This prospective cohort study was conducted between March 2021 and September 2023 at the Reproductive Medicine Center, Sanmenxia Hospital, Yellow River. The study was complied with the guidelines of the Declaration of Helsinki and relevant national regulations. The study protocol has been reviewed and approved by the hospital’s institutional ethics committee (No: 2024LLS20240914043), and all participants have provided written informed consent.

A total of 341 patients were initially assessed for eligibility. Following the application of inclusion and exclusion criteria, 172 patients met the eligibility requirements and were included in the study. Participants were then allocated into two groups using a 1:3 ratio: 43 patients were assigned to Group A (Long Luteal Phase Protocol) and 129 to Group B (Flexible Antagonist Protocol). During follow-up, 1 patient in Group A and 11 patients in Group B were lost to follow-up. As a result, 42 patients in Group A and 118 patients in Group B were included in the final analysis ( Figure 1 ). The 1:3 allocation ratio was based on the proportional use of protocols in clinical practice at our center during the study period. The flexible antagonist protocol (Group B) was more commonly utilized due to its shorter duration and lower risk of OHSS, whereas the long agonist protocol (Group A) was reserved for specific cases.

Inclusion Criteria were as follow: (1)Women included in this study were aged less than 40 years, with normal menstrual cycles (25–35 days) in the preceding three months, and AFC ranging from 8 to 15. (2)Baseline serum (FSH levels were ≤10 U/L, and AMH levels were >1.1 ng/mL. (3)All patients met the established criteria for IVF/ICSI treatment. Exclusion Criteria were as follow: (1) presence of chromosomal abnormalities in either partner, uterine abnormalities (e.g., unicornuate uterus, bicornuate uterus, or septate uterus), (2) any contraindications for IVF-embryo transfer (IVF-ET).

Data collection was meticulously carried out to record both biochemical and clinical outcomes following embryo transfer, including key demographic, clinical, and outcome variables. Baseline variables included patient age, body mass index (BMI), duration of infertility, AFC, AMH levels, and baseline serum FSH. During ovarian stimulation, data on gonadotropin dose, stimulation duration, endometrial thickness, estradiol and progesterone levels on trigger day, and the number of follicles ≥18 mm were recorded. Embryology-related variables included the number of oocytes retrieved, mature oocyte count, fertilization rate, and embryo quality on day 3.

Long Luteal Phase Protocol: this protocol involved the initiation of a GnRH agonist in the mid-luteal phase of the preceding menstrual cycle to achieve downregulation of the hypothalamic-pituitary-ovarian axis. After 14 days of downregulation, COH was commenced using recombinant human FSH (rFSH) or urinary-derived gonadotropins. The gonadotropin dose was adjusted based on patient age, body weight, and AFC. Ovulation was triggered with human chorionic gonadotropin (hCG) once at least three follicles reached a diameter of ≥18 mm, and oocytes were retrieved 36 hours after the trigger.

Flexible Antagonist Protocol: this protocol began with COH initiated on day 2 or 3 of the menstrual cycle. The starting dose of gonadotropins was individualized according to ovarian response as assessed by transvaginal ultrasound (B-ultrasound) and serum estradiol (E2) levels. When the leading follicle reached 12 mm in diameter, a GnRH antagonist was administered to prevent premature LH surges. Ovulation was triggered using a dual trigger (hCG combined with a GnRH agonist) once two or more follicles reached ≥18 mm, and oocyte retrieval occurred 36 hours post-trigger.

Embryos were assessed on day 3 post-fertilization according to standard morphologic criteria. Grading was based on the number of blastomeres, degree of fragmentation, and uniformity of blastomere size, with grades I and II considered high-quality embryos. Embryos of higher quality were prioritized for transfer. In cases of a high risk for OHSS— characterized by serum E2 levels exceeding 4000 pg/mL or retrieval of more than 15 oocytes— a freeze-all strategy was employed. This involved cryopreservation of all viable embryos and postponement of the embryo transfer to a subsequent frozen embryo transfer (FET) cycle to minimize the risk of OHSS.

Patients with confirmed clinical pregnancies were followed at regular intervals to ensure continued pregnancy viability and to monitor for complications. The first follow-up was conducted at 12 weeks of gestation, focusing on the viability of the pregnancy and the assessment of early fetal development. A second follow-up was carried out at 24 weeks of gestation to further assess fetal growth and maternal health. Finally, all patients were followed up 2 weeks postpartum to document delivery outcomes, including birth weight, gestational age at delivery, and any neonatal complications.

