GLP-1 Medicines and Suicidal Thoughts: What the FDA, EMA and Cohort Studies Found About Mood Risk

Do GLP-1 medicines raise the risk of suicidal thoughts? On the evidence available in 2026, no. The FDA, the European Medicines Agency, pooled trial analyses and several large cohort studies have all looked for a link between GLP-1 drugs and suicidal thoughts or depression, and none has found one.

The worry was reasonable. European regulators opened a review in July 2023 after case reports of suicidal thoughts and thoughts of self-injury in people using liraglutide and semaglutide. Until January 2026, the US labels for Wegovy, Zepbound and Saxenda carried a warning about suicidal behavior and ideation. Plenty of people starting a GLP-1 protocol still remember those headlines.

This article walks through what each regulator decided, what the STEP and SURMOUNT trial data show about depression scores, what the population studies found, and where the evidence is still thin. That last part matters most for people with a psychiatric history and for anyone using a lower dose, since no study has looked at mood on microdose regimens.

Why did the FDA remove the suicide warning from GLP-1 labels?

The January 2026 Drug Safety Communication

The FDA removed the warning because a comprehensive review found no increased risk of suicidal ideation or behavior with GLP-1 receptor agonists. The agency announced the change in a Drug Safety Communication in January 2026 and asked the makers of Saxenda (liraglutide), Wegovy (semaglutide) and Zepbound (tirzepatide) to take the language out of the Warnings and Precautions section.

The FDA review had three parts. First, a meta-analysis of 91 placebo-controlled GLP-1 trials covering 107,910 patients, 60,338 on a GLP-1 receptor agonist and 47,572 on placebo. It did not show an increased risk of suicidal ideation or behavior, and it did not show more anxiety, depression, irritability or psychosis either. Second, a study in the FDA's Sentinel claims system compared intentional self-harm in 1,161,983 new GLP-1 users and 1,081,155 new SGLT2 inhibitor users with type 2 diabetes, 2,243,138 people in all, and found no increased risk. Third, a review of published studies, which the FDA summarised by saying the totality of these studies does not support a causal relationship.

The warning was never specific to GLP-1 science. GLP-1 medicines approved only for type 2 diabetes did not carry it. The wording on the weight-loss products traced back to older weight-loss medicines, and the FDA's stated aim was to make the labeling consistent across all GLP-1 products.

What did European regulators conclude?

The EMA safety committee review

The European Medicines Agency concluded in April 2024 that the available evidence does not support a causal association between GLP-1 receptor agonists and suicidal or self-injurious thoughts and actions. The conclusion came from PRAC, the agency's Pharmacovigilance Risk Assessment Committee, and covered dulaglutide, exenatide, liraglutide, lixisenatide and semaglutide.

The review started in July 2023 after case reports from people using liraglutide and semaglutide. In November 2023 the committee asked the manufacturers for more data. After weighing non-clinical studies, clinical trials, post-marketing reports and two observational studies, PRAC decided no update to the product information was warranted. The manufacturers still have to keep monitoring these events and report new evidence, which is standard for any safety question that has been closed but not forgotten.

Does semaglutide or tirzepatide worsen depression scores in trials?

STEP and SURMOUNT post hoc analyses

No. In the large weight-loss trials, depression scores on semaglutide and tirzepatide ended slightly lower than on placebo, not higher. Both drugs were tested with the PHQ-9, a standard depression questionnaire, and the Columbia-Suicide Severity Rating Scale, which asks directly about suicidal thoughts and behavior.

Wadden and colleagues pooled the STEP 1, 2 and 3 trials of semaglutide 2.4 mg in JAMA Internal Medicine in 2024. Across 3377 participants, mean PHQ-9 scores started at 2.0 on semaglutide and 1.8 on placebo, which both mean no or minimal depression. At week 68 they were 2.0 and 2.4, an estimated treatment difference of -0.56. People on semaglutide were less likely to move into a more severe depression category, with an odds ratio of 0.63. On the suicide scale, 1% or fewer reported suicidal ideation or behavior, with no difference between groups. STEP 5, with 304 participants over 104 weeks, looked similar.

The same lead author ran a parallel analysis of SURMOUNT-1, -2 and -3 for tirzepatide at 5, 10 or 15 mg, published in Obesity in 2026. Across 4056 participants, PHQ-9 scores went from 2.7 to 1.9 on tirzepatide and from 2.6 to 2.4 on placebo by week 72, a difference of -0.6. The STEP authors described the change as small and not clinically meaningful, which is the honest reading for both drugs: the scores did not get worse, and they did not dramatically improve.

Who was left out of the trial safety data?

The limit that matters most

People with known major psychiatric conditions were left out. Both the STEP and SURMOUNT analyses were restricted to participants without known major psychopathology, and the baseline PHQ-9 scores confirm the groups started out with little or no depression. That design protects trial participants, but it means the trials cannot say much about someone with active depression, a past suicide attempt or a serious mental illness.

