BackgroundOvarian reactivity to gonadotrophin stimulation varies, and individual adjustments to the timing and dose of gonadotrophin-releasing hormone (GnRH) antagonist administration are necessary to prevent excessive increases and decreases in luteinizing hormone (LH) levels in patients with different ovarian response following the GnRH antagonist (GnRH-A) protocol. The present study aims to investigate optimal LH suppression thresholds for patients with normal ovarian response (NOR), high ovarian response (HOR), and poor ovarian response (POR) following the GnRH-A protocol respectively. MethodsA total of 865 in vitro fertilization (IVF) cycles using a flexible or fixed GnRH-A protocol were included. Patients were categorized into the HOR, NOR, or POR group according to their anti-Müllerian hormone (AMH) levels. Then, patients in each group were stratified into one of four subgroups according to the quartile (Q1-Q4) of the basal LH level to LH on triggering day ratio (bLH/hLH). The primary outcomes were the clinical pregnancy and live birth rates, and the secondary outcomes were the number of oocytes retrieved, MII oocytes, two pronucleus (2PN) embryos, and good-quality embryos. ResultsThere were 526 patients with NOR, 180 with HOR, and 159 with POR. Basal LH level, LH on triggering day and bLH/hLH were identified as independent predictors of clinical pregnancy rate and live birth rate by logistics regression analysis. Compared to those with NOR, patients with POR had the lowest embryo implantation rate (22.6% vs. 32.8%, P < 0.05), clinical pregnancy rate (32.3% vs. 47.3%, P < 0.05) and live birth rate (22.6 vs. 37.8%, P < 0.05) of fresh embryo transfer (ET). The embryo implantation, clinical pregnancy and live birth rates of frozen embryo transfer (FET) were not significantly different among the three groups. In the subgroup analysis, patients with HOR had the highest embryo implantation rate (51.6%, P < 0.05), clinical pregnancy rate (68.4%, P < 0.05) and live birth rate (52.6%, P < 0.05) of ET in Q3, with a bLH/hLH ratio of 2.40–3.69. In the NOR group, the embryo implantation rate (41.9%, P < 0.05), clinical pregnancy rate (61.5%, P < 0.05) and live birth rate (50.8%, P < 0.05) of ET and live birth rate (53.1%, P < 0.05) of FET were highest in Q2, with a bLH/hLH ratio of 1.29–2.05. Patients with POR had the highest clinical pregnancy rate (57.1%, P < 0.05) and live birth rate (42.9%, P < 0.05) of ET in Q2, with a bLH/hLH ratio of 0.86–1.35. ConclusionsIn the present study, the bLH/hLH ratio represented the LH suppression threshold. The subgroup analysis of HOR, NOR and POR showed that, the LH suppression threshold varies according to ovarian response. We recommend LH suppression thresholds of 2.40–3.69 for HOR, 1.29–2.05 for NOR, and 0.86–1.35 for POR to obtain the highest clinical pregnancy rate and live birth rate. This study provides comprehensive and precise references for clinicians to monitor LH levels individually during controlled ovarian stimulation (COS) according to the patient’s ovarian response following the GnRH-A protocol.