OBJECTIVE To examine whether changes in menstrual cycle length occurred in women who received 20 mg of fluoxetine per day and women who received 60 mg of fluoxetine per day as compared with women who received placebo. METHODS A two-cycle, single-blind placebo phase was followed by a six-cycle double-blind phase. Only placebo nonresponders were randomized to one of three treatment arms: placebo, fluoxetine 20 mg/day, or fluoxetine 60 mg/day. Cycle length was established using a prospective daily calendar. Cycle-length data from women who completed at least three treatment cycles were used in this analysis. Cycle-length change was defined a priori as cycle lengthening or shortening of 4 days or more (1 or more standard deviations from the mean change between baseline cycles). RESULTS Eleven of the 62 women receiving 60 mg of fluoxetine per day and seven of the 70 women receiving 20 mg of fluoxetine per day but only one of the 61 women in the placebo group demonstrated cycle-length change at the end of the first treatment cycle (P = .011). In total, 15% of the women in the fluoxetine 60 mg/day group demonstrated a cycle-length change during two of the three observed cycles compared with 6% of the women in the fluoxetine 20 mg/day group and 3% of the women in the placebo group (P = .048). Women with cycle-length changes demonstrated either shortening or lengthening only. There were no differences between the number of women with changes in either direction. CONCLUSIONS Cycle-length change occurred more frequently in the fluoxetine 60 mg/day group than in the fluoxetine 20 mg/day or placebo groups. The effects of selective serotonin reuptake inhibitors on menstrual cycle length are mostly unknown and warrant careful monitoring in women of reproductive age, who are the prime consumers of these medications.