Psychiatric safety concerns have followed these drugs for years, and the trial evidence is covered in the effort and reward trial. This study asked the plainer question in a population carrying substantial psychiatric burden: do depression scores get worse after starting [1].
Overall, no. Across 354 participants the change in questionnaire score was -0.1 points with a confidence interval from -0.7 to 0.5 — a tight interval around nothing, which is the most robust finding in the paper and a direct answer to the safety question that prompted it.
That caution is not a dismissal. A real improvement in severe depression is plausible — the randomized evidence on motivation points that way — and so is a small worsening in people who had nothing to improve. What this design cannot do is distinguish either from the pull toward the average, and only a control group could — the same limitation that shapes the clozapine cohort.
The population is specific and worth naming. These were people with HIV in routine care, mean age 54, with 78% carrying obesity and 60% diabetes, and their questionnaire scores were collected opportunistically during clinic visits rather than on a research schedule. The instrument is a screening questionnaire rather than a diagnosis, which limits what any change in it means clinically.
The useful conclusion is the narrow one the authors draw: starting semaglutide was not associated with worsening depressive symptoms in this group, across every subgroup examined including people already taking antidepressants. That sits alongside the other psychiatric findings on this site, the bipolar cohort and the mortality gap in serious mental illness, as reassurance rather than benefit.