The most publicized worry about these drugs in young people is suicidality. This study addresses it directly in 9,222 matched pairs of adolescents, and the reassuring finding holds however you cut it: no increase. The more interesting part is what happened when the authors changed who they compared against.
The reassuring answer first
Adolescents with overweight or obesity starting semaglutide, liraglutide or tirzepatide were matched against adolescents receiving lifestyle interventions, using global electronic health records, and followed from 30 days to about three years after starting. [1] Incretin therapy was not associated with increased risk of any psychiatric outcome measured.
Suicide attempt, insomnia, depression and eating disorders all showed no significant difference. That is the finding most families are looking for, and it is consistent with what the pooled randomized evidence in children found over its much shorter follow-up.
Then it looked protective
Against lifestyle intervention, the drugs were associated with lower risks: suicide-related events HR 0.74, 95% CI 0.57 to 0.95; suicidal ideation HR 0.73, 95% CI 0.56 to 0.96; and anxiety HR 0.92, 95% CI 0.84 to 0.99.
The reductions were significant among girls, HR 0.61, 95% CI 0.46 to 0.82, and among adolescents without type 2 diabetes, HR 0.63, 95% CI 0.47 to 0.85. They were not significant among boys or among those with type 2 diabetes. The study reports those splits without explaining them, and so does this page.
Why the comparator decides so much
An adolescent prescribed a GLP-1 and an adolescent enrolled in a lifestyle program differ before either starts. One has a prescriber willing to medicate, usually insurance, and a family navigating specialty care. The other may have been offered the cheaper option, or the only one available. Those differences predict mental health outcomes on their own.
Metformin narrows that gap. It is also a prescription, also given for metabolic reasons, also requiring a clinician. When the difference disappears under that comparison, the most economical explanation is that some of what looked like a drug effect was a difference in who receives drugs — the point comparator choice makes in almost every observational study this desk reads.
What to carry away
The safety question has a clear answer and the efficacy-adjacent claim does not. Nothing here suggests these drugs raise psychiatric risk in adolescents, across two comparators and seven outcomes. Nothing here establishes that they lower it either.
For adults the picture is separate and has its own longer history. And nobody on this roster ships to anyone under eighteen or offers a pediatric service, so this is a conversation for a family and a clinician — one in which the questions worth asking are the ones no intake form covers.