ObjectiveTo compare the efficacy of gonadotropin-releasing hormone antagonist versus gonadotropin-releasing hormone agonist long protocol in women belonging to POSEIDON groups 3 and 4. Study designA total of 380 patients with expected low ovarian response [antral follicle count < 5 and/or anti-Müllerian hormone < 1.2 ng/mL] were studied retrospectively. The efficiency of the gonadotropin-releasing hormone antagonist protocol and the gonadotropin-releasing hormone agonist long protocol was compared in patients from POSEIDON group 3 (age < 35 years) and group 4 (age ≥ 35 years), respectively. The primary outcome was the cumulative live birth rate. All patients underwent complete cycles of in vitro fertilization/intracytoplasmic sperm injection for the first time from January 2016 to June 2019. ResultsIn POSEIDON group 4, age, anti-Müllerian hormone, initial gonadotropin dose and induction protocols were significantly correlated with cumulative live birth by multivariate regression analysis. The optimum cut-off value of anti-Müllerian hormone for prediction of cumulative live birth was 0.785 by receiver operating characteristic analysis. Patients with higher anti-Müllerian hormone levels (anti-Müllerian hormone ≥ 0.785 ng/mL) who received the gonadotropin-releasing hormone agonist long protocol achieved significantly higher cumulative live birth rate than who received the gonadotropin-releasing hormone antagonist protocol, whereas no significant difference in cumulative live birth rate of the two protocols was found in patients with low anti-Müllerian hormone levels (anti-Müllerian hormone < 0.785 ng/mL). In POSEIDON group 3, only body mass index was significantly correlated with cumulative live birth. There was no correlation between cumulative live birth and induction protocols. ConclusionsPatients with higher anti-Müllerian hormone levels from POSEIDON group 4 are more likely to benefit from the gonadotropin-releasing hormone agonist long protocol than the gonadotropin-releasing hormone antagonist protocol.