Evaluation of suicidality during pharmacologic treatment of mood and nonmood disorders.

Double-blind, controlled clinical trial data were evaluated to assess a hypothetical relationship between fluoxetine and suicidality (suicidal acts and ideation) in patients with mood (n = 5,655) and nonmood disorders (n = 4,959) (Mantel-Haenszel incidence difference method). In mood disorders, act rates (suicide attempts/completions) were low (treatment differences nonsignificant). Substantial suicidal ideation emerged less frequently with fluoxetine than placebo and was comparable with fluoxetine and tricyclic antidepressants. Improvement in ideation was greater with fluoxetine than placebo; it was comparable with fluoxetine and tricyclic antidepressants (United States trials) and greater with tricyclic antidepressants than fluoxetine (international trials). In nonmood disorders, no suicides occurred. Act and emergent ideation rates were low (treatment differences nonsignificant). Results do not suggest a causal relationship between pharmacotherapy and emergence of suicidality. Fluoxetine or tricyclic antidepressants reduce suicidal ideation and may protect against the emergence of substantial suicidal ideation.