Introduction
Diminished ovarian reserve (DOR) refers to a reduction in the number and quality of follicles in the ovaries, which significantly affects fertility [1]. DOR results in poor-quality oocytes, fewer oocytes retrieved after ovulation induction, and lower clinical pregnancy rates, posing a challenge to fertility treatment [2]. The global prevalence of DOR has ranged from 10% to 26% in various studies [3,4]. Asian countries have a relatively high prevalence of 37.2%–38.4% [5], while in Pakistan, 22% of women had DOR, according to a 2020 report [6]. DOR is diagnosed through hormonal markers such as elevated follicle-stimulating hormone (FSH) and reduced anti-Müllerian hormone (AMH) levels, along with a diminished antral follicle count (AFC) [5,7].
As DOR has become a major research topic in reproductive endocrinology, various treatment modalities have been explored, including the use of growth hormone (GH) [8]. Assisted reproductive technology (ART) is an effective method for treating infertility in patients with DOR [9]. In 1988, GH was first used in the induction of ovulation to enhance ovarian sensitivity to gonadotropins, ultimately increasing the number of oocytes retrieved [10]. Previous studies also suggest that GH may increase the ovarian follicular pool by improving follicular development and enhancing sensitivity to gonadotropins, leading to better oocyte retrieval. However, the findings vary considerably, particularly regarding the timing and duration of GH administration. These variations may be attributable to non-personalized GH pretreatment strategies [11-13]. Furthermore, concerns about the metabolic side effects of GH, such as insulin resistance, remain a significant consideration in clinical practice, particularly for patients with obesity or insulin resistance [6].
Despite these promising observations, the current literature does not provide conclusive evidence on the optimal duration and timing of GH supplementation in patients with DOR. Consensus is also lacking regarding the personalization of GH treatment based on individual response markers. As such, existing research does not adequately address how early hormonal and metabolic responses can guide the continuation or modification of GH treatment protocols. This knowledge gap presents a critical barrier to refining treatment strategies for DOR, making it difficult for clinicians to determine the most effective GH protocol based on individual needs.
The primary aim of this study is to investigate the impact of different durations of GH pretreatment on ART outcomes in women with DOR, focusing on oocyte retrieval, embryo quality, and metabolic effects. By adopting a dynamic stratification approach based on early hormonal and metabolic responses, this study seeks to personalize the GH treatment protocol, providing real-time adjustments to optimize ART success for individual patients. We aim to address key gaps in the current understanding of GH treatment by: (1) comparing different durations of GH pretreatment (1 month, 2 months, and during ovulation induction); (2) incorporating metabolic markers such as homeostatic model assessment of insulin resistance (HOMA-IR) to evaluate the impact of GH on insulin sensitivity; and (3) developing a nomogram to predict the likelihood of oocyte retrieval success based on baseline hormonal and metabolic profiles.
Methods
We designed a prospective cohort study to evaluate the effects of different durations of GH pretreatment on ART outcomes in patients with DOR, in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [14]. Participants were monitored in real time from October 2023 to June 2025. After GH treatment, dynamic grouping based on early hormonal and metabolic responses was performed to ensure a more personalized approach to GH treatment outcomes. Early response refers to measurements taken after 1 month of GH pretreatment but before ovulation induction. These included estradiol (E2) levels, ultrasound-assessed follicular count and size, and insulin resistance assessed using HOMA-IR.
The Ethical Review Board of Shahida Islam Medical Complex approved this research (approval no. SIMC/ET.C./0007/25). The study was conducted according to the STROBE guidelines [14]. Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.
1. Sample selection
The sample size was calculated using the World Health Organization sample size calculator, with a 5% margin of error, a 99.999% confidence interval, and an assumed 22% population proportion for infertility [15]. The estimated sample size was 1,311; however, we recruited 2,000 participants to compensate for dropouts and improve the generalizability of the results.
where n=sample size, Z=the Z-score corresponding to the desired confidence interval (for 99.999%, Z=4.417), p=estimated population proportion, and d=margin of error.
