Jan Tesarik 1,2,5, André Hazout 3, Raquel Mendoza-Tesarik 1, Nicolas Mendoza 1 and Carmen Mendoza 1,4

Transcription

1 Human Reproduction Vol.21, No.10 pp , 2006 Advance Access publication August 22, doi: /humrep/del173 Beneficial effect of luteal-phase administration on embryo implantation after ICSI in both - and antagonist-treated ovarian stimulation cycles Jan Tesarik 1,2,5, André Hazout 3, Raquel Mendoza-Tesarik 1, Nicolas Mendoza 1 and Carmen Mendoza 1,4 1 MAR&Gen, Molecular Assisted Reproduction and Genetics, Granada, Spain, 2 Laboratoire d Eylau, 3 ARCEFAR, Paris, France and 4 Department of Biochemistry and Molecular Biology, Faculty of Sciences, University of Granada, Campus Universitario Funetenueva, Granada, Spain 5 To whom correspondence should be addressed at: cmendoza@ugr.es BACKGROUND: was recently suggested as a novel luteal-phase support that may act at different levels, including the pituitary gonadotrophs, the endometrium and the embryo itself. This prospective randomized study evaluates the effect of administered in the luteal phase on ICSI outcomes in both - and GnRH antagonist-treated ovarian stimulation protocols. METHODS: Six hundred women about to undergo ovarian stimulation for ICSI (300 using a long protocol and 300 using a GnRH antagonist protocol) were enrolled in this study. Patients treated with each of these two protocols were randomly assigned to receive a single injection of or placebo 6 days after ICSI. Implantation and live birth rates were the primary outcomes. RESULTS: Administration of 0.1 mg of triptorelin on day 6 after ICSI led to a significant improvement of implantation and live birth rates after ICSI as compared with placebo. In GnRH antagonist-treated ovarian stimulation cycles, luteal-phase also increased ongoing pregnancy rate. Moreover, luteal-phase GnRH agonist administration increased luteal-phase serum HCG, estradiol and progesterone concentrations in both ovarian stimulation regimens. CONCLUSIONS: administration enhances ICSI clinical outcomes after - and GnRH antagonist-treated ovarian stimulation cycles, possibly by a combination of effects on the embryo and the corpus luteum. Key words: corpus luteum function/embryo developmental potential//gnrh antagonist/luteal-phase support Introduction deficiency is a common problem in current assisted reproduction techniques (ARTs) and has been described in cycles using pituitary down-regulation with a as well as in those using GnRH antagonists (Macklon and Fauser, 2000; Pritts and Atwood, 2002; Beckers et al., 2003; Kolibianakis et al., 2003). To cope with this problem, different regimens of luteal-phase support were suggested (reviewed in Pritts and Atwood, 2002). Most of these regimens involve different application forms and doses of three therapeutic agents, HCG, vaginal or intramuscular progesterone and oral estradiol (E 2 ) (Pritts and Atwood, 2002). Some recent data, however, have suggested a beneficial effect of GnRH agonist administered in the luteal phase on ART outcomes (Tesarik et al., 2004; Pirard et al., 2005). The mechanism of the presumed beneficial effect of lutealphase administration is not clear and may be due to the drug action at multiple levels. It was hypothesized that may support the corpus luteum by stimulating the secretion of LH by pituitary gonadotroph cells or by acting directly on the endometrium through the locally expressed GnRH receptors (Pirard et al., 2005). However, the administration of a single dose of in the luteal phase was also shown to increase pregnancy, implantation, delivery and birth rates in recipients of donated oocytes in whom ovulation was suppressed and the corpus luteum was thus absent, suggesting a direct effect of on the embryo (Tesarik et al., 2004). To help clarify the mechanism of action in the luteal phase, this prospective randomized study evaluates the effect of luteal-phase administration on ICSI outcomes in the two most currently used ovarian stimulation protocols using, respectively, and GnRH antagonist for the prevention of a premature LH surge. Implantation 2572 The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 for luteal-phase support and live birth rates were the main outcome measures. Clinical pregnancy and ongoing pregnancy rates and luteal-phase serum concentrations of HCG (in conception cycles), LH and progesterone were secondary outcome measures. Materials and methods Study design and power calculation This prospectively randomized study was conducted between September 2003 and September It involves 600 infertile couples treated by ICSI and given two different protocols of ovarian stimulation, one using started in the luteal phase