High Peak Estradiol Predicts Higher Miscarriage and Lower Live Birth Rates in High Responders Triggered with a GnRH Agonist in IVF/ICSI Cycles

The Journal of Reproductive Medicine High Peak Estradiol Predicts Higher Miscarriage and Lower Live Birth Rates in High Responders Triggered with a GnRH Agonist in IVF/ICSI Cycles Ryan G. Steward, M.D., Cindy E. Zhang, M.D., Anish A. Shah, M.D., Jason S. Yeh, M.D., Chen Chen, M.S., Yi-Ju Li, Ph.D., Thomas M. Price, M.D., and Suheil J. Muasher, M.D. ORIGINAL ARTICLES OBJECTIVE: To investigate parameters predictive of pregnancy outcomes in high responders undergoing fresh, autologous, GnRH antagonist IVF/ICSI cycles using a GnRH agonist trigger. STUDY DESIGN: Retrospective cohort study of all patients deemed high-risk for ovarian hyperstimulation syndrome who underwent fresh, autologous IVF/ICSI using a GnRH agonist trigger at an academic fertility center from 2010 2012. RESULTS: A total of 71 first cycles were analyzed. Rates of clinical pregnancy, live birth (LB), and total (clinical plus biochemical) miscarriage (MC) were 52%, 38%, and 25%, respectively. Mean peak estradiol (E2) and the number of oocytes retrieved were 3,701 pg/ml and 15.2, respectively. Peak E2 was significantly higher in those cycles resulting in clinical MC (p=0.003). After adjusting for age, basal follicle stimulating hormone, and the number of oocytes retrieved, elevated peak E2 remained Clinicians should recognize this patient population as inherently in danger of this scenario and minimize the occurrence of peak E2 levels >5,000 pg/ml. associated with increased clinical MC (p=0.029) and trended towards a relationship with higher total MC (p=0.062). When peak E2 was treated as a binary variable based on the threshold value of >5,000 pg/ml, peak E2 above this value was associated with a higher rate of clinical MC (OR=16.14 with 95% CI 1.25 209.35, p=0.033) and total MC (OR=6.81 with 95% CI 1.12 41.54, p=0.037), as well as a lower LB rate (OR= 0.095 with 95% CI 0.01 0.90, p=0.041). CONCLUSION: Clinicians should recognize most IVF/ICSI patients triggered with a GnRH agonist as inherently in danger of excessively high serum E2 and avoid peak levels >5,000 pg/ml in order to avoid higher MC and lower LB rates. (J Reprod Med 2015;60: 463 470) Keywords: agonists, assisted reproductive techniques, assisted reproductive technology, con- From the Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, Duke University Medical Center; the Department of Biostatistics and Bioinformatics, Duke University Medical Center; and Duke Institute of Molecular Physiology, Duke University Medical Center, Durham, North Carolina. Presented as an oral presentation at the 69th American Society for Reproductive Medicine/International Federation of Fertility Societies Annual Meeting, Boston, Massachusetts, October 12 17, 2013. Address correspondence to: Ryan G. Steward, M.D., 5704 Fayetteville Road, Durham, NC 27713 (rgstew777@gmail.com). Financial Disclosure: The authors have no connection to any companies or products mentioned in this article. 0024-7758/15/6011-12 0463/$18.00/0 Journal of Reproductive Medicine, Inc. The Journal of Reproductive Medicine 463

464 The Journal of Reproductive Medicine trolled ovarian hyperstimulation, estradiol, GnRH, gonadotropin-releasing hormone, in vitro intracytoplasmic sperm injection, fertilization, live birth, miscarriage, ovarian hyperstimulation syndrome, pregnancy outcome, spontaneous abortion. Ovarian hyperstimulation syndrome (OHSS) is a well-recognized and potentially fatal iatrogenic complication of controlled ovarian hyperstimulation (COH) during in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) cycles. While its true incidence is probably vastly underestimated due to inconsistent reporting, a protocol development has arisen in the past decade allowing for a paradigm shift in the treatment of high-risk patients and a virtual elimination of this serious complication. 1 Triggering the induction of final oocyte maturation with a gonadotropin-releasing hormone agonist (GnRHa) instead of human chorionic gonadotropin (hcg) was first proposed over 20 years ago as a means to decrease the burden of OHSS. 2 While it achieves this end with great efficiency, subsequent studies demonstrated poorer pregnancy outcomes attributable to defective hormone production from GnRHa-induced luteolysis. 