Does the recovery of serum FSH influence the ovarian sensitivity index in the follicular-phase depot GnRH agonist protocol?

A total of 2286 patients were included in the study. Figure  1 shows the study procedure flowchart. The median age was 29 years, the median infertility duration was 3 years, the median AMH level was 6.22 ng/ml, and the median AFC was 18. The clinical pregnancy rate in this study was 68.83%, and other basic information is shown in Table  1 . Fig. 1 Flow chart of this research Flow chart of this research Table 1 Basic characteristics of the study population Project Value No. of cycles 2286 Age (years) 29 (20–34) Infertility duration (years) 3.00 (1.00–16.00) BMI(kg/M 2 ) 22.27 (15.23–35.88) AFC 18 (5–61) AMH(ng/ml) 6.22 (1.32–24.20) Baseline FSH(mIU/ml) 6.59 (0.64–11.85) Dosage of Gn (IU) 1962.50 (825.00-7762.50) Duration of Gn (days) 12.00 (6.00–20.00) FSH on trigger day (mIU/ml) 12.69 (4.20-37.39) LH on trigger day (mIU/ml) 0.82 (0.00-4.97) E 2 on trigger day (pg/ml) 2255.00 (412.00-9275.00) P on trigger day (ng/ml) 0.67 (0.08–2.58) No. of oocytes retrieved 10 (1–22) No. of mature oocytes 10 (0–20) No. of transferred embryos  1 767 (40.52%)  2 1126 (59.48%) The type of embryo transferred  Cleavage-stage embryo 534 (28.21%)  Blastocyst 1359 (71.79%) Clinical pregnancy  Unpregnant 590 (31.17%)  Pregnant 1303 (68.83%) Basic characteristics of the study population As shown in Table  2 , Patients in group T2 and T3 demonstrated higher AFC and AMH levels, and lower baseline FSH compared to the T1 group. Additionally, a slight but significant increase in maternal age was observed across the tertiles ( p  = 0.015), while the duration of infertility showed no significant difference among groups ( p  = 0.06). The distribution of infertility types (primary vs. secondary) and causes (particularly ovulatory disorders and male factor) also differed significantly among the tertiles ( p  < 0.05). BMI values showed a progressive increase from T1 to T3 groups ( p  < 0.001). OSI was significantly higher in the T3 group compared to T1 and T2 groups (Fig.  2 A). Hormonal levels on the trigger day showed distinct patterns: FSH and E₂ levels decreased significantly with increasing tertiles, while progesterone (P) was significantly lower in the T3 group( p  < 0.05) (Fig.  2 B-D). The T3 group required a significantly lower total gonadotropin dosage and shorter stimulation duration( p   T3, p  = 0.006), the clinical pregnancy rate was significantly higher in the T3 group (T2 < T3, p  = 0.009) (Fig.  2 G-H). No significant differences were observed in LH levels on trigger day, the number of oocytes retrieved, the number of embryos transferred, or the type of embryos transferred among the three groups( p  > 0.05). Other infertility causes including pelvic/tubal factors, endometriosis, immunological and unexplained infertility showed no significant differences across the groups( p  > 0.05). The results of post hoc pairwise comparisons are presented in Supplementary Table S1. Table 2 Comparison of baseline characteristics and outcomes among the 3 groups GROUP T1 (0.56) p -value Effect Size (η²) Post-hoc Analysis (Tukey HSD) No. of cycles 762 762 762 Continuous Variables, Mean ± SD  Age (years) 28.69 ± 2.89 28.83 ± 2.85 29.11 ± 2.79 0.015 0.004 T1 < T3 *  Infertility duration (years) 3.15 ± 1.99 3.33 ± 2.07 3.40 ± 2.18 0.06 0.002  BMI(kg/M 2 ) 21.29 ± 2.94 22.90 ± 3.10 23.93 ± 3.07 < 0.001 0.113 T1 < T2 * , T1 < T3 * , T2 < T3 *  AFC 18.07 ± 6.68 19.79 ± 7.65 20.98 ± 7.90 < 0.001 0.024 T1 < T2 * ,T1 < T3 * ,T2 < T3 *  AMH(ng/ml) 6.48 ± 3.43 7.27 ± 3.84 7.64 ± 4.15 < 0.001 0.015 T1 < T2 * , T1 < T3 *  Baseline