Evaluating the role of GnRH agonists in frozen embryo transfer cycles among polycystic ovary syndrome patients: a randomized clinical trial.

Frozen embryo transfer (FET) is an essential technique in assisted reproductive technology (ART), particularly for women with polycystic ovary syndrome (PCOS). FET allows better timing of embryo transfer and reduces the risk of ovarian hyperstimulation syndrome (OHSS) by separating the ovarian stimulation and embryo transfer phases [ 1 – 5 ] . Despite these advantages, successful outcomes strongly depend on optimal endometrial preparation, and no regimen has yet proven superior to standard hormone replacement therapy (HRT) protocols in PCOS patients [ 6 , 7 ] . Implantation failure in PCOS results from multiple factors. Hormonal imbalances, including hyperandrogenism and insulin resistance, disrupt the hypothalamic–pituitary–ovarian axis, leading to anovulation or irregular ovulatory cycles [ 8 – 10 ] . These disturbances also impair endometrial development and receptivity [ 11 , 12 ] . Elevated luteinizing hormone (LH) levels relative to follicle-stimulating hormone (FSH) further interfere with follicular maturation and increase the number of small antral follicles, contributing to excessive follicular development and OHSS [ 13 – 16 ] . HIGHLIGHTS This study determined whether GnRH agonists pretreatment can enhance clinical outcomes. Ongoing pregnancy rates showed no significant difference. Clinical pregnancy rates showed a nonsignificant trend toward more multiple gestations in the agonist group. This study determined whether GnRH agonists pretreatment can enhance clinical outcomes. Ongoing pregnancy rates showed no significant difference. Clinical pregnancy rates showed a nonsignificant trend toward more multiple gestations in the agonist group. Gonadotropin-releasing hormone agonists (GnRHa) may improve FET outcomes by inducing pituitary downregulation, synchronizing the endometrial lining, and creating a stable hormonal environment conducive to implantation [ 17 , 18 ] . They may also modulate cytokines such as interleukin-6 (IL-6) and IL-11, enhancing endometrial receptivity [ 19 ] . Previous studies have reported conflicting results: some demonstrated improved endometrial thickness and higher pregnancy rates with GnRHa pretreatment, while others found no significant benefit and highlighted increased treatment costs [ 20 – 23 ] . These inconsistent findings underscore the need for further investigation. Well-designed randomized trials are required to clarify the clinical efficacy and cost-effectiveness of GnRHa pretreatment in FET cycles among PCOS patients, thereby guiding evidence-based practice. The primary hypothesis of this randomized clinical trial is that, in women with PCOS undergoing FET, omitting GnRHa pretreatment will result in pregnancy outcomes comparable to those achieved with GnRHa pretreatment, while potentially reducing adverse effects associated with conventional protocols. This study addresses a critical gap in the current literature. While several studies have evaluated different endometrial preparation protocols for FET in women with PCOS, the specific role of GnRHa as an adjunct remains unclear and controversial. This trial is novel in directly assessing the efficacy of GnRHa administration prior to endometrial preparation in improving pregnancy outcomes in this high-risk subgroup, providing evidence that may help optimize FET protocols and enhance clinical outcomes for women with PCOS.