3-5 However, as a subject of significant investigative interest during the past several years, newer studies promoting further protocol changes have brought about greatly improved pregnancy outcomes while maintaining a very low incidence of OHSS. These include the addition of low-dose hcg, either at the time of GnRHa trigger ( dual trigger ) or on the day of oocyte retrieval, more intensive luteal phase supplementation (LPS) with estradiol (E2) and progesterone (P), and the planned cryopreservation of all embryos for transfer in a programmed cycle. 6-12 As a result, clinicians have become more comfortable using GnRHa trigger, now able to worry less about OHSS and impaired pregnancy rates. The association between the supraphysiologic hormone milieu of COH, endometrial receptivity, and pregnancy outcomes is another area of recent investigative interest. Most studies appear to find detriment with high hormone levels. 13-21 There is, however, little known about what stimulation parameters, if any, are important predictors in GnRHa-triggered cycles. One such study reported clinical outcomes to be related to luteinizing hormone (LH) level on the day of trigger and even improved in cycles with peak E2 >4,000 pg/ml. 22 In light of such conflicting data, we sought to attempt to elucidate what factors can be useful as clinical determinants in high responders undergoing IVF/ICSI cycles using a GnRHa trigger. This patient population, by definition, is predisposed to excessive hormone profiles during COH. If, indeed, defective endometrial receptivity results from high E2 levels in fresh cycles, patients triggered with a GnRHa, given their propensity for a high response, may be quietly in danger of a worse prognosis than would otherwise be expected. Materials and Methods Study Design and Participants This was a retrospective cohort study conducted at the Duke Fertility Center. Study approval was granted by the Duke University Institutional Review Board. The records of all patients who underwent fresh, autologous, IVF/ICSI using a GnRH antagonist protocol and a GnRHa trigger for oocyte maturation from January 2010 to December 2012 were reviewed. Seventy-one first cycles were included for analysis. All patients were deemed high risk for OHSS based on age, ovarian reserve testing, peak serum E2, and/or >13 follicles measuring >11 mm in diameter on the day of trigger. 23 Treatment Protocol The treatment and LPS protocols used were previously described by Engmann et al. 12 Baseline transvaginal ultrasound was performed following 2 3 weeks of oral contraceptive pill suppression. On menstrual cycle day 1 2, a daily subcutaneous injection of recombinant follicle stimulating hormone (rfsh) (Follistim, Merck, Whitehouse Station, New Jersey) was initiated. The usual starting dose of 100 225 IU was based on patient age, body mass index, antral follicle count, and anti-müllerian hormone level (available in 63/71 patients). Serial ultrasounds and serum E2 measurements were performed to assess response to stimulation and adjust medication dose at the discretion of the treating physician. When the lead follicle reached a mean diameter of 13 mm and/or the serum E2 exceeded 350 pg/ml, a daily subcutaneous injection of 250 µg of the GnRH antagonist ganirelix acetate (Ganirelix, Merck) was added to prevent premature ovulation. These were continued until the day of oocyte maturation induction. When 3 follicles reached 17 mm in mean diameter, 1 mg of leuprolide acetate (Lupron, TAP Pharmaceuticals, Lake Forest, Illinois) was administered subcutaneously between 8 and 10 o clock in the evening. Serum

Volume 60, Number 11-12/November-December 2015 465 E2, LH, and P levels were measured the following morning to ensure adequate surge. Transvaginal, ultrasound-guided oocyte retrieval was performed approximately 36 hours after trigger injection. Luteal Phase Supplementation All patients received 50 mg intramuscular (IM) P in oil daily and 0.3 mg E2 patches (Vivelle-Dot, Novartis Pharmaceuticals, East Hanover, New Jersey) every other day beginning on the day after oocyte retrieval and continuing until either 10 weeks of gestation or a negative pregnancy test. Serum E2 and P were measured 3 and 7 days after oocyte retrieval and then weekly until 10 weeks of gestation. To maintain serum E2 above 200 pg/ml, either patches were increased up to 0.4 mg every other day or 2 mg of twice-daily