FSH(mIU/ml) 7.31 ± 1.44 6.79 ± 1.28 5.98 ± 1.21  T2 * , T1 > T3 * , T2 > T3 *  OSI 5.55 ± 2.43 5.68 ± 2.59 6.16 ± 3.01 < 0.001 T1 < T3 * , T2 < T3 *  Dosage of Gn (IU) 2158.72 ± 733.46 2155.65 ± 783.71 1975.19 ± 786.33  T3 * , T2 > T3 *  Duration of Gn (days) 12.27 ± 1.43 11.96 ± 1.62 11.47 ± 1.81  T2 * , T1 > T3 * , T2 > T3 *  FSH on trigger day (mIU/ml) 15.01 ± 5.28 13.69 ± 4.72 11.84 ± 3.96  T2 * , T1 > T3 * , T2 > T3 *  LH on trigger day (mIU/ml) 0.95 ± 0.55 0.93 ± 0.52 0.93 ± 0.47 0.526 0  E 2 on trigger day (pg/ml) 2689.76 ± 1139.20 2457.79 ± 1118.07 2380.34 ± 1094.13  T2 * , T1 > T3 *  P on trigger day (ng/ml) 0.78 ± 0.39 0.77 ± 0.37 0.70 ± 0.35  T3 * , T2 > T3 *  No. of oocytes retrieved 10.70 ± 2.77 10.78 ± 2.89 10.47 ± 2.75 0.077 0.002  No. of mature oocytes 9.83 ± 2.69 9.69 ± 3.01 9.37 ± 3.07 0.008 0.003 T1 > T3 * Categorical Variables, n (%) T1 T2 T3 p -value Effect Size (φc/V) Infertility type, n (%) 0.036 0.054  Primary 488 (64.21) 456 (59.84) 441 (57.87) T1 > T3 *  Secondary 272 (35.79) 306 (40.16) 321 (42.13) Infertility cause, n (%) < 0.001 0.102  Pelvic and tubal factors 390 (51.18) 395 (51.84) 395 (51.84) 0.957  Ovulation disorder 121 (15.88) 171 (22.44) 204 (26.77) < 0.001 T1 < T2 * , T1 < T3 *  Endometriosis 19 (2.49) 15 (1.97) 19 (2.49) 0.734  Male factor 215 (28.22) 172 (22.57) 132 (17.32)  T2 * , T1 > T3 * , T2 > T3 *  Immunological infertility 0 (0.00) 2 (0.26) 1 (0.13) 0.367  Unexplained infertility 17 (2.23) 7 (0.92) 11 (1.44) 0.11 No. of transferred embryos 0.199 0.042  1 254 (43.05) 235 (38.59) 245 (38.70)  2 336 (56.95) 374 (61.41) 388 (61.30) The type of embryo transferred 0.779 0.017  Cleavage-stage embryo 168 (28.47) 176 (28.90) 172 (27.17)  Blastocyst 422 (71.53) 433 (71.10) 461 (72.83) Clinical pregnancy 0.01 0.071  Unpregnant 191 (32.37) 212 (34.81) 171 (27.01)  Pregnant 399 (67.63) 397 (65.19) 462 (72.99) T2 < T3 * Statistical analysis: Data for continuous variables (A-G) are presented as mean ± SD and were analyzed by one-way ANOVA followed by Tukey's post hoc test for multiple comparisons. Data for the categorical variable (H) are presented as % and were compared between groups using Chi-square test, with Bonferroni correction applied for pairwise post hoc analyses OSI Ovarian sensitivity index, FSH Follicle stimulating hormone, E2 Estradiol, P Progesterone, Gn Gonadotropin * p <0.05 Comparison of baseline characteristics and outcomes among the 3 groups Statistical analysis: Data for continuous variables (A-G) are presented as mean ± SD and were analyzed by one-way ANOVA followed by Tukey's post hoc test for multiple comparisons. Data for the categorical variable (H) are presented as % and were compared between groups using Chi-square test, with Bonferroni correction applied for pairwise post hoc analyses OSI Ovarian sensitivity index, FSH Follicle stimulating hormone, E2 Estradiol, P Progesterone, Gn Gonadotropin * p <0.05 Fig. 2 The post hoc analysis of clinical data in GnFSH/bFSH tertile groups (T1, T2, T3). A OSI. B FSH on trigger day. C E 2 on trigger day. D P on trigger day. E Duration of Gn. F Dosage of Gn. G No. of mature oocytes. H Clinical pregnancy rate. Statistical analysis: Data for continuous variables ( A - G ) are presented as mean ± SD and were analyzed by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Data for the categorical variable ( H ) are presented as % and were compared between groups using Chi-square test, with Bonferroni correction applied for pairwise post hoc analyses. Note: OSI, ovarian sensitivity index; FSH, follicle stimulating hormone; E 2 , estradiol; P, progesterone; Gn, gonadotropin The post hoc analysis of clinical