed treatment cessation.
Mechanism and Clinical Positioning
Dronabinol contains tetrahydrocannabinol, one of the primary psychoactive compounds found in cannabis. “There is evidence suggesting that during REM sleep—the intense dream phases in which the brain processes experiences and regulates emotions—THC reduces dream activity and thereby alleviates nocturnal stress responses,” Röpke explained.
Röpke emphasized that the drug should be positioned after existing treatments rather than replacing them. “If established psychotherapeutic and pharmacological treatments have not provided sufficient relief, our findings suggest that dronabinol could be considered as an additional treatment option for trauma-related nightmares,” he said.
The effect held across subgroups that often complicate PTSD care, including patients with complex PTSD and those with comorbid borderline personality disorder. “We found no evidence that efficacy differed between men and women, in patients with the ICD-11 diagnosis of complex PTSD, or in those with comorbid borderline personality disorder,” Röpke noted.
Limitations and Next Steps
The trial lasted only ten weeks, leaving questions about long-term durability, tolerance development, and prolonged safety unanswered. Blinding was imperfect: four out of five patients on dronabinol correctly guessed they had received the active drug, with most attributing their guess to perceived improvement. The authors urge caution regarding patient-reported outcomes, though the guesses did not interact significantly with the magnitude of benefit.
The sample was predominantly female civilians, so results may not generalize to combat-related PTSD. No objective sleep recordings were collected, and the sample size limits detection of rare adverse events.
“Our next priority is to establish whether the benefits persist over time and whether they remain after treatment is stopped,” Röpke said. Several other cannabinoid trials in PTSD are underway, including studies of inhaled formulations and one pairing the drugs with psychotherapy, though none targets nightmares alone as this trial did.
The investigator-initiated trial was funded by Bionorica SE through an unrestricted grant; the company supplied the medication but had no role in the trial's design, analysis, or publication.