Low serum luteinizing hormone levels on trigger day increase oocyte yield but elevate miscarriage rate: a differential analysis of agonist and antagonist protocols

The impact of serum luteinizing hormone (LH) levels during controlled ovarian stimulation (COS) on the assisted reproductive outcomes of in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) remains controversial. This study aims to investigate the association between different LH levels on the trigger day and assisted reproductive outcomes, and to clarify its clinical significance.

A retrospective cohort study design was adopted. Clinical data of 3227 patients treated with gonadotropin-releasing hormone (GnRH) agonist protocol (including long-acting follicular-phase long protocol and long-acting luteal-phase long protocol) and 1355 patients treated with GnRH antagonist protocol in our department from January 2020 to December 2024 were collected. The participants were stratified into 4 groups separately for each protocol based on the quartiles of LH levels on the trigger day. For the patients receiving GnRH agonist protocol, they were divided into Group A (LH <0.79 mIU/mL, n=802), Group B (0.79≤ LH < 1.22 mIU/mL, n=802), Group C (1.22≤ LH <2.01 mIU/mL, n=813), and Group D (LH ≥2.01 mIU/mL, n=810). Similarly, for the antagonist protocol, the 4 groups were Group A (LH <1.52 mIU/mL, n=336), Group B (1.52 ≤ LH <2.61 mIU/mL, n=341), Group C (2.61≤ LH <4.19 mIU/mL, n=339), and Group D (LH ≥4.19 mIU/mL, n=339). First, intergroup difference analysis was performed for general characteristics, clinical data, embryonic laboratory parameters, and clinical outcomes of fresh embryo transfer cycles among the 4 groups in each protocol to compare the distribution characteristics of the above indicators. Subsequently, multivariate linear and logistic regression models were constructed with LH levels on the trigger day as the independent variable, combined with differential analysis results and adjusted for confounding factors including age, body mass index (BMI), basal endocrine levels, to further clarify the impact of different LH levels on the trigger day on outcome indicators.

In the GnRH agonist protocol, univariate analysis showed statistical differences across the 4 groups in terms of female age, BMI, basal follicle-stimulating hormone (bFSH), basal LH, anti-Müllerian hormone (AMH), antral follicle count (AFC), trigger-day progesterone, and endometrial thickness on embryo transfer day (all P<0.05), whereas no significant differences were observed in major outcome indicators clinical pregnancy rate or live birth rate. Multivariate linear regression analysis indicated that, with Group A (LH <0.79 mIU/mL) as the reference, Group D (LH ≥2.01 mIU/mL) had significantly fewer retrieved oocytes (β=-2.61, 95%CI:-3.46 to-1.77, P<0.001) and fewer high-quality embryos (β=-0.64, 95%CI:-1.20 to-0.09, P=0.023), and obviously increased two-pronuclear (2PN) fertilization rate (β=0.03, 95%CI: 0.01 to 0.05, P=0.044). Multivariate logistic regression showed that, with Group A as the reference, Group C (1.22≤ LH <2.01 mIU/mL) had significantly lower early miscarriage rate (OR=0.45, 95%CI: 0.25 to 0.80, P=0.007) and total miscarriage rate (OR=0.52, 95%CI: 0.29 to 0.91, P=0.021). In the GnRH antagonist protocol, multivariate linear regression showed that, with Group A (LH <1.52 mIU/mL) as the reference, Group C (2.61≤ LH <4.19 mIU/mL, β=-1.72, 95%CI:-3.04 to-0.40, P=0.011) and Group D (LH ≥4.19 mIU/mL, β=-2.07, 95%CI:-3.54 to-0.60, P=0.006) had significantly fewer retrieved oocytes. Multivariate logistic regression analysis revealed no significant differences in clinical pregnancy rate, abortion rate, live birth rate, or other outcome indicators among the groups.

The distribution characteristics of trigger-day LH levels differ substantially between GnRH agonist and antagonist protocols. In the agonist protocol, LH levels are profoundly suppressed; LH <0.79 mIU/mL is associated with increased oocyte yield but poorer embryo quality and elevated miscarriage risk. In the antagonist protocol, trigger-day LH levels are markedly higher than those in the agonist protocol, and are only associated with oocyte number, with no significant correlation to embryo quality or clinical outcomes.