Does Wegovy Cause Depression? GLP-1 Mental Health Evidence Review

Wegovy does not cause depression. Neither does Zepbound, Ozempic, Mounjaro, Saxenda, or Foundayo. Across five large independent datasets — SELECT (17,604 patients, Lincoff 2023 NEJM^[1]), STEP 1/2/3/5 post-hoc (n=3,377, Wadden 2024 JAMA Intern Med^[4]), the Wang TriNetX EHR cohort (240,618 semaglutide users, Nat Med 2024^[3]), the Ueda Nordic registry (n=298,236, JAMA Intern Med 2024^[5]), and the McIntyre FAERS Bradford-Hill analysis^[6] — the signal points the other direction. PHQ-9 depression scores improved on semaglutide vs placebo. Incident suicidal ideation was lower in GLP-1 users. The EMA PRAC concluded the same in April 2024; the FDA removed the Section 5 suicidality precautionary warning from Wegovy, Saxenda, and Zepbound on January 13, 2026 after reviewing 91 placebo-controlled trials in 107,910 patients. This article covers the depression and anxiety evidence in detail, the medication interactions with SSRIs, SNRIs, bupropion, Contrave, lithium, and second-generation antipsychotics, the body image and disordered eating intersection, and when mental health stabilization should precede a GLP-1.

The honest summary

• No causal link between GLP-1 use and depression or suicidality. Five independent datasets agree. The 2023–2024 EMA review concluded “the available evidence does not support a causal association.”
• Depression scores improve on semaglutide. The Wadden 2024 post-hoc^[4] of STEP 1/2/3/5 (n=3,377, 68 weeks) found a PHQ-9 treatment difference of −0.56 favoring semaglutide (p<0.001). The effect is small in absolute terms but directionally consistent across four trials.
• Incident suicidal ideation is lower, not higher. Wang 2024^[3] reported HR 0.27 (95% CI 0.20–0.36) for incident suicidal ideation in overweight/obese semaglutide users vs comparators in the TriNetX EHR network, with similar results for recurrence (HR 0.44).
• No clinically significant interactions with SSRIs, SNRIs, bupropion, or lithium. Section 7 of every FDA-approved GLP-1 label (Wegovy, Zepbound, Ozempic, Mounjaro, Saxenda, Foundayo) does not list any of these drug classes as interacting agents.
• Contrave (naltrexone + bupropion) is the genuine exception. Its boxed warning, seizure threshold considerations, CYP2D6 interactions, and current/prior bulimia or anorexia nervosa contraindication are meaningfully different from a pure GLP-1.
• Antipsychotic-induced weight gain responds to GLP-1s. Larsen 2017 JAMA Psychiatry^[8] (n=103) showed liraglutide 1.8 mg produced −5.3 kg at 16 weeks in clozapine- or olanzapine-treated schizophrenia patients with prediabetes, vs +0.7 kg on placebo. The intervention does not interfere with antipsychotic efficacy.
• Active anorexia nervosa or unstable bulimia is a stop sign. Active eating disorders, untreated severe MDD, recent suicide attempt, or unstable bipolar disorder warrant stabilization before any weight-loss medication.

Why this article exists

“Does Wegovy cause depression” and the broader GLP-1-mental-health cluster attract roughly 1,800 monthly Google searches in the US. The concern is reasonable: the 2023 EMA review of suicidality reports, the FAERS signal that triggered it, and a media cycle through 2024 that conflated case reports with causal evidence have all combined to leave patients (and many primary-care prescribers) uncertain. The precautionary Section 5 warning on Wegovy, Saxenda, and Zepbound was not removed until January 2026.

The actual evidence base, read carefully, is unusually clean. Five datasets with different designs, different populations, and different sponsors all point the same direction. We have a dedicated companion piece on the FDA + EMA postmarket surveillance arc and one on the anhedonia / emotional blunting question. This article is the practical synthesis: depression, anxiety, medication interactions, eating-disorder safety, and the action plan.

What the trial data actually show

SELECT (Lincoff 2023 NEJM, PMID 37952131^[1]). 17,604 adults with overweight or obesity and established cardiovascular disease but without diabetes, randomized to semaglutide 2.4 mg weekly vs placebo, followed for a mean of 39.8 months. The primary cardiovascular endpoint (MACE) was reduced 20% on semaglutide. Psychiatric adverse events — including depression, anxiety, and suicidality — were tracked as part of safety reporting and did not differ meaningfully between arms. This is the largest and longest-duration RCT of semaglutide in overweight/obesity to date, and the absence of a psychiatric safety signal at 17,604 patients over nearly three and a half years is load-bearing evidence.

