Abstract
Background
Estradiol-to-oocyte ratio (EOR) was proposed as an assisted reproductive technology (ART) outcome predictor in general population. In women with polycystic ovary syndrome (PCOS), hormonal imbalances, especially estradiol and luteinizing hormone (LH) may affect the predictive value of EOR.
Objectives
We hypothesized that EOR plays a specific predictive role in PCOS and LH activity during ovarian stimulation modulates EOR and ART outcomes.
Design
Retrospective cross-sectional design.
Methods
The study included 276 patients with PCOS undergoing ART, stratified into four EOR groups: I (<466), II (467–683), III (684–886), and IV (>886). Additionally, patients were categorized based on LH activity during ovarian stimulation: no LH, recombinant LH, or LH-like activity. Key-outcome variables included counts of periovulatory follicles, retrieved oocytes, metaphase II (MII) oocytes, total number of embryos, good-quality (GQ) and frozen embryos, biochemical pregnancy rate (BPR), and clinical pregnancy rate (CPR). Comparisons were made across EOR and LH activity groups.
Results
Group I (lowest EOR) demonstrated the highest counts of MII oocytes, total number of embryos, GQ and frozen embryos. BPR and CPR showed a decreasing trend with increasing EOR, peaking in Group I. Multivariable analysis identified the absence of LH activity, higher periovulatory follicle count, and lower EOR as independent predictors of improved ART outcomes. The lowest EOR was observed in patients without LH supplementation, who achieved the best clinical results, though the differences between BPR and CPR were not statistically significant.
Conclusion
EOR is a valuable prognostic biomarker for ART outcomes in PCOS patients. Lower EOR is associated with more favorable outcomes. LH activity during stimulation influences EOR and ART success, particularly in patients with PCOS.
Keywords
Introduction
Ovarian stimulation during assisted reproductive technology (ART) usually leads to supraphysiological estradiol (E2) levels, 1 especially in women with polycystic ovary syndrome (PCOS). 2 Basic, biomedical, and clinical research papers revealed that high E2 in ART hurts ART outcomes by affecting endometrial receptivity, 3 gene expression during the window of implantation, and 4 abnormal stroma development and pattern of myometrial contractions. 5 Nevertheless, some studies delivered opposite results, indicating that high E2 levels on the day of trigger 6 and high E2 levels per mature follicle or retrieved oocytes, 7 significantly impact ART success. Considering the above-mentioned data and issues, attempts were made to predict ART outcomes by examining the relationship between E2 levels on the day of the trigger and the number of retrieved oocytes.8,9 One of the most important predictors of ART outcomes is the estradiol-to-oocyte ratio (EOR). In the general population of infertile women undergoing ART, clinical pregnancy rates (CPR) are the highest in women with EOR between 250 and 750 and decline as this ratio increases, independently of the age of the patients. 8 Furthermore, the lowest EOR 9 or both ends of the EOR spectrum values (less than 250 and higher than 750) carry lower ART success rates. 8
PCOS is the most prevalent endocrine disorder in reproductive-age women and an endocrine-related cause of infertility. PCOS is a challenging condition for ovarian stimulation in ART due to asynchronous follicular growth, poor ovarian response, and relatively high cancelation rates.10,11 In addition, due to achieving higher E2 levels than in other causes of infertility, PCOS is a risk factor for ovarian hyperstimulation syndrome (OHSS), a potentially life-threatening condition. Previous studies found that high E2 levels after ovarian stimulation and high EOR in ART are related to higher miscarriage rates 12 and increased risks of pregnancy complications.13,14 In contrast to other causes of infertility, a considerable number of women with PCOS have luteinizing hormone (LH) levels higher than follicle-stimulating hormone (FSH) levels. 15 The long-lasting rapid pulse of secretion of gonadotropin-releasing hormone (GnRH) is considered to be the cause of this phenomenon. This neuroendocrine symbol of PCOS favorably endorses pituitary LH over FSH synthesis and secretion, resulting in increased LH levels and LH/FSH ratio. 16 A high LH/FSH ratio, which often occurs in women with PCOS, disturbs ovulation and raises androgen synthesis. 