Endocrine Requirements for Oocyte Maturation Following hCG, GnRH Agonist, and Kisspeptin During IVF Treatment

Objective The maturation of oocytes to acquire competence for fertilization is critical to the success of in vitro fertilization (IVF) treatment. It requires LH-like exposure, provided by either human chorionic gonadotropin (hCG), or gonadotropin releasing hormone agonist (GnRHa). More recently, the hypothalamic stimulator, kisspeptin, was used to mature oocytes. Herein, we examine the relationship between the endocrine changes following these agents and oocyte maturation. Design Retrospective cohort study. Methods Prospectively collected hormonal data from 499 research IVF cycles triggered with either hCG, GnRHa, or kisspeptin were evaluated. Results HCG-levels (121 iU/L) peaked at 24 h following hCG, whereas LH-levels peaked at ~4 h following GnRHa (140 iU/L), or kisspeptin (41 iU/L). HCG-levels were negatively associated with body-weight, whereas LH rises following GnRHa and kisspeptin were positively predicted by pre-trigger LH values. The odds of achieving the median mature oocyte yield for each trigger were increased by hCG/LH level. Progesterone rise during oocyte maturation occurred precipitously following each trigger and strongly predicted the number of mature oocytes retrieved. Progesterone rise was positively associated with the hCG-level following hCG trigger, but negatively with LH rise following all three triggers. The rise in progesterone per mature oocyte at 12 h was greater following GnRHa than following hCG or kisspeptin triggers. Conclusion The endocrine response during oocyte maturation significantly differed by each trigger. Counter-intuitively, progesterone rise during oocyte maturation was negatively associated with LH rise, even when accounting for the number of mature oocytes retrieved. These data expand our understanding of the endocrine changes during oocyte maturation and inform the design of future precision-triggering protocols.


INTRODUCTION
The World Health Organization recognizes subfertility as the fifth most serious global disability, affecting 1 in 6 couples (1). In vitro fertilization (IVF) is a supraphysiological process that simulates many of the physiological processes apparent during the natural menstrual cycle (2). The number of IVF cycles carried out annually is increasing (3), however 11.8% of cycles commenced did not progress to oocyte retrieval (3). Indeed, a recent international priority setting partnership designated the variation in oocyte number following IVF treatment as one of the top ten unresolved research uncertainties (4).
The "trigger" of oocyte maturation replicates the function of the mid-cycle ovulatory luteinizing hormone (LH) surge of the natural cycle (2) and provides LH-like exposure such that oocytes resume meiosis and advance to the metaphase II stage of development to acquire competence for fertilization (5). This LH-like activity can be provided by human chorionic gonadotropin (hCG), gonadotropin releasing hormone agonist (GnRHa), or kisspeptin. These agents have distinct mechanisms of action; namely hCG acts directly at ovarian LH receptors, GnRHa stimulates gonadotropin release from the pituitary gland, and kisspeptin stimulates the hypothalamus to induce release of endogenous GnRH (2). The hormone used to provide this LH-like exposure plays a determinant role in key outcomes affecting the success and safety of IVF treatment, including the ability to retrieve mature oocytes, luteal phase characteristics (impacting on pregnancy rates) and the occurrence of "ovarian hyperstimulation syndrome" (OHSS) (2).
HCG is the most widely used trigger in current practice, being applied in more than three quarters of cycles (6). As it shares the same alpha subunit, and has 85% homology of the beta subunit as native LH, hCG activates the LH receptor (7). HCG has a greater affinity for the LH receptor than native LH and activates distinct intracellular signaling, with a five-fold increased potency for cAMP activity in granulosa cells, whereas LH preferentially activates extracellular signal-related kinase 1/2 and protein kinase B (8). Overall, hCG has a greater steroidogenic action consistent with a critical role in supporting pregnancy, whereas LH has a stronger anti-apoptotic signal (2). Furthermore, hCG (t 1/2 28-29 h) has a longer half-life than native LH (t 1/2~2 0 min), and thus also risks OHSS (2). GnRHa activates the pituitary gland to induce a shorter duration of LH exposure and thus has a lower risk of OHSS than hCG (9). Thus, GnRHa is usually reserved for patients at increased risk of OHSS, as more intensive luteal phase support is required to maintain pregnancy rates (9,10).
More recently, kisspeptin has been used to safely induce oocyte maturation even in women at increased risk of OHSS (11)(12)(13). Kisspeptin stimulates hypothalamic GnRH neurons to release endogenous GnRH (14). Not all GnRH receptors on pituitary gonadotrophs are contiguous with GnRH neuronal terminals, and thus kisspeptin induces a tempered LH rise compared to equimolar doses of GnRH (15). Indeed, kisspeptin induced even fewer signs and symptoms of OHSS in women at increased risk of OHSS (16). This may in part be mediated through an additional direct ovarian action to reduce ovarian vascular endothelial growth factor (VEGF) production (13).
The ability to accurately quantify oocyte maturation facilitates the assessment of the minimum endocrine requirements for oocyte maturation. Different measures have been used to quantify oocyte maturation, including the absolute number of mature oocytes and the "oocyte maturation rate" (proportion of oocytes retrieved that are mature). However, suboptimal oocyte maturation can lead to fewer oocytes being retrievable, leading to a reduction in the denominator as well as the numerator, and thus the "oocyte maturation rate" is often preserved even in the context of suboptimal oocyte maturation. Consequently, we advocate the use of the "mature oocyte yield" defined by the number of mature oocytes expressed as a proportion of the number of follicles on the day of trigger most likely to yield a mature oocyte (17). While estimates for this follicle size have ranged from 10 to 14 mm (18)(19)(20), we have determined that follicles with a diameter of 12 to 19 mm on day of trigger are most likely to yield a mature oocyte (17). In the present study, we used this measure to evaluate the minimal endocrine requirements for LH-like activity (either LH or hCG level) for oocyte maturation following hCG, GnRHa and kisspeptin triggers.

