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

Key Points

Question  Does treatment with semaglutide, 2.4 mg, once weekly affect the risk of developing symptoms of depression or suicidal ideation/behavior vs placebo in people with overweight or obesity without known major psychopathology?

Findings  In this post hoc analysis of the STEP 1, 2, and 3 trials and STEP 5 trial over 68 and 104 weeks, respectively, there were no clinically meaningful differences in mean Patient Health Questionnaire 9 scores throughout the trials between the semaglutide, 2.4 mg, and placebo groups. The proportion of participants reporting suicidal ideation/behavior (as assessed by the Columbia-Suicide Severity Rating Scale) was also similar between groups.

Meaning  The results of this post hoc analysis suggest that the risk of developing symptoms of depression or suicidal ideation/behavior was similar between semaglutide, 2.4 mg, and placebo.

Importance  Obesity is associated with numerous psychosocial complications, making psychiatric safety a consideration for treating people with obesity. Few studies have investigated the psychiatric safety of newly available antiobesity medications.

Objective  To evaluate the psychiatric safety of subcutaneous semaglutide, 2.4 mg, once weekly in people without known major psychopathology.

Design, Setting, and Participants  This post hoc analysis of pooled data from the randomized, double-blind, placebo-controlled, multicenter phase 3a STEP 1, 2, and 3 trials (68 weeks; 2018-2020) and phase 3b STEP 5 trial (104 weeks; 2018-2021) included adults with overweight or obesity; STEP 2 participants also had type 2 diabetes. Trial designs have been published previously.

Interventions  Semaglutide, 2.4 mg, vs placebo.

Main Outcomes and Measures  Depressive symptoms and suicidal ideation/behavior were assessed using the Patient Health Questionnaire (PHQ-9) and Columbia–Suicide Severity Rating Scale, respectively. Psychiatric and nervous system disorder adverse events were investigated.

Results  This analysis included 3377 participants in the STEP 1, 2, and 3 trials (2360 women [69.6%]; mean [SD] age, 49 [13] years) and 304 participants in STEP 5 (236 women [77.6%]; mean [SD] age, 47 [11] years). In the STEP 1, 2, and 3 trials, mean (SD) baseline PHQ-9 scores for the semaglutide, 2.4 mg, and placebo groups were 2.0 (2.3) and 1.8 (2.3), respectively, indicating no/minimal symptoms of depression. PHQ-9 scores at week 68 were 2.0 (2.9) and 2.4 (3.3), respectively; the estimated treatment difference (95% CI) between groups was −0.56 (−0.81 to −0.32) (P < .001). Participants treated with semaglutide vs placebo were less likely to shift (from baseline to week 68) to a more severe category of PHQ-9 depression (odds ratio, 0.63; 95% CI, 0.50-0.79; P < .001). Based on the Columbia–Suicide Severity Rating Scale, 1% or fewer of participants reported suicidal ideation/behavior during treatment, with no differences between semaglutide, 2.4 mg, and placebo. Psychiatric disorder adverse events were generally balanced between groups. Similar results were observed in STEP 5.

Conclusions and Relevance  The results of this post hoc analysis suggest that treatment with semaglutide, 2.4 mg, did not increase the risk of developing symptoms of depression or suicidal ideation/behavior vs placebo and was associated with a small but statistically significant reduction in depressive symptoms (not considered clinically meaningful). People with obesity should be monitored for mental health concerns so they can receive appropriate support and care.