of the preceding cycle for pituitary down-regulation (300 couples) and the other using a GnRH antagonist for the prevention of a premature LH surge (300 couples). Patients to be treated by each of the two protocols were randomly assigned to two groups. In one group, women were given a single dose of 6 days after ICSI (3 days after embryo transfer). Women in the other group received placebo at the same time after ICSI. The decision of which couples would be treated by each of the two different ovarian stimulation protocols was subjective, and the study is thus not randomized as to comparison of ICSI outcomes with ovarian stimulation protocols using and GnRH antagonist, respectively, and the results can only serve to evaluate the usefulness of luteal-phase administration in each of the two protocols. Implantation and live birth rates were chosen as the main outcome measures. On the basis of previously published data (Tesarik et al., 2004) showing that the administration of 6 days after ICSI to recipients of donated oocytes increases implantation rate from 24.7% to 36.2%, it was hypothesized that a similar increase can be achieved with this method in ICSI patients undergoing ovarian stimulation. Assuming a significance level of 0.05 and a power of 0.80, it was calculated that some 100 treatment cycles were needed in each branch to detect this increase. It was decided to increase the number of cycles enrolled in each group to 150 to obtain a better chance to detect significant differences in some of the secondary outcome measures of this study; these were pregnancy and delivery rates, and luteal-phase serum concentrations of HCG (in conception cycles), LH and progesterone. The study protocol was approved by a competent institutional review board, and all participants signed an informed consent. Randomization protocol Couples selected for an ICSI attempt using each of the two ovarian stimulation protocols were randomized between a luteal-phase GnRH agonist group and a placebo group. Randomization was done with the use of a computer-generated randomization list. Sealed envelopes with treatment allocation instructions were opened on the day of embryo transfer by a nurse who assigned participants to their groups. The doctor and the biological team performing the ART were blinded to group assignment. Participants Six hundred infertile couples were enrolled in this study. Exclusion criteria were female age of >40 years and non-obstructive azoospermia requiring testicular sperm retrieval. Three hundred couples were assigned to an ovarian stimulation protocol using long and 300 to a protocol using GnRH antagonist. The decision between each of these two protocols was subjective and depended on the clinical context. Usually, the GnRH antagonist protocol was chosen in women presenting symptoms of polycystic ovaries and in those with a history of ovarian hyperstimulation syndrome (OHSS), but also in some women with expected good ovarian response without these features with the aim to reduce treatment cost and patient discomfort. The protocol was mostly, but not exclusively, used in women with expected normal or low ovarian response. Ovarian stimulation Long protocol Pituitary down-regulation was achieved with the use of daily injections of 0.1 mg of triptorelin (Decapeptyl 0.1 mg; Ipsen Pharma, Barcelona, Spain) starting in the luteal phase of the cycle preceding ovarian stimulation and reduced to 0.05 mg daily dose from the first day of vaginal bleeding. The gonadotrophin therapy was started 2 7 days after the beginning of the bleeding. This therapy was carried out with the use of recombinant human FSH (Puregon; Organon, Oss, The Netherlands or Gonal F; Serono, Rome, Italy) and HMG (Menopur; Ferring Pharmaceuticals, Langley, Berkshire, UK). The doses of FSH and HMG were continuously adapted according to serum concentrations of E 2 and LH and to the dynamics of ovarian follicular growth as described (Tesarik and Mendoza, 2002) with the aim to maintain serum LH concentration in the range of IU l 1 throughout the stimulation period. Ovulation was induced with 250 μg of human recombinant HCG (Ovitrelle, Serono) when at least three follicles reached a mean diameter of 18 mm. The treatment with triptorelin was stopped on the day of HCG administration (last triptorelin injection on the day preceding HCG administration). GnRH antagonist protocol Recombinant human FSH and HMG administration was started on day 2 of menstrual bleeding following withdrawal of a contraceptive pill (Tri-Minulet; Wyeth Lederle, Madrid, Spain) administered during the cycle preceding ovarian stimulation. The doses of FSH and HMG were continuously adapted