This single-blind, noninferiority randomized clinical trial was performed based on a parallel design of a one-to-one ratio. This work has been reported in line with the CONSORT criteria ( https://www.consort-statement.org ) [ 24 ] . A total of 297 women with PCOS were initially assessed for eligibility. A total of 13 participants were excluded due to not meeting the inclusion criteria, declining participation, or other reasons. The remaining 284 patients were randomized equally into 2 groups (142 per arm). After attrition, 270 participants completed the trial and were included in the final analysis (133 in the GnRHa group and 137 in the non-GnRHa group; see CONSORT flowchart, Fig. 1 ). Participants were recruited based on predefined inclusion and exclusion criteria. Inclusion criteria were as follows: Diagnosis of PCOS according to the Rotterdam Consensus (presence of at least two of the following: oligomenorrhea, clinical or biochemical hyperandrogenism, and polycystic ovarian morphology on ultrasound); age ≥20 years; 5-day (blastocyst-stage) embryo freezing; normal uterine cavity; undergoing intracytoplasmic sperm injection (ICSI) cycles. Exclusion criteria included: age >39 years; FSH level ≥12 mIU/mL; history of endometriosis; history of moderate or severe adenomyosis; history of more than three previous failed embryo transfer cycles, sex selection cycles, oocyte donation cycles, or surrogacy; presence of hydrosalpinx on ultrasonography; presence of any uterine structural abnormalities; presence of submucosal myoma >2 cm; diagnosis of Asherman’s syndrome. Figure 1. Flow chart for patient enrollment, randomization, and retention. Flow chart for patient enrollment, randomization, and retention. All patients in this study underwent a standard ovarian stimulation regimen using a GnRH antagonist protocol. GnRHa was administered to induce oocyte maturation when at least three follicles reached a diameter of ≥18 mm. Oocyte retrieval was performed 34–36 hours after GnRHa administration. At this center, all PCOS patients undergoing FET cycles are pretreated with Duphaston and GnRHa by default. Participants were interviewed regarding their obstetric and gynecological history, including previous failed pregnancies, miscarriages, ICSI attempts, and causes of infertility. Demographic information was also collected, including age, height, weight, education level, number of children, duration of marriage, and place of residence. A total of 284 eligible patients who provided informed consent were randomly assigned to two groups: one group received Duphaston alone, and the comparison group received Duphaston plus GnRHa. Both groups received 10 mg of Duphaston orally daily from days 15 to 25 of their menstrual cycle. In the GnRHa group, in addition to Duphaston, patients were administered half a vial of Leuprorelin Acetate (3.75 mg) subcutaneously on day 21 of the menstrual cycle. After completion of Duphaston, HRT was initiated from the third day of menstruation in both groups, consisting of daily administration of 6 mg of estradiol (E2) valerate tablets. The GnRHa dosage and timing were selected based on national clinical protocol and a previous study demonstrating efficacy in FET cycles for patients with PCOS [ 25 ] . Specifically, the chosen dose was intended to achieve optimal pituitary suppression while minimizing potential side effects. A transvaginal ultrasound was performed 7–10 days later to assess endometrial thickness. When the endometrial thickness reached ≥7 mm, daily intramuscular injections of 50 mg progesterone were started, and blastocyst embryo transfer was performed 5 days later. According to institutional protocol, one blastocyst was transferred in women under 30 years of age, and two blastocysts in women aged 30–40 years. On the first day of E2 administration, serum levels of E2, progesterone, LH, and FSH were measured. Patients residing in Tehran had laboratory tests performed at the Avicenna Infertility Clinic, while patients from other cities had tests conducted at local laboratories. All endometrial preparation cycles and embryo transfers were performed by a single physician (the project manager), who was blinded to group allocation and whether the patient received the GnRHa. In all patients, following embryo transfer, E2 valerate was continued at the same dose used during endometrial preparation. Luteal phase support consisted of progesterone suppositories (400 mg) administered vaginally every night, in combination with intramuscular progesterone injections (50 mg) every 3 days. Pregnancy was assessed