Across three SURMOUNT trials pooling 4,056 adults, tirzepatide did not raise depression. Scores on the PHQ-9 questionnaire went from 2.7 to 1.9 on the drug and from 2.6 to 2.4 on placebo over 72 weeks, and 0.6% of each group reported suicidal thoughts [1]. Anxiety is the less settled half of the question, because the trials did not score it and the health-record studies that did found a small rise against semaglutide and a fall against an older weight-loss pill.
For someone who asks it, the question can arrive a few weeks in, somewhere around the second or third pen. Appetite has gone quiet, and a person notices they feel flatter, or more on edge, than they did before the first dose. Whether the drug did that is a fair question, and it has a better answer for depression than for anxiety. The class-wide picture, drawn mostly from semaglutide, is set out in whether GLP-1 drugs cause depression or anxiety; this page stays with tirzepatide.
What the SURMOUNT trials measured over 72 weeks
A post hoc analysis went back through SURMOUNT-1, SURMOUNT-2 and SURMOUNT-3 and pulled out every psychiatric measure the trials had collected along the way [1]. Depressive symptoms were scored with the PHQ-9, a nine-item questionnaire, and suicidal thoughts were asked about directly with the Columbia-Suicide Severity Rating Scale.
Both groups started with scores around 2.7 and 2.6, which is the range of someone with essentially no depressive symptoms. By week 72 the tirzepatide group sat at 1.9 and the placebo group at 2.4, a treatment difference of 0.6 points in the drug's favor, a small gap that nonetheless points the reassuring way rather than the worrying one.
The more useful number for someone counting their own weeks is who got worse. By the end, 18.2% of people on tirzepatide had moved into a more severe PHQ-9 category against 24.3% on placebo. That is fewer than one person in five on the drug drifting upward, against about one in four on the dummy injection.
Suicidal ideation, mostly rated low risk, was reported by 0.6% of each group. Nonfatal suicidal behavior appeared in 0.1% of people on tirzepatide and in nobody on placebo, a difference too small to read either way at these numbers. The authors describe the rates as similar to those of other incretin drugs, a comparison this site follows further in the cohorts on suicidal ideation.
Where anxiety shows up, and against what
Anxiety was not scored on a scale in SURMOUNT; it appears in that analysis only inside the adverse event tallies, which the authors call generally similar between groups [1]. The anxiety numbers people quote come from health-record studies instead, and each one depends heavily on which drug tirzepatide was being measured against.
One US record network followed adults with obesity for 24 months after they started tirzepatide or semaglutide, and matched them one to one with people starting naltrexone-bupropion, phentermine, or phentermine-topiramate [2]. Against naltrexone-bupropion, tirzepatide came with fewer new anxiety and depression diagnoses, whether or not a person had type 2 diabetes.
| Tirzepatide against | Outcome | Hazard ratio (95% CI) |
|---|---|---|
| Naltrexone-bupropion, with diabetes | Anxiety disorder | 0.55 (0.40–0.76) |
| Naltrexone-bupropion, without diabetes | Anxiety disorder | 0.78 (0.68–0.89) |
| Naltrexone-bupropion, with diabetes | Depression | 0.49 (0.33–0.72) |
| Naltrexone-bupropion, without diabetes | Depression | 0.57 (0.49–0.66) |
| Semaglutide, without diabetes | Anxiety disorder | 1.12 (1.06–1.18) |
| Semaglutide, without diabetes | Insomnia | 1.13 (1.05–1.21) |
The row that matters to someone choosing between the two injections is the fifth one. Among adults without diabetes, tirzepatide users picked up an anxiety diagnosis about 12% more often than semaglutide users, and insomnia about 13% more often. The authors add that these findings do not show cause and remain open to residual confounding and to diagnoses being recorded wrongly.
A second, larger record study matched 85,546 pairs of tirzepatide and semaglutide starters and found the two drugs level on a combined psychiatric outcome in both years, with hazard ratios of 0.984 and 1.002 [3]. A nominally higher anxiety hazard of 1.052 surfaced in the second year only, with an interval running from 1.001 to 1.106, and the authors warn it should be read cautiously given how many comparisons they ran. Put together, the two studies suggest that any extra anxiety relative to semaglutide is small, if it is there at all, while the side-by-side on weight and tolerability is in tirzepatide against semaglutide.
The comparator problem is the same one this site traced in a depression estimate that flipped with its comparison group. People steered to a pill such as naltrexone-bupropion may differ from people handed an injection, in ways a diagnosis code never records.
What the adverse event reports and the label say
Spontaneous reports tell a similar story from a noisier angle. A search of the FDA's adverse event database from 2004 to 2024, with 15,597 reports for tirzepatide, found significant depression and suicide-related signals for semaglutide only, at reporting odds ratios of 1.87 and 1.73, and none for tirzepatide [4]. Reports are counts without a denominator, so they flag questions rather than measure risk, which is why pooling them with trial data causes trouble.
A cohort focused on suicide itself matched 16,321 pairs of adults with overweight or obesity, tirzepatide against non-GLP-1 weight-loss medicines, and followed them for a median of a year [5]. Suicidal ideation or attempts were recorded in 17 people on tirzepatide and 33 on the other drugs, roughly one and two in every thousand, an adjusted hazard ratio of 0.52.
The regulator has moved in the same direction. The current Zepbound prescribing information lists its suicidal behavior and ideation warning as removed in February 2026, among its recent major changes [6].
The anxiety that may be low blood sugar
One ordinary source of jitteriness is written into the label itself. The patient information lists anxiety, irritability, or mood changes among the signs of low blood sugar, alongside sweating, shakiness and a fast heartbeat [6]. The label puts the risk higher when Zepbound is taken with a medicine that can lower blood sugar, such as insulin or a sulfonylurea. So a person with type 2 diabetes who feels a wave of unease an hour after a missed lunch has a checkable explanation.
That is a narrower thing than an anxiety disorder, and it has a meter reading attached. Tiredness and low mood can also travel with eating far less, a thread this site follows in whether GLP-1 drugs make you tired.
What is still unresolved
The trial evidence covers people who began without a psychiatric diagnosis, and it measured depression carefully and anxiety barely at all. Nobody has randomized people with established depression, anxiety disorder or bipolar disorder to tirzepatide and followed their symptoms. That leaves the people most likely to be asking this question without trial evidence of their own, the same gap that runs through the bipolar disorder cohort.
Until that trial exists, the honest reading is two-sided. Tirzepatide has not been shown to cause depression in people who start without it, and the anxiety signal against semaglutide is small and observational. Neither finding tells someone with an existing diagnosis what their own next six months on the drug will look like.