17 Presumably, due to the issue of high LH levels, exogenous recombinant LH (rec-LH) treatment should not be routinely used in combination with FSH for ovarian stimulation in women with PCOS undergoing ART. 18 However, everyday clinical practice in ovarian stimulation reveals relatively frequent noncompliance with recommendations from the international evidence-based guidelines for the management of PCOS. Hence, GnRH agonist ovarian stimulation protocol and the use of LH bioactivity from different human chorionic gonadotropin (hCG) or recombinant LH are still used in ovarian stimulation of women with PCOS. Nevertheless, it has been demonstrated that follicle development and pregnancy rates achieved in women with PCOS receiving recombinant FSH (rec-FSH) alone were similar to those receiving a combination of rec-FSH and rec-LH. 19 However, disagreeing findings were found in a study by Kugelman and colleagues. They revealed that LH supplementation in women with PCOS undergoing GnRH antagonist (GnRH-an) stimulation protocols might impair cumulative live birth rates despite lowering gonadotropin dosages and shortening the duration of ovarian stimulation. 20
Given that PCOS is an endocrine-related cause of infertility, often with hyperestrogenic and hyperandrogenic features, it is reasonable to assume that optimal EOR values for ART outcomes may differ in PCOS compared to other causes of infertility. However, the optimal ranges of EOR values have not been estimated in women with PCOS. Therefore, our research aimed to conduct a preliminary investigation into which ranges of EOR values in PCOS are associated with the highest pregnancy rates, as well as the greatest number of obtained embryos, good quality (GQ) embryos, and frozen embryos. Additionally, we sought to determine whether LH activity in ovarian stimulation (LH-like activity, recombinant LH, or the absence of LH activity) is linked to the optimal EOR values for ART outcomes.
Methods
Setting and aim of the study
This study with a retrospective cross-sectional design was conducted in the Clinic for Gynecology and Obstetrics “Narodni Front”, Belgrade, Serbia from April 2022 to April 2025. It aimed to compare the laboratory and clinical outcomes of ART between EOR groups in women with PCOS, with a special focus on the absence or presence of LH or LH-like activity in ovarian stimulation. The study was approved by the Ethical Committee of Clinic for Gynecology and Obstetrics “Narodni Front” (Approval number 2208/2024/006989 on the 3rd of April 2024), and the reporting of this study conforms to the STROBE statement. 21 The requirement for written informed consent was waived due to the retrospective design and the use of deidentified data.
Study design and subjects
The study evaluated the ART outcomes of fresh, autologous cycles of infertile women diagnosed with PCOS. For the diagnosis of PCOS, the Rotterdam criteria were applied. Women were classified as PCOS patients if they met a minimum of two of the following criteria: multi-follicular ovarian aspect on transvaginal ultrasonography, oligo- or amenorrhea, biochemical or clinical hyperandrogenism. 22 Women aged greater than 45 years and women with a body mass index greater than 30 were excluded from the study to ensure a homogeneous patient population. After applying these inclusion and exclusion criteria, a total of 276 women with PCOS underwent ART in our clinic during the study period.
For each study participant, EOR was calculated by dividing the estradiol concentration on the triggering day by the total number of retrieved oocytes. Until now, it has not been determined which EOR values in women with PCOS are linked to favorable ART outcomes. Since the cut-off values for these outcomes have not been defined, we divided all study participants into four EOR groups based on quartiles of EOR values. This grouping method is statistically rational and obtains balanced groups as follows: EOR group I (<466), EOR group II (467-683), EOR group III (684-886), and EOR group IV (>886). Different types of gonadotropins were used to stimulate all study participants. Ovarian stimulation in study participants was performed either without LH activity or with added LH activity. One group of women was stimulated exclusively with recombinant FSH (rec-FSH), without LH or LH-like activity (defined as “absent”). These patients received follitropin delta (Rekovelle®, Ferring GmbH, Kiel, Germany) according to the individualized dosing algorithm provided by the manufacturer. In the remaining participants, ovarian stimulation included either recombinant LH combined with rec-FSH (“rec-LH”; Pergoveris®, Merck KGaA, Darmstadt, Germany) or LH-like activity (“LH-like”), achieved by the use of highly purified human menopausal gonadotropins (Menopur®, Ferring Pharmaceuticals, Saint-Prex, Switzerland, or Merional®, IBSA Institut Biochimique S.A., Lugano, Switzerland).