STEP 1 (Wilding 2021 NEJM, PMID 33567185^[2]). The pivotal Wegovy obesity trial: 1,961 adults, −14.9% body weight at 68 weeks on semaglutide 2.4 mg vs −2.4% on placebo. PHQ-9 depression screening was performed throughout. No signal of worsened depression emerged.

Wadden 2024 STEP 1/2/3/5 post-hoc (JAMA Intern Med, PMID 39226070^[4]). Pooled psychiatric safety analysis across four STEP trials, n=3,377, in people without known major psychopathology. PHQ-9 depression scores were lower on semaglutide than placebo at week 68 (treatment difference −0.56, p<0.001), suicidal ideation by Columbia Suicide Severity Rating Scale (C-SSRS) was rare and did not differ between arms, and no completed suicides occurred. The PHQ-9 effect size is modest but directionally consistent with weight-loss-improves-mood literature.

Wang 2024 (Nat Med, PMID 38182782^[3]). TriNetX EHR network analysis of 240,618 patients with overweight/obesity prescribed semaglutide vs non-GLP-1 comparators (e.g., other anti-obesity medications). Adjusted hazard ratio 0.27 (95% CI 0.20–0.36) for incident suicidal ideation; HR 0.44 for recurrent suicidal ideation in patients with prior history. Observational and unblinded, but the magnitude and direction are robust across multiple sensitivity analyses including propensity matching by age, sex, BMI, prior psychiatric history, and antidepressant use.

Ueda 2024 Nordic registry (JAMA Intern Med, PMID 39226030^[5]). Sweden + Denmark binational registry, 298,236 adults using GLP-1 receptor agonists (semaglutide and liraglutide) vs SGLT2 inhibitor comparators. Hazard ratio 0.98 (95% CI 0.79–1.21) for suicide death, i.e., no association. Self-harm and incident depression likewise showed no signal.

McIntyre 2024 FAERS Bradford-Hill (Expert Opin Drug Saf, PMID 38087976^[6]). A formal application of the Bradford-Hill causal-inference criteria to the FAERS suicidality reports. The conclusion was explicit: “no causal link between GLP-1 RAs and suicidality exists.” The FAERS signal that triggered the EMA review was attributable to confounding by indication and notoriety reporting, not pharmacology.

Anxiety: what we know and what we do not

Anxiety has been studied less rigorously than depression in the GLP-1 trial program. The Wadden 2024 post-hoc^[4] included GAD-7 anxiety screening alongside PHQ-9 and found no worsening on semaglutide. The SELECT psychiatric safety reporting^[1] tracked anxiety adverse events and showed no excess. Patient-reported anxiety symptoms during early titration are common and almost always reflect either (1) acute GI tolerability stress, (2) hypoglycemia in patients also on insulin or sulfonylureas, or (3) anticipatory anxiety about side effects — not a pharmacologic anxiety effect of the medication itself. For separate context on how anxiety and weight loss interact biologically, see our dedicated anxiety + weight-loss evidence review.

SSRIs, SNRIs, and GLP-1s: the interaction question

Section 7 Drug Interactions of every FDA-approved GLP-1 label (Wegovy, Zepbound, Ozempic, Mounjaro, Saxenda, Foundayo) does not list SSRIs, SNRIs, mirtazapine, trazodone, or bupropion as interacting agents. There is no clinically significant pharmacokinetic interaction between any GLP-1 and the standard antidepressant classes. The only general caveat in every GLP-1 label is the gastric-emptying warning for narrow-therapeutic-index oral drugs — warfarin, digoxin, levothyroxine, certain immunosuppressants — and antidepressants are not in that category.

Two practical considerations are worth flagging anyway:

• Additive nausea during titration. SSRIs carry their own GI side-effect profile in the first 2–4 weeks: sertraline ~26% nausea, escitalopram ~15–18%, paroxetine ~20–26%. GLP-1 titration adds 18–44% nausea on its own. Starting an SSRI and a GLP-1 simultaneously is tolerable for most patients but predictably rougher than staggering them by 4–6 weeks.
• Weight effect of the SSRI matters separately. The Serretti 2010 meta-analysis (J Clin Psychiatry, PMID 21062615^[7]) of 116 antidepressant studies puts paroxetine in the weight-gain cluster with mirtazapine and amitriptyline; fluoxetine is weight-neutral to mildly weight-reducing short-term; sertraline and escitalopram are essentially weight-neutral; bupropion is mildly weight- reducing. If a patient on long-term paroxetine is struggling on a GLP-1, a switch to sertraline or escitalopram (in coordination with the prescribing psychiatrist) can remove a contributing weight-gain pressure. We have detailed evidence reviews on paroxetine and fluoxetine for the per-drug detail.

