Assisted Reproduction Unit, American Hospital of Istanbul, Guzelbahce Sokak No. 20, Nisantasi, Istanbul 34365, Turkey

Human Reproduction Vol.23, No.3 pp. 668 673, 2008 Advance Access publication on January 12, 2008 doi:10.1093/humrep/dem421 GnRH agonist protocol administration in the luteal phase in ICSI ET cycles stimulated with the long GnRH agonist protocol: a randomized, controlled double blind study B. Ata 1, K. Yakin, B. Balaban and B. Urman Assisted Reproduction Unit, American Hospital of Istanbul, Guzelbahce Sokak No. 20, Nisantasi, Istanbul 34365, Turkey 1 Correspondence address. Tel: þ90-212-3112000; Fax: þ90-212-3112339; E-mail: barisata@hotmail.com. BACKGROUND: GnRH agonist administration in the luteal phase was reported to beneficially affect the clinical outcome of intracytoplasmic sperm injection (ICSI) and embryo transfer (ET) cycles. This double blind, randomized, placebo controlled trial evaluates whether a single dose GnRH agonist administered 6 days after ICSI increases ongoing pregnancy rates following ET in cycles stimulated with the long GnRH agonist protocol. METHODS: Five hundred and seventy women undergoing ET following controlled ovarian stimulation with a long GnRH agonist protocol were included. In addition to routine luteal phase support with progesterone, women were randomized to receive a single 0.1 mg dose of triptorelin or placebo 6 days after ICSI. Randomization was done on the day of ET according to a computer generated randomization table. Ongoing pregnancy rate beyond 20th week of gestation was the primary outcome measure. The trial was powered to detect a 12% absolute increase from an assumed 38% ongoing pregnancy rate in the placebo group, with an alpha error level of 0.05 and a beta error level of 0.2. RESULTS: There were 89 (31.2%) ongoing pregnancies in the GnRH agonist group, and 84 (29.5%) in the control group (absolute difference 11.7%, 95% confidence interval 25.8% to 19.2%). Implantation, clinical pregnancy and multiple pregnancy rates were likewise similar in the GnRH agonist and placebo groups. CONCLUSIONS: Single 0.1 mg triptorelin administration 6 days after ICSI following ovarian stimulation with the long GnRH agonist protocol does not seem to result in an increase 12% in ongoing pregnancy rates. Clinicaltrials.gov Trials registration number NCT 00516490. Keywords: luteal phase support; GnRH agonist; intracytoplasmic sperm injection; ongoing pregnancy; implantation rate Introduction Luteal phase support (LPS) is an integral part of assisted reproduction treatment (ART). Defective luteal phase in assisted reproduction cycles has been attributed to adverse effects of controlled ovarian stimulation, suppression of the pituitary luteinizing hormone (LH) release by gonadotropin releasing hormone (GnRH) analogs, and to depletion of granulosa cells due to follicle aspiration (Garcia et al., 1981; Smitz et al., 1988). Controlled ovarian stimulation has been shown to advance endometrial maturation thus disrupting the delicate mechanism of embryo-endometrium interaction (Fauser and Devroey, 2003). It has been long recognized that supporting the luteal phase with progesterone or human chorionic gonadotrophin (hcg) is associated with higher pregnancy and delivery rates (Daya and Gunby, 2004). The initial agent of choice to support the luteal phase has been hcg, however, due to an increased risk of ovarian hyperstimulation syndrome, it has been largely replaced by progesterone. Estrogen has been advocated as an adjuvant to progesterone for LPS. However, trials investigating the efficacy of estrogen supplementation revealed contradictory results (Lukaszuk et al., 2005; Fatemi et al., 2006; Engmann et al., 2007). Recently, the co-administration of a single dose GnRH agonist in the midluteal phase was reported to significantly increase implantation and live birth rates in women undergoing intracytoplasmic sperm injection (ICSI) and embryo transfer (ET) (Tesarik et al., 2006). The authors suggested a possible direct action of the drug on the embryo and proposed a novel situation where the early human embryos were drug targets. Despite stressing the importance of this finding, properly conducted trials to evaluate the reproducibility of these results were advised before widespread adoption of this simple yet seemingly effective strategy (Lambalk and Homburg, 2006). In this report, we present a double-blind, randomized, placebo-controlled trial aimed to asses the effect of a single 0.1 mg dose of triptorelin (Decapeptyl, Ferring GmbH Kiel, Germany), administered 6 days after ICSI, on the probability of ongoing pregnancy rates following ET in cycles stimulated with a long GnRH agonist protocol. 668 # The Author 2008. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