data in GnFSH/bFSH tertile groups (T1, T2, T3). A OSI. B FSH on trigger day. C E 2 on trigger day. D P on trigger day. E Duration of Gn. F Dosage of Gn. G No. of mature oocytes. H Clinical pregnancy rate. Statistical analysis: Data for continuous variables ( A - G ) are presented as mean ± SD and were analyzed by one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. Data for the categorical variable ( H ) are presented as % and were compared between groups using Chi-square test, with Bonferroni correction applied for pairwise post hoc analyses. Note: OSI, ovarian sensitivity index; FSH, follicle stimulating hormone; E 2 , estradiol; P, progesterone; Gn, gonadotropin After adjustment for potential confounders (age, BMI, AMH level, AFC, infertility cause, infertility type), the relationship between the basal GnFSH/bFSH ratio and OSI was analyzed (Fig.  3 ). Table  3 presents the threshold effect analysis for the association between the GnFSH/bFSH ratio and OSI. The table contains two models (Models I and II) and their respective adjusted beta coefficients (95% CI) and P values. In Model I (linear analysis), the one-line slope has an adjusted β of 2.60 with a 95% CI of (2.10, 3.09) and P  < 0.0001. In Model II (nonlinear analysis), the turning point is 0.8, with a slope l of 2.96 (95% CI: 2.37, 3.56) and P  < 0.0001 for values below 0.8 and a slope 2 of 0.19 (95% CI: -2.02, 2.39) and P value of 0.8686 for values above 0.8. The LRT test results indicated that there was a significant difference between Models I and II, with P values of 0.028, suggesting a nonlinear relationship between the GnFSH/bFSH ratio and OSI. Fig. 3 Association between GnFSH/bFSH and OSI. All adjusted for Age, BMI, AMH, AFC, infertility cause, Infertility type. Red line represents the smooth curve fit between variables and the blue dotted curves represents the 95% of confidence interval Association between GnFSH/bFSH and OSI. All adjusted for Age, BMI, AMH, AFC, infertility cause, Infertility type. Red line represents the smooth curve fit between variables and the blue dotted curves represents the 95% of confidence interval Table 3 Threshold effect of the relationship between the GnFSH/bFSH ratio and OSI Models OSI Adjustedβ(95%CI) p value Models I  One line slope 2.60 (2.10, 3.09) < 0.0001 Models II  Turning point 0.8  <0.8 slope1 2.96 (2.37, 3.56) 0.8 Slope2 0.19 (-2.02, 2.39) 0.8686  LRT test 0.028 Models I, linear analysis; Models II, non-linear analysis. LRT test, Logarithmic likelihood ratio test (p value<0.05 means Models II is significantly different from Models I, which indicates a non-linear relationship). Adjusted for Age, BMI, AMH, AFC, infertility cause, Infertility type Threshold effect of the relationship between the GnFSH/bFSH ratio and OSI Models I, linear analysis; Models II, non-linear analysis. LRT test, Logarithmic likelihood ratio test (p value<0.05 means Models II is significantly different from Models I, which indicates a non-linear relationship). Adjusted for Age, BMI, AMH, AFC, infertility cause, Infertility type In the multiple linear regression model, after full adjustment (for age, BMI, AMH level, AFC, infertility cause, infertility type), each one-unit increase in the GnFSH/bFSH ratio was related to increased OSI values (β = 2.59, 95% CI 2.10, 3.09, P   0.56) had increased OSI values (β = 0.49, 95% CI 0.24, 0.74, P  = 0.0001 and β = 1.25, 95% CI 0.99, 1.51, P  < 0.0001, respectively) when compared with patients who had ratio values in T1 (< 0.37) (Table  4 ). Table 4 Association between the GnFSH/bFSH and OSI in unadjusted and adjusted regression models Variables Un Adjusted Model Adjusted Model