OBJECTIVE To test the efficacy of late-luteal phase dosing of sertraline hydrochloride in women with moderate-to-severe premenstrual dysphoric disorder. This highly prevalent disorder often causes significant psychosocial impairment. METHODS Double-blind, crossover trial of each 2-menstrual cycle of baseline, sertraline treatment, and placebo. Randomization to sertraline treatment vs placebo occurred after a 2-cycle, drug-free period. METHODS A large outpatient multispecialty clinic in central Texas. METHODS Fifty-seven women aged 19 to 49 years with a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, diagnosis of premenstrual dysphoric disorder. METHODS Late-luteal phase treatment with sertraline hydrochloride in daily doses of 50 mg (cycle 1) followed by 100 mg (cycle 2) vs placebo. METHODS The 22-item calendar of premenstrual experiences was completed daily and constituted the primary outcome measure, consisting of a total score and behavioral and physical factor scores. RESULTS A repeated-measures analysis of variance for crossover designs found a significant beneficial effect from sertraline treatment in improving the calendar of premenstrual experiences total (P < .01), behavioral factor (P < .01), and physical factor (P < .04) scores. Most women improved when taking sertraline, 50 mg, although a dose increase to 100 mg yielded further improvement in approximately 25% of women. Use of sertraline was extremely well tolerated; the only adverse event reported by 10% or more of women was insomnia in 8 (14%) of them. CONCLUSIONS Luteal phase treatment with sertraline was a safe and effective treatment for moderate-to-severe premenstrual dysphoric disorder. Further controlled studies are needed to confirm the results of this preliminary study.