GnRH agonist in the luteal phase support Materials and Methods Study design and study population This was a double-blind, randomized placebo-controlled trial to test the hypothesis that a single 0.1 mg dose of triptorelin administered s.c. 6 days after ICSI would increase the ongoing pregnancy rate by 12% following ET in women stimulated with a long GnRH agonist protocol. The study was conducted between September 2006 and July 2007 in the assisted reproduction unit of the American Hospital of Istanbul. There were 570 consecutive patients undergoing ET who met the inclusion criteria and were included in the study. Each patient was included with only one cycle. The last patient was recruited in February 2007. The study protocol was approved by the Institutional Review Board and informed consent was obtained from all participants. Inclusion criteria (i) Couples undergoing ART with their own gametes (ii) Women stimulated with a long GnRH agonist protocol (iii) Couples having at least one embryo available for transfer. Exclusion criteria (i) Participation in another clinical trial that was being conducted in our unit at the same time (ii) Preimplantation genetic screening (PGS) cycles. Randomization protocol and data management Women were randomized according to a computer generated randomization list prepared by the chief investigator. Study subjects were randomized in blocks of 10; i.e. of every 10 subjects randomized, five were allocated to the GnRH agonist, and five were allocated to the placebo arms in a random manner. Opaque envelopes, which were numbered and sealed, containing the allocation information were given to the hospital pharmacy. Women meeting the inclusion criteria were enrolled by the assisted reproduction center nurse coordinator and invited for injection of the study medication 6 days after ICSI. The nurse coordinator informed the pharmacy regarding the number of patients that would be included in the study and received study medication every morning. Patients were assigned to the study groups in the hospital pharmacy. Syringes containing either triptorelin or sterile saline, according to the allocation information in the envelopes opened in an orderly fashion, were prepared in the pharmacy. Syringes were labeled with only the patient identification number, and sent to the assisted reproduction center nurse daily. Allocation information was kept in a file in the pharmacy. Study medications were administered to each woman by the assisted reproduction nurse, who matched the study subject with the syringe labeled with her patient identification number. Both the nurse injecting the study medication and women receiving injections were blinded for allocation. Patient records other than allocation information were kept in the assisted reproduction center. Hence, outcome assessors who performed the pregnancy tests and ultrasonographic examinations to determine if the patient was pregnant were also blinded for allocation. The allocation code was broken upon completion of the 20th gestational week of the last pregnant subject. Interventions Pituitary suppression was achieved with daily s.c. injections of triptorelin 0.1 mg (Decapeptyl, Ferring GmbH) starting on the 21st day of the preceding cycle. Triptorelin dose was reduced to 0.05 mg/day on commencement of menstrual bleeding and continued at the same dose until hcg (Pregnyl, Organon, Istanbul, Turkey) injection. Controlled ovarian stimulation was affected with recombinant human follicle stimulating hormone (rfsh) (Gonal-F, Serono, Bari, Italy) started on the third day of menstrual bleeding. The daily rfsh dose ranged between 150 and 300 IU, depending on body mass index and age of the women, and the anticipated ovarian response. Dose adjustment was done according to follicular development and serum estradiol levels. hcg 10.000 IU was administered intramuscularly when the leading follicle reached 20 mm in the mean diameter accompanied by 2 follicles.16 mm. Stimulation protocols and the indications for hcg injection or cycle cancellation did not change throughout the study period. Oocyte retrieval was undertaken 36 h after the administration of hcg. Fertilization was achieved with ICSI in all couples. ET was performed on Day 3. Cleavage stage embryos were graded according to Hardarson et al. (2001). A maximum of three embryos were transferred under ultrasound guidance using the Wallace catheter (Sims Portex Ltd., Hythe, Kent, UK). Six days after ICSI women allocated to GnRH agonist group were injected 0.1 mg triptorelin s.c. (Decapeptyl, Ferring GmbH) and women allocated