ART procedures
Ovarian stimulation protocols, ART practices, and procedures were previously issued.10,23 In summary, a flexible GnRH-an protocol for ovarian stimulation was used. Usually, when the leading follicle reaches ≥ 13 mm in diameter, GnRH-an is introduced in ovarian stimulation. As previously disclosed, follitropin delta was individually dosed according to the manufacturer’s dosing algorithm, which takes into consideration women’s anti-Mullerian hormone (AMH) levels and body weight of women stimulated with other gonadotropins, the gonadotropin dosage per day varied between 100 and 300 IU, depending on ovarian reserve and previous ovarian response in those women with more than one IVF attempt. In this group, high gonadotropin dosages were given only to a few women with poor ovarian response in current or previous ovarian stimulations during ART. After the beginning of ovarian stimulation, the first ultrasound assessment was performed on the fifth day of ovarian stimulation, and further ultrasounds were performed daily until the introduction of GnRH-an. Generally, at least three follicles with a diameter greater than 17 mm were the criterion for triggering. However, in women with poor ovarian response, the trigger was introduced if there was a unique follicle with a diameter of> 17 mm. Oocyte pickup under transvaginal ultrasound guidance was performed 35 hours after triggering the injection. Afterward, GQ embryo transfer was performed. Luteal phase support was done by vaginal application of micronized progesterone 600 mg daily and intramuscular application of 250 mg of hydroxyprogesterone caproate every fifth day.
Data collection
The data were gathered from paper and electronic health records stored in Meditex software (Fertility database system for therapy documentation in ART and assurance of quality for reproductive medicine, CRITEX GmbH, Regensburg, Germany). Several parameters were obtained from patients’ medical histories. Baseline parameters assessed shortly before or at the beginning of ovarian stimulation encompassed age, weight, smoking habit, menarche, cycle duration, antral follicle count (AFC), FSH, LH, E2, and AMH levels. Age was categorized into three clinically relevant groups (<35, 35–39, and ≥40 years) based on established reproductive medicine thresholds and the concept of advanced maternal age as described by the American Society for Reproductive Medicine (ASRM). 24
Apart from statistically justified choice of quartile-based classification of EOR groups, prior studies additionally rationalise quartile stratification. Previous research point to a non-linear, frequently parabolic relationship between the EOR and follicular efficiency, 8 endometrial receptivity, implantation 25 and pregnancy rates. 26 Furthermore, quartile-based groups provide a framework for assessing “follicular efficiency” in a such a heterogeneous group of infertile women such as those women with PCOS. Dividing the cohort into quartiles allows researchers to account for the “U-shaped” relationship and the heterogeneity of ovarian response. Quartile stratification transforms a continuous, highly variable biological marker into discrete categories, making it easier to identify specific threshold effects that might be obscured in a simple linear regression. While high responders among women with PCOS face the challenge of quantity over quality and systemic safety, poor-responder PCOS patients represent a “silent” subset where the high antral count is deceptive, masking an underlying follicular dysfunction that requires aggressive or alternative stimulation strategies. Moreover, researchers claim that absolute E2 levels are deceptive since they are dependent on total oocyte yield. By employing quartiles of the EOR, we might provide a more precise proxy for per-follicle health, allowing for a nuanced comparison between those women with PCOS who are “high-responders” and those considered as “low-responders” among women with PCOS (due to follicular growth asynchrony) regardless of the total number of eggs retrieved. Finally, from clinical predictive value perspective, establishing specific cut-off points (e.g., the 25th vs. 75th percentile) offers a practical diagnostic framework. This allows clinicians to categorize patients into distinct “metabolic phenotypes,” facilitating more precise decisions regarding cycle segmentation (Freeze-all) or trigger adjustments.
Other collected parameters from patients’ medical history are referred to as the outcomes of ovarian stimulation and ART and are presented as follows.
Outcomes of ovarian stimulation and ART
The outcomes of ovarian stimulation and ART were compared across different EOR groups and groups with various kinds of LH activity (absent, rec-LH, and LH-like). Analyzed outcomes of ovarian stimulation encompassed the number of periovulatory follicles, E2 levels at trigger day, the number of retrieved oocytes, and the number of retrieved metaphase II (MII) oocytes. Evaluated ART outcomes were described as the total number of embryos, GQ embryos, frozen embryos, and biochemical and clinical pregnancy rates (BPR and CPR, respectively). Periovulatory follicles were defined as follicles measured more than 15 mm on the last day of ovarian stimulation. 10 Oocytes MII are considered when they contain a polar body and show no signs of post-maturity. 27 Embryos of GQ were classified as either top-quality or GQ embryos based on the criteria established during the Istanbul Consensus Workshop on Embryo Assessment 28 surplus GQ embryos after embryo transfer were frozen.