oral, micronized E2 (Estrace, Bristol-Meyers, New York, New York) was added, as necessary. Similarly, IM P was increased to 75 mg daily, as needed, to keep serum levels above 20 ng/ml. 24 Outcome Variables The primary outcome was live birth (LB). Secondary outcomes included rates of clinical pregnancy (CP), clinical miscarriage (MC), total MC, and moderate to severe OHSS. Clinical pregnancy was defined as the presence of an intrauterine gestational sac (GS) at 6 7 weeks gestation coincident with a rising serum β-hcg level. Clinical MC was defined as pregnancy loss after ultrasound visualization of an intrauterine GS. Total MC included clinical plus biochemical MC, the latter defined as pregnancy loss prior to ultrasound identification of a GS. Data Analysis The SAS statistical package (Release 9.3, SAS Institute, Cary, North Carolina) was used for statistical analysis. Summary statistics for baseline and cycle variables were computed and are presented as means (standard deviation [SD]). Univariable analysis to compare the equivalence of each clinical variable between outcome statuses was performed using two-sample t tests for continuous variables and χ 2 tests for categorical variables. Multiple logistic regression modeling was performed for each variable of interest with covariate adjustment for patient age, basal FSH, and the number of oocytes retrieved. Estimated odds ratios (ORs) and their 95% confidence intervals (CIs) are reported. A nominal significance level (p<0.05) was used to declare statistical significance. Results Baseline and cycle variables are shown in Table I. The mean age was 33.4 years. The average anti- Müllerian hormone level and antral follicle count were 5.0 ng/ml and 21.2, respectively. All patients demonstrated adequate surge after trigger. The rates of CP, clinical MC, total MC, and LB were 52.1%, 14.1%, 25.4%, and 38%, respectively. The rate of twin gestation was 15.5%, and none of these pregnancies ended in miscarriage. There were no cases of high-order multiples or OHSS. The mean peak E2 and number of oocytes Table I Baseline and Cycle Variables Variable N Mean (SD) Minimum Maximum Age (yrs) 71 33.41 (3.94) 26 42 BMI (kg/m 2 ) 71 27.12 (6.72) 18.7 49 AMH (ng/ml) 63 5.01 (3.59) 0.47 19 AFC (n) 71 21.24 (10.39) 3 51 D3 FSH (miu/ml) 47 7.16 (1.99) 0.6 11.8 Days of stimulation (n) 71 9.81 (1.46) 7 14 Peak E2 (pg/ml) 69 3,700.65 (1,669.47) 1,042 7,208 Posttrigger P4 (ng/ml) 69 9.67 (4.91) 1.8 23 Posttrigger LH (IU/L) 68 37.75 (20.62) 9.7 98.1 Oocytes retrieved (n) 71 15.18 (6.90) 1 39 2PN (n) 71 8.63 (4.34) 1 28 Cleaved (n) 71 8.52 (4.27) 1 28 Transferred (n) 71 1.99 (0.43) 1 3 Cryopreserved (n) 71 5.45 (4.20) 0 26 BMI = body mass index, AMH = anti-müllerian hormone, AFC = antral follicle count, D3 FSH = day 3 follicle stimulating hormone, E2 = estradiol, posttrigger P4 = progesterone on day after trigger, posttrigger LH = luteinizing hormone on day after trigger, 2PN = two pronuclear zygote.

466 The Journal of Reproductive Medicine retrieved were 3,701 pg/ml and 15.2, respectively. Pregnancy outcomes per peak E2 level for the entire cohort are shown in Table II. While not statistically significant, peak E2 was lower in pregnancies resulting in LB. It was, however, significantly higher in those resulting in clinical MC (p=0.003). Post-trigger LH and P4 were not related to the investigated outcomes (Table III). After adjusting for age, basal FSH, and the number of oocytes retrieved in the multiple logistic regression model, peak E2 remained associated with increased clinical MC (p=0.029) (Table III) and trended towards a relationship with higher total MC (p=0.062). We then examined the frequency of clinical MC for each 1,000 pg/ml peak E2 interval (Figure 1). Due to the abrupt increase in clinical MC above 5,000 pg/ml, we chose this value as the threshold around which to dichotomize the data (peak E2 status). That is, clinical MC rate ranged from approximately 6 14% in each of the five 1,000 pg/ml groups below 5,000 pg/ml but increased to 33% (5/15 patients) when peak E2 reached 5,000 pg/ ml. Once again controlling for age, basal FSH, and the number of oocytes retrieved, peak E2 5,000 pg/ml was associated with significantly higher rates of clinical MC (OR 16.14 with 95% CI 1.25 209.35, p=0.033) and total MC (OR 6.81 with 95% CI 1.12 41.54, p=0.037), as well as fewer LBs (OR 0.010 with 95% CI 0.01 