according to serum concentrations of E 2 and LH and to the dynamics of ovarian follicular growth in the same way as in the long protocol (discussed above). GnRH antagonist (Orgalutran; Organon or Cetrotide; Serono) was started on day 5 of gonadotrophin therapy and continued, at a daily dose of 0.25 mg, until at least three follicles reached a diameter of 18 mm when ovulation was induced with 250 μg of human recombinant HCG (Ovitrelle). ARTs Mature (metaphase II) oocytes were incubated at 37 C in G-FERT medium (Vitrolife, Göteborg, Sweden) equilibrated with 6% CO 2 in air for 3 6 h. They were then freed from the cumulus oophorus and corona radiata cells, subjected to ICSI as described (Tesarik and Mendoza, 2002) and incubated in G1 medium (Vitrolife) under the same gas phase. Fertilization was assessed h after ICSI. Zygotes were scored as having good morphology or poor morphology (see section Embryo selection for transfer ). After a medium change, embryos were cultured further in G1 medium until day 3 after ICSI with a medium change every day. At the time of medium change on day 2, embryo quality was assessed again (see section Embryo selection for transfer ). Embryo transfer was performed 3 days after ICSI with the use of K-JETS-7019-SIVF embryo transfer set (Cook, Queensland, Australia) and under transabdominal ultrasound guidance. One to three embryos were transferred in each cycle. If suitable, supernumerary embryos were frozen for later transfer. This study, however, only deals with fresh embryo transfers. Embryo selection for transfer Zygotes were evaluated according to criteria based on the assessment of the number and distribution of nucleolar precursor bodies in the pronuclei as described previously (Tesarik and Greco, 1999). The 2573

3 J.Tesarik et al. scoring system was simplified by grouping together all abnormal patterns (Tesarik et al., 2000), so as all zygotes were allocated to one of two groups good-morphology zygotes and poor-morphology zygotes. Cleaving embryos were evaluated twice, first at the time of medium change 2 days after ICSI and second min before transfer to the uterus 3 days after ICSI as described (Tesarik and Greco, 1999). Briefly, embryos that had more than three cells 2 days after ICSI and more than 6 cells 3 days after ICSI, that showed <10% of their space occupied by cell fragments and that developed from good-morphology zygotes were considered to have good morphology and were selected for transfer where available. supplementation Irrespective of whether was used as luteal-phase support or not, all women were given 4 mg of E 2 valerate (Progynova; Schering, Madrid, Spain) daily and 400 mg of vaginal micronized progesterone (Utrogestan; Laboratoires Besins-Iscovesco, Paris, France) daily from the day of oocyte recovery for 17 days. In addition, all women received an injection of 250 μg of human recombinant HCG (Ovitrelle) on the day of embryo transfer. Hormone assays Serum concentrations of 17β-E 2, progesterone and β-hcg were determined with the use of commercial enzyme immunoassay kits (Diagnostic Product Corporation, Los Angeles, CA, USA). The intraand inter-assay coefficients of variation (CVs) were, respectively, 3.8 and 5.6% for E 2, 3.7 and 4.0% for progesterone, and 3.5 and 4.1% for β-hcg. All measurements were taken according to the manufacturer s instructions. Statistics Differences between groups were assessed by two-tailed χ 2 -test with Yates correction or Fisher s exact test for categorical variables, and by Mann Whitney U-test or Welch t-test for continuous variables. The level of significance was set to P < Efficacy analyses were performed on an intention-to-treat (ITT) basis as appropriate, taking into account all randomized patients. Multidimensional statistics were performed using multivariate logistic regression analysis. All analyses were performed using the Statistica 5.0 package (Statsoft Version 5.1, Hamburg, Germany). Results Patient flow through the trial The trial was performed according to CONSORT guidelines. Infertile couples continued to be assessed for eligibility study until 150 patients were eligible in both the treatment and the placebo arm in each of the two ovarian stimulation protocols. Among the patients in whom the long ovarian stimulation protocol was indicated, 13 did not meet the inclusion criteria and eight refused to participate in the study (Figure 1A). Among the patients in whom the GnRH antagonist ovarian stimulation protocol was indicated, 11 did not meet the inclusion criteria and seven refused to participate in the study (Figure 1B). Among the patients treated with the long stimulation protocol and assigned to the