Liraglutide, a GLP-1 receptor agonist, regulates appetite via receptors in the brain. Because of concerns regarding the potential of centrally-acting anti-obesity medications to affect mental health, pooled neuropsychiatric safety data from all phase 2 and 3a randomized, double-blind trials with liraglutide 3.0 mg were evaluated post hoc.Data from the liraglutide weight-management programme were pooled. Across trials, individuals with a body mass index ≥30 or ≥27 kg/m with weight-related comorbidities were randomized to once-daily subcutaneous liraglutide 3.0 mg (n = 3384) or placebo (n = 1941), both with a 500 kcal/d deficit diet, plus exercise. Adverse events related to neuropsychiatric safety were collected in all trials. Additionally, in the phase 3a trials, validated mental-health questionnaires were prospectively and systematically administered.In the pooled analysis of 5325 randomized and exposed individuals, rates of depression (2.1 vs 2.1 events/100 person-years) and anxiety (1.9 vs 1.7 events/100 person-years) through adverse event reporting were similarly low in liraglutide and placebo groups. Nine (0.3%) individuals receiving liraglutide and 2 (0.1%) receiving placebo reported adverse events of suicidal ideation or behaviour. In phase 3a trials, mean baseline Patient Health Questionnaire-9 scores of 2.8 ± 3.0 vs 2.9 ± 3.1 for liraglutide vs placebo improved to 1.8 ± 2.7 vs 1.9 ± 2.7, respectively, at treatment end; 34/3291 individuals (1.0%) receiving liraglutide 3.0 mg vs 19/1843 (1.0%) receiving placebo reported suicidal ideation on the Columbia-Suicide Severity Rating Scale.Results of this exploratory pooled analysis provide no cause for concern regarding the neuropsychiatric safety of treatment with liraglutide 3.0 mg in patients similar to those included in the examined trials. Although there was a small numerical imbalance in suicidal ideation with liraglutide through adverse event reporting, no between-treatment imbalances in suicidal ideation/behaviour or depression were noted through prospective questionnaire assessments.