Do we finally have positive data on GLP-1’s in depression?
STUDY: Gill H et al, JAMA Psychiatry 2026; doi:10.1001/jamapsychiatry.2026.0594
STUDY TYPE: Randomized, double-blind, placebo-controlled trial (secondary analysis)
FUNDING: Physicians’ Services Incorporated (PSI) Foundation
Background
Recently, I reported on the first randomized trial of a GLP-1 agonist in depression. It didn’t work. Now the authors return with a re-analysis of the data, asking whether it improved anhedonia (loss of motivation and pleasure).
Normally a journal with JAMA’s prestige would not publish a secondary analysis of a small trial, but it’s a hot and contested topic. For example, there are anectodal reports of apathy on GLP-1 agonists, raising questions about whether these meds might worsen anhedonia. As we’ll see, the results have more to due with persistence than response to rewards.
The Study
- 72 overweight adults with major depressive disorder (BMI 25 or higher)
- 16-week, double-blind, placebo-controlled trial
- Randomized to oral semaglutide (14 mg/day, titrated from 4 mg) or placebo, added to their existing treatment
- Anhedonia measured by Effort-Expenditure for Rewards Task (EEfRT), in which participants choose between low-effort tasks for small rewards versus high-effort tasks for larger ones (also a test of persistence)
On semaglutide, patients became progressively more willing to exert effort when the expected reward was higher, a pattern not seen in the placebo group. Semaglutide reduced “effort discounting,” meaning patients perceived the cost of physical effort as lower relative to the reward on offer. Sensitivity to reward probability was unchanged, suggesting the drug targeted effort cost specifically, not general optimism about outcomes. These effects were independent of any changes in overall depression severity.
Limitations: Small, secondary analysis, authors note there was likely functional unblinding (participants could probably guess whether they were on active drug).
Practice Implications
- Is this a random finding? Like a gambler who wants to re-roll the dice, secondary analyses break the laws of statistics. Interpret it as observational, not truly randomized.
- The original study was negative for depression and cognition, and another new analysis found benefits in well-being but not depression on GLP-1 agonists.
— Chris Aiken, MD
Director, Psych Partners
Editor in Chief, Carlat Psychiatry Report
