GLP-1 and depression/suicidality: the evidence

In 2023, a handful of patient reports set off one of the biggest safety scares in the history of weight-loss drugs: were Ozempic, Wegovy, and similar GLP-1 medications pushing people toward dark thoughts? Regulators on two continents launched investigations, and the question hung over millions of prescriptions. As of January 2026, after reviewing 91 clinical trials and more than 100,000 patients, the U.S. Food and Drug Administration asked drugmakers to pull the suicide warning off these labels entirely. This article walks through exactly what the evidence says, where it's strong, where it's shaky, and what it means if you or someone you love takes one of these drugs.

Where the scare came from

The story starts in July 2023. Iceland's medicines agency flagged a few cases of suicidal thoughts and self-harm in people taking liraglutide (Saxenda) and semaglutide (Ozempic, Wegovy). That's a tiny number of reports. But when a signal like this surfaces, regulators are obligated to dig in.

The European Medicines Agency (EMA) opened a review. The FDA followed. News outlets ran headlines. And patients who were already nervous about a new class of drugs now had a fresh worry stacked on top of nausea and "Ozempic face."

Two things made the signal worth taking seriously. First, these drugs cross into the brain and act on regions tied to mood and reward, so a biological mechanism for psychiatric effects isn't far-fetched. Second, weight-loss drugs as a class have a troubled history here. An older obesity drug, rimonabant, was pulled from the market because it raised the risk of depression and suicidal thinking. Regulators were not going to repeat that mistake by ignoring early reports.

So the question was never silly. The job was to figure out whether the signal was real cause and effect, or noise.

How GLP-1 drugs touch the brain

GLP-1 (glucagon-like peptide-1) is a gut hormone. The drugs that mimic it — semaglutide, tirzepatide, liraglutide — were built to lower blood sugar and curb appetite. But GLP-1 receptors aren't only in the pancreas and gut. They sit in the brain too.

Receptors show up in the hippocampus, the amygdala, the dorsal raphe nucleus, and parts of the reward system. Those are the same regions that go haywire in depression and addiction. In animal studies, GLP-1 drugs do several things that, on paper, should help mood rather than hurt it. They calm neuroinflammation by dialing down signaling molecules like IL-6 and TNF-alpha. They boost BDNF, a protein that supports the growth of new brain cells. They nudge the dopamine reward circuit, which is why researchers are testing the same drugs for alcohol and nicotine cravings.

Here's the catch. Mechanism cuts both ways, and animal brains aren't human brains. The reward-pathway effects that might blunt a craving for a drink could, in theory, also blunt the pleasure people get from food, sex, or hobbies — and a flattened sense of reward is a hallmark of depression. There's also the weight loss itself. Rapid changes in body weight, blood sugar, and eating patterns can ripple into mood, energy, and sleep. So the biology gives you a plausible story for benefit and a plausible story for harm. Mechanism alone can't settle the question. Only the outcome data can.

For a fuller tour of the side-effect landscape beyond mood, see our complete guide to GLP-1 side effects.

The actual evidence, graded honestly

There are four kinds of evidence here, and they don't all carry the same weight. From weakest to strongest:

Case reports and pharmacovigilance databases (FAERS, VigiBase, EudraVigilance) — these collect spontaneous reports of bad events. They generate hypotheses but cannot prove cause and effect. No denominator, heavy reporting bias.
Observational cohort studies — large databases of real patients. Better, but vulnerable to confounding (the reasons a person gets prescribed a drug can also drive their outcomes).
Randomized controlled trial (RCT) meta-analyses — pooled data from trials where treatment was assigned by chance. This is where causal claims get strong.
Regulatory meta-analyses — the FDA and EMA pooling the raw trial data themselves, with access to the original records.

A reader who only saw the pharmacovigilance headlines would walk away scared. A reader who only saw the RCT data would walk away reassured. The honest reading needs both. Here's the breakdown.

The pharmacovigilance signals (weakest evidence)

This is where the alarming numbers live. Analyses of the World Health Organization's VigiBase and the European EudraVigilance system found "disproportionate" reporting of suicidal ideation and depression for semaglutide. One EudraVigilance analysis found depression was the most common psychiatric event reported (about 50% of psychiatric reports), followed by anxiety (about 39%) and suicidal ideation (about 20%). A signal for semaglutide-related suicidal ideation showed a reporting odds ratio around 1.45, and the signal was much stronger in people also taking antidepressants or benzodiazepines.