I Adjusted Model II GnFSH/bFSH β (95%CI) P value β (95%CI) P value β (95%CI) P value Continuous 1.27 (0.76, 1.78) < 0.0001 2.59 (2.10, 3.08) < 0.0001 2.59 (2.10, 3.09) < 0.0001 Categories  T1 0 0 0  T2 0.13 (-0.14, 0.40) 0.3437 0.48 (0.23, 0.73) 0.0002 0.49 (0.24, 0.74) 0.0001  T3 0.60 (0.33, 0.87) < 0.0001 1.23 (0.97, 1.49) < 0.0001 1.25 (0.99, 1.51) < 0.0001 When the GnFSH/bFSH ratio was treated as a continuous independent variable: The Unadjusted Model was a simple linear regression without any covariates. Adjusted Model I was a multivariable linear regression adjusted for Age, BMI, AMH, and AFC. Adjusted Model II was a multivariable linear regression further adjusted for infertility cause and infertility type in addition to all covariates in Model I. The complete results of these multivariable models, including coefficients for all covariates, are presented in Supplementary Table S2–Table S6 When the GnFSH/bFSH ratio was analyzed as a categorical variable: The categorical analysis employed a multivariable linear regression framework in which the GnFSH/bFSH ratio groups were incorporated using dummy variables, with T1 serving as the reference category. The coefficients for T2 and T3 represent the adjusted mean differences in OSI relative to T1, controlling for the same sets of covariates as in the corresponding continuous model. Full results are provided in the Supplementary Material Association between the GnFSH/bFSH and OSI in unadjusted and adjusted regression models When the GnFSH/bFSH ratio was treated as a continuous independent variable: The Unadjusted Model was a simple linear regression without any covariates. Adjusted Model I was a multivariable linear regression adjusted for Age, BMI, AMH, and AFC. Adjusted Model II was a multivariable linear regression further adjusted for infertility cause and infertility type in addition to all covariates in Model I. The complete results of these multivariable models, including coefficients for all covariates, are presented in Supplementary Table S2–Table S6 When the GnFSH/bFSH ratio was analyzed as a categorical variable: The categorical analysis employed a multivariable linear regression framework in which the GnFSH/bFSH ratio groups were incorporated using dummy variables, with T1 serving as the reference category. The coefficients for T2 and T3 represent the adjusted mean differences in OSI relative to T1, controlling for the same sets of covariates as in the corresponding continuous model. Full results are provided in the Supplementary Material

From January 2018 to January 2023, patients who underwent IVF/ICSI in the Reproductive Medicine Center of Shiyan People’s Hospital Affiliated with Hubei University of Medicine were analyzed. All patients were treated with the follicular-phase depot GnRH agonist protocol. A total of 2286 patients were included in this study according to the inclusion and exclusion criteria. The inclusion criteria were as follows: all patients who met the diagnostic criteria for infertility, had undergone their first IVF cycle, were treated with the follicular-phase GnRH agonist protocol, were aged < 35 years, and had an FSH level < 25 mIU/ml. The exclusion criteria were as follows: patients with severe medical diseases such as hypertension, diabetes, gastrointestinal diseases, thyroid disease, hyperprolactinemia, and congenital adrenal hyperplasia and those who underwent PGT. In the follicular-phase GnRH agonist protocol [ 17 ], the patients received a single 3.75-mg intramuscular injection of long-acting triptorelin acetate (Decapeptyl; Ferring, SaintPrex, Switzerland) on Day 3 of the cycle. After 30–42 days of