luteal-phase GnRH agonist group, two discontinued the ovarian stimulation and three had no embryos for transfer. Of the remaining 145 patients who received allocated intervention, one was lost to 2574 follow-up, three were excluded from analysis because of failure to follow correctly the protocol of luteal-phase examinations, and 141 terminated the study and were analysed (Figure 1A). Among the patients treated with the long stimulation protocol and assigned to the placebo group, two were cancelled because of inadequate response to ovarian stimulation and five had no embryos for transfer. Of the remaining 143 patients who received allocated intervention, one was excluded from analysis because of failure to follow correctly the protocol of luteal-phase examinations, and 142 terminated the study and were analysed (Figure 1A). Among the patients treated with the GnRH antagonist stimulation protocol and assigned to the luteal-phase group, four had no embryos for transfer. Of the remaining 146 patients who received allocated intervention, one was excluded from analysis because of failure to follow correctly the protocol of luteal-phase examinations, and 145 terminated the study and were analysed (Figure 1B). Among the patients treated with the GnRH antagonist stimulation protocol and assigned to the placebo group, one was cancelled because of inadequate response to ovarian stimulation and four had no embryos for transfer. Of the remaining 145 patients who received allocated intervention, one was excluded from analysis because of failure to follow correctly the protocol of luteal-phase examinations, and 144 terminated the study and were analysed (Figure 1B). Effects of luteal-phase administration in the long ovarian stimulation protocol The patients in the luteal-phase and placebo groups did not differ in their basic demographical characteristics, and the features of their ovulation stimulation cycles were also comparable (Table I). The concentrations of E 2, determined on day 7 after ICSI in both groups, did not differ significantly (Table II), although there was a strong trend towards higher values in the group (P = 0.053). However, serum concentration of E 2, determined on day 15 after ICSI, and that of progesterone, determined on both days 7 and 15 after ICSI, were higher in the luteal-phase group as compared with the placebo (Table II). Moreover, the patients given 6 days after ICSI, who subsequently developed a clinical pregnancy, had a higher serum β-hcg concentration on day 15 after ICSI as compared with pregnant patients in the placebo group, and this difference was also evident when only those patients in whom a single gestational sac was subsequently detected were taken into account (Table II). With a similar number and morphology of embryos transferred, patients in the luteal-phase group had higher implantation and live birth rates than those in the placebo group, whereas the differences in the clinical pregnancy and ongoing pregnancy rates were not significant (Table III). There were 22 multiple births (21 twin and 1 triplet) in the group as compared with only five multiple births (4 twin and 1 triplet) in the placebo group. Using multivariate analysis, we observed that the female age and the number of oocytes recovered were independent predictors of

4 for luteal-phase support A Assessed for eligibility (n = 321) Excluded (n = 21) - Not meeting inclusion criteria (n = 13) - Refused to participate (n = 8) - Other reasons (n = 0) Randomized (n = 300) GnRHa Allocated to intervention (n = 150) - Received allocated intervention (n = 145) - Did not receive allocated intervention (n = 5) Allocated to intervention (n = 150) - Received allocated intervention (n = 143) - Did not receive allocated intervention (n = 7) - Lost to follow up (n = 1) - Discontinued intervention (n = 0) - Lost to follow up (n = 0) - Discontinued intervention (n = 0) Analysed (n = 141) - Excluded from analysis (n = 3) Analysed (n = 142) - Excluded from analysis (n = 1) B Assessed for eligibility (n = 318) Excluded (n = 18) - Not meeting inclusion criteria (n = 11) - Refused to participate (n = 7) - Other reasons (n = 0) Randomized (n = 300) GnRHa Allocated to intervention (n = 150) - Received allocated intervention (n = 146) - Did not receive allocated intervention (n = 4) Allocated to intervention (n = 150) - Received allocated intervention (n = 145) - Did not receive allocated intervention (n = 5) - Lost to follow up (n = 0) - Discontinued intervention (n = 0) - Lost to follow up (n = 0) - Discontinued intervention (n = 0) Analysed (n = 145) - Excluded from analysis (n = 1) Analysed (n = 144) - Excluded from analysis (n = 1) Figure 1. Patient flow through the randomized trial. (A) Patients with ovarian stimulation protocol using. (B) Patients with ovarian stimulation protocol using GnRH antagonist. 2575

