GLP-1 drugs and depression: a meta-analysis finds a signal, but the heterogeneity is huge
An open-access meta-analysis reports higher odds of depression-related adverse events among GLP-1 receptor agonist users (OR 1.49), but between-study heterogeneity is extreme (I² ~99%), which limits how literally the pooled number should be taken.
The question “do GLP-1 receptor agonists affect mood?” keeps bouncing between three worlds that rarely speak the same language:
- randomized cardiometabolic trials (where psychiatric outcomes are usually not primary)
- pharmacovigilance and case reports (where reporting is biased but can catch rare signals)
- observational datasets (where confounding can be heavy)
A new systematic review and meta-analysis tries to pull the evidence together. The headline result suggests higher odds of depression-related outcomes with GLP-1 receptor agonist use, but the studies do not agree with each other very well.
What this study is
- systematic review + random-effects meta-analysis
- searched multiple databases for studies published 2000–2025
- included both randomized trials and observational studies
- population focus: adults with overweight or obesity reporting depression/mood-related outcomes
What they found
They included 19 studies.
Pooled estimate (random effects):
- odds ratio for depression associated with GLP-1RA use: 1.49
- 95% CI: 1.18 to 1.88
The big caveat is the heterogeneity:
- I² = 99.21%
That level of between-study variation is a warning label: it means the pooled number is summarizing studies that likely differ in important ways (population, baseline psychiatric risk, comparator treatments, how “depression” is defined or captured, duration, and surveillance intensity).
What we know vs what we don’t
What we can say:
- this paper argues there may be an association between GLP-1RA use and depression-related outcomes
- the estimate is not subtle, but it is also not stable across studies
What we cannot say from this alone:
- whether the relationship is causal
- whether risk is concentrated in specific subgroups (prior depression, specific drugs, rapid weight loss, concomitant meds)
- whether the signal reflects detection bias (people interacting more with healthcare systems are more likely to have mood symptoms recorded)
Why it matters
GLP-1 receptor agonists deliver real metabolic benefits, and they are being used by tens of millions of people. Even a small psychiatric effect, if real, could matter at population scale.
The practical takeaway is not panic. It’s measurement:
- if a patient has a history of depression or psychiatric vulnerability, monitor mood symptoms deliberately
- if a patient reports mood changes after starting a GLP-1RA, treat it as a real adverse event report worth documenting and discussing, not as “internet noise”
Source
Bi Z, et al. Glucagon-like peptide-1 receptor agonists and risk of depression: a systematic review and meta-analysis. BMC Psychiatry. (2026).
- Article (open access): https://link.springer.com/article/10.1186/s12888-026-08016-x
- DOI: https://doi.org/10.1186/s12888-026-08016-x