BACKGROUND The coincidence of major depressive disorder in bulimia nervosa ranges from 35% to 80%. Because of this comorbidity and because suicidality (suicidal acts and ideation) is an inherent part of depression, assessment of the risk of suicide in patients with bulimia nervosa is of considerable interest. METHODS Data from United States Investigational New Drug double-blind, placebo-controlled fluoxetine clinical trials in bulimia nervosa were analyzed comprehensively to assess the potential association between fluoxetine treatment and suicidality in 785 patients with DSM-III-R bulimia nervosa. Patients were predominantly women (98%), aged 17 to 63 years; of the randomly assigned patients, 16.9% exhibited 17-item Hamilton Rating Scale for Depression (HAM-D) total scores of 17 or greater at baseline (range, 0-31). Incidence of suicidality was analyzed by the incidence difference method. RESULTS No fatal suicidal acts occurred; 9 (1.15%) of 785 patients made nonfatal attempts; 24 (3.06%) experienced emergent (text-defined) suicidal ideation. No statistically significant increases in the incidence of suicidal acts or suicidal ideation were observed among fluoxetine-treated compared with placebo-treated patients. A smaller percentage of fluoxetine-treated (2.0%) than placebo-treated (3.8%) patients experienced emergence of substantial suicidal ideation (change in baseline HAM-D Item 3 [suicide item] score of 0 or 1 to 3 or 4 during therapy). A statistically significantly greater proportion of fluoxetine-treated than placebo-treated patients experienced improvement in suicidal ideation (decrease in HAM-D Item 3 score) from baseline to endpoint (p = .026). CONCLUSIONS Analyses of the incidence of suicidal acts and suicidal ideation did not indicate an increased risk of suicidality in patients with bulimia nervosa treated with fluoxetine compared with those treated with placebo.