to control group were injected 0.1 ml sterile saline s.c. All women were given 90 mg vaginal progesterone gel (Crinone 8%, Serono, Bedfordshire, UK) starting from the day of oocyte collection. LPS was continued until the pregnancy test performed 12 days after ET. Women with a positive pregnancy test continued the vaginal progesterone gel until the 10th week of gestation. Outcome measures Pregnancy was confirmed by measuring serum beta-hcg levels 12 days after ET. Clinical pregnancy was defined as the presence of a fetus with a heart beat at six weeks of gestation, multiple pregnancy was defined as a gestation with more than one fetus, and ongoing pregnancy was defined as pregnancy proceeding beyond the 20th gestational week. Implantation rate was calculated separately for each woman as number of gestational sacs divided by number of transferred embryos multiplied by 100. Statistical analysis Sample size calculation Ongoing pregnancy rate was the primary outcome measure. In a previous trial, it has been reported that a single dose of GnRH agonist injection in the midluteal phase of ICSI ET cycles increased ongoing pregnancy rate per ET from 38% to 46.8% (Tesarik et al., 2006). More than 1000 treatment cycles would be required to detect an absolute 8.8% increase from 38% with an alpha error level of 0.05 and beta error level of 0.2. This figure was not considered feasible for a single center trial which was planned to be completed in a year, and therefore the trial was designed to include 570 patients, which would enable detection of an absolute increase in ongoing pregnancy rate by 12% from the assumed 38% in the control group with the same alpha and beta error levels. 669

Ata et al. It should be noted this is a post hoc calculation and the difference of 12% was arbitrarily defined in order to complete the trial in a year. However, 12% was compatible with the 95% confidence interval (CI) of the difference between ongoing pregnancy rates per ET (22.67% to 20.27%) in the previous report (Tesarik et al., 2006). Statistical tests Efficacy analysis was done according to the intention-to-treat principle. Differences between the study groups were assessed with chi-squared test with Yates correction or Fisher s exact test for categorical variables, and independent samples t-test for continuous variables. For continuous and binary variables, as well as for differences between outcome variables, 95% CIs were calculated. Results The trial was completed as planned and the results are reported in accordance with the CONSORT statement. Patient enrolment was done on the day of ET; hence, there were no drop-outs due to inadequate ovarian response or fertilization failure; hence, all women (n ¼ 801) proceeded to ET. Two hundred and thirty-one women who did not meet the inclusion criteria or met the exclusion criteria were excluded. Finally, 570 women were enrolled and randomized. Three women in the GnRH agonist arm and four women in the control arm were lost to follow-up. These women were included in the final analysis in accordance with the intention to treat principle. They were assumed to be not pregnant. The number of women who were lost to follow-up was small with regard to overall study population, and was balanced between the groups. Consequently, this was not expected to skew the results in either way. The study flowchart is presented in Fig. 1. Two groups were comparable in terms of basic demographic characteristics and the features of ART cycles (Table I). Women in both groups received a similar number and quality of embryos (Table II). There were 89 (31.2%) ongoing pregnancies in the GnRH agonist group, and 84 (29.5%) in the control group (absolute difference þ1.7%, 95% CI 25.8% to 9.2%). There were 122 (42.8%) and 120 (42.1%) clinical pregnancies in the GnRH agonist and control groups, respectively (absolute difference þ0.7%, 95% CI, 27.4% to 8.8%). Multiple pregnancies constituted 32.8% of clinical pregnancies in the GnRH agonist group, whereas 30.8% of clinical pregnancies were multiples in the control group (absolute difference þ2%, 95% CI 29.7% to 13.7%). Implantation rates were 21.1% and 20.1% in the GnRH agonist and control groups, respectively (absolute difference þ1%, 95% CI 23.6% to 5.6%). Ongoing pregnancy, clinical pregnancy, implantation and multiple pregnancy rates were not significantly different between the groups (Table II). Figure 1: Flowchart of the study 670