The SURMOUNT analysis also reported nonfatal suicidal behavior in 0.1% of tirzepatide users and in no one on placebo. With rates that low, a handful of events can create or erase an apparent difference, and the authors judged the rates similar to other incretin-based therapies. They still wrote that further study in people with significant psychiatric illness may be warranted, which is the right caveat to carry forward.

Funding is worth knowing about too. Novo Nordisk employees co-authored the STEP analysis and Eli Lilly employees co-authored the SURMOUNT analysis. The FDA and EMA reviews, and the independent cohort studies below, are the counterweight to that.

What do real-world studies show about suicidal ideation on semaglutide?

Health records from hundreds of thousands of patients

The largest real-world study of semaglutide found a lower, not higher, rate of suicidal ideation. Wang, Volkow and colleagues, writing in Nature Medicine in 2024, used TriNetX electronic health records to compare 240,618 people with overweight or obesity who were prescribed semaglutide or a non-GLP-1 anti-obesity medicine. The semaglutide group had a mean age of 50.1 years and was 72.6% female.

Over 6 months, semaglutide was linked with a hazard ratio of 0.27 for first-time suicidal ideation and 0.44 for recurrent suicidal ideation, consistent across sex, age and ethnicity. The pattern was replicated in 1,589,855 people with type 2 diabetes. Electronic health record studies cannot rule out that healthier or better-supported people were more likely to get semaglutide, so a lower rate here is not proof of a protective effect. It does make a large hidden harm unlikely.

Do GLP-1 users die by suicide more often?

Nationwide registers from Sweden and Denmark

The best data on suicide deaths show no statistically significant increase, and suicide among GLP-1 users was rare. Ueda and colleagues, in JAMA Internal Medicine in 2024, followed 124,517 adults in Sweden and Denmark who started a GLP-1 receptor agonist and 174,036 who started an SGLT2 inhibitor, another diabetes drug class with no known mood effects. Most had type 2 diabetes. GLP-1 users had a mean age of 60, and 45% were women.

Over a mean follow-up of 2.5 years there were 77 suicide deaths among GLP-1 users and 71 among SGLT2 inhibitor users. The weighted rates were 0.23 and 0.18 per 1000 person-years, a hazard ratio of 1.25 with a confidence interval of 0.83-1.88, which crosses 1 and so does not show a difference. The authors calculated that any true increase was no more than 0.16 events per 1000 person-years.

The broader outcomes pointed the other way. Suicide death or nonfatal self-harm combined had a hazard ratio of 0.83, and new depression or anxiety disorders 1.01, essentially even. A separate genetic study by Nguyen and colleagues in Clinical Obesity in 2024 used Mendelian randomisation, which uses natural gene variants in the GLP-1 receptor as a stand-in for the drug, and did not find evidence that GLP-1 receptor activity raises suicide attempts. The authors noted the result awaits confirmation.

Can GLP-1 medicines improve mood or quality of life?

The 80-trial meta-analysis

GLP-1 medicines did not change depressive symptoms on average, but they did modestly improve quality of life and eating behavior. That is the finding of Pierret and colleagues in JAMA Psychiatry in 2025, who pooled 80 double-blind placebo-controlled trials with 107,860 patients with overweight, obesity or diabetes, mean age 60.1.

There was no significant difference in serious or nonserious psychiatric adverse events, and the effect on depressive symptoms was close to zero (g = 0.02). Quality-of-life scores did improve: mental health-related quality of life by g = 0.15, and weight-related quality of life by 0.27. Restrained eating (0.35) and emotional eating (0.32) showed the largest shifts. In plain terms these are small to moderate effects.

The eating-behavior result fits what users often describe as quieter food noise, a topic covered in our article on food noise on a GLP-1 protocol. For the broader brain-chemistry picture, see GLP-1s and mental health. None of this makes a GLP-1 an antidepressant, and nobody should use one as a substitute for mental health care.

Does a GLP-1 microdose change the mood risk?

What can and cannot be said

Nobody knows directly, because no study has measured mood or suicidality on microdose regimens. Every figure in this article comes from standard doses: semaglutide 2.4 mg in STEP, tirzepatide 5 to 15 mg in SURMOUNT, and whatever doses patients were prescribed in the cohort studies.

Since the full-dose evidence does not show a raised risk, there is no specific reason to expect a lower dose to be worse. That is an inference, not a tested finding. The evidence gap for people with existing psychiatric conditions applies at every dose, and lower doses do not close it.

The practical points hold regardless of dose. Anyone with a history of depression, a past suicide attempt or a serious mental illness should mention it before starting. Rapid weight change, low energy intake and stress can all affect mood, as our article on cortisol and stress eating explains, so a mood change is worth reporting even if the drug is not the cause. The FDA advises patients to tell a healthcare professional about new or worsening depression, suicidal thoughts or unusual changes in mood, and not to stop the medicine without discussing it first. Anyone in the US who is in crisis can call or text 988.

Frequently Asked Questions