Meta-analysis · 2026
Ketamine and substance use disorders: pooled results from 15 randomized trials
- Original title
- Ketamine for substance use disorders: a systematic review and meta-analysis.
- Authors
- Fang SP, Yang X, Zhao DC, et al.
- Journal
- Frontiers in psychiatry, 2026
- Molecules
- Ketamine
- Conditions
- Substance use disorder
- Participants
- 798
- Evidence grade
- Moderate
- Source
- Read the original paper
The question
Does ketamine help people with substance use disorders stay abstinent, and is it well tolerated?
What they did
The researchers searched several research databases for randomized controlled trials testing ketamine as a treatment for substance use disorder. Fifteen trials with 798 participants in total met their criteria. Seven of those trials reported abstinence data that could be pooled statistically. The team also compared rates of side effects and of people dropping out between ketamine and control groups.
What they found
In the short term (under one month), people given ketamine were more likely to be abstinent than controls (odds ratio 3.27, 95% confidence interval 1.55 to 6.92, p < 0.01). At one to six months the difference was no longer statistically significant (odds ratio 1.74, 95% confidence interval 0.91 to 3.30, p = 0.09). Dropout rates did not differ between groups (odds ratio 0.74, 95% confidence interval 0.46 to 1.21, p = 0.23). No significant difference in adverse events was found, but the authors say the safety data were too limited to draw firm conclusions.
The fine print
Only seven of the fifteen trials contributed to the main abstinence analysis, so the pooled sample is small. The trials likely differed in the substances targeted, ketamine doses and any accompanying therapy, which the abstract does not describe. Safety reporting was sparse, and longer-term benefit was not demonstrated. This summary is based on the abstract alone.
What it means
Pooled trial data point to a short-lived signal that ketamine may help with abstinence in substance use disorder, but the effect was not clear beyond a month and the evidence is not strong enough to settle questions of benefit or safety.
Evidence grade: Moderate. A meta-analysis of randomized trials, but only seven of the fifteen trials provided data on the main outcome and the total number of participants was modest.
This summary is based on the study’s published abstract.