Sounds damning. But here's why you can't trust it as proof. Pharmacovigilance databases have no denominator — you don't know how many people took the drug without any problem. They're plagued by "stimulated reporting": once the media runs scary headlines, more people file reports, which makes the signal look bigger regardless of the truth. And the people drawn to weight-loss drugs are, on average, already at higher risk for depression because obesity and depression travel together. The fact that the signal was concentrated in people already on psychiatric meds is the tell — those are people who already had mental health conditions before they ever touched a GLP-1 drug.

Grade: hypothesis-generating only. Useful for raising a flag. Useless for proving cause.

The cohort studies (mixed, mostly reassuring)

When researchers turned to large real-world patient databases, the picture flipped. The biggest study, published in Nature Medicine in January 2024 by Wang and colleagues, looked at electronic health records from a network covering 240,618 people with overweight or obesity and roughly 1.6 million with type 2 diabetes. Semaglutide was associated with a lower risk of new suicidal ideation (hazard ratio 0.27, 95% CI 0.20–0.36) and recurrent ideation (HR 0.44, 95% CI 0.32–0.60) compared with other weight or diabetes drugs. In plain terms: people on semaglutide had fewer suicidal-thought events, not more.

A separate propensity-weighted cohort study in Diabetologia (Hurtado et al., 2024) found no increased risk of suicidal ideation or self-injury — but with wide confidence intervals (HR 1.04, 95% CI 0.35–3.14 in the per-protocol analysis). The authors were careful: because these events are so rare, their data couldn't rule out a threefold increase either. That's an honest limitation.

A France-wide case-time-control study (Bezin et al., eClinicalMedicine, 2025) using the national health database also found no convincing association between GLP-1 use and suicide or suicide attempt.

Grade: moderate. Consistently points toward "no increased risk" and possibly lower risk, but cohort studies can't fully escape confounding by indication. The Nature Medicine protective finding is striking but should be read as "no signal of harm," not "proof these drugs prevent suicide."

The randomized-trial meta-analyses (strongest evidence)

This is the data that should carry the most weight, because randomization breaks the confounding that haunts the other studies.

A 2025 meta-analysis in JAMA Psychiatry (Ebrahimi et al.) pooled 27 randomized controlled trials — 32,354 people on a GLP-1 drug versus 27,042 on placebo. Suicide and self-harm events were extremely rare in both groups, with no significant difference (rate ratio 0.76, 95% CI 0.48–1.21). The authors concluded an increase in these very rare events is unlikely, while still urging ongoing monitoring as use expands.

A 2025 meta-analysis in the Journal of Diabetes (covering 25 RCTs and over 81,000 participants) reached the same place: no significant difference in suicidal behavior between GLP-1 and control (risk ratio 0.84, 95% CI 0.54–1.32), with consistent results across suicidal ideation, attempts, and completed suicide.

A broader systematic review in Diabetes/Metabolism Research and Reviews (Bushi et al., 2025) found no significant link, while flagging very high statistical heterogeneity (I-squared of 98%) — meaning the underlying studies varied a lot, so the pooled number should be read with caution.

Grade: strong and consistent. Across tens of thousands of randomized patients, GLP-1 drugs do not raise suicidal behavior.

The regulatory verdict (the tiebreaker)

In April 2024, the EMA concluded its review and stated it found no evidence that GLP-1 receptor agonists cause suicidal thoughts or actions. The FDA's January 2024 update reached the same preliminary conclusion.

Then, on January 13, 2026, the FDA went further. After a comprehensive meta-analysis of 91 placebo-controlled trials including 107,910 patients (60,338 on a GLP-1 drug, 47,572 on placebo), it found no increased risk of suicidal ideation or behavior — and no increased risk of related events like anxiety, depression, or irritability. The FDA formally asked the makers of Saxenda (liraglutide), Wegovy (semaglutide), and Zepbound (tirzepatide) to remove the suicide warning from their labels.

That is about as definitive as drug safety conclusions get. A regulator does not pull a warning label lightly.

