To accrue systematic information in different ovulatory disorders on the precise relationship among endocrine response, clinical outcome, and the occurrence of complications, we treated 114 patients with pulsatile GnRH (2.5-5.0 micrograms, iv, every 60 min) for 187 cycles and compared them to 20 normal menstrual cycles. Thirty of these patients had primary hypogonadotropic amenorrhea (PHA; 40 cycles), 33 had other forms of hypogonadotropic hypogonadism (HH; 55 cycles), and 51 had polycystic ovary syndrome (PCOS; 92 cycles). Daily blood samples were drawn for hormone determinations. In PCOS, 50 cycles were preceded by GnRH analog suppression. PHA treatment cycles were characterized by the reestablishment of a normal endocrine pattern, almost no dose-related endocrine differences, elevated ovulatory (93%) and conception rates (23%), and no multiple pregnancies. In the HH subjects the ovulatory (91%) and pregnancy rates (31%) were high; however, while the lower GnRH dose elicited a normal endocrine pattern, the 5-micrograms dose induced excessive folliculogenesis and high estradiol levels and was associated with most of the multiple pregnancies of this study (three of four). GnRH analog suppression was successfully used to avoid recurrence of ovarian over-stimulation in two HH subjects. Finally, GnRH analog suppression in PCOS permitted normalization of the follicular phase endocrine pattern, achievement of good ovulatory (76%) and pregnancy (28%) rates, and avoidance of multiple pregnancies; however, luteal phase steroid secretion was abnormal, and the abortion rate remained elevated (43%). Obesity was associated with a reduced ovulatory rate in PCOS, but not in hypogonadotropic, subjects. Thus, we can conclude that in pulsatile GnRH ovulation induction: 1) a profound hypogonadotropic condition, whether spontaneous as in PHA or induced with GnRH analogs as in other ovulatory disorders, is associated with optimal menstrual cycle restoration, high ovulatory and conception rates, and virtually absent risks of multiple pregnancy; 2) residual hypothalamic activity in HH may be responsible for supraphysiological pituitary-ovarian stimulation and result in multiple pregnancy unless a low GnRH dose (2.5 micrograms/bolus) or GnRH analog pretreatment is employed; 3) obesity does not affect treatment outcome in hypogonadotropic patients; and 4) the high spontaneous abortion rate in PCOS may be related to corpus luteum dysfunction.