Semaglutide and Depression: What Clinical Data Shows

Millions of people now take semaglutide for weight loss or type 2 diabetes. And millions of those people, quite reasonably, want to know: will this drug make me depressed?

The short answer from clinical trials is no -- at least for people without pre-existing major depression. But the full picture is more complicated than a single reassuring sentence can capture. Between the STEP trial data, pharmacovigilance signals, real-world cohort studies, and the first trials in patients who actually have depression, there is a lot of evidence to sort through.

This article does exactly that. No hype, no hand-waving. Just what the data shows.

The STEP Trials: The Largest Dataset We Have

The most important evidence comes from a post hoc analysis of the STEP clinical trial program, published in JAMA Internal Medicine in November 2024 by Thomas Wadden and colleagues at the University of Pennsylvania.

The researchers pooled data from STEP 1, 2, and 3 (each 68 weeks) and STEP 5 (104 weeks), covering 3,377 participants in the first three trials and 304 in STEP 5. All participants had overweight or obesity; STEP 2 also included people with type 2 diabetes. The analysis measured depressive symptoms using the Patient Health Questionnaire-9 (PHQ-9) and suicidal ideation using the Columbia-Suicide Severity Rating Scale (C-SSRS).

What the PHQ-9 Scores Showed

At baseline, mean PHQ-9 scores were 2.0 in the semaglutide group and 1.8 in the placebo group -- both in the "no/minimal depression" range. At week 68:

Semaglutide group: PHQ-9 score of 2.0
Placebo group: PHQ-9 score of 2.4
Estimated treatment difference: -0.56 (95% CI: -0.81 to -0.32; P < .001)

In plain terms, semaglutide users had a statistically significant but small improvement in depressive symptoms compared to placebo. The researchers were careful to note this difference is not clinically meaningful -- it would not be noticeable to an individual patient. But the direction matters: semaglutide pushed depression scores down, not up.

Fewer People Shifted Toward Worse Depression

The analysis also looked at whether participants moved into worse PHQ-9 categories over time. Semaglutide users were less likely to shift toward more severe depression (odds ratio 0.63; 95% CI: 0.50-0.79; P < .001).

Only 2.8% of semaglutide-treated participants ever scored 15 or above on the PHQ-9 (the threshold for moderately severe depression) during the trial period, compared to 4.1% on placebo.

Suicidal Ideation Was Rare in Both Groups

On the C-SSRS, 1% or fewer participants in either group reported suicidal ideation or behavior, with no difference between semaglutide and placebo.

What the STEP Data Can and Cannot Tell Us

To put these numbers in context: the PHQ-9 is a 27-point scale where 0-4 indicates no/minimal depression, 5-9 is mild, 10-14 is moderate, and 15-19 is moderately severe. The baseline scores of 1.8-2.0 mean these trial participants were, on average, about as far from clinically depressed as the scale can measure.

The trial populations were also generally healthy: mean age around 47-49, roughly 70-78% female, with BMIs in the obesity range but relatively few medical complications beyond what the trials were designed to treat. These were people with weight problems but not, by and large, people with psychiatric problems.

This matters because it defines the population to which these results apply. The STEP data is strong evidence for a specific group. It is not a universal safety guarantee.

The Big Caveat: Who Was Not in These Trials

Here is the part that gets less attention than it deserves.

The STEP trials excluded people with current major depressive disorder, other serious mental illness like schizophrenia, or a history of suicide attempts. Anyone with a PHQ-9 score of 15 or above at screening was not enrolled. Wadden himself flagged this directly: "Further study is needed of the psychiatric safety of semaglutide 2.4 mg when used by persons with current major depressive disorder, other serious mental illness, or a history of suicide attempt."

This exclusion creates a blind spot in the evidence. The estimated 21 million American adults with major depression are significantly more likely to have obesity than the general population -- the comorbidity rate is roughly 43%. These are precisely the people who might benefit from semaglutide's weight loss effects and who are most likely to ask their doctors about depression risk. And they are precisely the people about whom the STEP data has the least to say.

Semaglutide in People Already Taking Antidepressants

A separate post hoc analysis by Kushner and colleagues (2024) addressed a related question: does semaglutide work differently in people already on antidepressants?

Across STEP 1, 2, 3, and 5, about 539 of the 3,683 randomized participants (roughly 15%) were taking antidepressants at baseline. The findings:

Weight loss was similar regardless of antidepressant use. In STEP 1, semaglutide users on antidepressants lost 15.7% of body weight vs. 0.2% on placebo. The pattern held across all four trials.
Adverse event rates were comparable between semaglutide and placebo users who were on antidepressants.

This is reassuring for clinicians making prescribing decisions. A patient who is stable on an antidepressant can expect to respond to semaglutide normally, without added psychiatric risk.