downregulation, gonadotrophin (Gn, r-hFSH, Serono, Switzerland, 75 IU/day) and/or hMG (Lipo Biochem, 75 IU/day) were added for ovulation induction. When there were more than 2 follicles with a diameter of 18 mm, r-hfsh and hMG were stopped, 250 µg recombinant human chorionic gonadotrophin (r-HCG, Serono, Switzerland) was injected, and ultrasound-guided aspiration was performed 36 h late. On the 3rd to 5th day after oocyte retrieval, embryo morphology was scored according to morphological parameters [ 16 ], and high-quality embryos were selected for embryo transfer. Intrauterine transfer was performed under the guidance of ultrasound. Clinical pregnancy was defined as a positive serum hCG level with ultrasound evidence of a gestational sac at approximately 6 weeks of gestation [ 5 ]. Serum FSH, LH, E 2 , and P levels on the 3rd day of menstruation and the trigger day were measured. All samples were assessed using an FSH kit and E 2 kit (chemiluminescence method). In the range of [1, 160] mIU/mL for FSH, the correlation coefficient (R) of the dose‒response curve was not lower than 0.9900 when fitted by the double logarithmic mathematical model. Repeatability (CV%) was not higher than 15.0%, and the difference between batches (CV%) was not higher than 20.0%. In the range of [0, 3000] pg/ml for E 2 , the absolute value of the dose‒response curve correlation coefficient (r) was ≥ 0.9900. The CV of the high- and low-concentration samples was not greater than 10% after 10 repeated tests. The CV among the three lot kits was not greater than 15% for the same sample. Serum FSH recovery refers to the percentage of FSH recovered on the Gn initiation day compared to the FSH level on the pituitary suppression day. In this study, we used the GnFSH/bFSH ratio to represent FSH recovery, which was calculated as follows: GnFSH/bFSH = FSH on the Gn initiation day/baseline FSH level (FSH level on the pituitary suppression day). The primary outcome was the ovarian sensitivity index (OSI) [ 5 ], which was calculated as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{OSI}^{5}=\text{Retrieved oocytes}\times{1000}/\text{total Gn dose}$$\end{document} Statistical analyses were performed using EmpowerStats [ 17 ] ( http://www.empowerstats.com ; X&Y Solutions, Inc., Boston, MA, USA) and SPSS (version 27.0.1, Chicago, IL, USA) and GraphPad Prism (version 9.0, California, USA). Normally distributed continuous variables are expressed as the mean ± standard deviation (x¯±s), and categorical variables are expressed as percentages (%). Continuous variables were expressed as mean ± standard deviation (SD). Comparisons among the three groups were performed using one-way analysis of variance (One-way ANOVA). If the ANOVA result showed an overall significance ( p  < 0.05), post-hoc pairwise comparisons were further conducted using the Tukey’s honestly significant difference (Tukey HSD) test (when variances were homogeneous) or the Games-Howell test (when variances were heterogeneous). Categorical variables are presented as numbers and percentages [n (%)]. The overall differences between groups were assessed using the chi-square test or Fisher’s exact test, as appropriate. If the overall test result was significant ( p  < 0.05), post hoc pairwise comparisons were conducted using chi-square tests with Bonferroni correction applied to adjust the significance level. Multiple linear regression analysis was used to study the relationship between the OSI and GnFSH/bFSH. The relationship between the OSI and GnFSH/bFSH ratio was fitted by smooth curve fitting and the threshold effect value. A P value < 0.05 was considered statistically significant.

