Abstract
OBJECTIVE: The aim of this study was to compare the effects of different endometrial treatment regimens on the outcome of frozen embryo transfer (FET). METHODS: A retrospective analysis of 1,454 FET cycles was conducted. Subgroups were stratified by primary infertility diagnosis (tubal factor, polycystic ovary syndrome (PCOS), and endometriosis) and by the number of embryos transferred. Multivariable logistic regression was employed to identify factors influencing live birth. RESULTS: Our study found that endometrial thickness on the day of embryo transfer was thicker in the gonadotropin-releasing hormone agonist (GnRHa)-hormone replacement therapy (HRT) group compared to the HRT group (p = 0.000), the levels of estrogen and progesterone in GnRHa-HRT group were lower than HRT group, and there were no statistically significant differences in other baseline characteristics between the two groups. The live birth rate in the single embryo transfer GnRHa-HRT group was higher than that of HRT(OR = 1.489, 95%CI:1.070-2.073). Within the tubal/male factor cohort, spontaneous abortion rate was slightly more frequent in the GnRHa-HRT group (p = 0.899). In the PCOS-endometriosis (EMs) subgroup, the spontaneous abortion rate had a decreasing trend, but the difference was not statistically significant (p > 0.05). Binary logistic regression analysis further confirmed the negative influence of age on live birth. CONCLUSION: The GnRHa-HRT significantly improves live birth rates compared to the conventional HRT in single-embryo transfer cycles. Downregulation protocols may be a priority for patients with endometriosis and single embryo transfer. For patients with PCOS and tubal factors, the HRT program is more cost-effective and takes less time to achieve pregnancy. These findings provide evidence-based guidance for individualized endometrial preparation regimen selection in FET cycles, though the retrospective design and lack of long-term follow-up limit the generalizability. Future prospective randomized controlled trials with larger sample sizes are warranted to validate these conclusions.