GnRH agonist in the luteal phase support Table I. Patient and Cycle Characteristics. Parameter (Units) GnRH agonist Placebo Number of patients 285 285 Etiology of infertility (%, 95% CI) Ovulatory 23 (8, 5 12) 20 (7, 5 11) Tubal 25 (9, 6 13) 22 (8, 5 11) Male 137 (48, 42 54) 145 (51, 45 57) Unexplained 43 (15, 11 20) 52 (18, 14 23) More than one factor 57 (20, 16 25) 46 (16, 12 21) Age (years) a 31.3 (30.7 31.9) 31.5 (30.9 32.1) Body mass index (kg/m2) a 23.7 (23.1 24.1) 24.4 (23.9 25) Duration of COH (days) a 9.9 (9.7 10.1) 9.8 (9.6 10) Total gonadotropin dose (IU) a 3005.7 (2861 3150) 2804 (2648 2960) Mean number of oocytes a 9.5 (9 10) 10.2 (9.7 10.7) Mean number of MII Oocytes a 6.9 (6.5 7.3) 7 (6.7 7.3) Mean number of 2PN zygotes a 5.2 (4.9 5.5) 5.3 (5 5.6) Fertilization rate (%) a 77.8 (75.9 79.7) 77.9 (75.6 80.3) a Values are mean (95% CI). Discussion The present double-blind, randomized, placebo controlled study was performed to assess the effect of a single 0.1 mg dose of triptorelin, administered 6 days after ICSI, on the probability of ongoing pregnancy rate following ET in women stimulated with a long GnRH agonist protocol. Although ongoing pregnancy, clinical pregnancy and multiple pregnancy rates were 1% higher in the GnRH agonist group, when compared with controls, the observed differences were short of reaching statistical significance. However, the trial was powered to detect an absolute 12% increase in ongoing pregnancy rate from an assumed 38% in the control group. Hence, differences,12% in absolute value cannot be excluded, since the present trial was not adequately powered to detect such differences. The 95% CI of the difference between ongoing pregnancy rates does not exclude differences between 5% and 9%, which means there s a probability of single dose GnRH agonist administration increasing the ongoing pregnancy rate by 5%, and this would be regarded as clinically significant. On the other hand, in the general population of women stimulated with the long protocol, a negative effect of the intervention tested cannot be refuted as well. The use of GnRH analogs in the context of LPS has been investigated previously by two separate groups of researchers (Tesarik et al., 2004, 2006; Pirard et al., 2005, 2006). Whether the use of GnRH analogs alone could provide adequate luteal support in assisted reproduction cycles was investigated by Pirard et al., whereas Tesarik et al. analyzed the effects of GnRH analog administered as an adjunct to progesterone and estradiol valerate during the luteal phase on implantation and live birth rates. Tesarik et al. (2004) investigated whether 0.1 mg triptorelin injection 6 days after ICSI to recipients in an oocyte donation program increased implantation rates. The form of luteal support was not mentioned in the article. The authors reported significantly increased implantation rates with GnRH agonist administration when compared with placebo. However, the clinical pregnancy rate was not found to be significantly increased in the GnRH agonist group. The same group reported similar results in a subsequent trial that included women undergoing ICSI ET with their own gametes (Tesarik et al., 2006). Subjects were randomly allocated to receive either 0.1 mg triptorelin or placebo 6 days after ICSI. Women in all groups received luteal support in the form of recombinant hcg (rhcg) injection, vaginal micronized progesterone and estradiol valerate tablets in addition to the study medication. Implantation and live birth rates per transferred embryo were significantly increased by GnRH agonist administration in women stimulated either with the long protocol or with a GnRH antagonist. Although the mechanism underlying the proposed favorable effect of GnRH agonist administration in the luteal phase was not clear, the authors speculated that it might have been due to effects on the embryo, endometrium and/or, unlike the donation model in the first trial, on the corpora lutea. Embryonic effects were thought to be more pronounced as implantation rates were increased more significantly than the pregnancy rates. There are several points that raise concern in the analysis of the data presented in the previous studies by Tesarik et al. (2004, 2006). In both trials, sample size analyses were based on embryo implantation rates, and as the women were randomized rather than the embryos, study designs introduced a unit-of-analysis error, which could account for the observed discrepancies. Moreover, unlike live birth rate or ongoing pregnancy rate, implantation rate is an intermediate outcome of ART (Arce et al., 2005). Implantation rate as an intermediate outcome may help to interpret primary outcome results, but it is not appropriate to form the basis of sample size calculation on implantation rates in an efficacy trial in the field of ART. The a priori assumptions made by the authors also raise concern about the sample size