Biochemical pregnancy was confirmed by measuring serum β-human chorionic gonadotropin (β-hCG) in the patient’s blood 14 days after embryo transfer. BPR was calculated by dividing the number of positive β-hCG (level greater than 50 mIU/mL) by the number of performed embryo transfers. Clinical pregnancy was defined as the detection of a gestational sac or fetal pole on an ultrasound assessment performed 35 days after embryo transfer, including ectopic pregnancies and single and multiple gestational sacs. CPR was calculated by dividing the number of patients with clinical pregnancy by the number of performed embryo transfers.
Statistical analysis
Results are presented as count (%), means ± standard deviation, or median (25th-75th percentile) depending on data type and distribution. Groups are compared using parametric (ANOVA) and nonparametric (Pearson chi-square, Mantel-Haenszel chi-square test for trend, Kruskal-Wallis test) tests. General linear model (linear regression) and logistic regression were performed to evaluate the relationship between dependent and independent variables. All p-values less than 0.05 were considered significant. All data were analyzed using SPSS 29.0 (IBM Corp. Released 2023. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp.) and R 3.4.2. (R Core Team, 2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/).
Results
Characteristics of participants according to EOR quartiles.
Results are presented as number (%), mean±SD or median (25-75th percentile); AFC - Antral Follicle Count; AMH - Anti-Müllerian Hormone; EOR – Estradiol-to-oocyte ratio; FSH - Follicle-Stimulating Hormone; GQ – good-quality; LH - Luteinizing Hormone; No – Number.
The table shows significant differences regarding smoking rates, AMH, AFC, and the number of periovulatory follicles across EOR groups. The MII oocytes, total embryos, GQ embryos and frozen embryos are the highest in the EOR group I. A one-way ANOVA indicated a significant difference in the number of GQ embryos across the four groups of EOR, F(3, 272) = 4.97, p = 0.002. Polynominal contrast revealed a significant linear trend, F(1, 272) = 14.43. p<0.001, with no significant deviation from linearity, F(2, 272) = 0.24, p = 0.789, which is presented in Figure 1. When analyzing the trend of pregnancy, the significant linear trend was observed as EOR increases (β = −0.330, p = 0.007), with the highest values in both BPR and CPR in the EOR group I (value of <466 pmol/oocyte) (Figure 2). Mean number of GQ embryos across EOR groups. EOR – Estradiol-to-oocyte ratio; GQ – good quality; No – Number. Clinical pregnancy rate across EOR categories. EOR – Estradiol-to-oocyte ratio.
Characteristics of participants according to the absence/presence of LH in ovarian stimulation.
Results are presented as number (%), mean±SD or median (25-75th percentile); AFC - Antral Follicle Count; AMH - Anti-Müllerian Hormone; EOR – Estradiol-to-oocyte ratio; FSH - Follicle-Stimulating Hormone; GQ – good-quality; LH - Luteinizing Hormone; No – Number.
Table 2 shows that cycle duration, FSH and AMH levels, and AFC significantly differed across the mentioned groups. The highest EOR, a significant finding, was present in the group with LH-like activity in ovarian stimulation. At the same time, the lowest EOR was achieved in the group of women whose ovarian stimulation was without LH. These findings carry substantial implications for our understanding of ovarian stimulation. The MII oocytes and total number of embryos were not significantly different across the groups, but the highest numbers were found in the group without LH. Similarly, the number of GQ embryos and frozen embryos was the highest in the group without LH activity.
The highest BPR and CPR were also noted in the group without LH (Figure 3), but the difference between the groups did not reach statistical significance. Biochemical and clinical pregnancy rates across different types of ovarian stimulation. BPR – Biochemical pregnancy rate; CPR - Clinical pregnancy rate; LH - Luteinizing Hormone; rec-LH – recombinant LH.
Statistical models with EOR, total number of embryos, MII oocytes, GQ, and frozen embryos.
AFC - Antral Follicle Count; Cycle duration L - logarithmic duration; Cycle duration value; EOR – Estradiol-to-oocyte ratio; GQE– good-quality embryos; PF – periovulatory follicles.
The outcomes are as follows: EOR, the total log-transformed data to obtain stable frozen embryos. The model is performed with original and log-transformed data to stable and small variances. The results are compared, and it is reassuring to note that they are very similar. For simplicity purposes, the model with original data is presented because the coefficients are interpretable, while models with log-transformed data are slightly more accurate.