0.90, p=0.041) (Table IV). Discussion Our findings demonstrate that extremely supraphysiologic E2 levels appear to be associated with worse pregnancy outcomes in fresh, autologous, GnRH antagonist IVF/ICSI cycles when using a GnRHa trigger. After adjusting for confounders, fewer LBs and more MCs were observed with elevated peak E2, especially when it exceeded 5,000 pg/ml. These results contribute to a recent focus in the literature attempting to clarify defective en- dometrial receptivity from supraphysiologic hormone levels in fresh cycles as the causative mechanism for poorer obstetric outcomes. Indeed this association may be even more pronounced in cycles using a GnRHa trigger, owing to the added corpora lutea (CL) dysfunction from the agonist, and there is currently a paucity of literature on the combination of these two topics. Our study suggests that, similar to IVF/ICSI cycles using hcg trigger, optimal pregnancy outcomes are achieved in GnRHatriggered cycles either by avoiding excessively high E2 levels during fresh cycles or by cryopreserving all embryos for subsequent frozen embryo transfer (FET). Moreover, it seems imperative to recognize this cohort of patients as inherently at risk for this scenario given their tendency for high response. hcg has historically been the therapeutic means to replace endogenous LH to induce oocyte maturation and ovulation in IVF/ICSI cycles. Due to their similar structure and biological activity, hcg and LH act on the same receptor. 25 However, the half-life of LH is approximately 60 minutes, whereas that of hcg is >24 hours. 26,27 Thus, hcg produces a much more sustained effect on the CL, predisposing normal and high responders to a greater risk of OHSS. Specifically, LH receptor stimulation on the ovarian theca cells stimulates production of androstenedione, which then diffuses into the granulosa cells and serves as the primary precursor for E2 production. It is the exaggeration of this process, particularly in high responders, that contributes to the pathogenesis of OHSS and altered endometrial receptivity. With the advent of GnRH antagonist protocols in the late 1990s it became possible to induce oocyte maturation with a GnRHa. Doing so causes an initial surge of LH and FSH followed by pituitary suppression with continued use. But the rise and fall of this flare is much more abbreviated than in a natural cycle, and thus the total amount of gonadotropin released from the pituitary after Table II Pregnancy Outcomes per Peak Estradiol Level Peak estradiol (pg/ml) <2,000 2,000 3,000 3,000 4,000 4,000 5,000 >5,000 Outcome n=12 n=17 n=14 n=11 n=17 + Pregnancy test (%) 50 71 71 100 53 Clinical pregnancy (%) 17 71 64 73 35 Live birth (%) 17 65 50 55 6 Miscarriage (%) 17 6 21 45 41

Volume 60, Number 11-12/November-December 2015 467 Table III Comparison of Hormone Variables Between Pregnancy Outcomes Live birth Mean (SD) Clinical miscarriage Mean (SD) Category (N) LB (27) No LB (44) P a P b CM (10) No CM (61) P a P b Peak E2 (pg/ml) 3,310 (1,998) 3,952 (1,956) 0.120 0.306 5,218 (1,879) 3,473 (1,526) 0.003 0.029 Posttrigger LH (IU/L) 38.8 (21.3) 37.1 (20.4) 0.738 0.459 36.6 (17.1) 37.9 (21.2) 0.860 0.678 Posttrigger P4 (ng/ml) 10.3 (5.9) 9.3 (4.2) 0.375 0.314 10.1 (3.8) 9.6 (5.1) 0.805 0.413 SD = standard deviation, E2 = estradiol, LH = luteinizing hormone, P4 = progesterone. a p values derived from t test. b p values derived from multivariate logistic regression which adjusted for age, basal FSH, and the number of oocytes retrieved. GnRHa is significantly less. 1,28 One potential benefit of GnRHa trigger, however, is the simultaneous release of FSH, akin to the natural cycle. 3 FSH has been shown to induce LH receptor formation on granulosa cells and promote nuclear maturation and cumulus expansion, 29-32 which may explain why some studies using GnRHa trigger have reported a significantly higher number of mature oocytes compared to hcg trigger. 4,33 In 1988 Itskovitz was the first to propose the use of GnRHa for oocyte maturation as a way to decrease the incidence of OHSS in IVF cycles. 2 In the past decade over a dozen randomized controlled trials (RCTs) have proven the efficacy of GnRHa in eliminating OHSS. 1 As alluded to above, during the same time period multiple welldesigned studies demonstrated no difference, or even an increase, in the rate of metaphase II oocytes, as well as no detriment to fertilization or embryo quality with GnRHa trigger. 