Women's Health

GLP-1s, mood and antidepressants: what the labels say now, and what the suicidality evidence actually shows

By the US Health Digest editorial team · Published August 28, 2026 · Every claim linked to its primary source

The January 2026 change to the GLP-1 labels was a removal, not an addition. On January 13, 2026 the FDA asked manufacturers to strike the Suicidal Behavior and Ideation warning from Saxenda, Wegovy and Zepbound, saying its review "did not show an increased risk for SI/B or for other relevant psychiatric adverse events." The current DailyMed labels for Wegovy, Zepbound and Saxenda each record the section as "(Removed) 02/2026"; none of the four US weight-management GLP-1 labels now carries a mood or suicidality warning in Section 5 or Section 17. The European regulator reached the same conclusion 21 months earlier, and no antidepressant is named anywhere in any of the four labels.

If you are in crisis right now: call or text 988, or use the chat at the 988 Suicide & Crisis Lifeline. It is free, confidential and available 24 hours a day, every day. SAMHSA administers it, and the FDA's own safety communication carries the same number.

This page reports what labels and studies say. It is not medical advice and not a reason to start, stop, skip or change any medicine, including an antidepressant — that decision belongs with your prescriber.

One framing point governs everything below. When a regulator says it "did not identify an increased risk", that is a finding about rates across very large populations — not a forecast about any individual, and not a claim that nobody on these drugs experiences a change in mood. If your mood changes on any medicine, it is worth telling your prescriber whether or not a label tells them to ask.

The premise, corrected

The question circulating as "the FDA's January 2026 GLP-1 label change" is real, but runs the opposite way from how it is usually told. Nothing was added. The January 13, 2026 Drug Safety Communication states that the FDA "is requesting that drug application holders remove information regarding the risk of suicidal ideation and behavior (SI/B) from the labeling of glucagon-like peptide-1 receptor agonist (GLP-1 RA) medications that currently include such language." The labels changed the following month. Those section numbers have since been reused — Saxenda's 5.8 now heads Hypersensitivity Reactions — which is why an old cross-reference lands somewhere unrelated.

The FDA also explained where the warning came from: the language had been present "at the time of the original FDA approvals" and is "based on reports of such events observed with a variety of older medicines used or studied for weight loss." By FDA's account, GLP-1 products approved for diabetes rather than weight never carried it at all.

What the labels used to say, word for word

A note on sourcing: neither FDA nor DailyMed serves superseded labels. Every pre-removal quote below is therefore read from an Internet Archive capture of the original FDA PDF, and each link goes to that capture, not to a live FDA page. All four were re-checked on August 28, 2026.

The Wegovy label revised 10/2025 — the last edition before removal — read in full: Suicidal behavior and ideation have been reported in clinical trials with other weight management products. Monitor patients treated with WEGOVY for the emergence or worsening of depression, suicidal thoughts or behavior, and/or any unusual changes in mood or behavior. Discontinue WEGOVY in patients who experience suicidal thoughts or behaviors. Avoid WEGOVY in patients with a history of suicidal attempts or active suicidal ideation. Its patient-counselling section added: Advise patients to report emergence or worsening of depression, suicidal thoughts or behavior, and/or any unusual changes in mood or behavior. Inform patients that if they experience suicidal thoughts or behaviors, they should stop taking WEGOVY. That instruction no longer appears in any current GLP-1 label. Zepbound's Section 5.9, revised 10/2024, opened with the identical sentence about other weight management products. Neither cited any event from semaglutide's or tirzepatide's own trials.

Saxenda was the exception, and the difference matters

Liraglutide's warning was not borrowed. It cited its own trial data from the original December 2014 label onward, and by the April 2023 revision read: In SAXENDA adult clinical trials, 9 (0.3%) of 3384 SAXENDA-treated patients and 2 (0.1%) of the 1941 placebo-treated patients reported suicidal ideation; one of these SAXENDA-treated patients attempted suicide. In a SAXENDA pediatric clinical trial, 1 (0.8%) of the 125 SAXENDA-treated patients died by suicide. There was insufficient information to establish a causal relationship to SAXENDA. That last sentence is the label's own; truncating it would misrepresent what the FDA approved. The section then closed with the same monitor-and-discontinue instruction quoted from Wegovy above.