The first model is used to assess the factors that influence EOR, while the other models include EOR as one of the predictors of the number of embryos. It is crucial to note that the most significant predictors of ART success are the type of ovarian stimulation without LH, the number of periovulatory follicles, and EOR.
Discussion
Our study, the first of its kind, has evaluated the impact of EOR on ART outcomes specifically in the population of women with PCOS undergoing ART. Previous studies have addressed this topic in the general population of women undergoing ART. Our results not only support our assumption that PCOS, as an endocrine-related cause of infertility, differs in terms of optimal EOR values for the best ART outcomes, but also provide practical insights that can guide future research and clinical practice.
The first study evaluating and mentioning EOR as the potential predictor of ART outcomes was conducted by Loumaye and colleagues in 1997. 29 The range of EOR values linked with the lowest ART outcomes was between 0 to 250 pmol/oocyte. Furthermore, they have indicated that the EOR range associated with the highest CPR ranges from 250 to 500 pmol/oocyte. However, their results suggested that the further increase of EOR values (above 500 pmol/oocyte) results in a decrease in CPR. This study comprised women with tubal disease, mild endometriosis, and unexplained infertility, while women with PCOS were excluded. Although a similar number of study participants were included in the study (303 vs. 276), it is hard to compare their results with ours since inclusion and exclusion criteria were quite the opposite, and different ovarian stimulation protocols were applied (GnRH agonist vs. GnRH-an). GnRH agonist and GnRH-an protocols differ significantly in EOR values, with those being higher in GnRH agonist protocols. 30 The largest study ever that addressed this issue, performed in Ireland and evaluated all IVF cycles (all causes of infertility were included) with 9109 oocyte retrievals performed for 5499 patients, demonstrated that the highest CPR was present also in EOR ranges from 250 to 500 pmol/oocyte. Like CPR, oocytes of poorer quality are more frequently retrieved as the EOR increases. Hence, they suggested an optimal EOR of 250–750 pmol/oocyte that results in the best outcomes. 8 Opposing the first study about EOR as a predictor of ART success that found the lowest EOR values (below 250 pmol/oocyte) linked with the lowest CPR, 29 we found that our lowest EOR range was associated with the highest CPR.
Several explanations could enlighten our findings. In general, E2 and estrone in follicular fluid have significant roles in regulating follicular development and oocyte maturation. 31 In women with PCOS, hormonal imbalances leading to high LH, relatively low FSH and androgen excess 32 affect granulosa cell differentiation and aromatase expression, reducing the conversion of androgens to estradiol within developing follicles. 33 This endocrine milieu plays a central role in disrupting normal follicular development and oocyte maturation and endorses excessive follicular recruitment without proportionate estradiol synthesis per follicle, leading to a reduced EOR. Such an inequity reflects impaired follicular steroidogenic capacity and suboptimal intrafollicular conditions, which are critical determinants of coordinated oocyte maturation and developmental competence. 32 PCOS-associated hyperandrogenism, dysregulation of gonadotropin secretion, and insulin resistance affect granulosa cell function and steroidogenesis. Reduced or insufficient FSH activity impairs aromatase expression in granulosa cells, conversion of androgens to estradiol within individual follicles, and consequently, despite the recruitment of multiple follicles, produces relatively low estradiol production, yielding a reduced EOR. 34 This phenomenon reflects asynchronous follicular development and incomplete granulosa cell maturation, both of which are critical for optimal oocyte competence. 35 Moreover, hyperandrogenism can exert direct detrimental effects on follicular microenvironment and oocyte quality by altering intra-follicular signaling, oxidative stress balance, and meiotic progression. 5 Although total circulating estradiol levels may appear normal or elevated due to the high number of developing follicles, the estradiol output per oocyte remains suboptimal, indicating compromised follicular efficiency rather than enhanced functional maturity. 36 Achieving lower values of EOR in PCOS suggests that ovarian stimulation was more optimal since the follicles were developing without excessive E2 levels, leading to higher-quality oocytes and embryos. There are several possible explanations for this. This could be due to the phenomenon entitled “Natural selection, quality for quantity” by Zeev Blumenfeld. 