by measuring serum beta-human chorionic gonadotropin (ß-hCG) levels on days 14 and 16 post-transfer. In patients with a positive ß-hCG result, luteal phase support was continued, and a transvaginal ultrasound was performed 5 weeks after embryo transfer. Outcomes were evaluated based on chemical pregnancy, clinical pregnancy, and ongoing pregnancy. All patients were followed for 12 weeks post-transfer, with ongoing pregnancies defined as those demonstrating a fetal heartbeat and considered viable (Fig. 2 ). Figure 2. Schematic timeline of interventions in the GnRHa and nonagonist groups during frozen embryo transfer (FET) cycles. The diagram illustrates the sequence of hormonal treatments. Schematic timeline of interventions in the GnRHa and nonagonist groups during frozen embryo transfer (FET) cycles. The diagram illustrates the sequence of hormonal treatments. The primary and secondary outcomes of this study were evaluated based on chemical pregnancy (positive ß-hCG per embryo transfer), clinical pregnancy (presence of a gestational sac on ultrasound at 7 weeks of gestation), and ongoing pregnancy (detection of a fetal heartbeat on ultrasound at 12 weeks) [ 20 , 21 , 26 ] . The miscarriage rate in each group was calculated using the following formula: (Number of clinical pregnancies − number of ongoing pregnancies)/Number of clinical pregnancies × 100. The sample size was calculated based on a non-inferiority design with a 10% margin, a Type I error rate (α) of 0.025, and a Type II error rate (β) of 0.2. The estimated treatment success rates were 56% in the group without GnRHa and 61% in the group with GnRHa, resulting in a required sample size of 142 patients per group [ 22 ] . In this study, participants were randomly assigned to the intervention group (no GnRHa injection) or the control group (receiving GnRHa) using a permuted block randomization method with blocks of four. R software (version 4.4.1) was used to generate a random sequence of numbers from 1 to 6 to ensure balanced allocation across groups. If a generated number exceeded 6, a new random number was generated. The random sequence was prepared by an independent individual not involved in the study design and placed in sealed envelopes labeled with six-digit serial numbers. After collecting participants’ baseline information and completing examinations, the envelopes were opened to finalize group assignments. A trained staff member at the clinic, who was not involved in the trial and blinded to group assignments, distributed numbered medication packages to the participants. In this single-blind study, an independent individual, not involved in the study design, generated the random sequence and placed it in sealed envelopes. After collecting participants’ baseline information and completing examinations, a trained staff member at the clinic, who was not involved in the trial, distributed numbered medication packages to the participants. The envelopes were then opened to finalize group assignments. The project administrator, who performed all endometrial preparation cycles and embryo transfers, was blinded to the randomization and to whether patients received the GnRHa. Patients were aware of their group allocation. Both the outcome assessor and the data analyst were blinded: the outcome assessor recorded pregnancy outcomes without knowledge of the intervention type, and the data were provided to the analyst in coded form, ensuring that analyses were conducted without awareness of group assignments. Thus, all investigators involved in endometrial preparation, embryo transfer, outcome measurement, and data analysis remained blinded to group allocations throughout the study to minimize bias. In this study, statistical analyses were performed to compare demographics, clinical characteristics, treatment details, causes of infertility, and pregnancy outcomes between patients receiving GnRHa and those receiving nonagonist interventions. Categorical variables were presented as frequencies and percentages and compared between groups using Fisher’s exact test. Continuous variables with a normal distribution were summarized as mean ± standard deviation and compared using the independent t -test. Continuous variables that did not follow a normal distribution were reported as median with interquartile ranges and compared using the Mann–Whitney U test. All analyses were conducted using R (version 4.3.2) and SPSS (version 27). The statistical significance threshold was set at P <0.05 for all tests. To control for the false discovery rate arising from multiple comparisons, all P values were adjusted using the Benjamini–Hochberg correction method [ 27 ] . There were no missing data for the outcomes analyzed in this study; all 270 participants who completed the trial were included in the final analysis.

