Does Ozempic or Mounjaro Cause Anxiety? What the Trials and Labels Show

In the randomized trials, anxiety was no more common on a GLP‑1 than on placebo. Pooled across 86 trials and 133,378 people, the risk ratio for anxiety was 0.92, with an interval that includes no effect.^[1] FDA’s own 2026 review of 91 trials reached the same finding.^[2] In the weight-loss trials for Wegovy and Zepbound, anxiety was reported slightly less often on the drug.^[7][8] That is not the whole record. Some cohort studies find a small rise, the global adverse-event databases show a reporting signal for semaglutide, and no trial enrolled people with serious psychiatric illness. Two physical causes can feel like anxiety and are worth ruling out first: low blood sugar, which every label lists anxiety as a sign of, and a faster heartbeat.

What the four labels say about anxiety

None of the four adult labels lists anxiety as a side effect of the drug in adults. The word does appear in each, in the patient Medication Guide, as one of the warning signs of low blood sugar. The semaglutide labels put it there,^[3][5] and so do the tirzepatide labels.^[4][6] Wegovy is the only one with anxiety in an adverse-reaction table, and that table is for teenagers.

The adolescent figure comes from one 68-week trial of 201 teenagers: anxiety was reported by 4% of the 133 on Wegovy and 2% of the 67 on placebo.^[3] That is a handful of people, and the label lists it because it crossed the table’s threshold of 3% and was higher than placebo, not because a link was established.

The suicidal behavior and ideation warning that Wegovy and Zepbound used to carry is gone. Both labels record its removal in February 2026,^[3][4] after FDA asked for it to come off.^[2] Ozempic and Mounjaro never had it. The full history and the studies behind that decision are in what happened with the suicidal ideation warning.

Suicidal Behavior and Ideation (5.10)…(Removed) 02/2026^[3]

— WEGOVY prescribing information, Recent Major Changes (DailyMed)

What the randomized trials recorded

Randomized trials are the only design that can separate a drug effect from everything else going on in a person’s life. The obesity programs for semaglutide, tirzepatide and liraglutide all collected psychiatric adverse events, and the drug makers later pooled them.

The STEP figures cover the category the authors call anxiety disorders and symptoms; the paper’s printed percentages are shown as published.^[7] For tirzepatide, the authors describe a “numerically lower occurrence of anxiety symptoms” on the drug.^[8] Liraglutide, an older daily injection, came out level.^[9] None of these was a formal test of anxiety, and depression was the questionnaire endpoint in each analysis, not anxiety.

The wider pools agree. A 2026 meta-analysis of 86 randomized trials covering type 2 diabetes and obesity found no difference in anxiety (risk ratio 0.92, 95% CI 0.72–1.19). For psychiatric disorders overall, no subgroup changed the result: not drug, dose, trial length, or whether people with existing psychiatric conditions had been allowed in.^[1] FDA ran its own meta-analysis before asking for the suicidality warning to come off:

“There were 91 placebo-controlled GLP-1 RA medication trials in the meta-analysis that included 107,910 patients (60,338 treated with a GLP-1 RA and 47,572 treated with placebo). The results did not show an increased risk for SI/B or for other relevant psychiatric adverse events such as anxiety, depression, irritability, or psychosis.”^[2]

— FDA Drug Safety Communication, January 13, 2026

Cohort studies: the answer depends on the comparison

Outside trials, researchers compare people who started a GLP‑1 with people who started something else, using health records. Here the anxiety results scatter, and most of the scatter tracks one choice: what the GLP‑1 users were compared with.

Two patterns stand out. Against SGLT2 inhibitors, another newer class of diabetes pill, two studies find a small excess of anxiety diagnoses on GLP‑1s, about 9%.^[11][12] Against DPP‑4 inhibitors or sulfonylureas, that excess shrinks or disappears.^[11][12] The Korean authors draw the conclusion directly: the apparent signal shifts with the comparator.^[12] That pattern is what you would expect if differences between patients, not the drug, are driving part of it. It is also compatible with a small real effect. These designs cannot tell the two apart.

The study with the largest number, a 108% rise, compared people on a GLP‑1 with people who were not taking one at all.^[14] People who are prescribed a weight-loss drug see doctors more often and have more conditions recorded, so they collect more diagnoses of every kind. That is the weakest comparison in the table, and it is the one most likely to be quoted.

Two studies on our site go deeper on specific questions. For people who already have depression or anxiety, a Swedish national study found semaglutide linked to a 38% lower risk of worsening anxiety, measured as worsening of an existing condition rather than as new diagnoses.^[16] That study and its caveats are in GLP-1s and worsening mental illness. The tirzepatide year-two figure in the table clears no effect by one thousandth, and its own authors warn against reading much into it; see tirzepatide versus semaglutide on mood.^[17]

Adverse-event reports: a signal, not a rate

Spontaneous reporting databases collect reports from patients, doctors and manufacturers. They can show that a drug is mentioned alongside a symptom more often than expected. They cannot show how often it happens, because nobody knows how many people took the drug, and they cannot show cause.