Evidence summary table

Evidence type	Key source(s)	Main finding	Direction	Strength
Pharmacovigilance (FAERS, VigiBase, EudraVigilance)	EudraVigilance & WHO analyses	Disproportionate reports of depression, anxiety, suicidal ideation (semaglutide ROR ~1.45)	Suggests harm	Weak — no denominator, reporting bias
Real-world cohort, overweight/obesity	Wang et al., Nature Medicine 2024	Lower suicidal-ideation risk (HR 0.27)	Suggests benefit	Moderate
Propensity-weighted cohort	Hurtado et al., Diabetologia 2024	No increased risk (HR 1.04, wide CI)	Neutral	Moderate
National case-time-control	Bezin et al., eClinicalMedicine 2025	No association with suicide/attempt	Neutral	Moderate
RCT meta-analysis	Ebrahimi et al., JAMA Psychiatry 2025	No difference (RR 0.76)	Neutral	Strong
RCT meta-analysis	Journal of Diabetes 2025 (25 RCTs)	No difference (RR 0.84)	Neutral	Strong
Regulatory meta-analysis	FDA, January 2026 (91 trials, 107,910 patients)	No increased risk; warning removed	Neutral	Strongest

Could GLP-1 drugs actually help depression?

This is the flip side, and the evidence is genuinely early and mixed — so read it with skepticism.

Some meta-analyses of mood scores have found a small reduction in depression symptoms in people on GLP-1 drugs compared with controls. Researchers point to the neurogenesis, anti-inflammation, and reward-circuit effects seen in animals as the possible engine. There's also a simpler explanation: losing weight, sleeping better, and moving more tend to lift mood on their own.

But here's the honest caveat. The largest and most rigorous analysis to date — pooling 80-plus randomized trials and over 100,000 patients — found GLP-1 treatment was not associated with a meaningful change in depressive symptoms versus placebo, and no difference in serious psychiatric adverse events. So the "antidepressant" story is mostly hopeful signal from smaller studies, not established fact. Dedicated trials of GLP-1 drugs as depression treatments are only now getting underway.

Bottom line: don't take a GLP-1 drug expecting it to treat depression. The evidence isn't there yet. It's an active research question, nothing more.

Comparing the drugs and the alternatives

People often ask whether one GLP-1 drug is safer for mood than another. The honest answer is that the data don't support meaningful differences. The trial meta-analyses lumped the class together because event rates were too low to split reliably by drug. Pharmacovigilance signals appeared for semaglutide, liraglutide, and others alike — which again points to the underlying population rather than a specific molecule.

Option	Mood/suicidality evidence	Notes
Semaglutide (Ozempic, Wegovy)	Most-studied; no increased risk in RCTs; lower ideation risk in one large cohort	Carried the original signal because it's the most prescribed
Tirzepatide (Mounjaro, Zepbound)	No increased risk in trial data; included in FDA label-removal	Dual GIP/GLP-1 agonist
Liraglutide (Saxenda, Victoza)	Older drug; no increased risk on review	One of the first flagged in Iceland
Older weight-loss drugs (e.g., bupropion-naltrexone)	Some carry their own mood cautions	Bupropion itself has a black-box psychiatric warning
Rimonabant (historical)	Withdrawn for raising depression/suicide risk	The cautionary tale behind the whole scare

If mood is a real concern, the relevant comparison isn't GLP-1 versus nothing — it's GLP-1 versus the alternative weight or diabetes treatment, some of which carry their own psychiatric baggage. GLP-1 drugs look clean by that standard. For the broader risk-benefit picture, see our overview of GLP-1 side effects and risks.

Safety: what to actually watch for

"No increased risk across a population" is not the same as "zero risk for every individual." Here's the practical guidance the evidence supports.

The high-risk group is people with a psychiatric history. Almost every reassuring study had the same caveat: the suicidality signals clustered in people already taking antidepressants or with a history of mental illness. That doesn't mean the drug caused it — but it does mean these patients deserve closer monitoring, not exclusion.

Watch the early weeks and the rapid-weight-loss window. Big metabolic shifts can affect mood and sleep. If depression or anxiety worsens after starting, that's worth a call to your prescriber.

Don't stop antidepressants on your own. Some interactions matter, and GLP-1 drugs slow stomach emptying, which can change how other oral medications absorb.

Know the emergency signs. Any new or worsening thoughts of self-harm, hopelessness, or suicide are a medical emergency regardless of cause. In the U.S., call or text 988 (the Suicide and Crisis Lifeline).

The FDA removing the warning does not mean clinicians should stop asking about mood. It means the warning was based on a signal the data didn't confirm. Good prescribers still screen.