Effect size was determined using the Cohen d. A medium effect size of 0.5 was assumed for the primary outcome (number of oocytes retrieved per cycle) based on previous studies of GH pretreatment and ART outcomes [16]. This corresponds to an expected difference of about 0.5 standard deviations between groups, which is generally considered a clinically significant difference in medical studies. With a sample size of 2,000, the study was powered at 99.9% to detect statistical differences between groups.
Inclusion criteria: The study included female patients aged 15–40 years with a history of poor ovarian response (≤3 oocytes retrieved in a previous ART cycle) and DOR, as indicated by FSH >10 IU/L, AMH <1.1 ng/mL, or AFC <7.3 [2,9]. Participants had no history of severe endocrine or autoimmune diseases or other significant medical conditions.
Exclusion criteria: Patients with a history of adverse reactions to GH, such as edema or joint pain, were excluded. Individuals with musculoskeletal disorders or tumors (including pituitary tumors) were also excluded, as GH can exacerbate these conditions and may affect metabolic health. Participants with severe endocrine disorders such as diabetes or thyroid dysfunction, or a known history of uterine malformations or endometrial polyps, were excluded. Patients with prior pregnancy complications (e.g., preeclampsia, preterm birth) or those who had received non-GH pretreatments in the past were also excluded from participation in this study. In addition, patients with a history of ovarian surgery, prior chemotherapy or pelvic radiotherapy, or known genetic syndromes associated with ovarian insufficiency (e.g., fragile X premutation, Turner mosaicism) were excluded to ensure a cohort with functional, non-iatrogenic DOR.
2. Grouping and dynamic stratification
Patients were initially assigned to four primary groups based on the duration and timing of GH pretreatment. This assignment was random, with both clinicians and participants blinded to which group they would belong to after secondary stratification. At this stage, the groups were created primarily to distribute the sample equally into four arms at the time of selection. Group 1 (G1) received 1 month of GH treatment before ovulation induction, followed by ART; group 2 (G2) received 2 months of GH treatment before ovulation induction, followed by ART; group 3 (G3) received GH treatment only during the ovulation induction period, followed by ART; and group 4 (G4, control group) received no GH pretreatment and underwent only the standard ART protocol.
1) Secondary stratification based on baseline hormonal and metabolic profiles
Before initiating GH pretreatment, each patient underwent hormonal and metabolic profile assessment. Based on these profiles, patients were further stratified into subgroups to develop personalized GH treatment options for future candidates (Table 1). These biomarkers (AMH, luteinizing hormone [LH]/FSH ratio, and HOMA-IR) were selected because they are well-established in the clinical literature as significant predictors of ovarian reserve, response to ovarian stimulation [7,17], and insulin sensitivity [18], which are key factors influencing the efficacy of GH in ART. To further personalize GH treatment, anti-androgenic markers (testosterone and dehydroepiandrosterone sulfate [DHEA-S]) and ovarian reserve-related cytokines (inhibin B and interleukin 6 [IL-6]) were measured. Elevated levels of testosterone and DHEA-S are commonly associated with polycystic ovary syndrome (PCOS) and can indicate hyperandrogenism, which negatively impacts oocyte quality and follicular development [19]. Ovarian reserve-related cytokines (inhibin B and IL-6) play important roles in folliculogenesis and ovarian function [8]. Inhibin B is a marker of follicular health, while IL-6 is involved in inflammation and regulation of ovarian function. Elevated levels of IL-6 are often seen in women with ovarian dysfunction. Accordingly, these markers could provide additional insight into the ovarian response to GH and gonadotropin stimulation. Monitoring these cytokines can help determine the personalized effect of GH on ovarian reserve and follicular development. Subgroup stratification for personalized GH treatment was performed on the following basis [7,18,20]. When patients presented with discordant biomarker profiles (e.g., low AMH but moderate HOMA-IR), a rule-based hierarchical stratification system was applied. First, AMH and HOMA-IR were designated as primary indicators of ovarian reserve and metabolic status, respectively. If AMH and HOMA-IR both fell within the same stratification level (low, moderate, or high), that level was assigned. If they diverged, the patient was conservatively assigned to the more impaired (lower) category to avoid overestimation of GH responsiveness. Secondary markers—LH/FSH ratio, testosterone/DHEA-S, and inhibin B/IL-6—were then reviewed to confirm or modify the provisional classification, particularly in borderline cases.
