INTRODUCTION: Selective serotonin reuptake inhibitors (SSRIs) are known to induce delayed orgasm and delayed ejaculation, while their effect on other aspects of sexual function, such as sexual motivation, arousal, and erectile function are unclear. AIM: In order to evaluate the effect of chronic administration of two SSRIs, citalopram and fluoxetine, on normal sexual function, we studied the parameters of male sexual behavior, erectile function, and ejaculation on 48 healthy male volunteers, aged 29.5 +/- 4.9, in a randomized, placebo-controlled, double-blind, double-dummy study. Methods. The subjects were randomized to receive placebo (16 subjects), or fluoxetine (20 mg/day) (16 subjects) or citalopram (20 mg/day) for the first week, and 40 mg/day in the following 3 weeks (16 subjects). MAIN OUTCOME MEASURES: Sexual function was investigated at the screening and at the end of the study by means of test of penile erection (TPE) and masturbation ejaculation latency time (MELT) performed during visual erotic stimulation, and at each visit by self-filled questionnaires (International Index Erectile Function [IIEF-15] and Golombock Rust Inventory of Sexual Satisfaction [GRISS]). RESULTS: All the erectile parameters, evaluated by means of RigiScan Plus during TPE, were not significantly different when both fluoxetine and citalopram were compared with placebo. A delay in the ejaculation time was observed both during citalopram and during fluoxetine treatment when compared with placebo, reaching a statistical significance only with citalopram. During the treatment with citalopram and fluoxetine, the IIEF-15 score of all items decreased except for those items related to sexual desire; however, the scores were significantly lower only for the citalopram treatment. CONCLUSIONS: The treatment with citalopram or with fluoxetine was confirmed to delay ejaculation, but was significant only for citalopram. Citalopram and fluoxetine did not affect sexual desire. Citalopram and fluoxetine did not directly affect penile erection as objectively assessed by RigiScan, although an impairment in the subjective assessment of erectile function was observed, but was significant only for citalopram, and it was thought to be a possible consequence of the delayed ejaculation perceived as a trouble.