But it is worth noting that "taking antidepressants" and "currently depressed" are not the same thing. Many of these participants likely had well-controlled symptoms -- the trials required PHQ-9 scores below 15 for enrollment. A patient in the grip of a major depressive episode is a different clinical scenario than a patient who has been stable on sertraline for three years.

Real-World Evidence: Beyond the Clinical Trials

The Oxford Cohort Study

A propensity-score matched cohort study published in eClinicalMedicine in 2024 used electronic health records from the TriNetX US Collaborative Network (covering over 100 million patients) to compare neuropsychiatric outcomes in people with type 2 diabetes prescribed semaglutide versus other diabetes medications (sitagliptin, empagliflozin, or glipizide).

After propensity-score matching -- comparing 19,000 to 23,000 patients per comparison -- the researchers found semaglutide was not associated with a higher 12-month risk of adverse neuropsychiatric outcomes compared to any of the three comparators. The study even found potential protective signals against cognitive decline and nicotine misuse, though the evidence on depression specifically remained uncertain.

The Pharmacovigilance Picture

Pharmacovigilance databases -- which collect spontaneous reports of adverse events -- paint a more complicated picture.

An analysis of the EudraVigilance database (2024) reviewed 31,444 adverse event reports for semaglutide, liraglutide, and tirzepatide. Of these, 372 reports (1.18%) involved psychiatric adverse events. Depression was the most commonly reported psychiatric event (187 reports, or 50.3% of psychiatric reports), followed by anxiety (144 reports) and suicidal ideation (73 reports).

A 2025 VigiBase study using WHO global pharmacovigilance data extended these findings, examining psychiatric and psychological adverse effects across dulaglutide, semaglutide, and liraglutide. The multinational scope of this analysis added geographical diversity to the pharmacovigilance picture, though the fundamental limitations of spontaneous reporting databases remained.

Why Pharmacovigilance Signals and Clinical Trial Data Differ

If clinical trials show no increased depression risk but pharmacovigilance databases flag depression, who is right?

Probably both, for different reasons:

Reporting bias. People who experience mood changes on a drug are more likely to report them. People whose mood stays the same rarely file reports.
Population differences. Clinical trials excluded people with severe depression. The real world does not.
Confounders. People taking GLP-1 drugs often have obesity, which is itself associated with higher depression rates. Pharmacovigilance data cannot easily control for this.
Media attention. After headlines about GLP-1 drugs and mental health in 2023, reporting rates likely increased -- a well-documented phenomenon called "notoriety bias."

Neither source is wrong. But clinical trials are better at establishing causation, while pharmacovigilance is better at catching rare signals that trials might miss.

The First Trial in Patients With Major Depression

The most intriguing recent development is a randomized clinical trial from the University of Toronto that directly tested oral semaglutide (14 mg) in adults with major depressive disorder and cognitive impairment who were overweight or obese.

This 16-week, double-blind, placebo-controlled trial randomized 72 participants. The primary outcome was executive function, measured by standardized cognitive testing.

Key results:

Semaglutide did not improve executive function (the primary outcome; adjusted Z-score difference: 0.32, 95% CI: -0.92 to 1.58, P = 0.60).
In secondary analyses, semaglutide improved global cognition and produced clinically significant weight loss.
Critically, semaglutide appeared safe in this psychiatric population -- no notable psychiatric adverse events.

This is a small trial, and it was not designed to test semaglutide as an antidepressant. But it is the first to establish a safety signal in people who actually have depression, and that matters.

The Weight Loss Connection: Does Losing Weight Improve Mood?

One question that hangs over all of this data: when semaglutide users show small improvements in depressive symptoms, is that the drug itself -- or the weight loss?

There are reasons to suspect both.

The weight loss hypothesis: Obesity and depression are strongly correlated. Weight loss often improves mood, energy, mobility, sleep quality, and social engagement. The STEP trials showed substantial weight loss (roughly 15% of body weight), which alone could explain modest PHQ-9 improvements.

The direct brain effects hypothesis: GLP-1 receptors exist throughout the brain, including in regions involved in mood regulation like the hippocampus, amygdala, and prefrontal cortex. Preclinical research on GLP-1 and neurodegeneration has shown these receptors modulate neuroinflammation and neurotrophic factors. There is a plausible biological mechanism for direct antidepressant effects independent of weight loss.

The honest answer is we do not yet know how to separate these effects in humans. The STEP trials cannot disentangle the two because all participants who received semaglutide lost substantially more weight than placebo. To isolate a direct mood effect, researchers would need to study semaglutide in people who do not lose weight on it, or compare mood outcomes in matched patients with identical weight loss achieved through different means (surgery, diet, GLP-1 therapy). Those studies have not yet been done.

