Impact of trigger-day serum luteinizing hormone levels on embryo quality and pregnancy outcomes in overweight and obese women undergoing GnRH antagonist protocols: a retrospective cohort study.

The prevalence of infertility has steadily increased in recent years ( 1 ), posing a significant challenge to family planning and the reproductive health of women. Currently, in vitro fertilization/intracytoplasmic sperm injection-embryo transfer (IVF/ICSI-ET) serves as one of the most effective therapeutic strategies ( 2 ). Among various contributing factors, obesity severely impairs female fertility through multiple mechanisms, including ovulatory dysfunction, oocyte impairment, chronic low-grade inflammation, and compromised endometrial receptivity ( 3 ). Clinical studies demonstrate that overweight or obese patients generally exhibit a diminished response to ovarian stimulation. Despite receiving higher doses and prolonged durations of gonadotropins, their trigger-day serum estradiol levels, oocyte yield, high-quality embryo rate, and ultimate live birth rates remain significantly lower compared to normal-weight counterparts ( 4 – 6 ). Notably, overweight status is frequently associated with hypothalamic-pituitary-ovarian (HPO) axis dysfunction, predisposing these individuals to dysregulated luteinizing hormone (LH) secretion rhythms. Elucidating the regulatory mechanism of trigger-day serum LH levels on IVF outcomes under overweight conditions will facilitate the optimization of clinical management protocols. LH plays a central role in folliculogenesis, oocyte maturation, and the maintenance of endometrial receptivity ( 7 ). During IVF cycles, LH must be maintained within a specific “LH window” to provide optimal support for follicular development ( 8 ). Suboptimal LH levels impair theca cell function and steroidogenesis, leading to decreased fertilization rates and an elevated risk of embryo aneuploidy ( 9 , 10 ). Conversely, a premature LH surge during ovarian stimulation triggers premature luteinization, severely compromising oocyte quality and endometrial receptivity ( 11 , 12 ). Currently, clinical strategies primarily rely on gonadotropin-releasing hormone (GnRH) antagonists to prevent premature LH surges, or exogenous LH supplementation to correct LH deficiency, aiming to stabilize LH concentrations within this optimal window ( 13 ). The correlation between serum LH levels and IVF outcomes exhibits significant population heterogeneity. For instance, higher LH levels under the GnRH antagonist protocol may benefit patients with normal ovarian reserve or polycystic ovary syndrome (PCOS) ( 14 ), whereas elevated trigger-day LH predicts adverse pregnancy outcomes in patients with poor ovarian reserve (POR) ( 15 ). However, studies investigating the association between trigger-day LH levels and clinical outcomes in overweight patients remain scarce. Recent evidence suggests that obesity alters the metabolic response of granulosa cells to the preovulatory LH surge, thereby increasing the risk of fertilization failure and poor blastocyst development ( 16 ). Therefore, this study aimed to explore the specific effects of trigger-day LH levels on embryo quality and pregnancy outcomes in overweight and obese women.

A total of 7,269 clinical cycles from infertile women were initially screened in this study. To ensure the statistical independence of observations, inclusion was strictly limited to the first controlled ovarian stimulation (COS) cycle of each patient. Ultimately, the clinical data of 1,135 overweight and obese women (representing exactly 1,135 independent first COS cycles) were included in the statistical analysis. The inclusion and exclusion criteria of the cases are shown in Figure 1 . According to the serum LH level on the hCG trigger day, the subjects were divided into three groups: Group 1 (LH 4.19 IU/L, n = 283). There were no statistically significant differences among the three groups in terms of female age, male age, overweight/obesity (including overweight and obese proportions), years of infertility, primary infertility, infertility factors (fallopian tube factors and male factor), basal endocrine indicators (FSH, LH, E2, P, T), AFC, BMI, ovulation trigger protocol (hCG, triptorelin, dual trigger), IVF/ICSI, E2 on trigger day level, P on trigger day level, number of high-quality embryos transferred, endometrial thickness on transfer day, number of embryos transferred, clinical pregnancy rate, and live birth rate (all P > 0.05). Statistically significant differences were observed among the three groups in the proportion of PCOS patients, Gn duration (days), total Gn dosage, number of follicles ≥14 mm on the trigger day, number of retrieved oocytes, number of mature oocytes (MII), number of fertilized oocytes, number of normally fertilized oocytes (2PN), number of available embryos, and number of high-quality embryos (all P < 0.05), as detailed in Tables 2 , 3 . We noted that the median number of high-quality embryos was identical (1 [0–2]) across groups but showed a significant difference by Kruskal-Wallis H test. This discrepancy was due to distinct frequency distributions (0, 1, ≥2 embryos) and rank-based differences rather than median values. Detailed distribution patterns were analyzed using frequency histograms and box plots, as shown in Supplementary Figure 1 ; Supplementary Table 1 . Inclusion and exclusion criteria of the cases. Baseline characteristics of overweight and obese women stratified by LH group. Data are presented as mean ± SD, median (IQR), or n (%). ANOVA, Kruskal–Wallis H test, and chi-square test were used for comparisons.* P < 0.05 vs. Group 1; ** P < 0.01 vs. Group 1; # P < 0.05 vs. Group 2; ## P < 0.01 vs. Group 2. LH, luteinizing hormone; FSH, follicle-stimulating hormone; E2, estradiol; P, progesterone; AFC, antral follicle count; BMI, body mass index; Gn, gonadotropin; ICSI, intracytoplasmic sperm injection; MII, metaphase II; 2PN, two pronuclear. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. Clinical pregnancy outcomes of overweight and obese women stratified by LH group. Group 1: LH 4.19 IU/L (above the 75th percentile). Clinical pregnancy rate (CPR) and Live birth rate (LBR) were calculated based on the outcome of the first embryo transfer (either fresh or frozen-thawed) following the index oocyte retrieval cycle. P values were obtained using the chi-square test. In the subgroup comparison stratified by transfer type (446 fresh transfer cycles and 689 FET cycles derived from the 1,135 initial COS cycles), in the fresh embryo transfer subgroup, no significant differences were observed among the three groups in demographic characteristics (female age, male age), BMI, overweight/obese proportions, infertility characteristics (years of infertility, primary infertility, infertility factors), PCOS, basal endocrine indicators (FSH, LH, E2, P, T), AFC, Gn duration (days), total Gn dosage, IVF/ICSI, E2 on trigger day and P levels, transfer-related parameters (endometrial thickness, number of embryos transferred, number of high-quality embryos transferred), and pregnancy outcomes (clinical pregnancy rate, ectopic gestation rate, early miscarriage rate, live birth rate) (all P > 0.05). However, significant differences were observed in the number of follicles ≥14 mm on the trigger day, number of oocytes retrieved, number of mature oocytes, number of fertilized oocytes, number of normally fertilized oocytes (2PN), number of available embryos, and number of high-quality embryos (all P < 0.05), as detailed in Table 4 . Baseline characteristics of fresh embryo transfer cycles. A1, LH 4.19 IU/L. Normally distributed continuous data are presented as mean ± SD, and non-normally distributed data as median (P25, P75). * P < 0.05 vs. A1; ** P < 0.01 vs. A1; # P < 0.05 vs. A2; ## P < 0.01 vs. A2. Clinical pregnancy and live birth rates are calculated per embryo transfer cycle. Rates for ectopic gestation and early miscarriage were calculated per clinical pregnancy. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. In the FET subgroup, no significant differences were observed among the three groups in demographic characteristics (female age, male age), BMI, overweight/obese proportions, infertility characteristics (years of infertility, primary infertility, infertility factors), basal endocrine indicators (FSH, LH, E2, P, T), AFC, Gn duration (days), total Gn dosage, IVF/ICSI, E2 on trigger day and P levels, transfer-related parameters (endometrial thickness, number of embryos transferred, number of high-quality embryos transferred), and pregnancy outcomes (clinical pregnancy rate, ectopic gestation rate, early miscarriage rate, live birth rate) (all P > 0.05). However, significant differences were observed in the proportion of PCOS, number of follicles ≥14 mm on the trigger day, number of oocytes retrieved, number of mature oocytes, number of fertilized oocytes, and Number of normally fertilized oocytes (2PN) (all P < 0.05), as detailed in Table 5 . Baseline characteristics of frozen-thawed embryo transfer cycles. B1, LH 4.19 IU/L. Normally distributed continuous data are presented as mean ± SD, and non-normally distributed data as median (P25, P75). * P < 0.05 compared with B1; ** P < 0.01 compared with B1; # P < 0.05 compared with B2; ## P < 0.01 compared with B2.Clinical pregnancy and live birth rates are calculated per embryo transfer cycle. Rates for ectopic gestation and early miscarriage were calculated per clinical pregnancy. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. In the PCOS subgroup, there were no significant differences among the three groups in terms of demographic characteristics, baseline endocrine indicators, transfer-related parameters, and pregnancy outcomes (all P > 0.05); significant differences were observed in the number of follicles ≥14 mm on the trigger day, number of oocytes retrieved, number of mature oocytes, number of fertilized oocytes, number of normally fertilized oocytes (2PN), and number of transferable embryos (all P < 0.05), as shown in Table 6 . Baseline characteristics of PCOS patients in each LH group. C1, LH 4.19 IU/L. Normally distributed continuous data are presented as mean ± SD, and non-normally distributed data as median (P25, P75). * P < 0.05 compared with C1; ** P < 0.01 compared with C1; # P < 0.05 compared with C2; ## P < 0.01 compared with C2.Clinical pregnancy and live birth rates are calculated per embryo transfer cycle. Rates for ectopic gestation and early miscarriage were calculated per clinical pregnancy. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. All patients included in this table completed exactly one embryo transfer cycle (either fresh or frozen). Patients who never underwent any embryo transfer were excluded. In the non-PCOS subgroup, there were no significant differences among the three groups in terms of demographic characteristics, baseline endocrine indicators (except FSH), transfer-related parameters (except endometrial thickness), and pregnancy outcomes (all P > 0.05); significant differences were observed in female ages, years of infertility, basal FSH, Gn administration days, endometrial thickness on transfer day, number of follicles ≥14 mm on the trigger day, number of retrieved oocytes, number of mature oocytes (MII), number of fertilized oocytes, number of normally fertilized oocytes (2PN), number of available embryos, and number of high-quality embryos (all P < 0.05), as shown in Table 7 . Baseline characteristics of non-PCOS patients in each LH group. D1, LH 4.19 IU/L. Normally distributed continuous data are presented as mean ± SD, and non-normally distributed data as median (P25, P75). * P < 0.05 compared with D1; ** P < 0.01 compared with D1; # P < 0.05 compared with D2; ## P < 0.01 compared with D2. Clinical pregnancy and live birth rates are calculated per embryo transfer cycle. Rates for ectopic gestation and early miscarriage were calculated per clinical pregnancy. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. In the overweight subgroup, there were no significant differences among the three groups in terms of demographic characteristics, baseline endocrine indicators, transfer-related parameters, and pregnancy outcomes (all P > 0.05); significant differences were observed in the proportion of PCOS, Gn administration days, number of follicles ≥14 mm on the trigger day, number of oocytes retrieved, number of mature oocytes, number of fertilized oocytes, number of normally fertilized oocytes (2PN), number of available embryos, and number of high-quality embryos (all P < 0.05), as shown in Table 8 . Baseline characteristics of overweight patients (BMI 25.0-29.9 kg/m 2 ) in each LH group. E1, LH 4.19 IU/L. Normally distributed continuous data are presented as mean ± SD, and non-normally distributed data as median (P25, P75). * P < 0.05 compared with E1; ** P < 0.01 compared with E1; # P < 0.05 compared with E2; ## P < 0.01 compared with E2. Clinical pregnancy and live birth rates are calculated per embryo transfer cycle. Rates for ectopic gestation and early miscarriage were calculated per clinical pregnancy. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. According to the World Health Organization (WHO) criteria, this table specifically presents data for the overweight subgroup (BMI 25.0–29.9 kg/m 2 ). In the obese subgroup, there were no significant differences among the three groups in terms of demographic characteristics, baseline endocrine indicators, transfer-related parameters, and pregnancy outcomes (all P > 0.05); a significant difference was observed only in total Gn dosage ( P < 0.05), as shown in Table 9 . Baseline characteristics of obese patients (BMI ≥ 30.0 kg/m 2 ) in each LH group. F1, LH 4.19 IU/L. Normally distributed continuous data are presented as mean ± SD, and non-normally distributed data as median (P25, P75). * P < 0.05 compared with F1; ** P < 0.01 compared with F1; # P < 0.05 compared with F2; ## P < 0.01 compared with F2. Clinical pregnancy and live birth rates are calculated per embryo transfer cycle. Rates for ectopic gestation and early miscarriage were calculated per clinical pregnancy. An asterisk (*) in the overall P- value column indicates statistical significance ( P < 0.05) among the three groups. According to the World Health Organization (WHO) criteria, this table specifically presents data for the obese subgroup (BMI ≥ 30.0 kg/m 2 ). A dual-model binary logistic regression analysis was performed to evaluate the independent effect of trigger-day LH levels on the clinical pregnancy outcome of the first transfer. The low LH group (Group 1) served as the reference group. Based on the fully adjusted model (Model 2), which accounted for baseline, clinical, and cycle-specific embryo transfer characteristics, the results in the total population showed no statistically significant differences in the clinical pregnancy rate between the medium LH group (Group 2: aOR = 0.945, 95% CI: 0.649–1.376, P = 0.767), the high LH group (Group 3: aOR = 1.130, 95% CI: 0.820–1.557, P = 0.457), and the low LH group. This suggests that after comprehensively adjusting for confounders, trigger-day LH levels are not independently associated with clinical pregnancy outcomes in overweight and obese women ( Table 10 ). Stratified analyses by transfer type yielded consistent results. In the fresh embryo transfer subgroup, no significant differences were observed compared to the reference group (A2: aOR = 0.658, 95% CI: 0.356–1.218, P = 0.183; A3: aOR = 0.829, 95% CI: 0.498–1.379, P = 0.470; Table 11 ). Similarly, no significant independent associations were found in the frozen-thawed embryo transfer (FET) subgroup (B2: aOR = 1.225, 95% CI: 0.746–2.010, P = 0.422; B3: aOR = 1.480, 95% CI: 0.965–2.271, P = 0.072; Table 12 ). Further stratified analyses by pathological type and BMI category reinforced these findings. No statistically significant differences in clinical pregnancy rates were detected across LH groups in the PCOS subgroup (C2: aOR = 0.755, P = 0.445; C3: aOR = 0.930, P = 0.799; Table 13 ), the non-PCOS subgroup (D2: aOR = 1.056, P = 0.814; D3: aOR = 1.209, P = 0.353; Table 14 ), the overweight subgroup (E2: aOR = 0.923, P = 0.702; E3: aOR = 0.988, P = 0.947; Table 15 ), or the obese subgroup (F2: aOR = 0.786, P = 0.658; F3: aOR = 1.865, P = 0.148; Table 16 ). Finally, in the sensitivity analysis treating the trigger-day LH strictly as a continuous variable, the fully adjusted multivariable logistic regression model similarly did not demonstrate a statistically significant linear association with clinical pregnancy (aOR = 1.015, 95% CI: 0.958–1.076, P = 0.610; Supplementary Table 3 ). Binary logistic regression analysis of clinical pregnancy for the total population. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; PCOS, polycystic ovary syndrome; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance. Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, BMI, PCOS status, trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. Binary logistic regression analysis of clinical pregnancy for fresh embryo transfer cycles. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; PCOS, polycystic ovary syndrome; hCG, human chorionic gonadotropin; GnRH, gonadotropin-releasing hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance. Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, BMI, PCOS status, trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. In the Fresh ET subgroup, only one patient received GnRH agonist alone trigger; to ensure model stability (as fresh ET is generally avoided after GnRHa-only trigger), this single case was excluded. Consequently, the trigger methods adjusted in this model included only “hCG trigger alone” (reference) and “Dual trigger”. Binary logistic regression analysis of clinical pregnancy for frozen-thawed embryo transfer cycles. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; hCG, human chorionic gonadotropin; GnRH, gonadotropin-releasing hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance.Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, BMI, PCOS status, trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. Binary logistic regression analysis of clinical pregnancy for PCOS patients. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; hCG, human chorionic gonadotropin; GnRH, gonadotropin-releasing hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance.Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, BMI,trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. Binary logistic regression analysis of clinical pregnancy for non-PCOS patients. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; hCG, human chorionic gonadotropin; GnRH, gonadotropin-releasing hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance.Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, BMI, trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. Binary logistic regression analysis of clinical pregnancy for overweight women. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; PCOS, polycystic ovary syndrome; hCG, human chorionic gonadotropin; GnRH, gonadotropin-releasing hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance.Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, PCOS status, trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. Binary logistic regression analysis of clinical pregnancy for obese women. aOR, adjusted odds ratio; CI, confidence interval; BMI, body mass index; PCOS, polycystic ovary syndrome; hCG, human chorionic gonadotropin; GnRH, gonadotropin-releasing hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count.* P < 0.05 indicates statistical significance.Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, PCOS status, trigger method, fertilization method, years of infertility, basal FSH, and AFC. Model 2 was fully adjusted. It included all variables in Model 1, with the further inclusion of embryo transfer characteristics as covariates: endometrial thickness, transfer of high-quality embryo, and total number of transferred embryos. In the obese subgroup, only 3 patients received GnRH agonist alone trigger. To ensure the stability and validity of the multivariable logistic regression model, these 3 cases were excluded from the analysis. Consequently, the trigger methods adjusted in this model only included “hCG trigger alone” (reference) and “Dual trigger”. The study analyzed the correlation between serum LH level on the hCG trigger day and various clinical indicators. Before adjustment for multiple comparisons, serum LH level showed a weak positive correlation with basal FSH (r = 0.064, P = 0.031), and weak negative correlations with the number of follicles ≥14 mm on the trigger day (r = -0.156, P < 0.001), number of oocytes retrieved (r = -0.163, P < 0.001), number of mature oocytes (r = -0.173, P < 0.001), number of fertilized oocytes (r = -0.157, P < 0.001), number of normally fertilized oocytes (2PN) (r = -0.158, P < 0.001), number of transferable embryos (r = -0.106, P < 0.001), and number of high-quality embryos (r = -0.131, P < 0.001). No significant correlations were observed between serum LH level and female age, BMI, infertility duration, basal LH, basal E2, basal P, basal T, AFC, E2 on trigger day, or P on trigger day (all P > 0.05). After applying Bonferroni correction for the 18 variables tested (adjusted α = 0.05/18 = 0.0028), the correlation with basal FSH lost statistical significance (P = 0.031 > 0.0028), suggesting it may be a false positive finding. All correlations related to ovarian response and embryo quality remained statistically significant after correction (all P < 0.001). However, it should be noted that the absolute r s values of these significant correlations ranged from 0.106 to 0.173, explaining less than 3% of the variance (r s 2 < 0.03), indicating weak correlations. Although statistically significant due to the large sample size (n = 1,135), these associations have limited clinical relevance. Thus, with the increase of serum LH level on the hCG trigger day, the indicators related to ovarian reactivity and embryo quality show a weak downward trend, as shown in Table 1 .

Luteinizing hormone (LH) plays a pivotal role in regulating folliculogenesis and final oocyte maturation within a specific “LH window.” Suboptimal LH levels impair follicular energy metabolism and steroidogenesis, whereas abnormally elevated LH triggers premature luteinization and down-regulates LH receptors, ultimately compromising endometrial receptivity ( 8 ). Consequently, trigger-day serum LH levels serve as a critical reflection of the underlying endocrine and metabolic milieu. While some studies report that elevated LH in GnRH antagonist protocols reduces oocyte yield without compromising final pregnancy rates ( 20 – 22 ), others indicate detrimental effects on live birth rates, particularly in vulnerable populations such as women with advanced maternal age ( 23 ) or diminished ovarian reserve ( 15 ).These conflicting findings suggest that the biological consequences of elevated LH are profoundly modulated by individual endocrine and metabolic phenotypes. However, overweight and obese populations—characterized by peripheral insulin resistance and HPOA axis disruption—have