Recent systematic review: A systematic review and meta-analysis published in Diabetes, Obesity & Metabolism evaluated the relationship between glucagon-like peptide-1 receptor agonists (GLP-1 RAs) and psychiatric disorders. The analysis included 133,378 participants from 86 randomized controlled trials (RCTs) comparing GLP-1 RAs with placebo or other treatments.
What the trials showed: Pooled across all 86 trials, psychiatric adverse events were no more common on GLP-1 RAs than in the control groups (RR, 0.96; 95% CI, 0.85-1.08; I2 = 0%) — an absolute difference of -9 events per 10 000 person-years. Broken out by condition, not one of the eight categories reached statistical significance: depression (RR, 0.88; 95% CI, 0.70-1.10), suicide (RR, 0.90; 95% CI, 0.59-1.38), anxiety (RR, 0.92; 95% CI, 0.72-1.19), sleep disorder (RR, 0.98; 95% CI, 0.70-1.37), addictive disorder (RR, 0.89; 95% CI, 0.45-1.77), delirium (RR, 1.11; 95% CI, 0.74-1.66), bipolar disorder (RR, 1.19; 95% CI, 0.56-2.55) and schizophrenia (RR, 1.45; 95% CI, 0.61-3.47). Subgroup analyses by drug type, indication, dose, treatment duration, comparator, baseline BMI, age and pre-existing psychiatric illness found nothing either, and changes in blood glucose, HbA1c, weight or BMI did not shift the picture.
Where the evidence runs thin: The reassurance is firmest where events were common enough to count. For the rarest outcomes the confidence intervals are wide enough to hold a genuine increase — schizophrenia runs from a 39% reduction to a three-and-a-half-fold rise, bipolar disorder from 0.56 to 2.55. That is absence of evidence, not evidence of absence, and the two should not be confused. Worth knowing too: these psychiatric events were collected as treatment-emergent adverse events reported during the trials, not as diagnoses from structured psychiatric assessment. That is a coarser instrument, and it tends to under-record symptoms nobody thought to ask about. Drug trials also rarely run long enough to catch something that takes years to surface.
Why this question was asked at all: Reports of mood changes and suicidal thoughts in people taking GLP-1 drugs drew regulatory attention and a great deal of news coverage. Those reports came largely from spontaneous safety databases, where anyone can file a report and there is no comparison group. Pooling the randomized trials is the harder test, because it asks whether the same problems turn up more often on the drug than on placebo. Here they did not.
Practical takeaways: If you take one of these medicines, the trial evidence does not point to a raised risk of depression, anxiety or suicide, and that is worth knowing. It is not a reason to ignore a change in your own mood — the review measures what happened across a population, not what will happen to you, and it says nothing about how long the follow-up needs to be. The review evaluates the drug class as a whole; it does not compare individual products or brands, and nothing here speaks to any one of them.
Bottom line: Across 86 randomized trials and 133 378 people, psychiatric adverse events were no more frequent on GLP-1 receptor agonists than on comparison treatments. The finding is solid for the common outcomes and simply too imprecise to settle the rare ones.
Medical note. This article is for general education purposes only and is not intended as personal medical advice.