Trial Registration  ClinicalTrials.gov Identifiers: STEP 1 (NCT03548935), 2 (NCT03552757), 3 (NCT03611582), and 5 (NCT03693430)

Obesity is associated with multiple health complications, including an increased risk of depressive disorders and other psychiatric illness, particularly in people with a body mass index (calculated as weight in kilograms divided by height in meters squared) of 40 or greater.^1-4 Depression may partially result from the pervasive weight stigma directed at people with obesity,^5,6 although there is a bidirectional relationship between these conditions in which depression increases the risk of subsequent weight gain.^2,7,8 The increased inflammatory state in people with obesity, which is a known risk factor for depression, may also contribute to this relationship.^9,10 Obesity-related complications, including type 2 diabetes (T2D) and osteoarthritis, further impair psychosocial status and quality of life.^11-13

Concerns about potential psychiatric adverse events (AEs) in people with clinically significant depression have accompanied the development of medications for chronic weight management,¹⁴ which, except for orlistat,¹⁵ access brain regions that regulate energy intake and may interact with mood-influencing pathways.¹⁶ Rimonabant was approved for weight management in Europe in 2006¹⁷ but was later withdrawn due to associations with depression and suicidal ideation (SI).¹⁸ These risks are formally assessed by regulatory authorities,^19,20 and with most medications approved for chronic weight management, health care professionals are advised to monitor patients for emerging or worsening psychiatric symptoms, including unusual changes in mood.^16,21

Semaglutide is a glucagon-like peptide-1 receptor agonist (GLP-1RA) that is approved globally for weight management at a once-weekly dosage of 2.4 mg. After 68 weeks, semaglutide, 2.4 mg, produces an approximately 10% to 15% mean reduction in baseline body weight, which is associated with clinically meaningful improvements in numerous health outcomes,^21-26 including cardiovascular events in people with overweight or obesity and a history of a cardiovascular event (but not T2D).²⁷ Neither semaglutide, 2.4 mg, nor liraglutide, 3.0 mg, another GLP-1RA approved for weight management, was associated with increased psychiatric AEs vs placebo in phase 3 trials.^21,25,28-30 However, the European Medicines Agency (EMA) and US Food and Drug Administration (FDA) are actively monitoring the psychiatric safety of GLP-1RAs following postmarketing surveillance reports of depression, SI, and suicidal behavior in people prescribed these medications for T2D or weight management.^20,31,32 To date, neither agency has found evidence of a causal association.^20,32 This post hoc analysis examines psychiatric data from 4 Semaglutide Treatment Effect in People With Obesity (STEP) trials to assess symptoms of depression, SI, and suicidal behavior, which is critical for evaluating postmarketing reports of these events in the general population.

This post hoc analysis evaluated pooled data from the STEP 1, 2, and 3 trials and data from STEP 5 (Supplement 1, Supplement 2, Supplement 3, and Supplement 4). All trials were performed at clinical research sites per the Declaration of Helsinki and Good Clinical Practice guidelines. Respective protocols were approved by local institutional review boards and ethics committees.

Trial designs for STEP 1, 2, 3, and 5 have been reported.^22-24,33 These were 68-week phase 3a (STEP 1, 2, and 3) or 104-week phase 3b (STEP 5) double-blind, placebo-controlled, multicenter trials. Participants were randomized to receive subcutaneous semaglutide, 2.4 mg, once weekly or placebo (2:1 [STEP 1 and 3]; 1:1:1 to semaglutide, 2.4 mg; semaglutide, 1.0 mg; or placebo [STEP 2]; or 1:1 [STEP 5]) as an adjunct to a lifestyle intervention. Participants receiving semaglutide, 1.0 mg, in STEP 2 were excluded from this analysis. In STEP 3, semaglutide, 2.4 mg, was provided as an adjunct to intensive behavioral therapy.²³ In all trials, the primary end point was change in body weight from baseline to the end of treatment.

Adults with overweight or obesity were recruited and provided written informed consent; STEP 2 also included adults with T2D. Validated mental health questionnaires were used to assess depression (major depressive disorder [MDD]), SI, and suicidal behavior. The Patient Health Questionnaire (PHQ-9) is a self-reported, 9-item screening tool that assesses symptoms of depression during the previous 2 weeks.³⁴ Total scores range from 0 to 27. Scores of 0 to 4, 5 to 9, 10 to 14, 15 to 19, and 20 or greater suggest no/minimal, mild, moderate, moderately severe, and severe symptoms of depression, respectively.³⁴ The Columbia-Suicide Severity Rating Scale (C-SSRS)³⁵ is a clinician-administered questionnaire used to identify SI and suicidal behavior.³⁶ SI is defined by the C-SSRS as (1) a wish to be dead; (2) nonspecific active suicidal thoughts; (3) active SI with any method (no plan) without intent to act; (4) active SI with some intent to act without a specific plan; and (5) active SI with a specific plan and intent. Suicidal behavior was also assessed. Additional C-SSRS methods are provided in the eMethods in Supplement 5.

