Association between endogenous LH level prior to progesterone administration and live birth rate in artificial frozen-thawed blastocyst transfer cycles of ovulatory women.

Abstract

Is there an association between serum LH levels prior to progesterone administration and live birth rate (LBR) in artificial frozen-thawed embryo transfer (FET) cycles? : Low serum LH levels on the day before progesterone initiation in artificial frozen-thawed blastocyst transfer cycles of ovulatory women are associated with a lower LBR. In artificial FET cycles, exogenous oestrogen and progesterone are administered sequentially to mimic the serum hormone pattern similar to the natural cycle. In oestrogen-only phase, the supplemental oestrogen causes thickening of the endometrium and is sometimes accompanied by a rise in serum LH. However, whether the endogenous LH level in artificial FET cycles is related to clinical outcomes remains unclear. A retrospective cohort study including 3469 artificial frozen-thawed blastocyst transfer cycles was conducted at a tertiary-care academic medical centre between February 2014 and January 2020. A total of 3469 frozen blastocyst transfer cycles were stratified into four groups based on the quartiles of serum LH level before progesterone initiation: <25th percentile (LH < 8.79 mIU/ml), 25-50th percentile (8.79 ≤ LH ≤ 13.91 mIU/ml), 51-75th percentile (13.91 < LH ≤ 20.75 mIU/ml) and >75th percentile (LH > 20.75 mIU/ml). The serum LH level >75th percentile group was considered as the reference group. Patients with polycystic ovarian syndrome or other ovulatory disorders were excluded from the study. We also excluded cycles with an endometrial thickness <7 mm before progesterone initiation and patients with intrauterine adhesions and uterine abnormalities. In order to avoid the interference of BMI, all patients were divided into two categories based on the overweight threshold: BMI <25 kg/m2 and ≥25 kg/m2, and the impacts of serum LH levels on LBR were investigated separately. Univariable and multivariable logistic regression analysis were performed to adjust for potential confounders. EmpowerStats software and R-project were used to build smooth curve fitting models. Compared with the reference group, the implantation rate significantly decreased with low LH levels (<25th percentile) on the day before progesterone initiation (odds ratio [OR] = 0.74; 95% CI, 0.64-0.86; P = 0.001). Accounting for major covariates, low LH levels were associated with a relatively lower LBR (adjusted OR = 0.649; 95% CI, 0.531-0.794; P < 0.001), mainly due to a lower implantation rate, lower clinical pregnancy rate and higher pregnancy loss rate. Moreover, in the patients with BMI <25 kg/m2, low LH was associated with a lower LBR (P < 0.001); while in the overweight subgroup, LBR and LH were not correlated (P = 0.823). The main limitation of this study is its retrospective design. Owing to the relatively small number in the overweight group, the results of the overweight subgroup should be interpreted with caution. The evidence provided in this study shows the importance of serum LH levels on the day before progesterone initiation in patients undergoing artificial FET cycles. Hypothalamic dysfunction may be one of the important causes of a relatively low LH, which is related to impaired pregnancy outcomes. Serum LH levels may be used as one of the clinical indicators to predict pregnancy outcomes. No funding and no competing interest were involved in this study. NA.