37 Excessive serum and follicular fluid E2 levels in nature and ovarian stimulation discard flawed follicles with aneuploidy or suboptimal genetics. Ovarian stimulation with lower gonadotropin dosages in ART decreases the risk of embryo aneuploidy in humans. Baart and colleagues have demonstrated in a preimplantation genetic screening experiment that ovarian stimulation with lower doses of gonadotropins causes a decreased proportion of aneuploid embryos compared to stimulation with higher doses. 38 Additionally, the results of a study performed by Arce and colleagues imply that above the gonadotropin dose threshold level (dependent on the AMH value) at the beginning of ovarian stimulation, further increase of dosage has no result in increasing the number of retrieved competent oocytes. 39 In the group with the lowest EOR values (with the best ART outcomes), applied dosages of gonadotropins were the lowest, according to previously mentioned literature data. Furthermore, in a study group with stimulation without LH, the applied doses of gonadotropin for each patient were determined by the algorithm designed by the manufacturer of the applied recombinant FSH. When calculating the individualized dose, this algorithm considers each patient’s body weight and AMH. Therefore, in this group, the individualized dosage of rec-FSH provided the needed gonadotropin dose threshold and avoidance of surplus gonadotropin dose that could only increase E2 levels without increasing the number of GQ oocytes and GQ embryos, but increasing the risks of OHSS. To account for potential confounding due to differences in gonadotropin dosing between groups—individualized in the no-LH group versus standardized LH supplementation in the other groups—we used EOR as the primary measure of ovarian response. EOR inherently reflects both the number of retrieved oocytes and circulating E2 levels, allowing fair comparison of ovarian response and ART outcomes. This approach supports the interpretation that lower or no LH supplementation can be sufficient in women with PCOS when EOR is monitored, while individualized dosing likely contributed to optimal EOR and a favorable safety profile in the no-LH group.
Apart from BPR and CPR, other differed significantly across EOR groups. The groups differed considerably in gonadotropin dosage, E2 levels on the trigger day, number of periovulatory follicles, MII oocytes, and total, GQ, and frozen embryos.
In women with PCOS, noticeably elevated E2 levels during and at the end of ovarian stimulation are present compared to normal and poor responders. At the same time, high E2 levels are a risk factor for developing OHSS. 40 The lowest E2 levels on the trigger day among our study participants were found in the lowest EOR group and the group of women with ovarian stimulation containing rec-LH. Apart from the best ART outcomes, the lowest EOR group had the best safety profile. The E2 to follicle ratio (EFR) is a surrogate marker for EOR. Previously, it was established that the lowest E2 to follicle ratio is associated with the lowest risk of the development of OHSS. 41 Furthermore, according to previously published data, the lowest gonadotropin dosage was applied in the EOR group with the lowest values.8,29 These findings suggest that for women with PCOS, ovarian stimulation without LH and with lower EOR values may lead to better ART outcomes and a reduced risk of OHSS, providing valuable insights for clinical practice.
In clinical practice, as a dynamic indicator reflecting the ratio between circulating E2 and the number of mature oocytes, EOR may mirror granulosa cell function and the degree of follicular synchronization. A balanced EOR is likely associated with optimal oocyte maturity and competence, supporting the decision to proceed with a fresh embryo transfer. Conversely, a low EOR, even in the presence of a high oocyte yield, may indicate suboptimal oocyte quality and impaired embryo development, whereas a disproportionately high EOR may reflect supraphysiologic E2 levels that could compromise endometrial receptivity. In such cases, a freeze-all strategy or deferred embryo transfer may be clinically justified. EOR assessment throughout the IVF cycle could also guide whether to include or adjust LH in the stimulation protocol. In women with low EOR, LH supplementation may be considered to improve follicular maturation, particularly in long GnRH agonist or GnRH antagonist protocols, where endogenous LH may be suppressed. In women with balanced or moderate EOR, minimal or no LH may be preferable to avoid excessive estradiol production and reduce OHSS risk. For cycles where EOR is trending high during stimulation, clinicians may opt to reduce LH or total gonadotropin dose, adjust trigger timing, or consider a GnRH agonist trigger in antagonist cycles to improve safety. Monitoring EOR throughout stimulation also allows retrospective and prospective optimization. It provides a measure of ovarian response after the cycle (retrospective), while EFR measured during stimulation can predict final EOR and allow mid-cycle adjustments.
