CBD Gel for Bruxism: Why Similar Results Do Not Prove Equal Strengths

forensic toxicology working group members October 28, 2026 Forensic Toxicology

A small controlled trial reported improvements with both 5% and 10% CBD gels, without a significant difference between strengths. That is not a formal demonstration that the two strengths are therapeutically equivalent.

The intervention and outcomes

Sixty participants with bruxism-associated muscular pain were assigned equally to placebo, 5% CBD gel or 10% CBD gel. The gels were applied intraorally over a 30-day period. Outcomes included sleep-quality and migraine-disability questionnaires, with muscle-activity assessments also reported.

This was a specific localized formulation in a selected clinical population. Its results should not be generalized to all CBD creams, ordinary swallowed CBD or every cause of migraine and poor sleep.

A positive comparison and a null comparison

The CBD groups showed significant questionnaire improvements compared with control. The comparison between CBD strengths was not significant. Those statements answer different questions: whether each treatment differed from placebo, and whether the study detected a difference between the active treatments.

Equivalence requires a justified margin and an analysis capable of excluding clinically important differences. Failure to detect a difference is insufficient, especially in a small sample.

Sample-size and mechanism limitations

The methods acknowledge a recalculation indicating approximately 26 participants per group would have been needed under the stated planning assumptions, compared with 20 actually studied. A post hoc power calculation does not erase uncertainty in the active-strength comparison.

Associations between reduced muscle activity and improved questionnaires do not establish the entire mechanism. The trial also did not establish long-term benefit, a systemic CBD concentration or a driving-impairment outcome. Its findings support further investigation of the particular intervention, with the strength of each inference matched to the design.