To determine the risk for secondary amenorrhea after pulse cyclophosphamide therapy in premenopausal women with systemic lupus erythematosus. Controlled, retrospective clinical study. Government referral-based research hospital. Thirty-nine women younger than 40 years treated with pulse cyclophosphamide therapy for active lupus nephritis or neuropsychiatric lupus. Sixteen women who received pulses of intravenous methylprednisolone were controls. Sixteen patients received pulse cyclophosphamide (0.5 to 1.0 g/m2 body surface area) monthly for a total of 7 doses (short-CY), and 23 patients received 15 or more doses (long-CY). Control patients were treated with monthly pulses of methylprednisolone (1.0 g/m2) for a total of nine doses. Rates of amenorrhea were evaluated according to duration of treatment (number of doses) and age at the initiation of pulse therapy. Two of 16 patients (12%) in the Short-CY group and 9 of 23 (39%) in the long-CY group developed sustained amenorrhea (P = 0.07). Rates of sustained amenorrhea (short- and long-CY) according to age at the start of pulse therapy were: < or = 25 years, 2/16 (12%); 26 to 30 years, 4/15 (27%); > or = 31 years, 5/8 (62%) (P = 0.04). The increased risk for sustained amenorrhea in patients treated with long-CY was most evident in patients older than 25 years (short-CY [2/12] compared with long-CY [7/11]; P = 0.03). Three other patients with short-CY had reversal of amenorrhea fewer than 12 months after cessation of therapy. Amenorrhea was not observed in any of the 16 control patients. Intermittent pulse cyclophosphamide therapy in patients with systemic lupus erythematosus is associated with sustained amenorrhea, which is related to both age and number of doses of cyclophosphamide.