Weight-loss jabs show no link to suicidal behaviour across 25 trials

GLP-1 receptor agonists have become some of the most widely used drug classes in the world for type 2 diabetes and obesity, and questions have circulated about whether they might be linked to suicidal behaviour. For anyone weighing up these medications, that is among the most serious safety concerns imaginable, and it deserves an answer from randomised evidence rather than anecdote or isolated reports. Individual trials are too small to judge a rare outcome like this, which is why pooling them matters. This analysis set out to test whether GLP-1 exposure is associated with suicidal behaviour across the trial record.

The authors searched PubMed, Web of Science, the Cochrane Library and ClinicalTrials.gov from the start of each database, gathering randomised controlled trials of GLP-1 receptor agonists in people with type 2 diabetes or obesity. Twenty-five trials qualified for inclusion. They calculated risk ratios comparing suicidal behaviour in people assigned to a GLP-1 drug against those assigned to a comparator. They then ran subgroup analyses by condition, by age group, by type of suicidal behaviour, by individual drug and by type of comparator.

The headline result was a risk ratio of 0.84, meaning no significant difference in suicidal behaviour between the GLP-1 group and controls. Every subgroup told the same story: no significant difference in people with type 2 diabetes or with obesity, in adolescents or in adults, and no difference for suicidal ideation, suicide attempts, depression-related suicides or completed suicides. The null also held for each drug examined, including dulaglutide, exenatide, semaglutide, lixisenatide and liraglutide, and against both placebo and other comparators. Strikingly, the included trials showed no disagreement with each other on this question.

This pooled analysis speaks directly to a safety concern that has followed GLP-1 drugs into mainstream use, and its answer is consistently null across 25 randomised trials. That is meaningful reassurance, though it comes with limits. Suicidal behaviour is a rare event, so even large pooled datasets can struggle to detect small increases in risk, and trial participants may differ from the broader population now taking these drugs. The findings strengthen, rather than finally settle, the case that no association exists.

If concern about suicidal thoughts has been part of your decision-making around GLP-1 medication, the randomised trial record offers no support for that worry, in adults or adolescents, and for any of the drugs studied here. That is worth knowing when weighing the benefits and risks of these treatments. It does not replace an individual conversation with a prescriber, but it removes one reason for alarm.