analysis in both trials (Tesarik et al., 2004, 2006). Both trials were powered to detect an Table II. Outcome measures. Parameter GnRH agonist Placebo P value Number of transfer cycles 285 285 Mean number of transferred embryos per transfer a 2.65 (2.58 2.72) 2.78 (2.72 2.84) 0.09 Mean number of good quality embryos transferred per transfer a 2.37 (2.26 2.48) 2.51 (2.42 2.60) 0.21 Clinical pregnancy rate (%, 95% CI) 122/285 (42.8, 37 49) 120/285 (42.1, 37 48) 0.86 Multiple pregnancy rate (%, 95% CI) 40/122 (32.8,25 42) 37/120 (30.8, 23 40) 0.74 Implantation rate (%) a 21.1 (17.8 24.4) 20.1 (16.9 23.2) 0.67 Ongoing pregnancy rate (%, 95% CI) 89/285 (31.2, 26 37) 84/285 (29.5, 24 35) 0.65 a Values are mean (95% CI). 671

Ata et al. increase in implantation rates from 25% to 35% which corresponds to an actual 40% increase. This would at best be described as an optimistic assumption. Overoptimistic assumptions in the sample size analysis result in recruitment of relatively small numbers of subjects in clinical trials. Although the major problem with less than adequately sized trials is failure to demonstrate, as statistically significant, a treatment difference which in fact is present, yet another important problem is observation of an exaggerated treatment effect which could be decreased once the appropriate sample size has been reached. Additionally, implantation rates have been calculated as total number of gestational sacs with a heart beat divided by the total number of transferred embryos, a method which does not address multiple implantations in a patient (Tesarik et al., 2004, 2006). Results could have been different had the authors calculated implantation rates per patient and compared mean implantation rates in each group as continuous variables. Live birth rates have been calculated in an unconventional way as well; the authors have divided the total number of live born babies by the total number of transferred embryos (Tesarik et al., 2004, 2006). When live birth rates are calculated per patient undergoing ET after correction for multiple births, in the long protocol arm of the study the figures are 66/150 and 54/150 in the GnRH agonist and placebo groups, respectively, and the difference is not statistically significant (P ¼ 0.15) (Tesarik et al., 2006). Finally, contrary to the authors conclusion, in women stimulated with the long protocol, a beneficial effect of a single dose GnRH agonist administration in the luteal phase, with regard to clinical pregnancy, ongoing pregnancy and live birth rates was not demonstrated by their previous trial (Tesarik et al., 2006). With regard to absolute values, our results differ widely from previous results reported by Tesarik et al., however after considering significance tests, the findings are consistent; none of the trials demonstrated a significant improvement of the relevant outcome measures such as ongoing pregnancy and delivery rates (Tesarik et al., 2004, 2006). Although, this discrepancy in absolute values may be a reflection of sampling differences, there are other points that need to be considered. The study population in our trial consisted of consecutive women stimulated with a long GnRH agonist protocol, rather than different protocols determined at physicians discretion. Unlike the trial conducted by Tesarik et al. rhcg injections during the luteal phase were not administered in our study (Tesarik et al., 2006). Supporting the luteal phase with rhcg in addition to progesterone and estradiol is not a widely adopted strategy, and such an approach would limit the external validity of our results. There is no evidence to suggest that the effect of GnRH agonist would be enhanced by rhcg injections, so it is not plausible that the lack of significant differences within the study groups in our present trial, or that the differences between overall outcome measures between the two trials, is due to withholding rhcg injections. Lower ongoing pregnancy rates in our study population when compared with Tesarik et al. s may be due to differences in definitions of ongoing pregnancy. We defined an ongoing pregnancy as a pregnancy proceeding beyond 20 completed 672 weeks of gestation, in order to exclude first trimester abortions, and to more closely approximate term live birth rates. Neither the definition of an ongoing pregnancy nor the timing of live births was specified in the report by Tesarik et al. (2006). There may be several explanations for the ineffectiveness of a single dose of triptorelin administered during the luteal phase, including the continuing down-regulated state of the GnRH receptors in reproductive organs. The period between the last GnRH agonist