This conservative classification ensured consistent and reproducible stratification. It also reduced the risk of false positives in identifying ‘moderate’ or ‘high’ responders, increasing the precision of GH personalization (Table 2).
2) Dynamic grouping based on early response to GH pretreatment
After 1 month of GH pretreatment, the initial response to GH was evaluated in the G1, G2, and G3 groups using pre-retrieval markers: serum E2 levels, ultrasound-based follicular growth, and insulin resistance (HOMA-IR). These clinically validated parameters served as early response indicators for dynamic stratification. Direct assessment of oocyte and embryo quality was not performed at this stage. Oocyte and embryo parameters were recorded only after oocyte retrieval and fertilization and were later correlated retrospectively with early hormonal and metabolic changes to evaluate their predictive value. Based on these parameters, participants from G1 were categorized as good responders or poor responders. Patients meeting at least two of the three criteria—a significant increase in E2 levels (≥15%), follicular development (≥3 follicles ≥10 mm), and improvement in insulin resistance (a reduction of ≥0.5 from baseline or an absolute value <2.5)—were classified as good responders; those not meeting these criteria were classified as poor responders and reassigned accordingly. Poor responders were then switched to a longer duration of GH pretreatment (from G1 to G2), while good responders continued with their current protocol. This dynamic reassignment ensured that each patient received the optimal duration of GH pretreatment based on early hormonal and metabolic feedback.
The criteria for early response—E2 levels, follicular growth, and insulin resistance—were chosen based on their established role in ovarian response during ART. E2 levels are a direct measure of follicular maturation and oocyte quality, and higher E2 levels are generally correlated with better ovarian stimulation. Follicular growth tracked by ultrasound is a reliable indicator of ovarian response and maturation, while insulin resistance (HOMA-IR) is crucial for assessing metabolic dysfunction, which can negatively impact ovarian response. These markers are clinically validated to predict long-term ART success. Elevated E2, adequate follicular growth, and improved insulin resistance suggest a positive response to GH treatment.
Although participants were prospectively assigned to fixed GH duration groups, a pre-specified adaptive reassignment protocol was employed after 1 month of GH treatment. Based on early endocrine and metabolic response (E2 levels, follicular development, and HOMA-IR), participants were reassigned to optimize GH exposure before ART. This biomarker-driven reassignment occurred before ovarian stimulation or oocyte retrieval, in alignment with adaptive clinical trial principles. Both intention-to-treat and per-protocol analyses were conducted to preserve statistical and methodological integrity.
3. Data collection
After written informed consent was obtained, patients’ demographic characteristics, body mass index, and medical history were recorded using a study-designed questionnaire. Regular blood samples were collected at specific time points during the ART cycle to monitor key hormonal and metabolic parameters. At the start of the study, all participants underwent baseline testing (5 mL of venous blood) for AMH, FSH, LH, testosterone, insulin resistance (HOMA-IR), and a lipid profile (triglycerides, low-density lipoprotein, high-density lipoprotein) using commercially available enzyme-linked immunosorbent assay (ELISA) kits (Supplementary Table 1). These baseline measurements established each patient’s hormonal and metabolic profile before GH pretreatment began. During the ovulation induction phase, 5-mL blood samples were taken again on day 3 of the cycle to measure FSH, LH, E2, testosterone, and HOMA-IR to assess ovarian response to gonadotropin stimulation. Additionally, blood samples were collected on the day of human chorionic gonadotropin (hCG) injection (after follicular development), focusing on E2 and testosterone levels to monitor maturation and readiness for oocyte retrieval. Continuous glucose monitoring (CGM) was performed in selected patients who had high insulin resistance or PCOS, as they were more likely to benefit from continuous tracking of glucose fluctuations throughout the ART cycle. CGM provided real-time data on insulin sensitivity, which is a critical factor affecting oocyte quality and embryo development. This personalized approach ensured that CGM was used only in patients who needed it most, optimizing insulin resistance management and offering a deeper understanding of its influence on ART outcomes. Patients were monitored throughout the ART cycle and postpartum to track both maternal and fetal outcomes. Follow-up visits were planned for immediately after delivery to assess maternal complications and 1 month post-delivery for continued maternal health monitoring.
