The aim of the present study was to evaluate the efficacy of short-term low-dose intravenous augmentative citalopram (10 mg/d) versus clomipramine (25 mg/d) versus placebo in a sample of patients with MDE and partial or no response to selective serotonin reuptake inhibitors (SSRIs). Fifty-four patients with a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision, MDE and partial or no response to SSRIs per os (21-item Hamilton Depression Rating Scale [HAM-D21] score reduction, <50% or < or =25%, respectively, compared with pretreatment scores) were selected and randomized to citalopram (n = 18), clomipramine (n = 18), or placebo (n = 18) intravenous augmentation. The augmentation regimen lasted 5 days during which patients were maintained on their previous treatment with oral SSRIs. Analyses of variance with repeated measures on HAM-D(21), collected daily in blind-raters design, were performed to detect any change of depressive symptoms between the 3 groups. In addition, the number of responders and remitters was computed in the 3 groups of treatment. At end point, a significant treatment effect (F= 4.57; P = 0.015) and time-by-treatment effect (F = 11.22; P < 0.0001) were found on HAM-D21 total scores in favor of citalopram and clomipramine versus placebo, with a superiority of citalopram over clomipramine on overall symptoms (P = 0.05) as well as on anxiety-somatization symptoms (P = 0.027). The number of responders was significantly superior in the active treatment groups versus the placebo group ([chi](2)(2) = 16.36; P < 0.0001). The same result was found, considering the number of remitters ([chi](2)(2) = 13.50; P < 0.0001). Present findings suggest that both clomipramine and citalopram intravenous augmentation at low doses and for a short period are well tolerated and superior to placebo in major depressives with partial or no response to oral SSRIs with a possible superiority of citalopram over clomipramine with regard to anxiety-somatization symptoms. The lack of double-blind conditions and the limited sample size may limit the confidence in the reported results, and larger randomized controlled trials are warranted to confirm the present findings.