injection in the long protocol and the subsequent 0.1 mg injection in the luteal phase was 8 days in our study. In women stimulated with a long GnRH agonist protocol, downregulation of GnRH receptors may not be totally abated 8 days after cessation of GnRH agonist injections (Smitz et al., 1988; Broekmans et al., 1992; DiLuigi and Nulsen, 2007). Pirard et al. (2006) investigated the effect of buserelin on the luteal phase in women stimulated with a GnRH antagonist protocol. Although recovery after the GnRH antagonist suppression is expected to occur more rapidly, when compared with recovery after the GnRH agonist down-regulation in a long protocol, even daily administration of 0.1 mg buserelin was not found to adequately support the luteal phase. Median serum progesterone levels in women receiving daily 0.1 mg buserelin injections did not reach 20 ng/ml at any point in time (Pirard et al., 2006). Consequently, Pirard et al. s findings do not suggest that even daily 0.1 mg GnRH agonist injections would not have a significant effect on corpus luteum function. GnRH receptors have been shown to be present in the human uterus (Reshef et al., 1990). Although there are molecular studies suggesting a direct effect of GnRH agonists on endometrial function, research has failed to show any clinically relevant effect (Takeuchi et al., 1998; Fauser and Devroey, 2005). Additionally, prolonged GnRH agonist administration can be expected to cause down-regulation of endometrial GnRH receptors similar to pituitary receptors, and render them, at least relatively, refractory to a subsequent single dose administered 1 week after down-regulation. There are several reports suggesting that GnRH agonists can enhance embryo development in vitro. GnRH agonist added to the culture medium was found to support in vitro development of porcine and murine embryos in a dose dependent fashion, and the presence of GnRH and GnRH receptor mrnas were demonstrated in preimplantation human embryos (Casan et al., 1999; Raga et al., 1999; Nam et al., 2005). However, a direct effect on embryos of GnRH agonist administered subcutaneously to the mother is highly unlikely. Unlike in vitro culture conditions, the GnRH agonist either needs to pass to the endometrial cavity to directly affect embryos that have not attached to the decidua, or must be present in the maternal circulation when the early uteroplacental circulation is established. Establishment of uteroplacental circulation requires invasion of the endometrium, arterioles and spiral arteries by the syncytiotrophoblast, and formation of lacunae within the syncytiotrophoblast. Formation of the lacunae is reported to occur 8 days after fertilization in a naturally occurring pregnancy (Moore and Persaud, 2003). This corresponds to the eighth day post-icsi or insemination in an assisted reproduction cycle. Provided that the half-life of triptorelin is 3 4 h, it is difficult to suggest any systemic effect at the time of

GnRH agonist in the luteal phase support establishment of uteroplacental circulation (Shoham, 2004). However, any indirect action of triptorelin on embryos or on embryo endometrium cross-talk cannot be completely refuted. In conclusion, our results fail to demonstrate an unequivocal beneficial effect of a single 0.1 mg dose of triptorelin, administered 6 days after ICSI, with regard to ongoing pregnancy rates in women stimulated with the long GnRH protocol. The observed 1.7% increase in ongoing pregnancy rates could have reached statistical significance had the sample size been larger, but from a clinical point of view it is questionable whether this would be regarded significant. The idea of significantly increasing pregnancy rates with such a simple and relatively inexpensive method is alluring, but to the best of our knowledge the available data in the literature, including the present trial, do not support this theory yet. In our opinion, this strategy should not be recommended as routine practice unless its efficacy and safety, especially with regard to developing embryos, are supported by further trials. Author s Contribution Design and institution of the study protocol, collection and analysis of data, drafting the article, final preparation of the article and approval of the final version B.A. Review of the article, approval of the final version K.Y. Institution of the study protocol, review of the article, approval of the final version B.B. Design of the study protocol, review and final preparation of the article, approval of the final version B.U. References Arce JC, Nyboe Andersen A, Collins J. Resolving methodological and clinical issues in the design of efficacy trials in assisted