1. Baseline demographic and clinical characteristics were largely comparable between the GnRHa and nonagonist groups, except for a slightly higher prevalence of secondary infertility in the GnRHa group. 2. Although GnRHa pretreatment significantly reduced serum FSH, LH, and E2 levels at the start of E2 administration, E2 dose requirements, and endometrial thickness obtained were similar between groups. 3. Pregnancy outcomes including chemical pregnancy, clinical pregnancy, ongoing pregnancy, and miscarriage rates did not differ significantly between the GnRHa and non-agonist groups, indicating no additional clinical benefit of GnRHa pretreatment (Fig. 3 ). Figure 3. Results summary of frozen embryo transfer cycles with and without GnRH agonists in patients with polycystic ovary syndrome. Results summary of frozen embryo transfer cycles with and without GnRH agonists in patients with polycystic ovary syndrome. Table 1 summarizes the demographic and clinical characteristics of the 270 participants, comparing the GnRHa group ( n = 133) with the non-agonist group ( n = 137). The mean age was 32.12 ± 4.72 years (GnRHa: 31.68 ± 4.86 vs. nonagonist: 32.54 ± 4.57; P = 0.13), and the mean BMI was 25.98 ± 4.09 kg/m 2 ( P = 0.50). The average duration of marriage was 7.67 ± 4.03 years (GnRHa: 7.66 ± 4.21 vs. nonagonist: 7.69 ± 3.86; P = 0.95). Baseline AMH levels ( P = 0.62) and the number of embryos transferred ( P = 0.60) were similar between groups. Educational levels were comparable, with 41.11% of participants holding a diploma ( P = 0.65). Consanguineous relationships were reported in 14.81% of participants, and distant familial relationships in 3.33%. Causes of infertility (female, male, or both) were generally similar ( P = 0.39). Secondary infertility was slightly higher in the GnRHa group ( P = 0.02), which is the only significant baseline difference. Overall, the groups were well balanced at baseline, with no other statistically significant differences observed (Table 1 ). Table 1 Demographic and clinical characteristics of patients undergoing GnRHa and nonagonist interventions at baseline Variable Levels Total ( n = 270) GnRHa intervention P -value No ( n = 137) Yes ( n = 133) Age (years) – 32.12 ± 4.72 32.54 ± 4.57 31.68 ± 4.86 0.137 BMI (kg/m 2 ) – 25.98 ± 4.09 26.15 ± 3.94 25.80 ± 4.25 0.500 Duration of marriage (years) – 7.67 ± 4.03 7.69 ± 3.86 7.66 ± 4.21 0.950 Education Illiterate 3 (1.11) 1 (0.73) 2 (1.50) 0.656 Middle school 49 (18.15) 30 (21.90) 19 (14.29) Diploma 111 (41.11) 55 (40.15) 56 (42.11) Bachelor 74 (27.41) 35 (25.55) 39 (29.32) Master 18 (6.67) 9 (6.57) 9 (6.77) PhD 12 (4.44) 5 (3.65) 7 (5.26) Affinal relationship Cousin 40 (14.81) 14 (5.19) 22 (8.15) 0.509 Distant familial relation 9 (3.33) 3 (2.19) 6 (4.51) Duration of infertility (months/years) 5 (2.5, 8) 4.5 (3, 8) 5 (2.5, 7) 0.762 Baseline AMH level (ng/mL) – 5.85 (3.67, 9.28) 6.00 (3.60, 9.85) 5.5 (3.7, 8.8) 0.624 Cause of infertility Female 61 (22.59) 35 (25.55) 26 (19.55) 0.393 Male 157 (58.15) 74 (54.01) 83 (62.41) Female + male 35 (12.96) 17 (12.41) 18 (13.53) Type of infertility Primary 217 (80.37) 109 (79.56) 108 (81.20) 0.022 Primary + secondary 17 (6.30) 13 (9.49) 4 (3.01) Secondary 27 (10.00) 9 (6.57) 18 (13.53) Number of previous pregnancies 0 1 (0.37) (0.00) 1 (0.75) 0.549 1 70 (25.93) 38 (27.74) 32 (24.06) 2 26 (9.63) 12 (8.76) 14 (10.53) 3 13 (4.81) 7 (5.11) 6 (4.51) 4 7 (2.59) 3 (2.19) 4 (3.01) 5 2 (0.74) 2 (1.46) (0.00) 6 2 (0.74) (0.00) 2 (1.50) History of live birth No 23 (8.52) 9 (6.57) 14 (10.53) 1.000 Yes 3 (1.11) 1 (0.73) 2 (1.50) History of delivery No 36 (13.33) 17 (12.41) 19 (14.29) 1.000 Yes 3 (1.11) 1 (0.73) 2 (1.50) Type of previous delivery Cesarean section (CS) 26 (9.63) 9 (6.57) 17 (12.78) 0.101 Normal vaginal delivery (NVD) 14 (5.19) 9 (6.57) 5 (3.76) History of previous preterm birth No 12 (4.44) 7 (5.11) 5 (3.76) 0.462 Yes 1 (0.37) (0.00) 1 (0.75) Menstrual status Regular 149 (55.19) 74 (54.01) 75 (56.39) 0.714 Irregular 119 (44.07) 62 (45.26) 57 (42.86) History of previous miscarriage No 64 (23.70) 33 (24.09) 31 (23.31) 0.28 Yes 32 (11.85) 18 (13.14) 14 (10.53) Type of previous embryo transfer Fresh 9 (3.33) 4 (2.92) 5 (3.76) 0.754 Fresh + freeze 15 (5.56) 9 (6.57) 6 (4.51) Freeze 145 (53.70) 80 (58.39) 65 (48.87) Number of embryos transferred 1 32 (11.85) 14 (10.22) 