Dronabinol Cuts PTSD Nightmares in New German Trial
A Charite Berlin trial of more than 170 patients found that THC drops eased PTSD nightmares more than placebo did after ten weeks. PTSD itself and depression were unaffected.
THC drops available on prescription in Germany eased nightmares in people with post-traumatic stress disorder (PTSD) more than a placebo did, a clinical trial led by Berlin’s Charite university hospital has found. The paper was published on August 5, 2026, in the journal Nature Medicine. The trial enrolled more than 170 patients with frequent, intense nightmares, ran for ten weeks, and produced its clearest result on the scale that measures how much nightmares burden a patient.
Nightmares are among the most treatment resistant PTSD symptoms. Patients relive the traumatic event at night, wake up in a state of panic, and over time can develop a fear of going to bed at all. No drug is approved in Germany specifically for this indication. Doctors instead prescribe antidepressants and blood pressure medication, with only partial success. The best known drug in that second group is prazosin, used for years in war veterans, but a large US trial published in 2018 in the New England Journal of Medicine found it performed no better than placebo. After that setback, researchers turned their attention to a different target in the brain.
That target was pursued by a team led by Prof. Dr. Stefan Repke of Charite’s Department of Psychiatry and Neurosciences, on the Benjamin Franklin campus, in the endocannabinoid system, which helps regulate sleep, stress, and the processing of emotional memories.
“There are indications that during REM sleep, the intense dreaming phase in which the brain processes experiences and regulates emotions, THC dampens dream activity and thereby eases nighttime stress reactions,” Repke explained.
How the Trial Was Designed
Every evening before bed, patients took dronabinol, THC derived from the cannabis plant, in drop form. The control group received a liquid that looked identical and tasted like cannabis, so taste alone could not reveal who was on the drug. Neither patients nor doctors knew who was in which group, making this a double blind, randomized, placebo controlled trial. Besides Charite, the study involved the University Medical Center Hamburg Eppendorf and the Central Institute of Mental Health in Mannheim.
Nightmare burden was measured on a scale from zero to eight points. After ten weeks, the dronabinol group’s score fell by an average of 3.7 points, against 2.2 points for the placebo group.
“More than a third of patients treated with dronabinol reported that after ten weeks they no longer had nightmares at all. Another 21 percent said their nightmare burden had at least been cut in half.”
About 83 percent of patients in the drug group, that is five out of six, rated their health as markedly improved.
How Much of That Is Really the Drug
The placebo group dropped by 2.2 points over the same ten weeks, about three fifths of the improvement measured on the drug. The difference attributable to dronabinol itself comes to 1.5 points on the eight point scale. That is still a shift a patient would feel, but the number shows how strong the placebo response is in sleep disorders, and why open label studies without a control group make it hard to draw conclusions about nightmares.
The safety data call for careful wording. Serious adverse events were recorded in seven of 87 patients on dronabinol, or 8 percent, versus none in the placebo group. Researchers did not link any of those cases to the drug. Mild and moderate effects, dizziness, headache, and increased appetite, occurred more often in the drug group. There were no withdrawal symptoms after the evening doses stopped.
Does This Mean THC Treats PTSD?
It does not. The cannabinoid acted on nightmares and sleep, but not on the disorder’s other symptoms. The severity of PTSD itself, and the depression that often accompanies it, did not change convincingly. The benefit lies in interrupting the nightly reliving of the trauma, so the patient can actually sleep.
The trial’s limits are clear. Follow up lasted ten weeks, so nothing is known about longer term use, including whether tolerance to the drug develops over time. That is the next step researchers say they plan to take. This is a single trial without replication. The study was funded by the company Bionorica SE, which does not invalidate the result but does factor into how it should be weighed. Finally, dronabinol is not an approved drug for this use. It is being prescribed off label, a decision made by the physician, not the patient.