Exclusion criteria included MDD during the previous 2 years or other severe psychiatric disorders, including, but not limited to, schizophrenia or bipolar disorder, or a screening PHQ-9 score of 15 or greater (moderately severe or greater depression), a lifetime suicide attempt, or SI with intent and with/without a specific plan or suicidal behavior (self-reported or in medical records) within 30 days before screening. Full inclusion and exclusion criteria have been reported.^22-24,33

The PHQ-9 and C-SSRS were administered at screening (1 week before randomization); at baseline (week 0); at weeks 12, 20, 36, 52, and 68 in STEP 1, 2, 3, and 5; and at weeks 84 and 104 in STEP 5 only. At each visit, the investigator reviewed the PHQ-9 and C-SSRS, and neuropsychiatric AEs were assessed through open-ended questioning. At any visit, participants who had a PHQ-9 score of 15 or greater or type 4 or 5 SI or suicidal behavior on the C-SSRS were referred to a mental health professional. Investigators could decide whether participants with a PHQ-9 score of 10 to 14 required referral.

AEs in the system organ classes of psychiatric disorders and nervous system disorders were identified using predefined Medical Dictionary for Regulatory Activities coding (version 22.1 [STEP 1, 2, and 3] or 23.1 [STEP 5]; International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use). AEs included those occurring during the in-trial period, plus 7 weeks of follow-up after treatment was discontinued.

STEP 1, 2, and 3 had the same trial duration, so data were pooled; STEP 5 had a longer trial duration, so data were analyzed separately. PHQ-9 scores, the proportion of participants reporting SI/suicidal behavior on the C-SSRS, and neuropsychiatric AEs were analyzed using the safety analysis set (all participants exposed to ≥1 dose of trial product) for the in-trial period. Changes in PHQ-9 scores while receiving treatment were reported at the end of treatment, as well as categorical shifts in PHQ-9 scores from baseline to end of treatment. For STEP 1, 2, and 3, the proportion of participants with AEs and event rates were adjusted using the Cochran-Mantel-Haenszel method to account for between-trial differences.

Changes in PHQ-9 scores and categorical shifts in PHQ-9 scores from baseline to the end of treatment were analyzed using analysis of covariance and logistic regression, respectively (with randomized treatment as a factor and baseline value as a covariate). A multiple imputation approach was used, and missing data were imputed from end of treatment measurements from participants in the same treatment group. No statistical analyses of SI/suicidal behavior were performed, as there were too few events. All analyses were post hoc and not adjusted for multiplicity. Statistical significance was defined as P < .05, with no correction for multiplicity. Statistical analyses were conducted using SAS, version 8.4 (SAS Institute).

This analysis included 3377 participants in STEP 1, 2, and 3 (pooled data) and 304 in STEP 5 (eFigure 2 in Supplement 5). Participants were predominantly female (2360 [69.9%] and 236 [77.6%], respectively) and White (2415 [71.5%] and 283 [93.1%], respectively; race was self-reported) (Table 1). Common comorbidities at screening included hypertension (1479 [43.8%] and 118 [38.8%], respectively) and dyslipidemia (1486 [44.0%] and 107 [35.2%], respectively).

In STEP 1, 2, and 3 and STEP 5, 533 (25%) and 49 (32%) participants receiving semaglutide, 2.4 mg, respectively, and 320 (25%) and 43 (28%) receiving placebo, respectively, reported concomitant psychiatric disorders (eTable 1 in Supplement 5). Information about antidepressant and other medication use, as well as concomitant nervous system disorders, is reported in eTable 1 in Supplement 5.