The question matters beyond academic interest. If semaglutide's mood effects are entirely mediated by weight loss, then the benefits disappear when weight is regained after stopping the drug. If there is a direct neurological effect, it could persist -- or it could create withdrawal-like mood changes when the drug is discontinued. Neither scenario has been studied systematically.

What About "Ozempic Sadness"? The Patient Experience Gap

Search online for "Ozempic depression" and you will find thousands of first-person accounts: people describing emotional blunting, loss of pleasure in food, feeling "flat," or experiencing new depressive episodes.

These experiences are real and should not be dismissed. But they are hard to interpret for several reasons:

Removing food as a coping mechanism. For people who used eating to manage stress, anxiety, or sadness, losing that coping tool can unmask underlying mood problems that were always there.
Body image disruption. Rapid physical changes can destabilize someone's sense of self, creating psychological distress that looks like depression.
Gastrointestinal side effects. Persistent nausea, which affects a substantial percentage of semaglutide users, can mimic or worsen depressive symptoms. Feeling sick every day would make anyone's mood worse.
Caloric restriction effects. Eating significantly less can affect serotonin production and energy levels.

The takeaway is not that these reports are wrong. It is that "semaglutide causes depression" and "some people feel worse on semaglutide for reasons we need to understand better" are very different claims.

Practical Guidance for Patients and Prescribers

Based on the current evidence:

For Patients Without a History of Depression

The STEP trial data is reassuring. Semaglutide does not appear to increase depression risk in this group. Monitor your mood as you would with any new medication, but there is no specific reason for concern.

For Patients With Current or Past Depression

The evidence here is thinner. The Toronto trial found semaglutide was safe in 72 patients with MDD over 16 weeks, but that is a small sample over a short period. If you take semaglutide and have depression:

Tell your prescriber about your psychiatric history.
Continue your current depression treatment.
Track your mood with a simple tool like the PHQ-9 (freely available online).
Pay attention to the difference between GLP-1 side effects and new depressive symptoms.

For Prescribers

The American Psychological Association noted in 2025 that psychologists are increasingly involved in supporting GLP-1 patients through the emotional and psychological dimensions of rapid weight loss. Screening for depression and anxiety at baseline and during treatment is reasonable, particularly for patients with psychiatric histories who were underrepresented in the trial evidence.

The Contrasting Signal: One Study to Consider

Not all the real-world data is reassuring. A 2024 study published in Scientific Reports found a 98% increased risk of any psychiatric disorders among GLP-1 RA users, including a 195% higher risk of major depression, 108% increased risk for anxiety, and 106% elevated risk for suicidal behavior.

These numbers are startling and deserve explanation. The study was a retrospective cohort analysis of over 162,000 patients, which gives it size. But it relied on diagnostic and medication codes, which can introduce inaccuracies -- a patient coded for "depression" might have a single screening visit, not a clinical diagnosis. It was also susceptible to the same confounding-by-indication problem that affects all observational studies: people prescribed GLP-1 drugs tend to be sicker in multiple ways than the general population, and sicker people have more psychiatric diagnoses.

The researchers themselves acknowledged these limitations. This study should not be dismissed, but it should be weighed against the larger body of evidence -- the STEP trials, the Oxford cohort, the French study -- that consistently finds no increased risk.

Where the Research Goes Next

Several lines of investigation are ongoing:

Larger trials in psychiatric populations. The Toronto trial opened the door; now bigger, longer studies are needed to confirm safety in people with depression, bipolar disorder, and other conditions.
Mechanistic studies. Neuroimaging research is beginning to map how semaglutide affects brain regions involved in mood regulation -- the same reward and emotional circuits implicated in its effects on eating behavior.
Post-marketing surveillance. Both the FDA and EMA continue monitoring psychiatric safety signals as tens of millions of people now take these drugs.
Comparative studies with tirzepatide. As the newer dual GIP/GLP-1 agonist gains prescriptions, comparative psychiatric safety data will be important.

The Bottom Line

The clinical trial evidence -- particularly the STEP program analysis by Wadden et al. -- shows that semaglutide does not increase depression risk in people without major psychiatric conditions. If anything, there is a small statistical signal in the other direction.

But the evidence has a significant blind spot: the very population most likely to be concerned about depression -- people who already have it -- was largely excluded from the studies that provide the most reassurance.

The emerging data from the Toronto trial and real-world cohort studies is encouraging. Semaglutide appears safe even in psychiatric populations, based on early and limited evidence. But "appears safe in a 16-week trial of 72 people" is a different level of certainty than "proven safe in 3,377 people over 68 weeks."

If you are taking semaglutide and your mood changes, do not panic -- but do not ignore it either. Talk to your doctor. The data says your risk is low, but your experience is always worth investigating.

This article is part of the GLP-1 and Mental Health series. For related topics, see our coverage of GLP-1 drugs and suicidal ideation, emotional eating and food noise, body image after weight loss, and GLP-1 agonists and alcohol reduction.