1) GH supplementation protocol
Recombinant human growth hormone (rhGH) in the form of commercially available Norditropin (Novo Nordisk) was used in the current study (Novo Nordisk; active ingredient: somatropin, which is biosynthetically produced and is identical to naturally occurring human GH in its amino acid sequence). A dose of 2 IU/day, equivalent to 0.67 mg/day of somatropin, was administered across all GH pretreatment groups (G1, G2, and G3) for their respective treatment durations. For the G3 group, 2 IU/day was administered starting at the beginning of the ovulation induction phase of the ART cycle and continued until the hCG injection day (36 hours before oocyte retrieval). GH injections were administered subcutaneously into the abdominal or thigh region (according to patient preference), as this is the most common method for rhGH supplementation. Patients were trained to self-administer rhGH using pre-filled pens, which make the process convenient and precise. They were also instructed to rotate injection sites to minimize irritation and discomfort. The pre-filled pens eliminate the need for reconstitution, ensuring ease of use and accurate dosing. The timing of daily injections was standardized in the evening to align with the natural GH secretion pattern. For the G1 and G2 groups, GH pretreatment began 1 and 2 months before the ART cycle, respectively. For the G3 group, GH supplementation started on day 3 of the menstrual cycle during gonadotropin stimulation and continued until hCG injection.
2) ART protocol and oocyte quality assessment
After the GH pretreatment period, patients underwent a standard ART cycle. The ART protocol involved ovarian stimulation, oocyte retrieval, fertilization, embryo culture, and embryo transfer. Ovarian stimulation was performed using FSH, LH, and hCG. The gonadotropin dosage was individualized based on each patient’s response to GH pretreatment. FSH was administered to stimulate follicular development, while LH was provided in combination with FSH to support follicular maturation. Once the follicles reached the desired size (typically 18 to 20 mm in diameter), hCG was used to trigger ovulation. Oocyte retrieval was performed 36 hours after hCG injection using ultrasound-guided aspiration. During the ART cycle, the number of mature oocytes retrieved was recorded, and embryos were cultured in G1/G2 media under controlled conditions (37 °C, 5% CO2). Embryos were then assessed for cleavage and blastocyst formation on day 3 and day 5, respectively.
The embryo quality grading system was used to assess both day 3 embryos and day 5 blastocysts. For day 3 embryos, a grading system classified embryos based on cell division, fragmentation, and morphology. Grade 1 embryos displayed uniform cell division with no fragmentation, indicating optimal embryo health. Grade 2 embryos had minor fragmentation, while grade 3 embryos exhibited noticeable fragmentation and asymmetric blastomeres. Grade 4 embryos, with extensive fragmentation and uneven division, were considered poor quality. For day 5 blastocysts, embryos were graded based on inner cell mass (ICM) quality, trophoblast development, and degree of expansion. Grade 1 blastocysts had a well-formed ICM and outer trophoblast, indicating the best potential for implantation. Grade 2 blastocysts had slightly less distinct ICM structure but remained viable for transfer, whereas grade 3 and grade 4 blastocysts were of suboptimal quality and were less likely to result in successful implantation.
Oocyte quality was assessed using confocal microscopy and fluorescent imaging techniques. Mitochondrial integrity was evaluated using MitoTracker (Thermo Fisher Scientific) to assess the oocytes’ energy function, which is critical for successful fertilization and development. Spindle morphology was examined under confocal microscopy to identify abnormalities in chromosome alignment during meiosis. Oocyte health was also evaluated by visual grading based on cytoplasmic consistency and the presence of polar bodies. Additionally, to provide an objective assessment of oocyte quality, morphokinetic analysis was incorporated. Using time-lapse imaging systems such as the EmbryoScope (Vitrolife), the behavior of the oocyte and its chromosomes was tracked during the early stages of development. This contributed an additional layer of precision in predicting fertilization rates and implantation potential.