In-trial changes in mean PHQ-9 scores over time were minimal in both groups (eTable 2 in Supplement 5). In STEP 1, 2, and 3, mean (SD) PHQ-9 scores at baseline were 2.0 (2.3) and 1.8 (2.3) for the semaglutide, 2.4 mg, and placebo groups, respectively, which at week 68 were 2.0 (2.9) and 2.4 (3.3) (eFigure 1A in Supplement 5). The estimated treatment difference between groups for change in PHQ-9 scores from baseline to week 68 was −0.56 (95% CI, −0.81 to −0.32; P < .001), favoring semaglutide, 2.4 mg (eTable 3 in Supplement 5). Mean scores in both groups increased marginally through week 52 (Figure, A) but remained in the range of no/minimal symptoms of depression (eFigure 1A, eTable 2A in Supplement 5). Results were broadly similar for participants in STEP 5, although not statistically significant (estimated treatment difference, −0.20; 95% CI, −0.98 to 0.59; P < .63) (Figure, B; eFigure 1B, eTable 2B, and eTable 3 in Supplement 5).

In STEP 1, 2, and 3, participants assigned to semaglutide, 2.4 mg, who lost less than 5% of their body weight from baseline to week 68 had a mean (SD) increase of 0.5 (3.5) points on the PHQ-9 at end of treatment, similar to the mean (SD) increase of 0.8 (3.5) points in placebo-treated participants. eTable 4 in Supplement 5 shows additional data by percentage body weight loss.

The categorical shifts in PHQ-9 scores from baseline to end of treatment for STEP 1, 2, and 3 and STEP 5 are summarized in Table 2. In STEP 1, 2, and 3, among the 1697 participants treated with semaglutide, 2.4 mg, who reported no/minimal symptoms of depression at baseline, 1520 (89.6%) reported the same at week 68, 147 (8.7%) reported mild symptoms, 22 (1.3%) moderate symptoms, and 8 (0.5%) moderately severe or severe symptoms (Table 2). A similar but somewhat less favorable pattern was observed in those receiving placebo. From baseline to week 68, participants receiving semaglutide, 2.4 mg, vs placebo were less likely to shift to a more severe category of depression on the PHQ-9 (odds ratio [OR], 0.63; 95% CI, 0.50-0.79; P < .001) (eTable 5 in Supplement 5). Similar but nonsignificant findings were observed in STEP 5 (OR, 0.74; 95% CI, 0.34-1.60; P = .44) (eTable 5 in Supplement 5).

In STEP 1, 2, and 3, examining maximum PHQ-9 scores occurring at any postbaseline visit, 60 participants (2.9%) assigned to semaglutide vs 54 (4.4%) receiving placebo had a PHQ-9 score of 15 or greater at 1 or more assessments (eTable 6 in Supplement 5). Semaglutide was associated with a reduced risk of this outcome (OR, 0.54; 95% CI, 0.37-0.80; P = .002) (eTable 5 in Supplement 5). Corresponding values in STEP 5 were 1 (0.7%) and 3 participants (2.0%), respectively (OR, 0.76; 95% CI, 0.10-5.80; P = .79). Across the trials, 12 (0.6%) and 17 participants (1.4%) receiving semaglutide, 2.4 mg, and placebo, respectively, had a PHQ-9 score of 15 or greater at the end of treatment.

In STEP 1, 2, and 3, 30 participants (1.4%) receiving semaglutide, 2.4 mg, and 28 (2.2%) receiving placebo had a history of low- to moderate-risk SI as determined by the C-SSRS; corresponding numbers in STEP 5 were 4 (2.6%) and 2 participants (1.3%), respectively (Table 3). At baseline in STEP 1, 2, and 3, 1 participant receiving semaglutide, 2.4 mg, and 2 receiving placebo reported SI (low risk) (eResults in Supplement 5). No participants in STEP 5 reported SI at baseline.

