Received July 18, 2016
Accepted January 30, 2016

Objective: In controlled ovarian hyperstimulation (COH) using antagonist cycles, an incomplete luteolysis could happen after an inefficient previous luteolysis. Since antagonist cycles are frequent today, this study aims to access the impact of serum progesterone in the beginning and at the end of stimulation, and pregnancy outcomes. Methods: single-center cohort study, 461 fresh embryo transfers in ICSI antagonist cycles. Serum progesterone levels was measured in the beginning of COH (P4i) and on hCG day (P4f) using threshold values of 1.5ng/mL. Four groups were created: Group 1, P4i and P4f ≤ 1.5; Group 2, P4i ≤ 1.5 and P4f > 1.5; Group 3, P4i > 1.5 and P4f ≤ 1.5 and Group 4, P4i and P4f > 1.5. The clinical pregnancy rate (CPR) and live birth rates (LBR) were the primary outcomes. Results: The number of cycles per group was: 393, 51, 6 and 11, respectively. Group 1 was considered the expected normal, while group 4 represented the persistence of higher levels. There was no difference in age, basal FSH and Estradiol, days of stimulation endometrium thickness and total amount of gonadotropins between group 1 versus group 4. However, significant differences occurred in embryological and clinical outcomes between these 2 groups. Conclusion: The impact of serum progesterone in the beginning of stimulation and pregnancy outcomes is a matter of concern. Basal elevated levels could help identify patients that will repeat it on hCG day, being probably a marker to define a freeze-all strategy to these cycles.