18 (13.53) 0.607 2 205 (75.93) 111 (81.02) 94 (70.68) 3 6 (2.22) 2 (1.46) 4 (3.01) 4 2 (0.74) 1 (0.73) 1 (0.75) Number of previous embryo transfers – 2.34 ± 1.66 2.33 ± 1.82 2.35 ± 1.45 0.946 Categorized variables were reported using frequency (percentage). Numeric variables were summarized using mean ± standard deviation. Fisher’s exact test and the independent t-test were used to evaluate the association between categorical and numeric variables and GnRHa intervention, respectively. Demographic and clinical characteristics of patients undergoing GnRHa and nonagonist interventions at baseline Categorized variables were reported using frequency (percentage). Numeric variables were summarized using mean ± standard deviation. Fisher’s exact test and the independent t-test were used to evaluate the association between categorical and numeric variables and GnRHa intervention, respectively. Table 2 summarizes the treatment details. Regarding basic treatment variables, there were no significant differences between groups in the duration of ovarian stimulation ( P = 0.08) or the number of embryos obtained ( P = 0.98). The GnRHa group exhibited significantly lower serum FSH, LH, and E2 levels on the day of E2 initiation compared with the non-agonist group ( P < 0.001 for FSH and LH; P = 0.043 for E2), whereas serum progesterone levels did not differ significantly ( P = 0.33). Despite the lower E2 levels on the day of E2 initiation in the GnRHa group, no significant differences were observed between groups in the total dose of E2 used for endometrial preparation ( P = 0.09) or endometrial thickness ( P = 0.11) (Fig. 4 ). Figure 4. Endometrial preparation parameters illustrate duration of ovarian stimulation, E2 dosage, and endometrial thickness, demonstrating comparable values between groups. Table 2 Comparison of treatment details in GnRHa compared with non-agonist ICSI cycles Variable Total ( n = 270) GnRHa intervention P -value Nonagonist ( n = 137) GnRHa ( n = 133) Duration of ovarian stimulation (days) a 10.09 ± 1.42 9.94 ± 1.16 10.25 ± 1.64 0.083 Number of embryos obtaind b 7 (5, 12) 7 (5, 12) 8 (4, 12) 0.986 Serum FSH on the day of estradiol start (mIU/mL) a 6.14 ± 2.41 7.16 ± 2.20 5.09 ± 2.17 <0.001 Serum LH on the day of estradiol start (mIU/mL) b 6.15 (3.70, 9.87) 7.30 (5.56, 12.35) 4.46 (2.42, 7.6) <0.001 Serum E2 on the day of estradiol start (pg/mL) b 39.80 (24.44, 63.85) 41.99 (28.85, 67.60) 34.40 (20.35, 58.35) 0.043 Serum progesterone on the day of estradiol starts (ng/mL) b 0.20 (0.11, 0.38) 0.19 (0.10, 0.32) 0.21 (0.11, 0.43) 0.332 E2 dosage (mg) a 92.09 ± 10.79 90.98 ± 10.10 93.27 ± 11.41 0.096 Endometrial thickness (mm) a 8.68 ± 1.36 8.55 ± 1.15 8.82 ± 1.55 0.113 a Normally distributed data are summarized as mean ± standard deviation. b Non-normally distributed data are summarized as median (interquartile range). For numeric variables, the independent t -test was used to compare means between the GnRHa intervention groups, while the Mann–Whitney U test was applied to compare medians for data that did not follow a normal distribution. Endometrial preparation parameters illustrate duration of ovarian stimulation, E2 dosage, and endometrial thickness, demonstrating comparable values between groups. Comparison of treatment details in GnRHa compared with non-agonist ICSI cycles Normally distributed data are summarized as mean ± standard deviation. Non-normally distributed data are summarized as median (interquartile range). For numeric variables, the independent t -test was used to compare means between the GnRHa intervention groups, while the Mann–Whitney U test was applied to compare medians for data that did not follow a normal distribution. Although GnRHa pretreatment significantly altered baseline hormonal profiles, key clinical outcomes including E2 dose requirements and endometrial thickness remained largely comparable between groups during endometrial preparation cycles. These findings suggest that both protocols provide similar endometrial preparation and implantation potential, indicating that GnRHa pretreatment may not confer additional clinical benefits in terms of embryo transfer outcomes. Table 3 summarizes pregnancy outcomes in patients undergoing ICSI with and without GnRHa intervention. There was no clinically significant difference in the chemical pregnancy rate – defined as the proportion of positive β-hCG tests between the GnRHa and non-agonist groups [48.12% (95% CI: 39.80–56.54) vs. 45.98% (95% CI: 37.86–54.33), P = 0.41]. The clinical pregnancy rate—defined as the proportion of pregnancies