Where the four labels stand today

Two survivals may mislead a reader scanning for "mood". First, adverse-reaction tables still report psychiatric terms as observed frequencies rather than warnings: Saxenda's list anxiety (2.0% versus 1.6% on placebo in adults) and depression (4% versus 2.4% in the pediatric trial), and Wegovy's pediatric table lists anxiety at 4% versus 2% on placebo. Second, all four Medication Guides name mood changes among the symptoms of low blood sugar — Wegovy's as "irritability or mood changes", the other three as "anxiety, irritability, or mood changes". Neither is a mood warning.

What the evidence actually shows

The disproportionality signal, and why it cannot settle anything

A disproportionality analysis of the WHO global adverse-drug-reaction database (JAMA Network Open, 2024) found 107 suicidal or self-injurious reports for semaglutide and 162 for liraglutide, and detected significant disproportionality for one thing only: semaglutide-associated suicidal ideation, reporting odds ratio 1.45 (95% CI 1.18–1.77). In a sensitivity analysis restricted to cases that also reported antidepressant use, that figure rose to 4.45 (2.52–7.86). That is the easiest number on this page to misread. It compares how often ideation was reported among people who had already filed an adverse-event report. It is not a measure of anyone's risk, and it does not say that a GLP-1 taken alongside an antidepressant multiplies anything by four. The authors' conclusion was that this "warrants urgent clarification" — not that a risk had been shown.

Plainly: such an analysis counts how often a side effect is reported for one drug versus all others. It has no denominator, because only reports enter the database; it cannot adjust for why the drug was prescribed; and it is exquisitely sensitive to publicity, since once a story runs, reports of that exact event rise for that exact drug. It flags where regulators should look. It cannot show that a drug caused anything — and the co-reported-antidepressant subgroup is its most fragile part, because those people are already being treated for a condition that itself carries this risk.

The cohort studies built to answer it

Wang and colleagues (Nature Medicine, 2024) compared 240,618 US patients with overweight or obesity (mean age 50.1, 72.6% women) in TriNetX records against users of non-GLP-1 anti-obesity drugs. Over six months semaglutide was associated with a lower rate of first-time suicidal ideation (hazard ratio 0.27, 95% CI 0.20–0.36) and of recurrent ideation (0.44, 0.32–0.60), replicated in 1,589,855 patients with type 2 diabetes. The authors are blunt — "This was a retrospective observational study, so no causal inferences can be drawn" — and name reverse causation directly. A protective-looking hazard ratio is as much a candidate for confounding as a harmful-looking one.

Ueda and colleagues (JAMA Internal Medicine, 2024) used the Swedish and Danish national registers: 124,517 people starting a GLP-1 against 174,036 starting an SGLT2 inhibitor, 2013–2021, mean follow-up 2.5 years. There were 77 suicide deaths in the GLP-1 group and 71 in the comparator — 0.23 versus 0.18 per 1,000 person-years, hazard ratio 1.25 (95% CI 0.83–1.88). For suicide death plus non-fatal self-harm the ratio was 0.83 (0.70–0.97); for incident depression and anxiety disorders, 1.01 (0.97–1.06). Their most useful line is about precision, not point estimates: "the upper limit of the confidence interval was compatible with an absolute risk increase of no more than 0.16 events per 1000 person-years." The population was mostly type 2 diabetes, mean age 60, 45% women — not the reader of this page.

What the randomized trials recorded, and what they did not

The most honest number here is in a 2025 JAMA Psychiatry systematic review: of 144 randomized placebo-controlled GLP-1 trials meeting inclusion criteria, only 27 systematically recorded suicide or self-harm events at all. Those 27 covered 32,354 people on a GLP-1 and 27,042 on placebo; events were rare in both arms (0.047 versus 0.042 per 100 person-years), rate ratio 0.76 (95% CI 0.48–1.21, P = .25). An increase is unlikely "within the context of RCTs" — a narrower claim than it sounds, because trials exclude those at highest baseline risk.