The Bollinger 2025 cohort (Dig Dis Sci, PMID 39604664^[10]) provides reassurance for patients on baseline mood or anxiety treatment: in a real-world MASLD cohort on weight-management therapy including GLP-1s, the presence of a mood or anxiety disorder did not attenuate weight loss. Patients on chronic antidepressant therapy achieved comparable weight outcomes to patients without a psychiatric history.

Bupropion (Wellbutrin) and Contrave

Bupropion is structurally and pharmacologically different from SSRIs. It is a norepinephrine and dopamine reuptake inhibitor (NDRI) approved for major depressive disorder, seasonal affective disorder, and smoking cessation. At 300–400 mg/day it is mildly weight-reducing rather than weight-neutral, which makes it the antidepressant most commonly considered for patients with depression and a weight concern, including ADHD patients where bupropion is also used off-label. There is no clinically significant interaction between bupropion and any FDA-approved GLP-1.

Contrave (naltrexone 8 mg + bupropion 90 mg per tablet, maintenance 32/360 mg/day) is the exception that requires care. The Contrave label carries an FDA boxed warning for suicidal thoughts and behaviors in patients under 25 (inherited from the bupropion class warning), and Contrave is explicitly contraindicated in seizure disorder, current or prior bulimia or anorexia nervosa, uncontrolled hypertension, chronic opioid use, and concurrent MAOI use. The COR-I trial (Greenway 2010 Lancet, PMID 20673995^[9]) established Contrave at −6.1% body weight at 56 weeks vs −1.3% on placebo — meaningful but well below the −14.9% (Wegovy) and −20.9% (Zepbound) GLP-1 magnitudes. Stacking Contrave with a GLP-1 is not standard practice and has not been studied in adequate RCTs; the seizure-threshold and CYP2D6 considerations argue against casual combination. Our Contrave evidence review has the full safety footprint.

Lithium and GLP-1 co-administration

Lithium has no pharmacokinetic interaction with GLP-1s, but the GI tolerability profile of a GLP-1 introduces a clinically relevant indirect concern: dehydration and reduced sodium intake during nausea, vomiting, or diarrhea can elevate serum lithium toward toxic levels. Lithium is renally cleared; volume depletion and reduced sodium intake both decrease lithium clearance. A patient stable on a maintenance lithium level who develops a GI flare on GLP-1 titration can move from therapeutic to toxic without a dose change.

Practical pattern that has worked well in clinical practice:

• Baseline lithium level immediately before initiating a GLP-1.
• Repeat level 2–4 weeks after each dose escalation, and any time the patient has >48 hours of vomiting, diarrhea, or reduced oral intake.
• Aggressive hydration counseling and a lower threshold for temporarily holding or down-titrating the GLP-1 during GI illness than for a patient not on lithium.
• Continued coordination with the prescribing psychiatrist on target lithium level and any dose adjustments.

This is not a contraindication. It is a monitoring layer. Lithium and GLP-1s can be co-administered safely with attention to hydration and serum-level monitoring.

Antipsychotic-induced weight gain

Second-generation antipsychotics — olanzapine, clozapine, quetiapine, risperidone — carry well- documented metabolic side effects: weight gain, insulin resistance, dyslipidemia, and elevated diabetes risk. Olanzapine and clozapine are the worst offenders. Patients on these medications often gain 5–15 kg in the first 12 months and have historically had few effective pharmacologic options because anti-obesity medications were either contraindicated (Contrave + seizure threshold) or untested in this population.

Larsen 2017 JAMA Psychiatry (PMID 28601891^[8]) was the first major RCT of a GLP-1 in antipsychotic-treated patients. 103 adults with schizophrenia spectrum disorder, prediabetes, and overweight/obesity on clozapine or olanzapine were randomized to liraglutide 1.8 mg vs placebo for 16 weeks. The liraglutide arm lost a mean of 5.3 kg; the placebo arm gained 0.7 kg (between-group difference 6.0 kg, p<0.001). Glycemic control improved, and crucially, antipsychotic efficacy was not compromised — PANSS (Positive and Negative Syndrome Scale) scores remained stable. Subsequent observational work and the more recent 2025 incretin/SGA cohort data continue to support GLP-1s and tirzepatide as reasonable interventions for antipsychotic-induced weight gain.