In the last 40 years, assisted reproductive technology (ART) has developed rapidly and has become an important method for the treatment of infertility. Sunderam showed that despite a gradual increase in clinical pregnancy rates with ART, the live birth rate was only 38.1% for each in vitro fertilization-embryo transfer (IVF-ET) cycle [ 1 ]. Controlled ovarian hyperstimulation (COH) is one of the most important techniques in ART [ 2 ]. Due to individual differences, individuals have different ovarian responses to COH, and two adverse outcomes (high or low ovarian response) may occur after COH [ 3 ]. Therefore, it is very important to accurately predict the ovarian response after IVF/ICSI (in vitro fertilization/intracytoplasmic sperm injection) [ 4 ]. In current studies, age, AMH (anti-Mullerian hormone) levels, InH B (Inhibins B), AFC (antral follicle count), basal FSH (follicle-stimulating hormone) levels and other indicators are commonly used to predict ovarian responsiveness [ 5 ]. Some scientists have proposed using the ovarian sensitivity index (OSI) to predict ovarian response [ 6 ], which is defined as the number of oocytes retrieved/total amount of gonadotropin. Subsequently, the OSI was confirmed to be related to the AMH level and AFC [ 7 , 8 ]. Pan [ 9 ] showed that when the OSI value was lower, the ovarian sensitivity and pregnancy rate was lower. GnRHa is a commonly used drug to prevent early LH surge during IVF/ICSI [ 10 ]. In China, the standard GnRHa long protocol has been widely used due to its high fresh embryo transfer (ET) clinical pregnancy rate and stable clinical pregnancy rate [ 11 ]. The mainstream protocol in our center is also the follicular-phase depot GnRH agonist protocol. It is undeniable that the protocol has many advantages include higher clinical pregnancy rate [ 12 ] and higher ongoing pregnancy rate [ 13 ]. Besides the follicular-phase depot GnRH agonist protocol may improve endometrial receptivity [ 14 ], and be more friendly to patients with repeated IVF failure [ 15 ]. Gonadotropin (Gn) is an essential drug in the process of ovulation induction. The dosage of Gn is related to the patient’s age, body mass index (BMI), ovarian reserve function, hormone levels, personal and family history, and lifestyle and living environment. To ensure the efficacy and safety of Gn in COH, we hope to find convenient and sensitive indicators that can help to predict the OSI. Therefore, this study mainly focused on the population undergoing the follicular-phase depot GnRH agonist protocol by analyzing the recovery of FSH on the day of Gn initiation compared with the basal FSH level on the day of downregulation and studying the correlation with the OSI to discover a new monitoring target for COH and to better guide reproductive doctors to predict the effect of COH and select the appropriate Gn dosage.

We found that after pituitary suppression, the recovery of serum FSH (GnFSH/bFSH) is an independent risk factor for increased OSI values. When the GnFSH/bFSH ratio was less than 0.8, the OSI increased with the increase in the GnFSH/bFSH ratio value. However, when the ratio was higher than 0.8, there was no significant relationship between the GnFSH/bFSH ratio and OSI. This conclusion provides some basis for clinicians to adjust the dosage of Gn during COH. However, our study has certain limitations. For example, we derived our findings from a relatively small number of individuals, which should be validated in a larger cohort of Chinese patients. Second, this study was limited by its retrospective nature and single center design. In the future, the sample size will be increased, or a multicenter study will be performed for further validation.

Controlled ovarian hyperstimulation (COH), as an important aspect of human-assisted reproductive technology, greatly affects the outcomes of IVF/ICSI-ET. The gonadotropin-releasing hormone agonist (GnRHa) has long been used in COH due to its role in preventing premature follicular luteinization. As one of the mainstream protocols in current ovarian hyperstimulation, the follicular-phase long GnRH agonist (GnRHa) protocol has many advantages, such as good inhibition of the endogenous LH peak, synchronization of follicles and follicles, synchronization of follicle development and hormone growth, synchronization of follicle and embryo development, and synchronization of embryo development and the endometrium to improve the pregnancy rate. It can also improve endometrial receptivity by promoting apoptosis of endometrial cells and regulating immune and signaling pathways [ 17 – 19 ]. The follicular-phase long GnRHa protocol is the mainstream protocol in our center, which mainly uses GnRHa long-acting preparations. As with other ovarian stimulation protocols, the dosage of Gn in the follicular-phase long GnRHa protocol also needs indicators related to ovarian response. The ovarian reserve is the main factor