Two sponsor-run post hoc analyses looked at psychiatric measures directly. In the STEP 1, 2 and 3 trials (3,377 participants, 69.6% women, mean age 49 — STEP 5 was a separate 304-participant pool reported alongside them), week-68 PHQ-9 scores were 2.0 on semaglutide 2.4 mg and 2.4 on placebo — estimated treatment difference −0.56 (95% CI −0.81 to −0.32), which the authors themselves call statistically significant but "not considered clinically meaningful". On the Columbia-Suicide Severity Rating Scale, 1% or fewer in either arm reported suicidal ideation or behaviour, with no difference between groups. The matching SURMOUNT-1, -2 and -3 analysis (N = 4,056, 2026) found the same shape: PHQ-9 difference −0.6; ideation in 0.6% of each group; non-fatal suicidal behaviour in 0.1% of tirzepatide participants versus none on placebo.

These are safety analyses of observed in-trial data, not efficacy estimands, and must not be set beside the weight-loss percentages from the same trials, which are reported under treatment-policy or treatment-regimen estimands. Both programmes also screened out the people most likely to be at risk, though by different rules: the STEP analyses excluded anyone with severe major depressive disorder in the two years before screening or a screening PHQ-9 of 15 or above, while the three SURMOUNT trials excluded any "history of significant active or unstable major depressive disorder (MDD) or other severe psychiatric disorder within the last 2 years" and "any lifetime history of a suicide attempt". So, as the SURMOUNT authors put it, "Further study of tirzepatide's safety in persons with significant psychiatric illness may be warranted."

The other direction

Reported well-being also moved the other way, and the same caution applies. In the STEP 1–4 patient-reported outcomes the confirmatory secondary endpoint was physical functioning, and superiority was confirmed there; other quality-of-life scales were numerically better but were not confirmatory endpoints, and SF-36v2 Role Emotional in STEP 2 was not improved. A drug that eases knee pain and breathlessness moves a well-being score without touching mood pharmacologically. Improvement is as open to confounding as harm.

FDA and EMA: the same conclusion, 21 months apart

The FDA's January 11, 2024 communication reported a preliminary FAERS review, said it "has not found evidence that use of these medicines causes suicidal thoughts or actions", but added that "we cannot definitively rule out that a small risk may exist". It promised a meta-analysis and a Sentinel study; both arrived in the January 13, 2026 communication. The meta-analysis pooled 91 placebo-controlled trials covering 107,910 patients (60,338 on a GLP-1, 47,572 on placebo); the Sentinel cohort covered 2,243,138 people with type 2 diabetes, comparing 1,161,983 GLP-1 starters with 1,081,155 SGLT2-inhibitor starters. Neither found an increased risk: "the totality of these studies does not support a causal relationship between the use of GLP-1 RAs and the occurrence of SI/B."

Europe got there first and did less. The EMA's Pharmacovigilance Risk Assessment Committee closed its review at its meeting of April 8–11, 2024 (highlights published April 12), finding that "the available evidence does not support a causal association" between five GLP-1 substances "and suicidal and self-injurious thoughts and actions", and that "no update to the product information is warranted." The regulators agree on the science; they differed only in what it obliged them to do. The FDA had US label text to remove; the EU did not. The current EMA product information for Saxenda and Wegovy mentions suicide, suicidal ideation and self-harm nowhere at all — we checked both documents — matching FDA's account that the warning began as a US weight-loss-labelling convention. We have not traced every historical EU revision, so that describes the product information as it stands today.

If you are already taking an antidepressant

No SSRI, SNRI, bupropion or any other antidepressant is named anywhere in the current Wegovy, Zepbound, Saxenda or Foundayo labels — not in Contraindications, not in Drug Interactions, not in Warnings and Precautions. In the three injectables those sections address two things only: blood sugar, and the absorption of swallowed medicines. Wegovy's Section 7.2 reads: WEGOVY causes a delay of gastric emptying and thereby has the potential to impact the absorption of concomitantly administered oral medications… Monitor the effects of oral medications concomitantly administered with WEGOVY. Consider increased clinical or laboratory monitoring for medications that have a narrow therapeutic index or that require clinical monitoring. Zepbound's adds warfarin as its worked example of a narrow-therapeutic-index drug to monitor. Saxenda's carries the same monitoring instruction for delayed absorption of oral medications. Foundayo differs in kind: it carries a CYP3A4 framework, capping its own dose alongside strong CYP3A4 inhibitors and advising against strong inducers. Whether a particular antidepressant falls into those categories is a question for a prescriber or pharmacist holding your full medication list. The labels do not answer it.