5 J.Tesarik et al. Table I. Basic demographic and ovarian stimulation cycle characteristics of patients treated with the long ovarian stimulation protocol Characteristics Patient group a Age (years) 34.8 ± ± 1.7 BMI (kg m 2 ) 23.9 ± ± 1.9 Duration of FSH therapy (days) 11.8 ± ± 0.6 Total amount of FSH given (IU) 2488 ± ± 118 Total amount of HMG given (IU) 477 ± ± 34 Peak serum E 2 (pg ml 1 ) 1745 ± ± 190 Number of oocytes recovered b 12.4 ± ± 4.5 Values are mean ± SD. The differences between the two groups are not significant (P > 0.05). a Intention-to-treat group (n = 300). b Per stimulated cycle. Table II. characteristics of patients treated with the long GnRH agonist ovarian stimulation protocol implantation and ongoing pregnancy in both the and the placebo groups (data not shown). Effects of luteal-phase administration in the GnRH antagonist ovarian stimulation protocol The patients in the luteal-phase and placebo groups did not differ in their basic demographical characteristics, and the features of their ovulation stimulation cycles were also comparable (Table IV). However, the concentrations of E 2 and progesterone were higher in the luteal-phase group than in the placebo group, both on days 7 and 15 after ICSI (Table V). Moreover, the patients given 6 days after ICSI, who subsequently developed a clinical pregnancy, had a higher serum β-hcg concentration on day 15 after ICSI as compared with pregnant patients in the placebo group, and this difference was also evident when only those patients in whom a single gestational sac was subsequently detected were taken into account (Table V). With a similar number and morphology of embryos transferred, patients in the luteal-phase group had Characteristics Patient group a Table IV. Basic demographic and ovarian stimulation cycle characteristics of patients treated with the GnRH antagonist ovarian stimulation protocol Serum estradiol (pg ml 1 ) Day 7 after ICSI 432 ± ± 58 Day 15 after ICSI 480 ± 74 c 462 ± 71 Serum progesterone (ng ml 1 ) Day 7 after ICSI 44 ± 5 c 39 ± 5 Day 15 after ICSI 47 ± 7 c 43 ± 6 Serum HCG (IU l 1 ) Day 15 after ICSI In all conception cycles b 66 ± 8 c 42 ± 7 In singleton pregnancies 53 ± 6 c 34 ± 5 Values are mean ± SD. a Group of patients having terminated the study (n = 283). b Only cycles that resulted in a clinical pregnancy are included. c Significantly different from the placebo group (P < 0.05). Characteristics Patient group a Age (years) 31.3 ± ± 2.2 BMI (kg m 2 ) 23.8 ± ± 2.4 Duration of FSH therapy (days) 11.0 ± ± 0.6 Total amount of FSH given (IU) 2150 ± ± 172 Total amount of HMG given (IU) 282 ± ± 23 Peak serum E 2 (pg ml 1 ) 1931 ± ± 215 Number of oocytes recovered b 12.1 ± ± 4.2 Values are mean ± SD. The differences between the two groups are not significant (P > 0.05). a Intention-to-treat group (n = 300). b Per stimulated cycle. Table III. Clinical outcomes of patients treated with the long ovarian stimulation protocol Outcome variable Patient group a Mean ± SD (median). b Significantly different from the placebo group (P < 0.05). Intention to treat Transfer procedures Embryos transferred Embryos per transfer a 2.3 ± 0.5 (2.0) 2.3 ± 0.5 (2.0) Good-morphology 2.0 ± ± 0.5 (2.0) embryos per transfer a Clinical pregnancy rate Per embryo transfer 51.1% (72/141) 41.5% (59/142) Per intention to treat 48.0% (72/150) 39.3% (59/15 Clinical implantation rate 29.8% (97/325) b 18.2% (60/330) Ongoing pregnancy rate Per embryo transfer 46.8% (66/141) 38.0% (54/142) Per intention to treat 44.0% (66/150) 36.0% (54/150) Live birth rate 27.4% (89/325) b 18.2% (60/330) Table V. characteristics of patients treated with the GnRH antagonist ovarian stimulation protocol Characteristics Patient group a Values are mean ± SD. a Group of patients having terminated the study (n = 289). b Only cycle that resulted in a clinical pregnancy are included. c Significantly different from the placebo group (P < 0.05). Serum estradiol (pg ml 1 ) Day 7 after ICSI 405 ± 52 c 372 ± 48 Day 15 after ICSI 420 ± 56 c 408 ± 46 Serum progesterone (ng ml 1 ) Day 7 after ICSI 42 ± 8 c 29 ± 7 Day 15 after ICSI 48 ± 9 c 41 ± 7 Serum HCG (IU l 1 ) Day 15 after ICSI In all conception cycles b 64 ± 9 c 41 ± 7 In singleton pregnancies 50 ± 6 c 32 ±