with a gestational sac at 7 weeks confirmed via ultrasound was also similar between groups [42.84% (95% CI: 34.76–51.35) vs. 41.59% (95% CI: 33.69–49.98), P = 0.55]. Ongoing pregnancy rates, assessed by the presence of a fetal heartbeat at 12 weeks, did not differ significantly [28.60% (95% CI: 21.56–36.79) vs. 30.70% (95% CI: 23.07–39.10), P = 0.52]. Table 3 Comparison of pregnancy outcomes by GnRHa intervention Outcome Levels Total ( n = 270) GnRHa intervention P -value a Nonagonist ( n = 137) GnRHa ( n = 133) Chemical pregnancy No 143 (52.96%) 74 (54.01%) 69 (51.87%) 0.41 Yes 127 (47.03%) 63 (45.98%) 64 (48.12%) Clinical pregnancy No 156 (57.8%) 80 (58.39%) 76 (57.14%) 0.55 Single 80 (29.6%) 44 (32.11%) 36 (27.06%) Twins 34 (12.6%) 13 (9.48%) 21 (15.78%) Ongoing pregnancy No heartbeat detected 190 (70.4%) 95 (69.3%) 95 (71.4%) 0.52 Single 57 (21.1%) 32 (23.4%) 25 (18.8%) Twins 23 (8.5%) 10 (7.3%) 13 (% 9.8) Categorized outcomes are reported using frequency (percentage). The comparison of categorical variables between groups was performed using Fisher’s exact test. a Corrected P values are calculated using the Benjamini–Hochberg correction method. Comparison of pregnancy outcomes by GnRHa intervention Categorized outcomes are reported using frequency (percentage). The comparison of categorical variables between groups was performed using Fisher’s exact test. Corrected P values are calculated using the Benjamini–Hochberg correction method. The miscarriage rate [(Number of clinical pregnancies − number of ongoing pregnancies)/Number of clinical pregnancies] × 100. In the two groups, miscarriage rates were 33.3% (95% CI: 21.4–47.0) in the GnRHa group and 26.3% (95% CI: 15.5–39.6) in the non-GnRHa group, with no statistically significant difference between groups ( P = 0.41) (Fig. 5 ). All participants in both groups completed the treatment protocol as planned, with no missing or incomplete data affecting the primary outcomes, ensuring that the comparisons accurately reflect the effect of GnRHa intervention. Figure 5. Pregnancy outcomes display chemical, clinical, and ongoing pregnancy for both groups, showing no significant differences. Pregnancy outcomes display chemical, clinical, and ongoing pregnancy for both groups, showing no significant differences.

Our study demonstrates that GnRHa pretreatment does not significantly affect pregnancy outcomes, including chemical, clinical, and ongoing pregnancy rates, or miscarriage rates, in PCOS patients undergoing FET. While GnRHa modifies serum hormone levels (FSH, LH, and E2) during endometrial preparation, these changes do not translate into improved clinical outcomes. This randomized trial adds to the growing evidence suggesting that GnRHa may not be necessary in all patients, highlighting the potential for more cost-effective, individualized ART protocols. Strengths of our study include the randomized design, adequate sample size, and investigator blinding to minimize bias. Limitations include the single-center setting and exclusion of certain patient subgroups, which may affect generalizability. These findings provide practical guidance for optimizing FET strategies in PCOS while avoiding unnecessary interventions. Women with PCOS often experience anovulatory infertility and may require ART. FET is generally preferred over fresh transfer due to a lower risk of OHSS [ 23 , 28 ] . Despite various treatment modalities aimed at improving endometrial function and reducing hyperandrogenism, miscarriage and implantation failure rates remain elevated in this population [ 29 , 30 ] , highlighting the need for optimized, individualized ART protocols. Given the influence of hyperandrogenism on endometrial dysfunction, administration of GnRHa may theoretically enhance implantation rates, particularly when combined with progesterone treatment within an HRT–FET framework [ 5 ] . Mechanistically, GnRHa may reduce endometrial inflammation, modulate ß-hCG, and increase the expression of adhesion molecules such as αvβ3 integrin and cytokines IL-6 and IL-11 in the endometrium, potentially facilitating improved embryo–endometrial interaction [ 19 , 31 ] . The endometrium also expresses GnRH receptors, predominantly during the luteal phase, which may further contribute to embryo attachment [ 32 ] . Previous studies have investigated GnRHa use, alone or in combination with progesterone, across different dosages and administration methods, particularly in fresh transfer cycles and natural cycle FET [ 33 – 35 ] . However, its use in artificial cycle FET remains debated. For example, Ye