reproductive technologies: a mini-review. Hum Reprod 2005;20:1757 1771. Broekmans FJ, Bernardus RE, Berkhout G, Schoemaker J. Pituitary and ovarian suppression after early follicular and mid-luteal administration of a LHRH agonist in a depot formulation: decapeptyl CR. Gynecol Endocrinol 1992;6:153 161. Casan EM, Raga F, Polan ML. GnRH mrna and protein expression in human preimplantation embryos. Mol Hum Reprod 1999;5:234 239. Daya S, Gunby J. Luteal phase support in assisted reproduction cycles. Cochrane Database Syst Rev 2004, CD004830. DiLuigi AJ, Nulsen JC. Effects of gonadotropin-releasing hormone agonists and antagonists on luteal function. Curr Opin Obstet Gynecol 2007;19:258 265. Engmann L, Diluigi A, Schmidt D, Benadiva C, Maier D, Nulsen J. The effect of luteal phase vaginal estradiol supplementation on the success of in vitro fertilization treatment: a prospective randomized study. Fertil Steril 2007, in press. Fatemi HM, Kolibianakis EM, Camus M, Tournaye H, Donoso P, Papanikolaou E, Devroey P. Addition of estradiol to progesterone for luteal supplementation in patients stimulated with GnRH antagonist/rfsh for IVF: a randomized controlled trial. Hum Reprod 2006;21:2628 2632. Fauser BC, Devroey P. Reproductive biology and IVF: ovarian stimulation and luteal phase consequences. Trends Endocrinol Metab 2003;14:236 242. Fauser BC, Devroey P. Why is the clinical acceptance of gonadotropinreleasing hormone antagonist cotreatment during ovarian hyperstimulation for in vitro fertilization so slow? Fertil Steril 2005;83:1607 1611. Garcia J, Jones GS, Acosta AA, Wright GL Jr. Corpus luteum function after follicle aspiration for oocyte retrieval. Fertil Steril 1981;36: 565 572. Hardarson T, Hanson C, Sjogren A, Lundin K. Human embryos with unevenly sized blastomeres have lower pregnancy and implantation rates: indications for aneuploidy and multinucleation. Hum Reprod 2001;16:313 318. Lambalk CB, Homburg R. GnRH agonist for luteal support in IVF? Setting the balance between enthusiasm and caution. Hum Reprod 2006;21: 2580 2582. Lukaszuk K, Liss J, Lukaszuk M, Maj B. Optimization of estradiol supplementation during the luteal phase improves the pregnancy rate in women undergoing in vitro fertilization embryo transfer cycles. Fertil Steril 2005;83:1372 1376. Moore K, Persaud T. Introduction to the developing human. In: Moore K, Persaud T (eds). The Developing Human Clinically Oriented Embryology. Philadelphia: Saunders, 2003, 4. Nam DH, Lee SH, Kim HS, Lee GS, Jeong YW, Kim S, Kim JH, Kang SK, Lee BC, Hwang WS. The role of gonadotropin-releasing hormone (GnRH) and its receptor in development of porcine preimplantation embryos derived from in vitro fertilization. Theriogenology 2005;63:190 201. Pirard C, Donnez J, Loumaye E. GnRH agonist as novel luteal support: results of a randomized, parallel group, feasibility study using intranasal administration of buserelin. Hum Reprod 2005;20:1798 1804. 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. Raga F, Casan EM, Kruessel J, Wen Y, Bonilla-Musoles F, Polan ML. The role of gonadotropin-releasing hormone in murine preimplantation embryonic development. Endocrinology 1999;140:3705 3712. Reshef E, Lei ZM, Rao CV, Pridham DD, Chegini N, Luborsky JL. The presence of gonadotropin receptors in nonpregnant human uterus, human placenta, fetal membranes, and decidua. J Clin Endocrinol Metab 1990;70:421 430. Shoham Z. Drugs used for controlled ovarian hyperstimulation: clomiphene citrate, aromatase inhibitors, gonadotropins, and gonadotropin-releasing hormone analogs. In: Gardner D, Weissman A, Howles C, Shoham Z (eds). Textbook of Assisted Reproduction Techniques. Laboratory and Clinical Perspectives. London: Taylor & Francis, 2004, 532. Smitz J, Devroey P, Camus M, Deschacht J, Khan I, Staessen C, Van Waesberghe L, Wisanto A, Van Steirteghem AC. The luteal phase and early pregnancy after combined GnRH-agonist/HMG treatment for superovulation in IVF or GIFT. Hum Reprod 1988;3:585 590. Takeuchi S, Futamura N, Minoura H, Toyoda N. Possible direct effect of gonadotropin releasing hormone on human endometrium and decidua. Life Sci 1998;62:1187 1194. Tesarik J, Hazout A, Mendoza C. Enhancement of embryo developmental potential by a single administration of GnRH agonist at the time of implantation. Hum Reprod 2004;19:1176 1180. Tesarik J, Hazout A, Mendoza-Tesarik R, Mendoza N, Mendoza C. Beneficial effect of luteal-phase GnRH agonist administration on embryo implantation after ICSI in both GnRH agonist- and antagonist-treated ovarian stimulation cycles. Hum Reprod 2006;21:2572 2579. Submitted on July 25, 2007; resubmitted on December 1, 2007; accepted on December 14, 2007 673