The primary outcomes included: (1) Clinical pregnancy rate: defined as the number of clinical pregnancy cycles per number of transfer cycles × 100%. Clinical pregnancy was confirmed by the presence of at least one gestational sac with a fetal heartbeat detected via transvaginal ultrasound at 6–8 weeks of gestation. This calculation includes both fresh embryo transfer cycles and the first frozen embryo transfer (FET) cycle for patients undergoing a freeze-all strategy, with each patient contributing one transfer cycle. (2) Live birth rate: defined as the number of live births per number of transfer cycles × 100%. A live birth is defined as the delivery of at least one living infant at or beyond 24 weeks of gestation. This calculation includes both fresh embryo transfer cycles and the first FET cycle for patients undergoing a freeze-all strategy, with each patient contributing one transfer cycle, representing a single-cycle liver birth rate. (3) Incidence of moderate to severe OHSS: classified based on standard criteria including symptom severity, ovarian enlargement, and biochemical markers.

The secondary outcomes included: (1) Oocyte metrics: total number of oocytes retrieved, mature oocyte rate (number of mature oocytes per total oocytes retrieved), and fertilization rate. (2) Blastocyst formation rate: The percentage of embryos progressing to the blastocyst stage. (3) Embryo quality: proportion of high-quality embryos (grades I and II). (4) Implantation rate: the number of gestational sacs per number of embryos transferred × 100%. (5) Miscarriage rates: early miscarriage rate (spontaneous loss of pregnancy within 12 weeks) and overall miscarriage rate (spontaneous loss of pregnancy at any gestational age).

The sample size was calculated by Pass 11.0 and aimed to detect a 15% difference in clinical pregnancy rate (50% for the long luteal phase protocol vs. 35% for the flexible antagonist protocol) based on prior studies (14, 15). Using a two-proportion z-test with 80% power, a two-sided α = 0.05, and a 1:3 allocation ratio, 144 patients (36 in Group A and 108 in Group B) were required. Accounting for a 10% dropout rate, the target enrollment was increased to 160 patients (40 in Group A and 120 in Group B).

Continuous variables were described as mean ± standard deviation (SD) for normally distributed data. For data that were not normally distributed, variables were expressed as the median and interquartile range (IQR). Comparisons between the two groups were conducted using independent t-tests for normally distributed variables and the Mann-Whitney U test for non-normally distributed variables. Categorical variables were expressed as percentages or proportions. Comparisons between groups were made using chi-square tests or Fisher’s exact test.

Multivariable logistic regression analysis was conducted to identify independent predictors of clinical pregnancy, adjusting for potential confounders. Variables included in the model (age, BMI, AFC, AMH, gonadotropin dose, number of oocytes retrieved, and embryo quality [Grade I-II]) were selected a priori based on their established clinical relevance to IVF/ICSI outcomes, and their potential associations in univariate analyses (P < 0.10). The grouping factor (long luteal phase vs. flexible antagonist protocol) was excluded from the model because the primary study objective was to compare protocol outcomes directly, and the exploratory analyses confirmed that the protocol type was not a significant predictor (P > 0.05). Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported.

A subgroup analysis by age (>20–25, 30–35, and 35–<40 years) was performed to assess clinical pregnancy and live birth rates within each age stratum, comparing Group A and Group B using chi-square tests or Fisher’s exact test. All statistical analyses were performed using the latest version of SPSS (IBM, Chicago, USA). A two-tailed P value of less than 0.05 was considered statistically significant.

Table 1 showed no significant differences in baseline characteristics between Group A and Group B. Age, BMI, duration of infertility, AFC, AMH, FSH levels, gonadotropin dose, stimulation duration, number of follicles ≥18 mm, oocytes retrieved, mature oocyte count, fertilization rates, and day 3 embryo quality were comparable across both groups (P>0.05).

The clinical pregnancy rate was similar between groups (54.8% vs. 56.8%, P=0.092), with fresh cycle rates of 53.3% vs. 56.5% and FET cycle rates of 58.3% vs. 57.6%. Likewise, the live birth rate showed no significant difference (47.6% vs. 52.5%, P=0.278), with fresh cycle rates of 46.7% vs. 52.9% and FET cycle rates of 50.0% vs. 51.5%. These findings indicate comparable efficacy of the two protocols in achieving clinical pregnancy and live births. Additionally, the incidence of OHSS was low, with no cases in Group A and 2.5% in Group B, but this difference did not reach statistical significance (P=0.055) ( Table 2 , Figure 2 ).

For the secondary outcomes, there were no significant differences in the number of embryos transferred (1.8 ± 0.4 vs. 1.9 ± 0.3, P=0.093) ( Figure 3 ), implantation rate (45.2% vs. 42.4%, P=0.447), or early miscarriage rate (4.8% vs. 7.6%, P=0.329) ( Figure 3 ). Similarly, the overall miscarriage rate (7.1% vs. 12.7%, P=0.127), the number of blastocysts formed (5.2 ± 2.0 vs. 5.4 ± 1.8, P=0.549), and the day 5 blastocyst formation rate (38.1% vs. 39.8%, P=0.643) were all comparable between the two groups ( Table 2 ).