MATERIALS AND METHODS
We conducted a comprehensive retrospective analysis of endocrine profiles prospectively collected from 499 IVF cycles triggered with either hCG, GnRHa, or kisspeptin. These cohorts were chosen as detailed endocrine data were collected following each trigger. The primary objective was to investigate the relationship between LH-like exposure after each trigger and the efficacy of oocyte maturation as quantified by the mature oocyte yield (MOY). The secondary objective was to assess determinants of the level of LH-like activity following each trigger and of the rise in progesterone during and after oocyte maturation.
conducted at My Duc Hospital in Vietnam, whereas kisspeptintriggered cycles were conducted at Hammersmith Hospital in the UK. Data on kisspeptin have been published in (11)(12)(13). Data on GnRH agonist triggered cycles have been published in (21). Data on hCG/hormonal levels during the luteal phase have been published in (22,23) and reviewed in (24).

Study Approvals
All subjects gave written informed consent in accordance with the Declaration of Helsinki and Good Clinical Practice. Data from GnRHa triggered IVF cycles were obtained from a singlecenter randomized controlled trial conducted at My Duc Hospital, Ho Chi Minh City, Vietnam (21

Stimulation Protocol
Follicular stimulation was conducted using a GnRH antagonist co-treated cycle. FSH stimulation was in the form of 150 to 300 IU of follitropin-b for hCG-triggered, corifolliotropin alfa (Elonva; Merck Sharp & Dohme, UK) for GnRHa-triggered, and recombinant FSH (112.5-150 IU Gonal F, Merck Serono, Geneva, Switzerland) for kisspeptin-triggered cycles. The triggers recombinant hCG (250 µg) or GnRHa triptorelin (0.2-0.4 mg) were administered as soon as two follicles reached a size of ≥17 mm. The trigger kisspeptin-54 (6.4-12.8 nmol/kg as a single subcutaneous bolus or 19.2 nmol/kg as a split bolus over 10 h, Bachem Holding AG, Bubendorf, Switzerland) was administered once three follicles reached ≥18 mm. Patients who received two doses of kisspeptin were not included in analyses of parameters potentially affected by the second dose (outcomes after 10 h).