Nothing in any of the four labels addresses whether nausea or reduced food intake affects a swallowed antidepressant. The gastric-emptying language above is the closest they come, and it concerns absorption pharmacology, not missed or vomited doses. If that is your situation it is a conversation to have with the prescriber before anything changes — none of this is a reason to alter an antidepressant on your own.

The one weight-loss drug where this is genuinely different

Contrave is a weight-management medicine containing bupropion — the same molecule as several antidepressants — and its current label still carries a boxed warning headed "WARNING: SUICIDAL THOUGHTS AND BEHAVIORS — SUICIDALITY AND ANTIDEPRESSANT DRUGS", which tells prescribers: In patients of all ages who are started on CONTRAVE, monitor closely for worsening, and for the emergence of suicidal thoughts and behaviors. Advise families and caregivers of the need for close observation and communication with the prescriber. Unlike the GLP-1s it addresses antidepressants explicitly: it is contraindicated with other bupropion-containing products and with monoamine oxidase inhibitors, and warns that coadministration with CYP2D6 substrates including certain antidepressants (SSRIs and many tricyclics)… should be approached with caution and should be initiated at the lower end of the dose range of the concomitant medication. That is what a real drug–antidepressant interaction section looks like.

What the evidence says about this group specifically

A post hoc subgroup analysis of STEP 1, 2, 3 and 5 identified 539 of 3,683 randomized participants taking an antidepressant at baseline. In STEP 1, mean change from baseline to week 68 was −15.7% on semaglutide versus −0.2% on placebo among those on antidepressants, and −14.7% versus −2.8% among those not; adverse events were "generally similar between semaglutide and placebo in participants on ADs at baseline". These are exploratory subgroup means, not the trials' primary results, which use a different analytical framework — and the same exclusions applied. If your depression is currently severe, this analysis does not describe you.

Questions for your prescriber

• Given my mental-health history, is there anything the removed warning used to cover that the label no longer prompts you to check?
• Will you screen my mood before I start and at follow-up visits, and how — PHQ-9, or something else?
• Are any of my medicines ones where delayed absorption would matter, and does anything I take interact through CYP3A4?
• What symptoms should make me contact you rather than wait, and who do I reach outside office hours?
• If my mood changes, or I cannot keep medicines down, who reviews what and in what sequence?

The FDA's instruction to clinicians survived the label change: "If individuals disclose that they are experiencing SI/B, refer them to mental health professionals for evaluation." The warning left the label. The monitoring conversation should not leave the appointment.

Related: side effects in women, the telehealth medical review, GLP-1s in menopause, label eligibility.