The univariate logistic regression analysis ( Table 3 ) revealed several significant predictors of clinical pregnancy (P < 0.05). Younger age (OR = 0.940, P = 0.035), lower BMI (OR = 0.950, P = 0.045), higher AFC (OR = 1.150, P = 0.015), higher AMH levels (OR = 1.400, P = 0.008), greater endometrial thickness (OR = 1.180, P = 0.018), higher gonadotropin dose (OR = 1.003, P = 0.044), higher estradiol on trigger day (OR = 1.002, P = 0.050), lower progesterone on trigger day (OR = 0.820, P = 0.031), greater number of oocytes retrieved (OR = 1.250, P = 0.020), higher mature oocyte count (OR = 1.220, P = 0.025), and better embryo quality (Grade I-II) (OR = 1.600, P = 0.010) were significantly associated with increased odds of clinical pregnancy. Additionally, duration of infertility (OR = 0.900, P = 0.090), FSH levels (OR = 0.920, P = 0.080), and fertilization rate (OR = 1.030, P = 0.070) showed marginal associations (P < 0.10).

Variables with P < 0.10 in the univariate logistic regression analysis were included in the multivariable logistic regression analysis, and the results ( Table 4 ) demonstrated that younger age (OR = 0.956, P = 0.042), higher AFC (OR = 1.127, P = 0.018), higher AMH levels (OR = 1.357, P = 0.005), greater endometrial thickness (OR = 1.162, P = 0.021), greater number of oocytes retrieved (OR = 1.234, P = 0.023), and better embryo quality (Grade I-II) (OR = 1.485, P = 0.002) were significant independent predictors of clinical pregnancy (P < 0.05). Gonadotropin dose (OR = 1.002, P = 0.050) showed trends toward significance but did not reach statistical significance (P > 0.05).

Table 5 presents the age-related subgroup analysis of clinical pregnancy and live birth rates for Group A and Group B, stratified into >20–25, 30–35, and 35–<40 years subgroups. In the >20–25 years subgroup, clinical pregnancy rates were 66.7% (Group A, n=6) vs. 72.2% (Group B, n=18, P=0.706), and live birth rates were both 66.7% (P=1.000). For the 30–35 years subgroup, clinical pregnancy rates were 57.1% (Group A, n=28) vs. 57.7% (Group B, n=78, P=0.953), and live birth rates were 50.0% vs. 53.8% (P=0.717). In the 35–<40 years subgroup, clinical pregnancy rates were 37.5% (Group A, n=8) vs. 40.9% (Group B, n=22, P=0.861), and live birth rates were 25.0% vs. 36.4% (P=0.684). No significant differences were observed between groups, but success rates decreased with increasing age, highlighting age as a key factor influencing IVF/ICSI outcomes.

This study’s prospective design and large sample size are key strengths, allowing for rigorous data collection and robust comparisons between the two protocols. The random assignment of patients to each protocol minimized selection bias and enhanced the generalizability of the findings to other patient populations with normal ovarian reserve (34). The randomization methhod ensured balanced baseline characteristics and minimized the selection bias. Additionally, the use of multivariable logistic regression further adjusted for potential confounders, reducing the risk of bias in identifying predictors of clinical pregnancy.

However, there are several limitations to consider. First, the study was conducted at a single center, which may limit the external validity of the findings. Multi-center studies are needed to confirm these results in more diverse populations and clinical settings. Second, while the sample size was sufficient for the primary outcome, it may be considered moderate for detecting smaller differences in secondary outcomes, such as implantation or blastocyst formation rates, potentially limiting the study’s power for these endpoints. Nevertheless, the observed non-significant differences in primary outcomes (P > 0.05) suggest that the sample size was appropriate for the study’s main objectives. Additionally, while the study focused on short-term outcomes such as clinical pregnancy and live birth rates, it did not assess long-term neonatal outcomes or maternal health beyond the postpartum period. These outcomes are crucial for evaluating the overall safety and effectiveness of ovarian stimulation protocols. Another limitation is the potential for unmeasured confounders, such as genetic factors or variations in laboratory techniques, that may have influenced the outcomes. Future studies should aim to control for these variables and include more detailed assessments of patient and embryological characteristics.

Future research should focus on further refining ovarian stimulation protocols to improve patient outcomes and safety (35, 36). Randomized controlled trials with larger, more diverse populations are needed to confirm the equivalence of the GnRH agonist and antagonist protocols in different patient groups, including those with diminished ovarian reserve and those undergoing multiple IVF cycles.