Statistical Analysis
Normality was assessed by D'Agostino and Pearson test and equality of variances by F-test for two groups or the Brown-Forsythe test for multiple groups. Parametrically distributed continuous variables were analyzed by t-test for two groups, or one-way ANOVA with post hoc Tukey's test for multiple groups, and non-parametrically distributed data by the Mann Whitney U test for two groups, or the Kruskal Wallis test with post hoc Dunn's test for multiple groups. Simple linear regression was used to analyze the relationship between two continuous variables. Categorical variables were analyzed by logistic regression. Neural net and random forest models were used to quantify (a) the accuracy of the data when used to predict mature eggs and (b) the relative importance of LH/hCG as a predictor variable. Data were analyzed using GraphPad Prism version 8.0 and STATA version 14. P-value < 0.05 was regarded as indicating statistical significance. Supplementary Table 1.

Serum LH/hCG Levels Following Trigger
Body-weight negatively predicted hCG levels at 24 h following hCG trigger (r = −4.2, r 2 = 0.22, P < 0.0001; Figure 2A). LH rise at 4 h (but not 12 h) was negatively predicted by BMI after GnRHa (r = −3.3, r 2 = 0.05, P = 0.008), whereas LH rise at 12 h was greater in women with body-weight >75 kg following kisspeptin ( Figure 2B). Kisspeptin was dosed using a weightbased regimen, which may in part explain this result. However, higher pre-trigger LH levels with increased body-weight could also contribute to the increased rise in LH in those with increased body weight ( Figure 2C). Indeed, the pre-trigger LH level was   Figure 2F) were positively predicted by pre-trigger LH level.

Endogenous LH Levels Following hCG Trigger
HCG levels at 24 h were negatively associated with endogenous LH levels at 24 h (r = −5.46, r 2 = 0.035, P = 0.018), but there was no significant association between hCG and LH levels at any later timepoint. Women with higher endogenous LH levels prior to hCG trigger continued to have higher LH levels after hCG-levels started to fall following hCG trigger (i.e. pre-trigger LH was positively associated with LH at 24 h following hCG; r 0.49, r 2 0.35, P < 0.0001).

Effect of LH/hCG on Oocyte Maturation
Notably, there was little to no association between levels of LHlike activity and the number of mature oocytes after all three triggers. Indeed, neither hCG levels at 12 h, 24 h (Supplementary Figure 1A) or 36 h after hCG, nor LH rise at 4 h (Supplementary Figure 1B), 12 h (Supplementary Figure 1C), or 24 h following GnRHa were associated with the number of mature oocytes retrieved. After kisspeptin, LH rise at 12 h (P = 0.048) (but not at 4 h or 36 h) was weakly associated with the number of mature oocytes retrieved (Supplementary Figure 1D). We analyzed the cumulative "mature oocyte yield" (MOY; number of mature oocytes divided by the number of follicles of 12-19 mm on the day of trigger) to identify the threshold of hCG/ LH level, beyond which there was unlikely to be any significant additional benefit to oocyte maturation from higher levels. Binary thresholds were~80 iU/L for hCG at 24 h,~25 iU/L at 12 h following GnRHa and~10 iU/L at 12 h following kisspeptin ( Supplementary Figures 1E-H). MOY was not significantly associated with the peak hCG level at 24 h ( Figure 3A). LH at 4 h ( Figure 3B) following GnRHa was associated with the mature oocyte yield (MOY), but not at 12 h ( Figure 3C), or later timepoints. LH > 10 iU/L at 12 h following kisspeptin (but not at 4 h or 36 h) was associated with increased MOY ( Figure 3D).
Random forest models outperformed neural network models when predicting the number of mature oocytes from baseline characteristics, hormone levels and number of 12 to 19 mm follicles (88% accuracy of prediction to within 3 mature eggs for random forests; 57% accuracy for neural networks). Random forest accuracy fell from 88% to 83% when data on LH/hCG levels were not included.
We evaluated the area under the curve over 36 h for hCG levels following hCG trigger, or for LH following GnRHa or kisspeptin triggers, however this measure was not associated with either the number of mature oocytes, nor the MOY (P>0.14 for all). There was no significant difference in "oocyte maturation rate" or "fertilization rate" by category of hCG/LH following any trigger (Supplementary Figure 2).