3-5,34 Despite these apparent accomplishments, widespread use of GnRHa trigger has been met with resistance, largely due to several initial reports of poorer implantation, lower ongoing pregnancy rates, and higher pregnancy loss rates in normal responders using standard LPS. 3-5 However, subsequent RCTs investigating normal and high responders showed no difference in LB rates, while maintaining very low OHSS rates, with the addition of one or multiple smaller doses of hcg or recombinant LH (rlh) added at the time of trigger and/or oocyte retrieval in order to augment oocyte maturation or luteal phase hormone production, respectively. 6-11 Additionally, despite the fact that one small study 8 found disparate results, a 2008 RCT demonstrated similar LB rates between GnRHa and hcg trigger in high responders with only the addition of more aggressive luteal E2 and P supplementation. 12 Taken all together, it seems that overcoming the luteolytic property of the GnRHa, one way or another, is key to preserving pregnancy potential after GnRHa trigger in autologous, fresh cycles. Current recommendations from the Copenhagen GnRH Agonist Triggering Workshop Group expand upon this idea as they make recommendations for the use of GnRHa trigger in several patient populations. 1 According to their 2011 publication, for all patients, good oocyte quality can be achieved with GnRHa trigger while practically eradicating the risk of OHSS. They advocate its universal use in oocyte donors, highlighting the added benefits of a shorter luteal phase and reduction in ovarian size and pelvic free fluid after oocyte retrieval. Finally, while the optimal LPS regimen is yet to be fully elucidated, ongoing pregnancy rates similar to hcg trigger can be achieved in one of two ways: (1) by adding one or more injections of low-dose hcg or rlh at the time of trigger ( dual trigger ) and/or in the luteal phase, in addition to modified E2 and P4 supplementation, or (2) by freezing all embryos and transferring them after thaw in a subsequent programmed cycle. 1 Figure 1 Clinical miscarriage frequency per 1,000 pg/ml peak E2 interval. CM = clinical miscarriage, E2 = estradiol.

468 The Journal of Reproductive Medicine At present, the relationship between the supraphysiologic hormonal milieu of fresh IVF cycles and endometrial receptivity continues to be debated. Many investigators have examined the issue over the past 25 years. Despite an assortment of differing study designs and populations that ultimately make definitive conclusions elusive, when considered collectively, the data seem to suggest an adverse endometrial effect of very elevated E2. Forman et al demonstrated significantly lower implantation and pregnancy rates in cycles with high E2. 13 Simon et al faulted altered endometrial receptivity after similar findings in normal and high responders with peak E2 >2,500 pg/ml, despite no difference in oocyte number or embryo quality. 14 Another study by the same author concluded improved receptivity in high responders by using an rfsh step-down protocol to achieve a lower peak E2. 15 Similar conclusions were drawn from two studies that found inferior pregnancy rates in fresh but not in frozen/thaw cycles when peak E2 exceeded 5,000 pg/ml. 16,17 A 2010 study of 455 fresh IVF cycles using hcg trigger suggested an optimal peak E2 threshold of 4,000 pg/ ml for women <38 years of age and 3,000 pg/ml for those 38. 18 Moreover, endometrial gene expression, including that for E2 and P receptors, has been shown to be significantly altered in stimulated cycles, indicating a state of premature endometrial advancement compared to natural cycles. 19-21 Perhaps the most compelling evidence in support of the notion of defective endometrial receptivity associated with supraphysiologic hormone levels in stimulated cycles is the growing number of studies showing better outcomes in freeze-all/ thaw cycles. In separate RCTs investigating normal and high responders, Shapiro et al demonstrated improved clinical and ongoing pregnancy rates after elective FET. 35,36 In another elegant study from the same group, 93 pairs of fresh and frozen/thawed blastocysts were matched on embryo parameters and patient age. 37 Single blastocyst transfer revealed a significantly better PR in the frozen/thaw group for day-6 blasts (54.3% vs. 17.1% in the fresh