High-CBD Anxiety Treatment: Why Four-Week Improvement Does Not Isolate CBD

forensic toxicology working group members October 28, 2026 Forensic Toxicology

Fourteen patients reported less anxiety after four weeks of a high-CBD cannabis-derived preparation, with improvements on selected cognitive measures. The open-label study supplies preliminary human evidence, but cannot isolate CBD from expectancy, repeated testing or other constituents.

The preparation was not pure CBD

Patients used a full-spectrum sublingual solution three times daily. The supplied formulation contained 9.97 mg/mL CBD and 0.23 mg/mL THC. The intended daily exposure was approximately 30 mg CBD and less than 1 mg THC.

Describing this as a trial of isolated CBD would change the intervention. Absence of reported intoxication does not establish absence of THC exposure.

Clinical and cognitive outcomes

Anxiety ratings improved significantly relative to baseline. The investigators also reported improvements in mood, sleep, quality of life and executive-function measures. There were no reported serious adverse events or intoxication; minor effects included sleepiness or fatigue.

The cognitive battery was repeated at baseline and week four. Without an appropriate blinded control group in this stage, practice and changes in anxiety itself remain plausible contributors to improved test performance.

What remains uncertain

This was a small, short, open-label stage, not the definitive placebo-controlled stage described by the authors. A response definition based on a specified percentage reduction is also not synonymous with complete remission.

The paper disclosed outside cannabis-related research and advisory interests. Those disclosures should accompany assessment of the preliminary findings.

For a functional assessment, the study cannot certify an individual’s driving or establish a blood concentration indicating benefit. Its central lesson is that clinical improvement, a cognitive score and the concentration of an individual cannabinoid answer different questions.