From long time ago, it is known that different types of applied gonadotropins during ovarian stimulation exhibit diverse effect on levels of steroids and levels of other components of follicular fluid. Studies by Westergaard and colleagues and by Hill and Osteen demonstrated among women stimulated with human menopausal gonadotropins significantly higher levels of LH, FSH, estradiol and androstenedione in follicular fluid, while concentrations of human chorionic gonadotrophin and progesterone were significantly lower, compared with those women treated with rec-FSH.42,43 There are different approaches regarding OS in women with PCOS. While some advocate the use of solely rec-FSH (without LH or LH-like activity), others promote supplementation of LH or LH-like activity. Those who are against supplementation of LH or LH-like activity in women with PCOS justify their viewpoint by increased levels of androstenedione caused by use of human menopausal gonadotropins.42,43 As a precursor to testosterone, androstenedione is already elevated in women with PCOS, regardless to OS. Further increase of androstenedione, could only worsen the syndrome by exacerbating hyperandrogenism, intensifying chronic follicular atresia, and delaying oocyte maturation. Thus, such OS could lead to poorer outcomes, such as lower oocyte quality and lower fertilization rates. On the other hand, proponents of LH or LH like activity supplementation rationalize their standing point by the fact that LH activity could reduce the risk of OHSS by sustaining the growth of selected follicles without recruiting new follicles into growth. 44
Studies with conflicting data compared to ours comprised a population of women with endometriosis, and that could explain such discordance. PCOS and endometriosis share a mutual link with high ovarian estrogen levels, leading to a disbalance of sex hormones, presented as high estrogen with progesterone resistance in endometriosis and high estrogen with high androgen in PCOS. 45 However, these high estrogen levels are reflected in different body compartments. Compared to PCOS, increased E2 levels are detected in menstrual period blood rather than in circulation in endometriosis. This finding suggests an abundance of focal E2 secretion, 46 which an elevated aromatase enzyme activity can explain 47. In contrast, in women with PCOS, there is a decrease in aromatase enzyme activity due to alterations of regulatory sequences of the CYP19A1 gene (which encodes aromatase enzyme), and consequently, overproduction of androgens. 48 Earlier studies from 1994 and 1999 demonstrate that both low and high E2 levels are negative prognostic factors for clinical pregnancy in women undergoing ART.49,50 Although the E2 level reflects the overall process of folliculogenesis and may indicate granulosa cell function, excessively high E2 levels do not convince us to improve ART outcomes. Moreover, it does not consider the ratio of follicles to oocytes, which limits its predictive value. Consequently, EOR may provide a more accurate indication of follicular activity than E2, considering that EOR may be used to assess the quality of the oocyte and its maturation. 51 With increasing EOR numbers, the MII oocytes decrease, and the differences between EOR groups are highly significant. This is according to other studies, regardless of the ovarian stimulation protocol used.8,9,29
One of the goals of our study was to improve management strategies in ovarian stimulation in a challenging population of women with PCOS. By requiring the presence of only two of three features, the Rotterdam criteria acknowledge the heterogeneity of women with PCOS, by allowing for different phenotypes of the PCOS to be diagnosed. Notwithstanding the universal adoption of the Rotterdam criteria as the benchmark for PCOS diagnosis, this framework encompasses inherent limitations that warrant critical consideration. For example, PCOM is present in up to 25% of women with normal reproductive function, particularly younger women, thereby reducing its specificity for PCOS. Moreover, the original antral follicle count threshold was established using older ultrasound technology, and therefore its validity is nowadays challenged. Additionally, variability in imaging techniques and operator expertise may additionally undermine diagnostic accuracy. Moreover, the Rotterdam criteria do not account for metabolic status, body mass index, insulin resistance, overweight, and obesity, common features in PCOS. As a result, the current diagnostic process remains prone to underdiagnosis, delayed diagnosis, and overdiagnosis. These limitations highlight the need to refine diagnostic and prognostic criteria to improve both sensitivity and specificity, thereby enabling more accurate diagnosis and targeted management. On the other hand, accumulating evidence indicates that AMH represents a promising diagnostic and prognostic marker for PCOS, with potential to enhance and refine the Rotterdam diagnostic criteria.52–56 Therefore, both renown professional societies 6 and prominent individuals 52 call for refinement of diagnostic criteria for PCOS, suggesting inclusion of AMH levels. Moreover, AMH levels are associated with metabolic parameters, BMI and insulin resistance. In women with PCOS, AMH positively correlates with metabolic markers, fasting insulin and HOMA-IR and negatively correlates with BMI. 57 Consistent with the findings of Vale-Fernandes et al., 58 who reported that both PCOS and excess body weight independently and synergistically affect hormonal and metabolic profiles but did not significantly impact fertilization or blastocyst formation rates, we also observed that body weight in our cohort did not show significant associations with any of the reproductive outcomes or with EOR in univariate analyses. This suggests that, while metabolic factors remain relevant in PCOS, variations in body weight alone did not influence ART outcomes in our population. Taking all these factors into account, we acknowledge the lack of an in-depth analysis of AMH and BMI and its influence on IVF outcomes in different EOR groups as a limitation of this study. Further studies are needed to evaluate this issue in order to improve management strategies for ovarian stimulation in a challenging population of women with PCOS. Moreover, forthcoming prospective studies should aim to stratify EOR groups by phenotype in an effort to refine predictive models.