• VigiBase, the WHO global database. Among reports for diabetes and obesity drugs up to December 2024, anxiety was reported disproportionately often with semaglutide (adjusted reporting odds ratio 1.26). In reports filed before June 4, 2021, the authors found no signals, and they link the later signal to semaglutide’s use for weight management.^[18] This is a reporting signal, not a measured risk.
• EudraVigilance, the European database. Of 31,444 reports for semaglutide, liraglutide and tirzepatide from January 2021 to May 2023, 372 (1.2%) involved a psychiatric event, and 144 of those mentioned anxiety.^[19] Again, these are reports, not confirmed cases.
• FDA’s FAERS database. A live count of FAERS reports (data through July 2026) found 1,222 semaglutide reports and 1,542 tirzepatide reports listing anxiety, out of roughly 73,000 and 161,000 reports for each drug. Panic attack appeared in 139 and 379 reports.^[20] Each report is one person’s account; a count of reports is not a count of cases the drug caused.

A signal that appears only once a drug is widely used and widely discussed is hard to interpret, because both raise reporting without any change in risk. That is our reading, not a finding of the studies. Where a reporting signal and the randomized trials disagree, the trials are the stronger evidence.

Ozempic and panic attacks

No study has measured panic attacks or panic disorder as an outcome of any GLP‑1. Panic attacks are not listed on the Wegovy or Ozempic labels,^[3][5] nor on Zepbound or Mounjaro.^[4][6] What exists is a few hundred FAERS reports that mention them,^[20] which cannot say whether the drug, something else, or nothing caused them.

A panic attack feels like a pounding heart, sweating, shaking, lightheadedness and a sense that something is badly wrong. Several of those are also the body’s alarm response to low blood sugar, and a racing heart can come from the drug itself. That overlap is the practical reason to look at the physical causes below before deciding an episode is purely psychological.

Two physical causes that feel like anxiety

Low blood sugar

Each of the four Medication Guides lists the signs of low blood sugar, and anxiety is on every list. Ozempic’s version reads:

“dizziness or light-headedness • blurred vision • anxiety, irritability, or mood changes • sweating • slurred speech • hunger • confusion or drowsiness • shakiness • weakness • headache • fast heartbeat • feeling jittery”^[5]

— OZEMPIC Medication Guide (DailyMed)

On a GLP‑1 alone, low blood sugar is uncommon. The labels tie the risk to taking it alongside insulin or a sulfonylurea such as glipizide.^[4][5] If you take either and feel suddenly anxious, shaky and sweaty, check your glucose before anything else. Who is actually at risk is covered in GLP-1s and low blood sugar.

A faster heart rate

GLP‑1 drugs raise resting heart rate. On Wegovy’s label the average rise over placebo was 1 to 4 beats a minute in adults, but a quarter of people on the drug had a jump of 20 or more at some visit, against 16% on placebo, and the label asks patients to report palpitations or a racing heart at rest.^[3] Feeling your own heartbeat can set off worry, and worry raises heart rate further. What is normal, and which palpitations point elsewhere, is in GLP-1s, heart rate and palpitations.

Other routes are plausible but unmeasured. Days of nausea, poor sleep, eating far less, extra coffee to fight fatigue, and the stress of a large change in body and routine can all raise anxiety on their own. These are inferences, not findings from a GLP‑1 study. Related reading: coffee and caffeine on a GLP-1 and sleep quality and mood.

Anxiety after stopping

One 2026 study looked at what happens when the drug ends. Using hospital records from Shanghai for people with type 2 diabetes, it found GLP‑1 risk of depressive and anxiety disorders was neutral or lower while people were on treatment, but higher after they stopped, compared with people who stopped DPP‑4 or SGLT2 inhibitors.^[21] The authors call for vigilance about anxiety after discontinuation. It is a single observational study in a diabetes population, and people who stop a drug often stop for reasons, such as side effects, cost or illness, that carry their own stress. It is a reason to tell your prescriber how you feel when coming off, not proof of a withdrawal effect. What else changes after stopping is in what happens when you stop a GLP-1.

Can a GLP-1 treat anxiety?

Not on current evidence. A 2026 systematic review devoted to anxiety found that GLP‑1 drugs consistently reduced anxiety-like behavior in animals, while human cohort studies gave “mixed but suggestive” results, and it called for randomized trials.^[22] A second 2026 meta-analysis of 82 studies of neuropsychiatric outcomes found no association with anxiety in either direction.^[23] No randomized trial has tested a GLP‑1 as a treatment for an anxiety disorder, and none is registered with an anxiety disorder as the condition being treated. A GLP‑1 is not a substitute for anxiety treatment, and how these drugs sit alongside an existing antidepressant or anti-anxiety medicine is covered in antidepressants and a GLP-1.

If you feel anxious on a GLP-1

• If you use insulin or a sulfonylurea, check your blood sugar when anxiety comes on suddenly with shaking, sweating or hunger.^[5]
• Note the timing. Whether it started after a dose increase, after a rough week of nausea, after more coffee, or after stopping is useful to whoever prescribes the drug.
• Tell your prescriber, especially with a history of an anxiety disorder, panic disorder or depression, the groups the trials mostly left out.^[7][8]
• Do not stop other psychiatric medicines on your own because you started a GLP‑1, and do not stop the GLP‑1 without a plan.
• If you have thoughts of suicide or self-harm, call or text 988 in the US. FDA’s guidance to clinicians is to refer anyone who discloses those thoughts to a mental health professional.^[2]