Practical implication: a patient stable on clozapine or olanzapine who has gained substantial weight is not disqualified from GLP-1 therapy. Coordination with the prescribing psychiatrist is required, but the data support the intervention.

Foundayo (oral orforglipron) and SSRIs

Foundayo is the first FDA-approved oral non-peptide GLP-1 receptor agonist (orforglipron 17.2 mg). Like Rybelsus (oral semaglutide), it has gastric-emptying physiology, but unlike Rybelsus it does not have the strict fasting + 30-minute-no- food administration window because it is not a peptide that needs SNAC absorption enhancement. The relevant SSRI consideration is the same gastric-emptying concern that applies to every GLP-1: narrow-therapeutic-index oral drugs may have altered absorption kinetics. SSRIs are not narrow- therapeutic-index drugs, and no specific Foundayo + SSRI interaction has been identified in the orforglipron Phase 3 program or in the FDA label.

The rare consideration worth being aware of: some patients report variability in SSRI onset of action when starting a GLP-1, which probably reflects altered gastric emptying affecting peak plasma concentrations rather than total bioavailability. The clinical consequence is essentially nil for SSRIs at steady state; the consideration matters more for as-needed psychiatric medications like benzodiazepines, where delayed Tmax can change the perceived effect.

Body image and disordered eating: a real caution

This is the area where clinical judgment matters most and where the trial data are least informative — STEP and SURMOUNT explicitly excluded patients with active eating disorders. The practical framework:

• Active anorexia nervosa is a stop sign. A patient with active restrictive eating, current BMI below normal, or recent significant unintentional weight loss should not be prescribed a GLP-1. The medication is appetite-suppressing in a population that does not need appetite suppression and may reinforce restrictive behaviors and amplify medical risk.
• Unstable or active bulimia nervosa is also a stop sign for Contrave (explicit FDA contraindication) and warrants extreme caution for any weight-loss medication. Stabilization through eating- disorder treatment first.
• History of restrictive eating in remission requires individualized assessment. A patient with a remote history of anorexia who has been in stable recovery for years and has developed obesity in midlife is a different clinical situation from a patient with a recent active illness. Coordination with the patient's eating-disorder treatment team is essential. The GLP-1 can unmask body-image distress in some patients as they approach a lower weight; pre-emptive psychological support is reasonable.
• Binge eating disorder (BED) is a different situation. BED is the most common eating disorder in patients with obesity. Several small studies suggest GLP-1s reduce binge frequency, consistent with the appetite-suppression mechanism. BED in stable treatment is not a contraindication; active untreated BED warrants coordinated care.
• Watch for emerging restrictive behaviors during rapid weight loss. The Wadden 2024 post-hoc^[4] tracked eating-disorder symptoms and did not find emergent eating disorders at the population level, but individual cases are reported. A patient who begins skipping meals, eliminating food groups, or fixating on weight numbers during GLP-1 treatment should be assessed.

When mental health stabilization should precede GLP-1 therapy

The clinical questions to answer before starting a GLP-1 in a patient with active psychiatric illness:

• Is the depression actively suicidal? Active suicidal ideation with intent or plan, or a suicide attempt within the prior 6 months, should be stabilized before adding any new medication. This is not because GLP-1s cause suicidality — the evidence is clear that they do not — but because the priority during an acute crisis is psychiatric stabilization, not a 68-week weight-loss trajectory.
• Is the MDD adequately treated? A patient with untreated severe MDD will not adhere to a GLP-1 regimen, will not engage with diet and activity recommendations, and will not benefit from the weight loss psychologically the way an adequately-treated patient will. Get the depression treated first.
• Is the bipolar disorder stable? Active mania or rapid cycling is not the time to add a new medication. Euthymic bipolar disorder on stable mood stabilizers is generally fine.
• Is there active substance use disorder? Active alcohol use disorder, opioid use disorder, or stimulant use disorder warrants treatment of the SUD first. Of note, GLP-1s have emerging signals for alcohol use reduction, but this is not yet an indication and does not change the stabilization priority.
• Is the eating disorder in stable remission? See above. Active anorexia or bulimia requires eating-disorder treatment first.

Once stabilized, the same evidence base applies. Patients with treated psychiatric illness can and do achieve the same weight outcomes as patients without (Bollinger 2025^[10]).