Progesterone Rise During Oocyte Maturation
Progesterone peaked at 24 h to 51.2 nmol/L following hCG trigger consistent with granulosa cell luteinization/oocyte maturation, before a subsequent greater rise at 108 h (4.5 days post-hCG) to 190.8 nmol/L corresponding to secretion from corpora lutea. Similarly, progesterone peaked to 69.4 nmol/L at 24 h following GnRHa, before a secondary rise to 66.4 nmol/L at 84 h (3.5 days) post GnRHa. Following kisspeptin, progesterone peaked to 22.3 nmol/L at 14 h (luteal progesterone levels were not assessed due to contamination from luteal phase support).
Endogenous LH rise at 12 h following hCG was negatively associated with serum hCG ( Figure 4A). Progesterone rise at 12 h was positively predicted by hCG level following hCG trigger ( Figure 4B) but was negatively predicted by LH rise following all triggers ( Figures 4C-F).
The strongest biochemical predictor of the number of mature oocytes retrieved was the rise in serum progesterone. Progesterone rise at 12 h predicted the number of mature oocytes following hCG (r 2 = 0.23, P < 0.0001), GnRHa (r 2 = 0.29, P < 0.0001) and kisspeptin (r 2 0.18, P < 0.0001) ( Figures  5A-D). The relationship between MOY and progesterone rise at 12 h was much weaker (r 2 hCG 0.07, GnRH 0.02, KP 0.02). To assess, whether each oocyte produced the same amount of progesterone during maturation after the different triggers we assessed progesterone rise per mature oocyte. Accordingly, progesterone rise at 12 h per mature oocyte was greater after GnRHa (3.18 nmol/L) than hCG (1.7 nmol/L) or kisspeptin (1.99 nmol/L) ( Figure 5H). Similarly, the progesterone rise per mature oocyte was greater with hCG level following hCG trigger, but lesser with LH rise at 12 h following other triggers ( Figures  5E-G).
Data comparing luteal estradiol and progesterone rises are presented in Supplemental Data.