Sources

1. FDA. FDA Requests Removal of Suicidal Behavior and Ideation Warning from Glucagon-Like Peptide-1 Receptor Agonist (GLP-1 RA) Medications, Drug Safety Communication, January 13, 2026. Page · PDF.
2. FDA. Update on FDA's ongoing evaluation of reports of suicidal thoughts or actions in patients taking a certain type of medicines approved for type 2 diabetes and obesity, Drug Safety Communication, January 11, 2024. PDF · Page. Note on the date: FDA's own two documents disagree. The 2026 communication describes this one as "issued on January 30, 2024", but the PDF's masthead, the communication's own FDA page and FDA's Drug Safety Communications index all give 01-11-2024 (checked August 28, 2026), so January 11, 2024 is the date used above.
3. DailyMed. WEGOVY (semaglutide) injection and tablets, current SPL v19, published June 30, 2026 — Recent Major Changes records Suicidal Behavior and Ideation as removed 02/2026. Label.
4. DailyMed. ZEPBOUND (tirzepatide) injection, current SPL v38, published May 6, 2026. Label.
5. DailyMed. SAXENDA (liraglutide) injection, current SPL v22, published June 15, 2026. Label.
6. DailyMed. FOUNDAYO (orforglipron) tablets, current SPL v10, published August 13, 2026. Label.
7. DailyMed. CONTRAVE extended-release (naltrexone/bupropion), current SPL v27, published November 17, 2025 — boxed warning on suicidal thoughts and behaviors; MAOI contraindication; CYP2D6 interaction language naming SSRIs and tricyclics. Label.
8. FDA. WEGOVY prescribing information revised 10/2025 (last edition carrying Section 5.10 Suicidal Behavior and Ideation), archived copy of the Drugs@FDA PDF. Archived PDF.
9. FDA. ZEPBOUND prescribing information revised 10/2024, Section 5.9, archived copy. Archived PDF.
10. FDA. SAXENDA original prescribing information, revised 12/2014, Section 5.8, archived copy. Archived PDF.
11. FDA. SAXENDA prescribing information revised 04/2023, Section 5.8 including the pediatric trial paragraph and its causality caveat, archived copy. Archived PDF.
12. European Medicines Agency. Meeting highlights from the Pharmacovigilance Risk Assessment Committee (PRAC), 8–11 April 2024, published 12 April 2024. Page.
13. Schoretsanitis G, Weiler S, Barbui C, Raschi E, Gastaldon C. Disproportionality Analysis From World Health Organization Data on Semaglutide, Liraglutide, and Suicidality. JAMA Netw Open. 2024;7(8):e2423385. PMID 39163046.
14. Wang W, Volkow ND, Berger NA, Davis PB, Kaelber DC, Xu R. Association of semaglutide with risk of suicidal ideation in a real-world cohort. Nat Med. 2024;30(1):168–176. PMID 38182782.
15. Ueda P, Söderling J, Wintzell V, et al. GLP-1 Receptor Agonist Use and Risk of Suicide Death. JAMA Intern Med. 2024;184(11):1301–1312. PMID 39226030. Erratum JAMA Intern Med. 2024;184(11):1396, retrieved August 28, 2026: it corrects a patient count in Figure 1 only (143,651 rather than 107,341 SGLT2-inhibitor new users in Sweden). No figure quoted above comes from Figure 1.
16. Ebrahimi P, Batlle JC, Ayati A, 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;82(9):888–895. PMID 40105856. Erratum JAMA Psychiatry. 2025;82(9):947 ("Errors in Numbers in Meta-Analysis"), retrieved August 28, 2026: person-years of exposure were misreported for one included study, and the abstract, text, Figures 2–3 and Tables 1–2 were corrected on July 16, 2025. The figures quoted above were re-read from the corrected article and are the post-correction values; the journal states the corrections do not affect the findings or conclusions.
17. Wadden TA, Brown GK, Egebjerg C, et al. Psychiatric Safety of Semaglutide for Weight Management in People Without Known Major Psychopathology: Post Hoc Analysis of the STEP 1, 2, 3, and 5 Trials. JAMA Intern Med. 2024;184(11):1290–1300. PMID 39226070.
18. Wadden TA, Oquendo MA, Kushner RF, et al. Psychiatric Safety of Tirzepatide in People With Obesity and No Known Major Psychopathology: A Post Hoc Analysis of SURMOUNT. Obesity (Silver Spring). 2026;34(3):565–578. PMID 41537305.
19. Kushner RF, Fink-Jensen A, Frenkel O, et al. Efficacy and safety of semaglutide 2.4 mg according to antidepressant use at baseline: A post hoc subgroup analysis. Obesity (Silver Spring). 2024;32(2):273–280. PMID 37989717.
20. Effect of semaglutide 2.4 mg on physical functioning and weight- and health-related quality of life in adults with overweight or obesity: Patient-reported outcomes from the STEP 1–4 trials. Diabetes Obes Metab. 2024;26(7):2945–2955. PMID 38698650. Correction Diabetes Obes Metab. 2025;27(3):1631, retrieved August 28, 2026: it revises a P value, some baseline characteristics and several confidence intervals and figure panels. None of the corrected values is quoted above.
21. 988 Suicide & Crisis Lifeline. 988lifeline.org · SAMHSA, 988 program page.