Fitness rose, endocannabinoids didn't budge

Exercise is routinely recommended alongside psychological therapy for depression, and the endocannabinoid system has become a favoured explanation for why movement lifts mood — the biological story that replaced endorphins in popular accounts. If that account is right, an exercise intervention that works should leave a fingerprint in circulating endocannabinoids. Testing this properly requires a controlled trial that measures both the fitness response and the biomarkers in the same people over the same period, which is what this study set out to provide.

Eighty adults with mild-to-moderate depression were randomly assigned 1:1 to a 12-week combined physical exercise and internet-based cognitive behavioural therapy intervention or to a control group. Outcomes included nine endocannabinoid system biomarkers — 2-arachidonoylglycerol and anandamide plus seven analogues — and physical fitness, with cardiorespiratory fitness assessed by the six-minute walking test and muscular strength measured separately. Measurements were taken at baseline, immediately after the intervention and again at an 8-week follow-up. The primary analysis used repeated-measures ANOVA on completers, with a mixed-effects intention-to-treat model as a sensitivity check.

Cardiorespiratory fitness improved significantly in the intervention group, with a between-group difference of 81.6 m on the six-minute walk at 12 weeks, a gap the authors note exceeds the established minimal clinically important difference. No significant time-by-group interaction appeared for any of the nine endocannabinoid biomarkers, including 2-AG and anandamide; both groups showed comparable within-group decreases in several of them. None of the exploratory muscular fitness outcomes reached statistical significance, and the arm curl test produced only a borderline non-significant interaction.

The fitness result confirms that a remotely delivered exercise and therapy package can produce a meaningful aerobic gain in people with depressive symptoms, which supports the case for online delivery where in-person services are stretched. The biomarker null complicates the popular narrative: within this trial, changes in circulating endocannabinoids did not distinguish the group that got the intervention from the group that did not. That does not disprove a role for the system, since plasma levels may not reflect what happens in the brain, and the trial was not powered as a definitive mechanistic test.

This is a reason to think an internet-delivered mix of exercise and CBT can measurably improve aerobic fitness in people with mild-to-moderate depression, and a reason to be sceptical of confident claims that blood endocannabinoids explain how exercise makes you feel. One trial of 80 people cannot settle a mechanism, and the absence of a biomarker signal says nothing about whether the intervention helps someone feel better. Treat mechanism stories in fitness marketing as provisional until they survive controlled tests like this one.