FDA Asks to Remove Suicidal Ideation Warnings From Three GLP-1 Drug Labels

What the FDA did on January 13, 2026

The FDA Drug Safety Communication requested that the relevant drug application holders remove information about the risk of suicidal ideation and behavior (SI/B) from the *Warnings and Precautions* section of three weight-management drug labels: Saxenda (liraglutide), Wegovy (semaglutide), and Zepbound (tirzepatide). FDA said its comprehensive evaluation did not identify an increased risk associated with GLP-1 receptor agonist use.

“Requested removal” is the precise regulatory description. FDA announced a requested labeling change directed to the companies that hold the approved applications; it did not withdraw the medicines, add a contraindication, or issue a new approval. Prescribing information can be revised through a formal labeling process, so readers checking a label around the announcement date should confirm the current FDA-approved version rather than assume every document changed simultaneously.

Why the warning was there—and why FDA investigated

According to FDA, SI/B language was included when Saxenda, Wegovy, and Zepbound were originally approved. Similar precautions appeared on other weight-loss drug labels because suicidal thoughts or behavior had been observed with a variety of older medicines used or studied for weight loss. That history supplied a cautious class-of-use precedent; it did not establish that these newer medicines caused the events.

In July 2023, after postmarketing reports in people taking GLP-1 medicines, FDA began a focused investigation. Spontaneous reports can identify a signal worth testing, but they cannot by themselves show causation: reports generally lack a randomized comparator, reporting is incomplete, and the underlying conditions, other medicines, and life circumstances may all affect risk. Obesity and diabetes can also coexist with depression, making careful comparison especially important.

FDA's preliminary review of clinical-trial and postmarketing information did not suggest a causal link. In its January 2024 update, however, the agency emphasized uncertainty because so few SI/B events occurred in individual trials. It could not then definitively rule out a small risk. The larger cross-program meta-analysis and the Sentinel analysis were designed to obtain a more precise and complementary answer.

The evidence FDA reviewed

FDA did not base the 2026 decision on one study. It combined controlled trial data, a large real-world comparison, and a review of published observational and pooled analyses. These sources answer somewhat different questions and have different weaknesses.

The placebo-controlled meta-analysis pooled development programs to overcome the imprecision produced by rare events in any one trial. FDA reported that it did not find an increased risk of suicidal ideation or behavior. It also found no increased risk for the other psychiatric adverse-event groupings it examined: anxiety, depression, irritability, and psychosis. This analysis gives the strongest protection against measured baseline differences because treatment assignment in the underlying trials was randomized.

The Sentinel analysis used administrative health-care claims from 10 data partners, covering new users between October 1, 2015, and September 20, 2023. It compared GLP-1 initiators with initiators of SGLT2 inhibitors, another medicine class used in type 2 diabetes. After controlling for measured baseline confounders, FDA found no increased intentional self-harm among GLP-1 users, either overall or in the subgroup recorded as having both type 2 diabetes and obesity.

What FDA concluded—and what it did not

FDA's regulatory conclusion was that the available evidence did not show an increased risk and that the totality of evidence did not support a causal relationship between GLP-1 medicines and SI/B. That supports removing a specific warning that implies a drug-related risk. It is appropriately reassuring at the class and population level.

It is not the same as proving that no person can experience depression, suicidal thinking, self-harm, or another psychiatric event while taking one of these drugs. A safety analysis can assess whether events are more frequent in treated groups than comparators; it cannot certify the cause of every future individual event. Nor does this decision establish that a GLP-1 medicine treats depression, prevents suicide, or improves mental health.

• Rare-event precision still has limits. Pooling 91 trials improves precision substantially, but very uncommon outcomes remain intrinsically difficult to quantify.
• Trial populations are selected. Eligibility rules, follow-up periods, doses, indications, and psychiatric assessment methods can differ among development programs.
• Claims capture recorded care. Sentinel can detect coded intentional self-harm, not every unreported thought or mood change; residual confounding may remain after statistical adjustment.
• The comparator matters. The large claims result concerns people with type 2 diabetes starting a GLP-1 medicine versus an SGLT2 inhibitor, not every weight-management population or every possible comparator.
• Safety surveillance continues. Removing a warning does not stop adverse-event reporting, pharmacovigilance, or future label changes if new evidence emerges.