been largely overlooked. Given these unique metabolic derangements, extrapolating findings from the general population is insufficient. Therefore, this study independently evaluated the impact of trigger-day LH in overweight and obese women. A structured comparison of key methodological differences across studies is provided in Supplementary Table 4 . Our results demonstrated that elevated trigger-day LH was negatively correlated with oocyte yield and embryo quality. Although statistically significant, the absolute correlation coefficients were weak (explaining <3% of the variance), suggesting that elevated LH is not the sole direct cause of reduced embryo quality. Instead, elevated LH acts as a comprehensive manifestation of a “dual defect” inherent to women with elevated BMI (≥ 25 kg/m 2 ). Neuroendocrinologically, these patients frequently exhibit HPOA dysfunction with accelerated LH pulse frequency. Metabolically, prevalent insulin resistance exacerbates hyperandrogenism, which synergistically disrupts granulosa cell function and weakens follicular responsiveness to gonadotropins ( 24 , 25 ). Collectively, these intertwined pathways—neuroendocrine dysregulation and metabolic insulin resistance—render this overweight demographic inherently more susceptible to erratic LH elevations, which may trigger premature final follicular maturation. Consequently, clinical management often requires higher doses of exogenous gonadotropins to initiate follicular recruitment, inadvertently increasing treatment costs and potential risks such as ovarian hyperstimulation syndrome (OHSS). Notably, our study reveals distinct population heterogeneity regarding the detrimental effects of elevated LH. Subgroup analysis uncovered a key clinical phenomenon: in non-PCOS patients with BMI ≥ 25 kg/m 2 , the high-LH group was significantly associated with advanced maternal age (P = 0.007) and elevated basal FSH levels (P = 0.033). Mechanistically, these data suggest that abnormally high serum LH on the hCG trigger day in this specific cohort is more likely an endocrine manifestation of “early decline in ovarian reserve” or “physiological ovarian aging” ( 26 ).These patients typically exhibit poor ovarian response and require prolonged stimulation, leading to passive LH accumulation ( 27 ). In stark contrast, overweight PCOS patients exhibited a fundamentally different pathway. Stratified by LH, PCOS patients showed no significant differences in age, basal FSH, or antral follicle count (all P > 0.05), yet still exhibited a significant reduction in retrieved oocytes and transferable embryos. This divergence strongly indicates that the embryological decline in overweight PCOS women stems not from diminished ovarian reserve, but from a “synergistic toxicity” between the inherent pathophysiology of PCOS (e.g., endogenous LH hypersecretion) and severe insulin resistance ( 28 ).The sustained high LH exposure driven by this dual endocrine and metabolic disorder directly impairs the local ovarian microenvironment. This excessive LH exposure can induce premature follicular luteinization or compromise granulosa cell function, thereby exerting a direct inhibitory effect on follicular synchrony and oocyte quality, entirely independent of ovarian reserve ( 29 ). Since their introduction into clinical practice in the late 1990s and early 2000s, GnRH antagonists have been widely used to inhibit pituitary activity, thereby preventing premature LH surges and premature ovulation before follicular maturation ( 30 – 32 ).Mechanistically, however, the precise impact of elevated LH on oocyte developmental competence remains a subject of debate. Some studies indicate that inappropriately high LH levels can negatively impact folliculogenesis by inhibiting FSH receptor expression, inducing follicular atresia, and reducing estradiol production ( 33 ). Furthermore, excessive LH may accelerate granulosa cell decline and disrupt the synchrony of oocyte maturation ( 16 ). However, LH receptors (LHR) are predominantly expressed in mural granulosa cells rather than cumulus cells, implying that elevated LH may not directly compromise the oocyte itself ( 34 ). Additionally, the chronic low-grade inflammatory state associated with obesity may weaken the regulatory efficacy of GnRH antagonists ( 35 ). These intertwined cellular mechanisms explain why population characteristics profoundly alter the clinical impact of LH fluctuations. Despite the significant decline in embryological parameters, trigger-day LH was not an independent predictor of cumulative pregnancy or live birth rates. This aligns with findings in general GnRH antagonist cycles ( 36 ), though it contrasts with specific DOR cohorts ( 15 ). This discrepancy suggests that the clinical impact of elevated LH is likely context-dependent, relying heavily on patient-specific physiological characteristics rather than acting uniformly across all populations. Mechanistically, the lack of significant impact on pregnancy outcomes in our cohort may be partially explained by the progesterone profile. Aberrant LH elevations typically compromise pregnancy by triggering premature progesterone (P) rises, which interfere with endometrial receptivity ( 37 ). However, our data showed no significant variance in trigger-day progesterone levels across the three LH groups (P = 0.528), indicating that high LH levels did not induce severe premature luteinization prior to trigger. Collectively, these findings suggest that, at least in this overweight cohort, elevated LH alone is not the primary driver of impaired endometrial receptivity. Instead, elevated LH appears to exert its detrimental effect primarily by reducing the initial oocyte/embryo pool—acting as a mediator rather than a direct endometrial toxin. After adjusting for embryological and endometrial parameters (e.g., total transferable embryos, endometrial thickness) in our regression models, LH remained non-significant. As anticipated, established predictors such as female age, BMI, PCOS status, and basal FSH remained the primary drivers of pregnancy outcomes ( 38 , 39 ). Building on this, trigger-day LH demonstrated limited clinical utility in guiding embryo transfer strategies. Subgroup analyses revealed no significant independent effect on pregnancy outcomes in either fresh or frozen-thawed embryo transfer cycles (P>0.05). This indicates that while elevated LH compromises initial embryonic parameters, this detriment does not translate into differential clinical outcomes across transfer methods. Furthermore, even when considering the distinct endocrine profile inherent to PCOS, elevated trigger-day LH did not emerge as an independent risk factor for pregnancy outcomes in either the PCOS or non-PCOS cohorts. To further explore the heterogeneity of LH effects across different BMI strata, we stratified our analysis by overweight (BMI 25–29.9 kg/m 2 ) and obese (BMI ≥30 kg/m 2 ) categories. Strikingly, while elevated trigger-day LH was associated with significantly reduced oocyte yield and embryo quality in the overweight cohort (all P 0.05). Several mechanistic explanations may account for this intriguing observation. First, as BMI increases beyond 30 kg/m 2 , the reproductive endocrine milieu becomes increasingly dominated by profound insulin resistance, hyperinsulinemia, and chronic low-grade inflammation ( 25 , 40 ).These metabolic disturbances may override or obscure the relatively modest contribution of LH to follicular development ( 3 , 41 ).Second, the expression and function of the LH receptor (LHR) may be downregulated in the ovarian tissue of clinically obese individuals, potentially due to chronic exposure to inflammatory cytokines ( 42 ), thereby rendering granulosa cells less responsive to further LH elevation ( 43 ).Third, the higher prevalence of PCOS in the overweight high-LH group (33.6% vs 21.4%, P = 0.011) may partly drive the observed associations. In contrast, although the prevalence of PCOS in the obese cohort appeared to vary across the low-, medium-, and high-LH groups (27.9%, 35.5%, and 47.9%, respectively), this distribution did not reach statistical significance (P = 0.129), suggesting that different pathophysiological mechanisms operate at these distinct ends of the BMI spectrum. Clinically, these findings imply that while monitoring trigger-day LH may have prognostic value for embryological outcomes in overweight patients (BMI 25–29.9), its utility is limited in frankly obese patients (BMI ≥30), where metabolic factors likely play a more dominant role. Future studies with larger sample sizes in the obese category are warranted to validate these preliminary observations. Regarding the primary outcome, our findings indicated that trigger day LH levels were not independently associated with clinical pregnancy rates. However, this null conclusion must be interpreted in the context of statistical power. Given our sample size distribution (e.g., n=272 in Group 1 and n=283 in Group 3) and an observed baseline pregnancy rate of approximately 50%, a post-hoc minimum detectable effect size (MDE) analysis revealed that our study was adequately powered (80% power, two-sided α=0.05) to detect an absolute difference in pregnancy rates of approximately 11.5% between groups. Therefore, we cannot rule out the possibility of a Type II error for smaller, subtle differences (e.g., <10%). Nevertheless, an absolute difference in pregnancy rates of less than 11.5% is arguably of limited clinical significance in routine practice. Thus, we conclude that trigger day LH is unlikely to have a clinically meaningful independent impact on pregnancy outcomes. However, this study has several limitations. (1) Due to the retrospective and single-center design, causality cannot be definitively established. (2) We relied solely on BMI (≥25 kg/m 2 ) without assessing body composition or metabolic syndrome criteria, meaning various metabolic phenotypes were pooled together despite potential differences in gonadotropin reactivity. (3) Our LH evaluation has methodological constraints: lack of standardized assay information, and sample-derived thresholds (1.45 and 4.19 IU/L) that require external validation. Although sensitivity analysis using LH as a continuous variable yielded consistent null results, we acknowledge that arbitrary trichotomization results in some information loss. Future studies using restricted cubic splines or established clinical thresholds (e.g., 5.0 IU/L) are needed. (4) Long-term neonatal outcomes were not evaluated. (5) We lacked longitudinal weight tracking; the interval between oocyte retrieval and frozen embryo transfers can span months or years, during which weight fluctuations may occur, introducing residual confounding. (6) Using WHO criteria may underestimate obesity in our population (Chinese guidelines define obesity as BMI ≥28 kg/m 2 ), and the median BMI of our cohort exceeded this threshold.

For overweight infertile women, during COS, elevated serum LH level on the hCG trigger day is associated with impaired follicular development and decreased embryo quality, but no significant impact on pregnancy outcomes was observed. The above results suggest that in the GnRH antagonist protocol, the serum LH level on the hCG trigger day may affect embryological outcomes more than the final pregnancy outcomes. Therefore, during clinical controlled ovarian stimulation, it is necessary to comprehensively evaluate the individual characteristics of patients and reasonably adjust the gonadotropin dosage and medication regimen to maintain a relatively stable endocrine environment, thereby helping to obtain more ideal embryo quality.

