OBJECTIVE: To evaluate the predictive value of ovarian response to the gonadotropin-releasing hormone agonist (GnRH-alpha) stimulation test (GAST) in anovulatory women during ovulation induction treatment. METHODS: Thirty-five patients entered and completed the study. Following administration of GnRH-alpha during the early follicular phase, the relationship between estradiol (E(2)) response, ovarian response and subsequent pregnancy was investigated. RESULTS: Three patterns of E(2) response were noted. Pattern A patients were hyper responders in induction treatment;their peak E(2) levels were the highest (14426.6 pmol/L) and they produced the most mature follicales (27.3 +/- 9.7 ). Thus pattern A patients had the highest risk of ovarian hyperstimulation syndrome (OHSS). (81.8 % ), cancellation cycle rate(72.7 % )and low pregnancy rate(18.2 % ). Pattern B patients had a favorable ovarian response with the highest pregnancy rate(62.5 % ). Pattern C was associated with a poor ovarian response. CONCLUSION: During ovulation induction treatment, GAST can be used to predict ovarian response of anovulatory women. This technique may enable us to design an ovarian induction protocol that would both reduce OHSS and increase pregnancy rate.