From randomization through end of treatment in STEP 1, 2, and 3, 8 participants (0.4%) receiving semaglutide, 2.4 mg, and 7 (0.6%) receiving placebo reported incident SI; all but 1 case was characterized as low to moderate risk (Table 3). In STEP 5, 1 (0.7%) and 2 participants (1.4%) receiving semaglutide, 2.4 mg, and placebo, respectively, reported incident SI. Twelve of 18 SI cases (67%) had resolved by the end of the trial, spread evenly across groups; time of onset, nature, severity, and resolution are reported in eTable 7 in Supplement 5.

One participant who received semaglutide, 2.4 mg, and acknowledged incident SI (on the C-SSRS at week 36 and as a serious AE) also reported suicidal behavior (suicide attempt) at week 36 on the C-SSRS (STEP 1; eTable 7 in Supplement 5). Thereafter, the participant experienced low-risk SI at week 52. All reports of SI and suicidal behavior were resolved by the end of the trial. At baseline, the participant’s PHQ-9 score suggested mild symptoms of depression (total score 6), increasing to moderately severe symptoms (total score 16) at week 36 and returning to 6 at the end of the trial.

During clinic visits in STEP 1, 2, and 3, 6 participants (0.3%) receiving semaglutide, 2.4 mg, and 2 (0.2%) receiving placebo reported AEs of SI (Table 4), 1 of which was classed as a serious AE in the semaglutide, 2.4 mg, group, as mentioned previously. No participants in STEP 5 reported these behaviors. One participant who received placebo reported a serious AE of suicide attempt during STEP 1 (Table 4). The participant had a history of depression and mood swings, and at the first visit had mild symptoms of depression (PHQ-9 score of 5). The participant did not show worsening on the PHQ-9 or report SI on the C-SSRS. At visits before and after the attempt, the PHQ-9 score remained at 2. The participant was hospitalized for 2 days.

eTable 6 in Supplement 5 summarizes the number of participants who reported “thoughts that you would be better off dead, or hurting yourself in some way” on the PHQ-9 postbaseline. In STEP 1, 2, and 3, smaller proportions of participants receiving semaglutide, 2.4 mg, than placebo acknowledged these thoughts (3.4% vs 5.0%). Conversely, in STEP 5, a slightly higher proportion of participants treated with semaglutide, 2.4 mg, experienced these thoughts vs placebo (2.6% vs 2.0%).

Reports of psychiatric disorder AEs were balanced between the semaglutide, 2.4 mg, and placebo groups in all trials (Table 4; eTable 8 in Supplement 5), including disorders of sleep, anxiety, and depressed mood. Slight imbalances were observed in nervous system disorder AEs between treatment groups, driven mainly by an increased incidence of headaches (STEP 1, 2, and 3) and dizziness (STEP 1, 2, 3, and 5) in the semaglutide, 2.4 mg, group (Table 4; eTable 9 in Supplement 5).

In this post hoc analysis of the STEP 1, 2, 3, and 5 trials, semaglutide, 2.4 mg, relative to placebo, was not associated with increases in clinically significant symptoms of depression or in SI or suicidal behavior. Occurrence of these events, as assessed by PHQ-9 and C-SSRS, was low in both groups. Across the 4 trials, only 2.8% and 4.1% of semaglutide-treated and placebo-treated participants, respectively, had a postbaseline PHQ-9 score of 15 or greater, which is suggestive of moderately severe or greater symptoms of depression and requires evaluation by a mental health professional. Both percentages are consistent with the known risk of depression in people with obesity in the general population.^2-4 Most participants in both groups reported no change or improvement in their mood during the in-trial period, results that are similar to those observed with behavioral weight loss interventions.^14,37,38 Although in some analyses semaglutide, 2.4 mg, showed statistically greater improvements in depressive symptoms than placebo, we do not consider these differences to be clinically meaningful.