group, p<0.001). This resulted in a significantly higher ongoing PR in the frozen/ thaw group as compared to the fresh (55.9% vs. 26.9%, respectively, p<0.001). The authors point toward impaired embryo implantation from altered endometrial receptivity in the setting of recent ovarian stimulation. They highlight their findings as support for embryo-endometrium asynchrony since the more slowly-developing blasts (day 6) appeared to be most affected; i.e., the slowergrowing embryos may miss a more advanced or narrow implantation window after recent COS. Finally, the theory of impaired endometrial receptivity has gone so far as to surface in literature regarding obstetric and perinatal outcomes. A 2012 metaanalysis of 11 observational studies comparing singleton pregnancies in fresh versus frozen/ thaw cycles demonstrated better outcomes in the latter group. 38 These include significantly lower relative risks of antepartum hemorrhage, preterm birth, small-for-gestational-age infants, low birth weight, and perinatal mortality. Once again, the authors postulate that very high steroid hormone levels during COH result in endometrial advancement causing asynchrony with early embryo development that ultimately leads to abnormal placentation that contributes to these complications. The results of the present study support a role for altered endometrial receptivity to explain poorer pregnancy outcomes in high responders whose cycles have resulted in very high peak E2 levels. Weaknesses of the study include its retrospective design and small sample size. Therefore, estimates of the outcomes of interest are based on relatively few observations, and type 1 error cannot be excluded. Furthermore, CP, LB, and MC, despite being clinically significant outcomes, are statistically correlated. By repeating them in a multivariate analysis, we cannot exclude the possibility of finding associations by chance. In addition, selection bias may exist from heterogeneity in medication dosing and patient diagnosis (for example, the cohort included some patients with polycystic ovary syndrome). However, our inclusion of only GnRH antagonist cycles using a GnRHa for oocyte maturation makes our findings unique to the vast majority of published data on the subject and draws attention to a particularly vulnerable subgroup of patients. In an era when comfort with GnRHa trigger is becoming more common thanks to improvements in LPS, including the use of dual trigger, the continuing trend of evidence linking the supraphysiologic hormone milieu of COH to impaired endometrial receptivity deserves added reflection. The use of GnRHa trigger itself, in fresh, autologous cycles, should probably serve as an immediate red flag because it identifies those patients most likely to demonstrate an exaggerated response to stimulation. With the realization that the inci-

Volume 60, Number 11-12/November-December 2015 469 Table IV Results of Multivariate Logistic Regression Analysis for Each Clinical Outcome Using Dichotomized Peak E2 Status (>5,000 pg/ ml Compared to <5,000 pg/ml), with Adjusted Covariates Peak E2 status Covariates (p value) Outcome OR 95% CI p Value Age D3 FSH Oocytes retrieved Clinical pregnancy 0.53 0.10 2.66 0.438 0.859 0.524 0.347 Clinical miscarriage 16.14 1.25 209.35 0.033 0.572 0.964 0.342 Total miscarriage 6.81 1.12 41.54 0.037 0.718 0.554 0.429 Live birth 0.10 0.01 0.90 0.041 0.860 0.509 0.729 OR = odds ratio. dence of moderate to severe OHSS after GnRHa trigger is extremely low, less importance is often placed on peak E2. That is, agonist trigger prevents OHSS even with extremely high E2 levels. Thus, it becomes easy to aggressively stimulate patients with less regard to rising E2, knowing the agonist trigger is such an effective guarantee against OHSS. In this way, the attempt to prevent one ill effect directly opens the door for another. Our findings suggest that such a practice is a mistake. The importance of this observation may be further underestimated when one considers the luteolytic property of the agonist, as it seems plausible that this may have an additive negative effect on endometrial receptivity. The detriment from embryo/endometrium asynchrony due to high hormone levels may be compounded by deficient E2 and P production from the CL, creating an even more fragile environment for proper implantation. In conclusion, very high levels of peak E2 