Limitations
Our study is the first to evaluate the significance of EOR in the population of women with PCOS. In addition, it is the first study addressing this issue that conducted a comprehensive multivariable logistic regression analysis to evaluate the relationship between the dependent and independent variables. These are the strengths of our study, demonstrating our thorough approach. However, we also acknowledge several limitations of the study. Firstly, we compared our results regarding EOR in women with PCOS with the literature data, which assessed the significance of EOR in other populations undergoing ART (the general population of women and the population of women with other causes of infertility). Furthermore, the retrospective nature of the study design did not allow the assessment of PCOS phenotypes (A, B, C, and D). Although all PCOS phenotypes had comparable E2 levels, 59 different PCOS phenotypes have diverse effects on serum LH levels 59 and ART outcomes. 60 Moreover, various study groups regarding LH activity were dosed differently. While the group stimulated without LH was dosed using an individualized approach (considering body weight and AMH value for each patient), groups stimulated with LH (rec-LH or LH-like activity) comprised women undergoing standard dosing of gonadotropins. Finally, we acknowledge that a formal power analysis was not performed prior to recruitment, which represents a limitation of the study.
Conclusions
Our study, to the best of our knowledge, is the first to demonstrate the significance of EOR in determining ART outcomes, specifically in women with PCOS undergoing ART. While previous studies have explored this topic in the general population, our findings highlight that PCOS, as an endocrine-related cause of infertility, requires different optimal EOR values for the best ART outcomes. Beyond confirming our hypothesis, our results provide practical insights that can shape future research and clinical practice. We also acknowledged several limitations. Hence, prospective validation in larger, well-designed cohorts is essential before routine clinical implementation. Future studies should focus on confirming these findings in prospective settings, comparing the predictive value of EOR across different infertility etiologies, and evaluating stimulation protocols with and without LH supplementation. Such investigations may facilitate the integration of EOR into evidence-based clinical algorithms and ultimately contribute to the development of more precise ART guidelines for women with PCOS.
Footnotes
ORCID iDs
Ethical considerations
The study was approved by the Ethical Committee of Clinic for Gynecology and Obstetrics “Narodni Front” (Approval number 2208/2024/006989 on the 3rd April 2024).
Consent to participate
The requirement for written informed consent to participate were waived due to the retrospective design and the use of de-identified data.
Consent for publication
The requirement for written informed consent for publication were waived due to the retrospective design and the use of de-identified data.
Author contributions
Bojana Salovic: Conceptualization, Data curation, Formal analysis, Investigation, Writing – original draft, Visualization; Ivan Soldatovic: Data curation, Formal analysis, Methodology, Validation; Aleksandar Stojsavljevic: Methodology, Writing – review and editing; Svetlana Dragojevic Dikic: Supervision, Project administration, Writing – review and editing; Ana Nikolic: Writing – original draft, Investigation; Vladimir Gerginic: Writing – original draft, Investigation; Sonja Zafirovic: Supervision, Methodology, Writing – review and editing; Esma R. Isenovic: Supervision, Project administration, Writing – review and editing, Funding acquisition; Milan Perovic: Conceptualization, Supervision, Project administration, Writing – review and editing.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was partially funded by the Ministry of Science and Technological Development and Innovation of the Republic of Serbia (Contract No#451-03-136/2025-03/200017) under the Research Theme “Hormonal regulation of expression and activity of nitric oxide synthase and sodium-potassium pump in experimental models of insulin resistance, diabetes, and cardiovascular disorders” (No.0802501 to ERI).
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The dataset generated and analyzed during the current study is available from the corresponding author on reasonable request.
Artificial intelligence policy
The authors declare that no generative artificial intelligence, including tools such as ChatGPT, was used in the creation of the manuscript, including text, references, figures, or any other content. The authors take full responsibility for the accuracy, originality, and integrity of the work.