Who this matters for

You have no psychiatric history: The data are reassuring. Trial evidence in people without major mental illness showed no link to depression or suicidality — in the SELECT-type trials, depression scores were no worse (and sometimes slightly better) on semaglutide than placebo. Proceed with normal monitoring.

You have depression, anxiety, or a history of suicidal thoughts: The drugs are not contraindicated, and the population-level evidence is reassuring, but you and your prescriber should monitor mood actively, especially in the first few months. Loop in your mental health provider. This is the group where vigilance pays off.

You're hoping it treats your depression: Don't count on it. Promising early signals exist, but the strongest evidence shows no meaningful antidepressant effect. Treat your depression with proven treatments and use the GLP-1 drug for what it's approved for.

The same "watch the at-risk subgroup, reassure everyone else" logic shows up across GLP-1 research — it's the pattern in studies of these drugs for alcohol use disorder and smoking cessation too, where central reward effects are the active area of study. The off-label mood and addiction story is still being written; see our look at semaglutide's off-label uses in PCOS for how cautiously these claims should be read.

Frequently Asked Questions

Do Ozempic and Wegovy cause suicidal thoughts?

Based on the strongest available evidence, no. After reviewing 91 placebo-controlled trials and 107,910 patients, the FDA found no increased risk of suicidal ideation or behavior and asked drugmakers to remove the suicide warning from GLP-1 labels in January 2026. The EMA reached the same conclusion in 2024. Early reports that raised the alarm came from databases that can't prove cause and effect.

Why was there a warning in the first place if the drugs are safe?

A few case reports from Iceland in 2023 flagged suicidal thoughts in patients on these drugs. Because an older weight-loss drug (rimonabant) really did raise suicide risk, regulators took the signal seriously and added cautionary language while they investigated. The deeper review showed the signal didn't hold up — the events clustered in people who already had mental health conditions.

Can GLP-1 drugs make depression worse?

For most people, the trial data show no worsening of depression compared with placebo. The exception worth watching is people with an existing psychiatric history, who should monitor their mood closely in the first few months. If your depression or anxiety gets worse after starting, contact your prescriber — it's worth evaluating regardless of the cause.

Can GLP-1 drugs treat depression?

Not proven. Some smaller studies found a small improvement in mood scores, and the biology is intriguing, but the largest and most rigorous analysis found no meaningful antidepressant effect versus placebo. Dedicated trials are underway. Don't take a GLP-1 drug expecting it to treat depression.

Is one GLP-1 drug safer for mental health than another?

The evidence doesn't support a meaningful difference. Suicide and self-harm events were too rare in trials to reliably separate semaglutide, tirzepatide, and liraglutide. All three were included in the FDA's label-removal request. The risk appears tied to the patient's underlying mental health, not the specific molecule.

This article is for general information and is not medical advice. Talk to a qualified healthcare provider before starting, stopping, or changing any medication. If you are having thoughts of suicide or self-harm, call or text 988 (the U.S. Suicide and Crisis Lifeline) or your local emergency number now.

Sources

FDA Drug Safety Communication, Removal of Suicidal Behavior and Ideation Warning from GLP-1 RA Medications (January 13, 2026)
Wang W et al., Association of semaglutide with risk of suicidal ideation in a real-world cohort, Nature Medicine, 2024
Ebrahimi P et al., Suicide and Self-Harm Events With GLP-1 Receptor Agonists in Adults With Diabetes or Obesity: A Systematic Review and Meta-Analysis, JAMA Psychiatry, 2025
Bushi G et al., Association of GLP-1 Receptor Agonists With Risk of Suicidal Ideation and Behaviour: A Systematic Review and Meta-Analysis, Diabetes/Metabolism Research and Reviews, 2025
Hurtado I et al., Association of GLP-1 receptor agonists with suicidal ideation and self-injury: a propensity-weighted, population-based cohort study, Diabetologia, 2024
Bezin J et al., Suicide and suicide attempt in users of GLP-1 receptor agonists: a nationwide case-time-control study, eClinicalMedicine, 2025
PubMed search: GLP-1 receptor agonist, suicidality and depression
PubMed search: GLP-1 receptor agonist antidepressant effect meta-analysis
Kornelius E et al., The risk of depression, anxiety, and suicidal behavior in patients with obesity on GLP-1 receptor agonist therapy, Scientific Reports, 2024
GLP-1 Receptor Agonists in Mood Disorders: A Psychiatric Perspective, 2025