in high responders triggered with a GnRHa should be avoided to maximize endometrial receptivity and improve pregnancy outcomes, irrespective of OHSS prevention. Clinicians should recognize this patient population as inherently in danger of this scenario and minimize the occurrence of peak E2 levels >5,000 pg/ml. Well-designed, prospective trials are needed to shed further light on the concept of endometrial receptivity in IVF/ICSI cycles using a GnRHa trigger. Acknowledgments The authors would like to thank the Duke Fertility Center staff for their support in generating this manuscript. References 1. Humaidan P, Kol S, Papanikolaou EG: GnRH agonist for triggering of final oocyte maturation: Time for a change of practice? Hum Reprod 2011;17:510-524 2. Itskovitz J, Boldes R, Barlev A, et al: The induction of LH surge and oocyte maturation by GnRH analogue (buserelin) in women undergoing ovarian stimulation for in vitro fertilization. Gynecol Endocrinol 1988;2(Suppl 2):165 3. Fauser BC, de Jong D, Olivennes F, et al: Endocrine profiles after triggering of final oocyte maturation with GnRH agonist after cotreatment with the GnRH antagonist ganirelix during ovarian hyperstimulation for in vitro fertilization. J Clin Endocrinol Metab 2002;87:709-715 4. Humaidan P, Bredkjaer HE, Bungum L, et al: GnRH agonist (buserelin) or hcg for ovulation induction in GnRH antagonist IVF/ICSI cycles: A prospective randomized study. Hum Reprod 2005;20:1213-1220 5. Kolibianakis EM, Schultze-Mosgau A, Schroer A, et al: A lower ongoing pregnancy rate can be expected when GnRH agonist is used for triggering final oocyte maturation instead of HCG in patients undergoing IVF with GnRH antagonists. Hum Reprod 2005;20:2887-2892 6. Humaidan P, Bungum L, Bungum M, et al: Rescue of corpus luteum function with peri-ovulatory HCG supplementation in IVF/ICSI GnRH antagonist cycles in which ovulation was triggered with a GnRH agonist: A pilot study. Reprod Biomed Online 2006;13:173-178 7. Humaidan P, Ejdrup Bredkjaer H, Westergaard LG, et al: 1,500 IU human chorionic gonadotropin administered at oocyte retrieval rescues the luteal phase when gonadotropin-releasing hormone agonist is used for ovulation induction: A prospective, randomized, controlled study. Fertil Steril 2010;93:847-854 8. Pirard C, Donnez J, Loumaye E: GnRH agonist as luteal phase support in assisted reproduction technique cycles: Results of a pilot study. Hum Reprod 2006;21:1894-1900 9. Acevedo B, Gomez-Palomares JL, Ricciarelli E, et al: Triggering ovulation with gonadotropin-releasing hormone agonists does not compromise embryo implantation rates. Fertil Steril 2006;86:1682-1687 10. Castillo JC, Dolz M, Bienvenido E, et al: Cycles triggered with GnRH agonist: Exploring low-dose HCG for luteal support. Reprod Biomed Online 2010;20:175-181 11. Papanikolaou EG, Verpoest W, Fatemi H, et al: A novel method of luteal supplementation with recombinant LH, when a GnRH-agonist is used instead of HCG for ovulation triggering: A randomized prospective proof of concept study. Fertil Steril 2011;3:1174-1177 12. Engmann L, DiLuigi A, Schmidt D, et al: The use of gonadotropinreleasing hormone (GnRH) agonist to induce oocyte maturation after cotreatment with GnRH antagonist in high-risk patients undergoing in vitro fertilization prevents the risk of ovarian hyperstimulation syndrome: A prospective randomized controlled study. Fertil Steril 2008;89:84-91 13. Forman R, Fries N, Testart J, et al: Evidence for an adverse effect of elevated serum estradiol concentrations on embryo implantation. Fertil Steril 1988;49:118-122 14. Simon C, Coro F, Valbuena D, et al: Clinical evidence for a detrimental effect on uterine receptivity of high serum oestradiol concentrations in high and normal responder patients. Hum Reprod 1995;10: 2432-2437