Clinical data of overweight and obese women who underwent IVF/ICSI-ET for assisted reproduction in the Department of Reproduction and Genetics, Affiliated Hospital of Shandong University of Traditional Chinese Medicine, from July 1, 2023, to July 30, 2024, were collected. The study involved a retrospective analysis. The study protocol was reviewed and approved by the Ethics Committee of the Affiliated Hospital of Shandong University of Traditional Chinese Medicine (Approval No. 2025-033-KY). The Ethics Committee of the Affiliated Hospital of Shandong University of Traditional Chinese Medicine evaluated and approved this study protocol and waived the requirement for informed consent due to the study’s retrospective and anonymous design. This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. The completed STROBE checklist is provided as Supplementary Table 5 . Inclusion criteria: (1) Age between 22 and 40 years; (2) BMI ≥25 kg/m 2 ; (3) Treated with the gonadotropin-releasing hormone antagonist (GnRH-ant) protocol; (4) Complete follow-up data. Exclusion criteria: (1) Patients with congenital uterine anatomical abnormalities (e.g., uterine septum, unicornuate uterus) or untreated intrauterine space-occupying lesions (e.g., intrauterine adhesion, endometrial polyps); (2) Patients with other reproductive system diseases (e.g., endometriosis, adenomyosis, ovarian cancer); (3) Patients with other systemic diseases that are not suitable for pregnancy (e.g., autoimmune diseases, severe cardiovascular and cerebrovascular diseases); (4) Chromosomal abnormalities in either spouse; (5) Patients with other severe systemic endocrine diseases (e.g., Cushing’s syndrome, congenital adrenal hyperplasia, hyperprolactinemia, uncontrolled thyroid dysfunction, etc.). However, PCOS, as a key overweight-related subtype of interest in this study, was retained and included as a variable for subgroup analysis; (6) Incomplete medical records; (7) Multiple oocyte retrieval cycles or multiple embryo transfer cycles from the same patient. To strictly ensure the statistical independence of observations, only the patient’s first controlled ovarian stimulation (COS) cycle and her very first subsequent embryo transfer event (either a fresh embryo transfer or the first frozen-thawed embryo transfer) were included; (8) Patients who met the freeze-all criteria underwent a frozen-thawed embryo transfer as their first transfer and were included in the frozen-thawed embryo transfer (FET) subgroup. Patients who never underwent any embryo transfer (e.g., no viable embryos) were excluded. All patients received treatment with the GnRH-ant protocol. When the target follicles reached 18 mm in diameter, trigger was performed at 21:00 with 4000–10000 IU of human chorionic gonadotropin (hCG, 2000 IU, Zhuhai Livzon) and/or 0.1–0.2 mg of triptorelin acetate injection (Diphereline ® , 0.1 mg, Ferring). According to the serum LH levels on the hCG trigger day, patients were stratified into three groups based on the 25th and 75th percentiles (P25, P75) of the study population: Group 1 (LH 4.19 IU/L, n = 283). The 25th and 75th percentiles were used to define LH thresholds to ensure balanced group sizes. However, these cutoffs are sample-specific and may not be generalizable to other populations. Binary logistic regression analysis was used to explore the correlation between trigger-day serum LH levels and pregnancy outcomes, and multivariate logistic regression analysis was performed to examine the potential impact of known confounding factors on pregnancy outcomes. Meanwhile, Subgroup analyses were stratified by transfer type, PCOS status, and BMI category based on the first embryo transfer cycle per patient. Of the 1,135 included patients (representing 1,135 first COS cycles), 446 underwent a fresh embryo transfer as their first transfer, and 689 underwent a frozen-thawed embryo transfer as their first transfer (representing a freeze-all strategy). Subsequent embryo transfers from the same oocyte retrieval were excluded to ensure statistical independence. These transfer cycles were stratified by transfer type into the fresh embryo transfer subgroup (A1: LH 4.19 IU/L, n = 109; total n = 446 cycles) and the frozen-thawed embryo transfer (FET) subgroup (B1: LH 4.19 IU/L, n = 174; total n = 689 cycles). Stratified by PCOS status into the PCOS subgroup (C1: LH 4.19 IU/L, n = 102) and the non-PCOS subgroup (D1: LH 4.19 IU/L, n = 181). According to WHO criteria ( 17 ), patients were divided into the overweight group (BMI 25.0–29.9 kg/m 2 ) and the obese group (BMI ≥ 30.0 kg/m 2 ). Stratified by BMI category into the overweight subgroup (E1: LH 4.19 IU/L, n = 235) and the obese subgroup (F1: LH 4.19 IU/L, n = 48). Controlled ovarian stimulation was initiated on days 2–3 of the patient’s menstrual cycle with an initial dose of 150–300 U/d of recombinant follicle-stimulating hormone (rFSH) injection (Puregon ® , 600 IU, Merck Sharp & Dohme; or Gonal-F ® , Merck-Serono SA, Switzerland) or urinary FSH (Lishenbao; Zhuhai Lizhu Pharmaceutical Co., Ltd.). The dose and type of Gn were determined based on the female’s age, BMI, basal FSH level, and AFC. Serum endocrine levels and follicular development were monitored every 2–4 days. When transvaginal ultrasound revealed that the dominant follicle reached 12–14 mm in diameter, 0.25 mg/d of ganirelix acetate (ORGALUTRAN ® , 0.25 mg, Merck Sharp & Dohme) was subcutaneously administered until the hCG trigger day. When the target follicle reached 18 mm in diameter, trigger was performed at 21:00 with 4000–10000 IU of human chorionic gonadotropin (hCG, 2000 IU, Zhuhai Livzon Pharmaceutical Group Co., Ltd.) and/or 0.1–0.2 mg of triptorelin acetate injection (Diphereline ® , 0.1 mg, Ferring Pharmaceuticals). Oocyte retrieval was performed under ultrasound guidance 34–36 hours later, followed by IVF or ICSI (in cases of severe oligozoospermia, asthenospermia, or teratozoospermia in males). Embryo transfer was performed 3 days after embryo culture. All embryos were cryopreserved for elective transfer if any of the following conditions were met: (1) Endometrial thickness ≤ 6 mm or ≥ 15 mm on the hCG trigger day, or presence of endometrial polyps or intrauterine fluid; (2) Progesterone