4. Data analysis
Statistical analyses were performed using SPSS ver. 26.0 (IBM Corp.) and R ver. 4.2.1 (R Foundation for Statistical Computing) for advanced modeling and nomogram construction. Normality was evaluated using the Shapiro–Wilk test, and homogeneity of variance was assessed using the Levene test. Descriptive statistics were expressed as mean±standard deviation for continuous variables. Baseline and clinical characteristics across the four primary groups (G1, G2, G3, and G4) were compared using one-way analysis of variance (ANOVA) for normally distributed continuous variables, with Bonferroni post hoc corrections applied for pairwise comparisons. For subgroup comparisons between good and poor responders, independent-samples t-tests were used.
A multivariate linear regression model was applied to identify predictors of oocyte retrieval. Variable selection was conducted using the backward stepwise elimination method based on the Akaike information criterion. Assumptions of linearity, multicollinearity (using a variance inflation factor <5), and residual normality were verified. The following variables were tested in the model: AMH, GH dose, HOMA-IR, inhibin B, IL-6, testosterone, and DHEA-S. Final model significance was confirmed using ANOVA for nested models, and adjusted R2 was reported to indicate explained variance. A clinical nomogram was developed in R using the rms package, based on the final multivariate regression model. Predictor coefficients (AMH, GH dose, inhibin B, and testosterone) were weighted and transformed into individual score components. The nomogram estimates the probability of a high oocyte yield per patient, facilitating individualized GH dosing decisions. Internal validation was performed using bootstrap resampling (1,000 iterations), and calibration plots were generated to compare predicted versus observed values. The C-index (concordance index) was calculated to assess model discrimination [21].
Results
The current study included demographic-matched research participants and baseline characteristics were compared across study groups (Table 2 and 3).
Poor responders were reassigned to the G2 group for an additional month of GH pretreatment (Table 4). At the end of this period, the data were compared among the original G1 group, the new G2 group, and the control group. The results indicate that GH pretreatment significantly improves E2 levels, follicular growth, oocyte retrieval, and embryo quality, with the 2-month GH pretreatment group (G2) showing the best outcomes, but with an associated increase in HOMA-IR (insulin resistance) (Table 5, Figure 1).
Multivariate regression analysis was conducted using the backward elimination method. The results revealed that GH dose and AMH were the strongest predictors of oocyte retrieval. Inhibin B and testosterone also played significant roles, with testosterone having a negative impact. Variables such as IL-6 and DHEA-S had weaker effects on oocyte retrieval and may not need to be prioritized in clinical decision-making (Table 6).
Based on the strongest predictors of oocyte retrieval, a nomogram was constructed to help clinicians estimate ART outcomes from the baseline profile and to guide selection of the appropriate GH pretreatment duration (Figure 2).
Discussion
The results of the present study demonstrate that GH pretreatment significantly enhances ART outcomes, including E2 levels, follicular growth, oocyte retrieval, and embryo quality. Among the treatment protocols, the G2 pretreatment group exhibited the most favorable outcomes, particularly in E2 levels, oocyte retrieval, and good-quality embryo formation, aligning with previous findings on the positive effects of longer GH exposure. However, this group also displayed an increase in insulin resistance (HOMA-IR), highlighting a potential association between favorable reproductive parameters and increased insulin resistance.
Our findings are consistent with prior studies, which have indicated that GH supplementation improves ovarian response and embryo quality. One study demonstrated that GH pretreatment increases ovarian responsiveness and improves oocyte quality in women with poor ovarian reserve [22]. Another reported that GH supplementation resulted in higher E2 levels and better oocyte retrieval rates in poor responders [23]. The positive effects of GH treatment observed in the present study further corroborate these findings.
However, a notable divergence from the previous literature is the increased HOMA-IR observed in the G2 group. In our study, the increase in insulin resistance was more pronounced in the 2-month group compared to 1-month GH (G1) and GH during ovulation induction. This finding aligns with studies that have reported a metabolic burden associated with prolonged GH use, particularly in populations with obesity or insulin resistance [24]. The potential adverse effects on insulin sensitivity due to prolonged GH exposure have not been consistently highlighted in other studies, suggesting a need for further investigation into the long-term metabolic effects of GH treatment in ART cycles. Research has suggested that while GH supplementation enhances fertility outcomes, its effect on insulin resistance can complicate the management of patients with metabolic disorders [13,25].