affecting the ovarian response to COH. In the past, clinical indicators, including age, AMH level, AFC, basal FSH level and FSH/LH levels, were commonly used to reflect ovarian reserve [ 20 , 21 ]. A poor ovarian response has been reported in patients with normal FSH levels [ 22 ]. Researchers have also used the combined application of AMH levels and AFC to assess ovarian response [ 23 ]. Mutlu [ 24 ] showed that AMH levels have a low sensitivity for predicting poor ovarian response. Overall, there are no specific markers to evaluate ovarian reserve and response individually, and they still need to be evaluated in combination. The results of the prospective study [ 6 ] showed that AMH levels were highly correlated with the OSI and could be used as a surrogate for AMH monitoring to predict ovarian responsiveness in patients undergoing IVF. Although the OSI has been recognized by many scholars to reflect ovarian response, it cannot provide clinicians with reference data before the initiation of ovarian stimulation. Therefore, the purpose of our study was to discover indicators that can predict ovarian response before the initiation of ovarian stimulation and that are closely related to the OSI. In the follicular-phase long GnRHa protocol, after full-dose depot GnRH-an injection, the level of sex hormones could be completely suppressed for 2 weeks, and FSH levels gradually recovered after 3–4 weeks [ 25 ]. In our study, all subjects received a GnRHa injection on the third day of menstruation, and the FSH level on the day of GnRHa injection was called the basal FSH level. Gn was given 30–42 days after GnRHa injection, and the FSH level on the day of Gn was called GnFSH. Since FSH levels gradually recovered after 3–4 weeks of downregulation, the GnFSH/bFSH ratio indicates the recovery of endogenous FSH, and the correlation between the recovery of endogenous FSH and the OSI has not been reported until now. In our study, we found that the GnFSH/bFSH ratio was positively correlated with the OSI. Yamashita [ 26 ], who studied the effect of the FSH/LH ratio on IVF-ET outcomes after pituitary downregulation, found that the FSH/LH ratio was positively correlated with the Gn dosage after pituitary downregulation. He also noted that the recovery of LH spontaneous release and stimulated release was slow and dose dependent, while the recovery of the FSH concentration and stimulated release was faster. As he evaluated the FSH/LH ratio, this study cannot confirm his view; however, his study suggested that the recovery of FSH after pituitary downregulation may affect the Gn dosage. De Ziegler D [ 27 ] also mentioned that FSH and LH release after pituitary downregulation by GnRHa suggested restoration of endogenous GnRH action in the pituitary gland. Hastie [ 28 ] conducted animal experiments with sheep and found that insulin-like growth factor binding proteins (IGFBPs) are involved in the regulation of IGF-I and IGF-II, thereby regulating the growth and development of ovine follicles. GnRHa administration may affect follicle development by changing the proliferation and differentiation of IGFBP levels in ovine follicles through changes in FSH and LH levels. These results suggest that the changes in FSH levels after pituitary downregulation may be related to follicular development and ovarian response. Considering that there are many factors affecting the OSI, after adjusting for other confounding factors, we used a smooth fitting curve and threshold saturation effect analysis and found that when the GnFSH/bFSH ratio was less than 0.8, the OSI level gradually increased with the recovery of endogenous FSH. In the process of COH with the follicular-phase prolonged protocol, the use of GnRH-a to downregulate pituitary function can reduce the circulating endogenous follicle stimulating hormone (FSH) level below the threshold to support follicular growth [ 29 ], and then supraphysiological doses of Gn can be administered to obtain more synchronous follicular development. However, both endogenous and exogenous FSH contribute to follicle growth during this process. The GnFSH/bFSH ratio indicates the percentage of endogenous FSH to basal FSH during pituitary recovery after downregulation. We can imagine that the total FSH level during COH ovulation is the sum of endogenous FSH and exogenous FSH. Assuming that the total FSH level is relatively constant, the higher the endogenous FSH level is, the lower the exogenous FSH level will be. The level of exogenous FSH is closely related to the dosage of Gn, so it determines the OSI. The ovarian sensitivity index (OSI) has been suggested as a measurement of ovarian sensitivity during the IVF procedure, which could reflect the potential of follicles to produce oocytes in response to exogenous Gn stimulation [ 30 , 31 ]. This may also explain why the OSI was increased with a GnFSH/bFSH ratio < 0.8.

Supplementary Material 1. Supplementary Material 1.