level ≥ 2 ng/mL on the hCG trigger day; (3) Estradiol level ≥ 4800 pg/mL on the hCG trigger day; (4) Number of oocytes retrieved ≥ 15, suggesting a tendency for ovarian hyperstimulation syndrome (OHSS). Progesterone injection was administered for luteal support after embryo transfer. If the serum hCG level was > 25 mIU/mL on day 14 after transfer, luteal support was continued. Clinical pregnancy was confirmed by transvaginal ultrasound performed 28–35 days after transfer, which showed a gestational sac in the uterus. The primary outcomes of this study were the clinical pregnancy rate (CPR) and live birth rate (LBR). CPR and LBR were defined as the proportion of patients who achieved a clinical pregnancy or live birth after their first embryo transfer (either fresh or frozen). The analysis included 1,135 patients, with each patient contributing exactly one transfer cycle to ensure statistical independence. The secondary outcomes included the number of oocytes retrieved, number of mature oocytes, and embryological outcomes. Fertilization was assessed 16–18 hours after insemination or ICSI. Total fertilized oocytes were defined as all oocytes showing two pronuclei (2PN) or abnormal fertilization (1PN or ≥3PN). Normal fertilization was specifically defined as the presence of two pronuclei (2PN) and two polar bodies. The total fertilization rate was calculated as the number of total fertilized oocytes divided by the number of MII oocytes, multiplied by 100%. The normal fertilization rate was calculated as the number of 2PN oocytes divided by the number of MII oocytes, multiplied by 100%. Additional outcomes included the number of transferable embryos and the number of high-quality embryos. Additionally, the high-quality embryo rate was defined as the number of high-quality embryos divided by the number of 2PN oocytes. Embryo quality was graded by professional embryologists: cleavage-stage embryos were graded according to the Istanbul Consensus ( 18 ), and blastocysts were graded according to the method described by Gardner et al. ( 19 ). Cleavage-stage embryos of Grade I and blastocysts of Grade 3BB or higher were considered high-quality embryos. Serum β-hCG was measured on day 14 after embryo transfer; if β-hCG > 25 mIU/mL, luteal support was continued. Transvaginal ultrasound was performed 28–35 days after transfer. Clinical pregnancy was defined as the presence of a gestational sac with a fetal pole and primitive heartbeat detected by ultrasound in the uterus. Live birth was defined as the delivery of a neonate with vital signs at ≥ 28 weeks of gestation; Early miscarriage as termination of pregnancy at < 28 weeks of gestation with a fetal weight < 1000 g. Early miscarriage rate was calculated per clinical pregnancy. SPSS 27.0 statistical software was used for data processing and analysis. Continuous variables with normal distribution were expressed as mean ± standard deviation. One-way analysis of variance (one-way ANOVA) was used for overall comparison among the three groups, and post-hoc test based on Bonferroni correction was used for pairwise comparison. Continuous variables with non-normal distribution were expressed as median and interquartile range [M (P25, P75)]. Kruskal-Wallis H test was used for comparison among multiple groups, and Dunn test with Bonferroni correction for P values was used for pairwise comparison. Categorical variables were expressed as percentage and frequency [% (n/N)]. Pearson χ 2 test or Fisher’s exact test was used for comparison between groups, and P value correction for multiple comparisons between groups was performed when necessary. Spearman rank correlation analysis was used to explore the correlation between trigger-day serum LH levels and various skewed continuous clinical indicators (such as basal endocrine indicators, controlled ovarian stimulation parameters, and embryological indicators). Binary logistic regression analysis was used to evaluate the independent effect of trigger-day serum LH levels on clinical pregnancy outcomes. To strictly control for potential confounders and address intermediate effects, we adopted a stepwise adjustment strategy constructing two models. Model 1 was adjusted for baseline and clinical characteristics, including female age, male age, BMI, PCOS status, trigger modality, fertilization method (IVF vs. ICSI), years of infertility, basal FSH, and AFC. Model 2 (the fully adjusted model) included all covariates in Model 1, with the further inclusion of cycle-specific embryo transfer characteristics: endometrial thickness on the transfer day, transfer of high-quality embryo (Yes vs. No), and total number of transferred embryos. Specifically, categorical variables such as PCOS status were coded as binary variables with the non-PCOS group serving as the reference (1 = PCOS, 0 = non-PCOS). For subgroup analyses, variables used for stratification (e.g., BMI, PCOS status) were appropriately excluded from the covariates of the respective models to prevent collinearity. Furthermore, in subgroups with extremely rare categorical events (e.g., only 1 patient receiving GnRH agonist alone trigger in the fresh ET subgroup), these specific cases were excluded to ensure model stability. The results are expressed as adjusted odds ratios (aORs) with 95% confidence intervals (CIs). A post-hoc power calculation was performed to determine the minimum detectable effect size for the primary outcome (clinical pregnancy rate) using a conventional two-sided α of 0.05 and a power of 80% (1-β = 0.80). For the correlation analysis in Table 1 , which involved 18 clinical and laboratory indicators, the Bonferroni correction was applied to adjust for multiple comparisons. Consequently, the threshold for statistical significance was strictly adjusted to a two-sided P- value<0.0028 (0.05/18).To assess the robustness of the primary findings and address potential information loss from data categorization, a sensitivity analysis was performed by analyzing LH strictly as a continuous variable in the multivariable logistic regression model, adjusting for the same confounding factors. Correlation analysis of serum LH level on the trigger day. ** P < 0.0028 (Statistically significant after Bonferroni correction for multiple testing).