Review pools DMT studies for substance misuse and finds large but unreliable effects

The question

Can DMT (and the related compound 5-MeO-DMT) help people reduce problem drinking or drug use? The authors pooled results from all the studies they could find.

What they did

The researchers searched four research databases (PubMed, PsycINFO, Web of Science and EBSCO) for studies of DMT interventions published between 1960 and 2024. They combined the results statistically to estimate an overall effect on substance use. They also ran extra analyses to see whether adding psychotherapy, the type of population, or the study design changed the results. The abstract does not say how many studies or people were included.

What they found

Pooled across studies, DMT was linked to a large reduction in substance use (effect size g = 0.94, 95% confidence interval 0.56-1.31, p < 0.0001). The effect looked bigger for drug use (g = 1.35, CI 0.63-2.07) than for alcohol use (g = 0.65, CI 0.31-0.99). Studies that paired the drug with psychotherapy showed larger effects (g = 1.38, CI 1.06-1.71) than those without it (g = 0.60, CI 0.09-1.12), a statistically significant difference (p = 0.0121). The authors also reported high risk of bias and very high inconsistency between studies (I2 = 96.9%), though no sign of publication bias.

The fine print

The abstract does not report the number of studies or participants pooled, so the size of the evidence base is unclear. The authors themselves flag substantial methodological limitations and high risk of bias in the included studies, plus extremely high heterogeneity, meaning results varied widely from study to study. The psychotherapy comparison came from exploratory subgroup analysis, which can generate ideas but not firm conclusions. This summary is based on the abstract only.

What it means

Pooled early studies point to a possible benefit of DMT-type psychedelics for reducing alcohol and drug use, especially alongside psychotherapy, but the underlying research is weak and inconsistent enough that the authors call the findings preliminary rather than established.

Evidence grade: Low. A meta-analysis whose included studies had high risk of bias and very high inconsistency between them, and the abstract does not report how many people were pooled.

This summary is based on the study’s published abstract.