Incident SI, as assessed by the C-SSRS, was low (≤1%) in all trials, and no imbalance was observed between treatment groups, or as ascertained by AE reporting or question 9 of the PHQ-9. More participants in both groups acknowledged suicidal thoughts when completing the self-reported PHQ-9 than when responding to the clinician-administered C-SSRS. This discrepancy appears attributable to patient discomfort in directly disclosing sensitive information to health care professionals.³⁹ There were 2 reports of suicide attempts in the trials, 1 per treatment group. The participant receiving semaglutide, 2.4 mg, reported depressive symptoms related to life stressors; the participant who received placebo had a history of depression and mood swings, but not SI/suicidal behavior, and the event was potentially linked to life stressors.

Our findings are consistent with a pooled analysis of 5 trials (N = 5325) investigating liraglutide, 3.0 mg, another GLP-1RA approved for chronic weight management.^28-30 No differences were observed in symptoms of depression, SI, or suicidal behavior between recipients of liraglutide, 3.0 mg, and placebo, as assessed by the same validated measures used in the STEP trials.²⁹ However, AE reporting revealed a minimally higher rate of SI/suicidal behavior with liraglutide, 3.0 mg, than placebo (0.3% vs 0.1%).²⁹ Other reports have identified a slightly higher frequency of SI in people receiving semaglutide and liraglutide compared with other GLP-1RAs or metformin/insulin, although no causality was suggested, and investigations were deferred to the EMA and FDA.^40,41

Randomized clinical trials are critical for understanding postmarketing reports of depression and suicidality in the general population prescribed GLP-1RAs for T2D or weight management. After reviewing several datasets, the EMA concluded that “the available evidence does not support a causal association between the GLP-1RAs and suicide and self-injurious thoughts and actions.”²⁰ The FDA, in a preliminary evaluation of similar psychiatric events in individuals prescribed GLP-1RAs or similar incretin-based medications, concluded that it had “not found evidence that use of these medicines causes suicidal thoughts or actions.”³² However, regulators were unable to “definitively rule out that a small risk may exist” due to the small number of SI/suicidal behavior events in controlled and observational trials; surveillance and evaluation will continue.³² We encourage clinicians to monitor mental health in people with overweight or obesity, including those prescribed semaglutide, 2.4 mg, or similar medications. This is essential for ensuring that these individuals obtain the psychiatric care they require to manage anxiety, depression, and related psychiatric complications, which may occur in the presence or absence of weight management interventions.⁴²

Our analysis of SI and suicidal behavior used descriptive statistics rather than formal hypothesis testing due to an inadequate number of events. Other investigators addressed this challenge by using large, aggregated electronic health records. In an observational, retrospective analysis of 52 783 people receiving semaglutide for weight management, semaglutide, compared with non–GLP-1RA antiobesity medications, was associated with a 73% reduction in the risk of SI during the first 6 months of treatment.⁴³ Similar effects of semaglutide in relation to incident anxiety and depression were reported in a nationwide study in Taiwan.⁴⁴

Other psychiatric safety concerns were not apparent in our analysis, based on the number and type of AEs reported. The incidents of anxiety disorders, depression, and sleep disorders and disturbances were broadly similar between groups. Some nervous system disorders, including dizziness and headaches (known AEs of this treatment), were generally more frequent with semaglutide, 2.4 mg, than placebo.^21-26 A recent 2-year European database study showed that among 31 444 AE reports submitted for semaglutide, liraglutide, and tirzepatide, only 1.2% were psychiatric AEs.⁴⁵

Strengths of our analysis included its relatively large sample size, use of randomized clinical trials, and inclusion of well-validated measures of depressive symptoms, SI, and suicidal behavior, which were administered frequently. The trials also had high retention in both treatment arms (eFigure 2 in Supplement 5), thus limiting the risk of bias due to discontinuations. Limitations included the exclusion of individuals with a lifetime suicide attempt, current high-risk SI, recently diagnosed MDD, and other clinically significant psychopathology. These individuals could be among the most vulnerable in the general population to experiencing psychiatric AEs while taking semaglutide, 2.4 mg, or other weight management medications. A recent pooled analysis found that semaglutide, 2.4 mg, was safe and effective in patients who at baseline took antidepressant medication.⁴⁶ Controlled investigations and continued postmarketing surveillance are needed to assess the safety and effectiveness of semaglutide, 2.4 mg, and similar medications in people with obesity and major psychopathology.