What patients and caregivers should do

FDA advised patients to continue taking their medication as prescribed and discuss concerns with a health-care professional. A labeling update is not a reason to change a dose, discontinue treatment, or switch products without clinical guidance. Abruptly abandoning a prescribed plan may also leave diabetes, obesity, cardiovascular risk, or another treated condition inadequately managed.

1. If you feel well but are worried: ask the prescriber or pharmacist how the decision applies to your medicine and check the current Medication Guide and prescribing information.
2. If mood or behavior changes: tell a health professional about new or worsening depression, suicidal thoughts, or unusual changes in mood or behavior. Include when symptoms began, medication and dose changes, other drugs, and relevant stressors.
3. If there is immediate danger: in the United States, call or text 988 for the Suicide & Crisis Lifeline, call emergency services, or go to the nearest emergency department. Do not leave a person at imminent risk alone.
4. Report suspected adverse events: patients and clinicians can submit a report through FDA MedWatch. A report helps surveillance even when causation is uncertain.

Clinicians should discuss FDA's evidence review accurately while still evaluating disclosed SI/B and referring people for mental-health assessment when appropriate. Clinical care remains individualized: personal psychiatric history, current symptoms, concurrent medicines, benefits, and known nonpsychiatric risks all matter. For broader medication context, see our educational guides to semaglutide and tirzepatide.

EMA and international context

The U.S. review was not isolated. The European Medicines Agency's Pharmacovigilance Risk Assessment Committee (PRAC) opened its own review in July 2023 following case reports involving liraglutide and semaglutide. On April 12, 2024, EMA reported that available evidence did not support a causal association between dulaglutide, exenatide, liraglutide, lixisenatide, or semaglutide and suicidal or self-injurious thoughts and actions.

EMA considered nonclinical studies, clinical trials, postmarketing surveillance, and observational studies, including an analysis using electronic health records. PRAC concluded that no update to European product information was warranted. It also said marketing authorization holders should continue monitoring and report new evidence through periodic safety updates. The broad direction therefore agrees with FDA's later conclusion, although the agencies reviewed partly different product sets and acted within different labeling systems.

What the decision does not change about GLP-1 safety

This communication addressed one potential risk: suicidal ideation and behavior. It did not erase other warnings, precautions, contraindications, or common adverse reactions in each product's approved labeling. It also did not change which patients qualify for treatment, how dosing should be escalated, or how clinicians should monitor known risks.

Likewise, a conclusion about FDA-approved products should not be automatically extended to unapproved or counterfeit products. Formulation, concentration, dosing instructions, and quality controls may differ. Patients should obtain prescription medicines through licensed channels and use the exact product and dose their clinician prescribed.

The most useful reading is narrow: extensive available evidence did not support keeping a GLP-1-specific suicidality warning on the three named weight-management labels. Benefit-risk decisions still require the complete, current prescribing information and an individual clinical assessment. Readers who want the underlying pharmacology can review GLP-1 biology and how GLP-1 medicines work for weight loss.

The bottom line

FDA moved from signal detection to progressively stronger evidence: postmarketing reports triggered review; the preliminary assessment found no association but retained uncertainty; a much larger pooled trial analysis, Sentinel cohort, and published literature then converged on no increased risk. On that record, the agency requested removal of SI/B language from Saxenda, Wegovy, and Zepbound labeling.

That is a meaningful safety and labeling conclusion, not a universal mental-health guarantee. Patients should neither panic because an event was once listed nor dismiss symptoms because the warning is being removed. Continue prescribed therapy unless a clinician advises otherwise, promptly disclose concerning changes, and seek immediate crisis support when needed.