The patient action plan

1. List every psychiatric medication. Include dose, duration on dose, prescriber, and target indication. Bring it to the obesity medicine visit.
2. Get current labs. Baseline TSH, comp metabolic panel, HbA1c, lipids. If on lithium: baseline lithium level. If on valproate or other anticonvulsants: baseline level.
3. Coordinate explicitly. Ask the obesity medicine prescriber to communicate with the prescribing psychiatrist. A brief note is sufficient: medication list, plan, monitoring schedule, when to escalate.
4. Establish red flags up front. Patient should know: any new-onset suicidal thoughts, severe worsening of depression or anxiety, emerging restrictive eating, or symptoms of lithium or valproate toxicity warrant immediate contact with the prescribing clinicians.
5. PHQ-9 and GAD-7 at baseline and 12-week intervals. Simple, validated, free. Document the trajectory.
6. Crisis resources visible. 988 (Suicide and Crisis Lifeline), text HOME to 741741 (Crisis Text Line). Save them in the patient's phone before any new medication is started.
7. Hold the GLP-1 during acute psychiatric decompensation. Not because the medication caused it, but because adding GI side effects to a mental health crisis helps no one.

Bottom line

• The five-dataset convergence is unusually clean: SELECT, STEP, Wang TriNetX, Wadden post-hoc, Ueda Nordic registry, and the McIntyre FAERS Bradford-Hill analysis all point the same direction. GLP-1s do not cause depression or suicidality.
• Depression scores improve modestly on semaglutide (Wadden 2024^[4]). Incident suicidal ideation is lower in EHR cohorts (Wang 2024^[3]). The Nordic registry shows no association with suicide death (Ueda 2024^[5]).
• No clinically significant interactions with SSRIs, SNRIs, bupropion, or lithium. Contrave is the genuine exception and carries a boxed warning + seizure threshold + eating- disorder contraindications that pure GLP-1s do not.
• Antipsychotic-induced weight gain responds to GLP-1 therapy (Larsen 2017 JAMA Psychiatry^[8]) without compromising antipsychotic efficacy.
• Lithium co-administration requires hydration counseling and serum-level monitoring during GI side effects. Not a contraindication, a monitoring layer.
• Active anorexia, unstable bulimia, suicidal ideation with plan, or untreated severe MDD warrants psychiatric stabilization first. Once stable, the same weight outcomes are achievable (Bollinger 2025^[10]).
• Coordinate. Document. Use PHQ-9 and GAD-7. Save 988 in the phone.

Related research and tools

• FDA + EMA postmarket surveillance arc on GLP-1 and suicidality — the detailed regulatory history and the Section 5 warning removal
• GLP-1 anhedonia and emotional blunting — the reward-blunting phenomenology that is real and is not depression
• SSRI + GLP-1 interactions and combined use — the per-drug verbatim FDA-label walkthrough
• Antidepressants and weight on a GLP-1 — the cross-class weight-effect comparison
• Does anxiety cause weight loss? — the bidirectional anxiety + weight question
• Does buspirone cause weight loss? — the weight-neutral anxiolytic option
• Contrave (naltrexone + bupropion) evidence review — the boxed warning, seizure threshold, and eating- disorder contraindications
• Bupropion (Wellbutrin) weight effects — the modest weight-reducing antidepressant option
• Paxil (paroxetine) weight effects — the worst-offender SSRI for weight gain
• Prozac (fluoxetine) weight effects — the relatively weight-neutral SSRI
• Stress, cortisol, and food noise on a GLP-1 — the broader psychological backdrop
• What to eat on a GLP-1: the protein-first guide — the meal-pattern evidence base

Important disclaimer. This article is educational and does not constitute medical advice. Patients with active or recent suicidal ideation should contact 988 (Suicide and Crisis Lifeline) or text HOME to 741741 (Crisis Text Line) immediately and should not initiate or adjust any weight-loss medication without coordinated psychiatric care. Patients on lithium, valproate, or other narrow-therapeutic- index psychiatric medications require serum-level monitoring when starting or escalating a GLP-1. Active anorexia nervosa or unstable bulimia nervosa is a contraindication to weight- loss pharmacotherapy. Decisions about adding, switching, or stopping psychiatric medications must be made with the prescribing psychiatrist. PMIDs were independently verified against the PubMed E-utilities API on 2026-05-28; FDA label statements were verified verbatim from DailyMed.

Last verified: 2026-05-28. Next review: every 6 months, or sooner if new RCT or large-cohort data on GLP-1 psychiatric safety are published.