To explore the optimal time of adding antagonist and the reasonable range of luteinizing hormone (LH) level on the human chorionic gonadotropin (hCG) injection day.

The data of 1 327 cases of GnRH-A protocol were retrospectively analyzed. According to the antral follicle count (AFC), the patients were divided into three groups: 1)poor ovarian responder (AFC≤5,n=278); 2) normal ovarian responder (6≤AFC≤15,n=756); 3) high ovarian responder (AFC≥16,n=293), and then each group was divided into two or three subgroups according to the LH level on antagonist administered day or the hCG injection day, the clinical outcomes were compared amomg the groups.

1) There were no significant differences in pregnancy outcomes between LH<5 IU/Land≥5 IU/L groupson the day of antagonist addition, but for the poor ovarian responders,administration ofthe antagonist when LH≥5 IU/L could get higher high-quality embryos rate; for the normal ovarian responders, administration of the antagonist when LH≥5 IU/L could reduce the duration of Gn used. 2) Regardless of ovarian response, the serum LH level on the hCG injection day did not affect the pregnancy outcome.

The LH level on the antagonist administered day and the hCG injection day did not affect the pregnancy outcomes. However, the optimum LH level of follicular phase may vary in different ovarian responders.

GnRH antagonist (GnRH-A) protocol; Luteinizing hormone (LH) level of follicular phase; Ovarian response; Pregnancy outcome