et al [ 36 ] reported no statistically significant differences in implantation rates, clinical pregnancy outcomes, or miscarriage rates in women under 35 years, whereas women aged 35–37 years showed lower implantation rates. Similarly, a retrospective cohort by Liu et al [ 22 ] found no significant differences in live birth, clinical pregnancy, miscarriage, or ectopic pregnancy rates between PCOS patients treated with GnRHa + HRT versus HRT alone, consistent with our findings. Conversely, a meta-analysis by Li et al [ 23 ] indicated that PCOS patients receiving GnRHa had higher implantation and clinical pregnancy rates, although live birth, multiple pregnancy, ectopic pregnancy, and miscarriage rates were similar between groups. These findings suggest that while GnRHa may affect biological markers such as endometrial thickness or molecular expression, its impact on key clinical outcomes is inconsistent. In our study, LH and FSH concentrations at the start of E2 administration were significantly lower in the GnRHa group compared with the control group; however, this did not translate into significant differences in implantation rates. Mechanistically, GnRHa may influence endometrial receptivity by regulating cytokine expression, such as IL-6 and IL-11, in endometrial stromal cells [ 19 ] . A meta-analysis suggested that GnRHa may slightly improve clinical pregnancy, implantation, and live birth rates, yet subgroup analyses of randomized controlled trials did not demonstrate statistically significant effects [ 37 ] . Our findings indicate that, despite some biological effects of GnRHa pretreatment, there are no substantial improvements in key clinical outcomes such as implantation and miscarriage rates. These results underscore the need to reconsider routine GnRHa use in FET protocols, particularly when potential benefits are weighed against additional costs and complexity. Personalized treatment plans that consider hormonal profiles, previous ICSI history, patient characteristics, and preferences should be prioritized when designing FET protocols. Such an approach can enhance patient well-being, optimize the overall ICSI experience, and promote better clinical outcomes. Future research should investigate the long-term effects of GnRHa on maternal and infant health, as well as variability in treatment efficacy across different populations and ethnicities, to develop more effective, individualized approaches for diverse patient groups. Several limitations should be noted. First, the relatively small sample size may reduce statistical power to detect subtle differences between groups. Second, the single-center design and referral nature of the clinic may limit generalizability to other populations or centers. Third, hormone measurements (LH, FSH, and E2) were performed in different laboratories for some patients residing in other cities, potentially introducing measurement heterogeneity. Additionally, we could not evaluate the effects of different GnRHa or estrogen–progesterone brands and dosing regimens, and live birth data were not available, which is a key endpoint. Despite these limitations, the randomized design and standardized procedures strengthen the validity of our findings. Future multicenter studies with larger cohorts are warranted to confirm these results, explore population-specific responses, assess long-term maternal and neonatal outcomes, and evaluate cost-effectiveness and patient-centered outcomes in FET protocols involving GnRHa.

The findings of this study contribute to the ongoing debate regarding the use of GnRHa in FET protocols for women with PCOS. While previous research suggested potential benefits, our randomized trial demonstrates that GnRHa pretreatment does not confer a significant advantage in pregnancy outcomes. These results underscore the importance of evidence-based clinical decision-making while also considering practical implications. Specifically, omitting GnRHa may simplify treatment, reduce costs, minimize drug-related side effects, and support more patient-centered, individualized care. Strengths of our study include the randomized design, standardized procedures, and investigator blinding, whereas limitations include the single-center setting, relatively small sample size, laboratory measurement heterogeneity, and the lack of live birth data. Future multicenter studies with larger cohorts are warranted to confirm these findings, explore population-specific responses, and assess long-term maternal and neonatal outcomes as well as cost-effectiveness and patient-centered benefits.