A comprehensive meta-analysis was performed to address the possible association of fluoxetine with violence or aggression. Data from the United States Investigational New Drug Clinical Trial Databases for approved and potential indications (depression, obesity, bulimia nervosa, obsessive-compulsive disorder, smoking cessation, alcoholism; n = 3992) were evaluated. Statistically significantly fewer fluoxetine-treated patients (0.15%) than placebo-treated patients (0.65%) experienced events suggestive of aggression (hostility, personality disorder, antisocial reaction). A relative risk analysis indicated that aggression events were four times more likely to occur in placebo-treated patients than in fluoxetine-treated patients. Although the possibility that some rare phenomenon was not detected cannot be excluded, this meta-analysis did not show fluoxetine to be associated with an increased risk of emergence of violent or aggressive behaviour.