470 The Journal of Reproductive Medicine 15. Simon C, Garcia Velasco JI, Valbuena D, et al: Increasing uterine receptivity by decreasing oestradiol levels during the preimplantation period in high responders with the use of a follicle stimulating hormone step-down regimen. Fertil Steril 1998;70:234-239 16. Ng EH: What is the threshold value for serum estradiol levels associated with adverse IVF outcomes? Fertil Steril 2000;73:1071-1072 17. Wu C-H, Kuo T-C, Wu H-H, et al: High serum estradiol levels are not detrimental to in vitro fertilization outcome. Taiwan J Obstet Gynecol 2007;46:54-59 18. Joo BS, Park SH, An BM, et al: Serum estradiol levels during controlled ovarian hyperstimulation influence the pregnancy outcome of in vitro fertilization in a concentration-dependent manner. Fertil Steril 2010;93:442-446 19. Groothuis PG, Dassen HH, Romano A, et al: Estrogen and the endometrium: Lessons learned from gene expression profiling in rodents and human. Hum Reprod Update 2007;13:405-417 20. Haouzi D, Assou S, Mahmoud K, et al: Gene expression profile of human endometrial receptivity: Comparison between natural and stimulated cycles for the same patients. Hum Reprod 2009;24:1436-1445 21. Papanikolaou EG, Bourgain C, Kolibianakis E, et al: Steroid receptor expression in late follicular phase endometrium in GnRH antagonist IVF cycles is already altered, indicating initiation of early luteal phase transformation in the absence of secretory changes. Hum Reprod Update 2005;20:1541-1547 22. Kummer N, Benadiva C, Feinn R, et al: Factors that predict the probability of successful clinical outcome after induction of oocyte maturation with a gonadotropin-releasing hormone agonist. Fertil Steril 2011;96:63-68 23. Papanikolaou EG, Pozzobon C, Kolibianakis EM, et al: Incidence and prediction of ovarian hyperstimulation syndrome in women undergoing gonadotropin-releasing hormone antagonist in vitro fertilization cycles. Fertil Steril 2006;85:112-120 24. Rosenwaks Z: Donor eggs: Their application in modern reproductive technologies. Fertil Steril 1987;47:895-909 25. Kessler MJ, Reddy MS, Shah RH, et al: Structures of N-glycosidic carbohydrate units of human chorionic gonadotropin. J Biol Chem 1979; 254:7901-7908 26. Yen SS, Llerena O, Little B, et al: Disappearance rates of endogenous luteinizing hormone and chorionic gonadotropin in man. J Clin Endocrinol Metab 1968;28:1763-1767 27. Damewood MD, Shen W, Zacur HA, et al: Disappearance of exogenously administered human chorionic gonadotropin. Fertil Steril 1989;52:398-400 28. Itskovitz J, Boldes R, Levron J, et al: Induction of preovulatory luteinizing hormone surge and prevention of ovarian hyperstimulation syndrome by gonadotropin-releasing hormone agonist. Fertil Steril 1991;56:213-220 29. Zelinski-Wooten MB, Hutchison JS, Hess DL, et al: Follicle stimulating hormone alone supports follicle growth and oocyte development in gonadotropin-releasing hormone antagonist-treated monkeys. Hum Reprod 1995;10:1658-1666 30. Yding Andersen C, Leonardsen L, Ulloa-Aguirre A, et al: FSHinduced resumption of meiosis in mouse oocytes: Effect of different isoforms. Mol Hum Reprod 1999;5:726-731 31. Stickland S, Beers WH: Studies on the role of plasminogen activator in ovulation: In vitro response of granulosa cells to gonadotropins, cyclic nucleotides, and prostaglandins. J Biol Chem 1976;251:5694-5702 32. Eppig JJ: FSH stimulates hyaluronic acid synthesis by oocytecumulus cell complexes from mouse preovulatory follicles. Nature 1979;281:483-484 33. Oktay K, Turkcuoglu I, Rodriguez-Wallberg KA: GnRH agonist trigger for women with breast cancer undergoing fertility preservation by aromatase inhibitor/fsh stimulation. Reprod Biomed Online 2010;20:783-788 34. Griesinger G, Diedrich K, Devroey P, et al: GnRH agonist for triggering final oocyte maturation in the GnRH antagonist ovarian hyperstimulation protocol: A systematic review and meta-analysis. Hum Reprod Update 2006;12:159-168 35. Shapiro BS, Daneshmand ST, Garner FC, et al: Evidence of impaired endometrial receptivity after ovarian stimulation for in vitro fertilization: A prospective randomized trial comparing fresh and frozenthawed embryo transfer in normal responders. Fertil Steril 2011;96: 344-348 36. Shapiro BS, Daneshmand ST, Garner FC, et al: Evidence of impaired endometrial receptivity after ovarian stimulation for in vitro fertilization: A prospective randomized trial comparing fresh and frozenthawed embryo transfers in high responders. Fertil Steril 2011;96: 516-518 37. Shapiro BS, Daneshmand ST, Restrepo H, et al: Matched-cohort comparison of single-embryo transfers in fresh and frozen-thawed embryo transfer cycles. Fertil Steril 2013;99:389-392 38. Maheshwari A, Pandey S, Shetty A, et al: Obstetric and perinatal outcomes in singleton pregnancies resulting from the transfer of frozen thawed versus fresh embryos generated through in vitro fertilization treatment: A systematic review and meta-analysis. Fertil Steril 2012;98:368-377