In contrast, the G1 and GH during ovulation induction (G3) groups displayed improved ovarian response and embryo quality, although these effects were less robust than in the G2 group. This finding suggests that short-term GH treatment still provides meaningful benefits for oocyte retrieval and embryo development. Another study observed that short-term GH enhanced follicular development in women with DOR [11]. However, our study indicates that the 2-month protocol results in superior outcomes, suggesting a potential dose-response relationship between GH duration and ART success. The superior outcomes observed in the G2 pretreatment group may be explained by cumulative insulin-like growth factor 1-mediated sensitization of granulosa cells, which enhances FSH receptor expression and follicular responsiveness over time. Extended GH exposure may also help overcome early follicular arrest and support sustained folliculogenesis. Importantly, the 2-month duration appears to strike a balance: long enough to potentiate ovarian responsiveness without reaching the point of GH receptor desensitization, which may occur in more prolonged protocols.
The results of this study indicate that embryos from patients without GH treatment demonstrated the slowest development and the least favorable embryo quality. This corroborates the existing literature, in which the absence of GH pretreatment is associated with poorer ART outcomes [26]. In contrast, the GH-treated groups consistently outperformed the control group in terms of oocyte retrieval and embryo development, supporting a beneficial role of GH in enhancing ovarian function and embryo viability. This observation is similar to findings by Cozzolino [27], who reported that GH supplementation significantly improved embryo quality compared with non-GH–treated cycles in poor responders.
The multivariate regression analysis in the present study identified GH dose and AMH as the most significant predictors of oocyte retrieval success, consistent with studies highlighting the importance of ovarian reserve markers such as AMH in predicting ART outcomes. Additionally, testosterone was found to have a negative effect on oocyte retrieval, which aligns with previous research suggesting that high testosterone levels, often seen in conditions like PCOS, may hinder oocyte maturation and embryo quality. These results are supported by previous studies identifying testosterone as a negative predictor of ART outcomes, particularly in patients with PCOS [19].
The nomogram developed in this study offers a useful tool for clinicians to predict oocyte retrieval outcomes based on baseline variables, particularly GH dose, AMH, HOMA-IR, and testosterone. This predictive tool may aid in personalizing ART plans by highlighting associations of baseline metabolic and hormonal profiles with ART outcomes. While promising, its clinical application should be considered hypothesis-generating rather than definitive.
GH pretreatment significantly improves ART outcomes, particularly regarding E2 levels, follicular growth, oocyte retrieval, and embryo quality. The G2 group displayed the best outcomes, including higher E2 levels, greater follicular development, and increased oocyte retrieval, but this benefit was associated with an increase in insulin resistance (HOMA-IR). G1 pretreatment and GH administration at ovulation induction (G3) also led to improved outcomes compared to the control group (G4); however, G2 consistently outperformed the others. Given the observed association between prolonged GH use and elevated HOMA-IR, patients undergoing GH pretreatment beyond 4 weeks may benefit from concurrent metabolic monitoring and individualized insulin-sensitizing therapy when indicated. For patients undergoing their first ART cycle, 1 month of GH pretreatment should be considered the initial strategy. The nomogram offers a practical tool for clinicians to predict ART success and determine the optimal GH pretreatment duration based on individual patient profiles.
A potential limitation of this study is the use of dynamic reassignment based on early treatment response, which may introduce bias. However, this was mitigated by applying a pre-specified, objective reassignment algorithm before ART initiation, not one based on outcomes. This approach reflects an adaptive prospective cohort model, which is increasingly relevant in reproductive medicine where early hormonal responses can guide treatment duration. Bias was further controlled through covariate balancing, dual analytic strategies (intention-to-treat and per-protocol), and transparent reporting. In addition, the non-randomized design represents a key limitation, and we recommend conducting randomized controlled trials to further validate and extend these findings.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