This post hoc analysis of the STEP 1, 2, 3, and 5 trials found that semaglutide, 2.4 mg, did not increase the risk of developing symptoms of depression, SI, or suicidal behavior relative to placebo in people with overweight or obesity without known major psychopathology. A small number of patients taking semaglutide, 2.4 mg, may experience depression or SI, which may be related to psychosocial complications of obesity or other factors, including life stressors.

Accepted for Publication: June 29, 2024.

Published Online: September 3, 2024. doi:10.1001/jamainternmed.2024.4346

Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License. © 2024 Wadden TA et al. JAMA Internal Medicine.

Corresponding Author: Thomas A. Wadden, PhD, Center for Weight and Eating Disorders, 3535 Market St, Ste 3029, Philadelphia, PA 19104 (wadden@pennmedicine.upenn.edu).

Author Contributions: Dr Goldman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Egebjerg, Frenkel, Goldman, Overvad.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Wadden, Brown, Egebjerg, Frenkel, Goldman, McGowan.

Critical review of the manuscript for important intellectual content: Wadden, Brown, Egebjerg, Frenkel, Goldman, Kushner, McGowan, Overvad, Fink-Jensen.

Statistical analysis: Goldman, Overvad.

Administrative, technical, or material support: Overvad.

Supervision: Frenkel, Overvad.

Other—planning some of the statistical analyses: Wadden.

Conflict of Interest Disclosures: Dr Wadden reported grants from Novo Nordisk during the conduct of the study and personal fees from Novo Nordisk, advisory board service for WeightWatchers and Eli Lilly, and grant support from Eli Lilly outside the submitted work. Dr Egebjerg reports employment with and being a shareholder in Novo Nordisk. Dr Brown reported personal fees from the Research Foundation for Mental Hygiene Inc and Oui Therapeutics Inc outside the submitted work. Dr Frenkel reported personal fees from Novo Nordisk during the conduct of the study and personal fees from and being a shareholder in Novo Nordisk outside the submitted work. Dr Goldman reported employment with and being a shareholder of Novo Nordisk during the conduct of the study. Dr Kushner reported personal fees from Novo Nordisk outside the submitted work. Dr McGowan reported personal fees from Novo Nordisk, Eli Lilly, and Astra Zeneca, grants from Novo Nordisk, and being a shareholder in Reset Health outside the submitted work. Dr Overvad reported employment with Novo Nordisk. Dr Fink-Jensen reported grants from Novo Nordisk A/S outside the submitted work. No other disclosures were reported.

Funding/Support: This analysis was funded by Novo Nordisk A/S.

Role of the Funder/Sponsor: Representatives of the sponsor (Novo Nordisk A/S) were involved in the design and conduct of the analysis; collection, management, analysis, and interpretation of the data; and preparation, review, and approval of the manuscript. The authors chose the journal and made the final decision to submit the manuscript for publication.

Meeting Presentation: Data from this paper were presented at the Annual Meeting of The Obesity Society; October 14, 2023; Dallas, Texas.

Data Sharing Statement: See Supplement 6.

Additional Contributions: We thank the trial participants, investigators, and trial site staff who conducted the trials, as well as Maria Oquendo, MD, PhD, Perelman School of Medicine at the University of Pennsylvania, who shared her expertise on the results of different methods used to assess suicidal ideation and behavior and was not compensated for her contributions. Medical writing support was provided by Abbie Richold and Ruth Brown of Apollo, OPEN Health Communications, which was funded by Novo Nordisk in accordance with Good Publication Practice guidelines. Medical writers, working in collaboration with the first author (T.A.W.) and coauthors, assisted with drafting the Methods, portions of the Results, and the References. They also collaborated with the statistical coauthor (B.G.) and other coauthors in formatting the tables and figures for the published manuscript and online supplement. They participated with all authors in editing the final draft of the entire manuscript.