DISCUSSION
We examined the hormonal responses following three triggers of oocyte maturation with distinct mechanisms of action to provide an insight into the endocrine requirements for oocyte maturation. HCG levels peaked at 24 h following hCG (albeit similar levels were encountered by 12 h), whereas LH-levels peaked sooner at 4 to 6 h following both GnRHa and kisspeptin. It is likely that a threshold level for LH-like activity needs to be reached to initiate the process of oocyte maturation. The timing of oocyte retrieval is precisely controlled to occur following oocyte maturation but prior to ovulation (2). Consequently, it is noteworthy that the same interval between trigger and oocyte retrieval is ordinarily used (36-37 h) following all three triggers despite the different times of peak LH-like activity. Although near maximal levels of hCG are achieved by 12 h, it is conceivable that the threshold to initiate oocyte maturation is exceeded even sooner after hCG administration, closer to the 4 h peak observed following GnRHa and kisspeptin to permit a similar duration for oocyte maturation to occur after each trigger. We investigated the threshold for LH-like activity required for oocyte maturation. A lower level of LH-like activity was sufficient for efficacious oocyte maturation in some women, however oocyte maturation was more reliable with higher levels. Indeed, the proportion exceeding the median mature oocyte yield for each trigger was increased by the level of LHlike activity achieved. Although the LH surge following kisspeptin was of lower amplitude than GnRHa (15), it is possible that kisspeptin could enhance oocyte maturation via an additional direct action at ovarian kisspeptin receptors to compensate for the lower LH rise (25). Indeed, kisspeptin enhances in vitro maturation of ovine (26) and porcine (27) immature oocytes.
GnRHa and kisspeptin triggers must be used in the context of a GnRH antagonist co-treated stimulation protocol. Thus, the GnRH antagonist must be competitively displaced from the GnRH receptor in order to generate a gonadotropin rise. Consequently, a low LH level prior to the trigger could reflect increased suppression from the GnRH antagonist, and there are data to suggest that escape from GnRH antagonist suppression varies between individuals, occurring more precipitously with greater BMI (28). Thus, the attenuated LH rise following GnRHa and kisspeptin with lower pre-trigger LH values, could reflect more pronounced suppression by the GnRH antagonist. However, it is also recognized that the LH value prior to native GnRH determines subsequent LH rise even in the absence of GnRH antagonist pre-treatment (29). Indeed, the risk of encountering an LH <15 iU/L at 12 h following GnRHa (a threshold commonly used to denote a suboptimal LH rise), occurred in 17.3% of those with a pre-trigger LH <0.5 iU/L but only 3.8% if pre-trigger LH was >2 iU/L. Thus, adjusting the timing and dose of the GnRH antagonist to avoid excessive suppression could further optimize the efficacy of the trigger (30).
Similarly, the hCG level following hCG trigger was negatively associated with body weight consistent with published reports (26). The cohort of women who received hCG in this study was from Vietnam and had a lower body weight in comparison to  western women. Thus, it is likely that western patients with higher body weights could encounter lower serum hCG levels more often, with a more detrimental impact on oocyte maturation than observed in the present study. Indeed, obese women have been reported to have an increased chance of encountering an hCG level <50 iU/L, with a subsequent increased risk of suboptimal oocyte maturation (31). Interestingly, progesterone rose immediately following administration of the trigger and the level by 12 h was strongly predictive of the number of mature oocytes that would be retrieved. This was the case after all three triggers notwithstanding the fact that oocytes would not be mature had they been collected at this time-point. Logically, one would hypothesize that a greater LH rise following the trigger would result in increased oocyte maturation and a greater rise in progesterone. In fact, we observed that the LH rise after the trigger was inversely associated with progesterone rise. Progesterone is known to exert negative feedback on GnRH and LH during the luteal phase of the natural cycle (32,33). As progesterone increased almost immediately following administration of the trigger, it is conceivable that progesterone tempered the LH rise following the trigger through negative feedback. Progesterone rise was strongly associated with the number of mature oocytes retrieved; therefore we examined the "progesterone rise per mature oocyte" and observed that this was greater for GnRHa than for either hCG or kisspeptin. This increased progesterone production during oocyte maturation may represent a specific characteristic of GnRHa triggering as the LH rise following GnRHa exceeds that by kisspeptin, which was associated with lower progesterone production.
Insufficient LH-like exposure increases the risk of "empty follicle syndrome"-a condition where no oocytes are retrieved (2). Early assessment of biochemical parameters to predict whether a mature oocyte is likely to be retrieved is therefore of value, as this can allow for re-administration of the trigger and rescheduling of oocyte retrieval to prevent a failure to retrieve oocytes. Logically, LH rise would be the measure that should confirm the successful deployment of the trigger. However, due to the inverse relationship of LH rise with progesterone rise and the closer relationship between progesterone rise and the number of mature oocytes retrieved, in fact progesterone rise  appears to be the most reliable predictive biochemical marker of successful oocyte maturation. Peak progesterone levels after hCG occurred at 3 days following oocyte retrieval, however 89% of women still exceeded the limit of detection for progesterone at 156 h (day 5 post oocyte retrieval) when a blastocyst transfer is most usually conducted, whereas progesterone levels after GnRHa were already low by this timepoint. The mid-luteal rise in progesterone after hCG far exceeded that generated during oocyte maturation by~4-fold, whereas estradiol rose during the luteal phase to~2.5-fold higher levels than trough levels at 60 h post administration. Nevertheless, this suggests that both estradiol and progesterone can be used to monitor corpora luteal function (34).
Strengths of the study include access to detailed endocrine data following each trigger. Limitations include the heterogenous nature of the study population. Further prospective study with direct comparison in the same population is indicated to verify the findings presented.
In summary, we evaluated the endocrine profile following each trigger and assess its impact on oocyte maturation. An unexpected negative association between LH and progesterone rises was observed during oocyte maturation. Moreover, progesterone rise appears to be the most reliable biochemical predictive marker for oocyte maturation following all triggers and the progesterone rise per oocyte was higher following GnRHa trigger compared to other agents. These findings further explicate our understanding of the endocrine changes during induction of oocyte maturation.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by The Institutional Review Board (IRB) at My Duc Hospital, Ho Chi Minh City, Vietnam and Hammersmith and Queen Charlotte's Research Ethics Committee, London. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
AA, TH, PH, LV, and WD designed the study. AA, TH, VH, SC,