6 for luteal-phase support Table VI. Clinical outcomes of patients treated with the GnRH antagonist ovarian stimulation protocol Outcome variable Patient group a Mean ± SD (median). b Significantly different from the placebo group (P < 0.05). Transfer procedures Embryos transferred Embryos per transfer a 2.2 ± 0.4 (2.0) 2.3 ± 0.5 (2.0) Good-morphology embryos 1.9 ± 0.4 (2.0) 2.0 ± 0.4 (2.0) per transfer Clinical pregnancy rate Per embryo transfer 47.6% (69/145) 37.5% (54/144) Per intention to treat 46.0% (69/150) 36.0% (54/150) Clinical implantation rate 27.1% (86/317) b 17.4% (57/328) Ongoing pregnancy rate Per embryo transfer 44.8% (65/145) b 31.9% (46/144) Per intention to treat 43.3% (65/150) b 30.7% (46/150) Live birth rate 25.2% (80/317) b 14.6% (48/3328) higher implantation, ongoing pregnancy and live birth rates than those in the placebo group (Table VI), and there was also a trend towards a higher clinical pregnancy rate in the GnRH agonist group (P = 0.083). There were 15 multiple births (all twins) in the group as compared with only two multiple births (both twins) in the placebo group. Using multivariate analysis, we observed that the female age and the number of oocytes recovered were independent predictors of implantation and ongoing pregnancy in both the and the placebo groups (data not shown). Discussion Several lines of evidence suggest that the luteal phase after embryo transfer is deficient in cycles stimulated with HMG or FSH, irrespectively of whether a (Macklon and Fauser, 2000; Pritts and Atwood, 2002) or a GnRH antagonist (Beckers et al., 2003; Kolibianakis et al., 2003) is used for the prevention of a premature LH surge. To solve this problem, most clinics use some form of luteal-phase support after embryo transfer. This can be done by stimulating progesterone secretion by the corpus luteum with exogenous HCG administration or by a direct supplementation of exogenous progesterone (reviewed in Pritts and Atwood, 2002). In this prospective randomized study, evaluating the effects of administered in a single dose 6 days after ICSI on luteal-phase characteristics and clinical outcomes, both HCG and progesterone supplementation were used in both the treatment and the placebo groups, in addition to oral E 2. Consequently, any effect of luteal-phase administration described in this study must be interpreted on the background of this basic luteal-phase support. The effects of luteal-phase administration were evaluated in the context of two different ovulation stimulation regimens, one using and the other GnRH antagonist. The results concerning ICSI clinical outcomes, namely clinical implantation and birth rates, confirm previous findings obtained in an oocyte donation model (Tesarik et al., 2004). The observed increase in clinical implantation rate in patients treated with in the luteal phase, as compared with the placebo group, indicates an increased risk of multiple pregnancies after this luteal-phase supplementation. Hence, it might be interesting to combine luteal-phase GnRH agonist administration with single embryo transfer. In addition, unlike patients in whom ovulation stimulation had been performed with the use of a long protocol, patients stimulated with the use of GnRH antagonist and receiving in the luteal phase had also a higher delivery rate and showed a trend towards a higher clinical pregnancy rate as compared with the corresponding placebo group. The relationship between the protocol of ovarian stimulation and the subsequent effects of luteal-phase administration of is not clear and remains an intriguing challenge for further research, namely in the context of the current efforts at optimization of the use of GnRH antagonists in ovarian stimulation (Fauser and Devroey, 2005). The fact that beneficial effects of luteal-phase on these variables was achieved here on the background of luteal-phase support with HCG, progesterone and E 2 supports the originally formulated hypothesis that exerts a direct beneficial effect on the implanting embryo (Tesarik et al., 2004). The hypothesis of a direct action of on the embryo (Tesarik et al., 2004) is further corroborated by the present observation of higher levels of serum β-hcg, both in the overall group of patients who achieved a pregnancy and after the exclusion of multiple pregnancies from the analysis, after luteal-phase administration of as compared with placebo. It was reported previously that GnRH increases serum HCG in pregnant women (Iwashita et al., 1993), presumably by acting on a placental GnRH receptor (Lin et al., 1995), and results in excessive binding of circulating GnRH by a yet-undefined GnRH-binding substance, possibly an auto-antibody reported in patients with repeated early pregnancy wastage (Siler-Khodr et al., 1997). However, the arguments in favour of a direct effect of on the embryo do not exclude other mechanisms of action acting in parallel. Unlike the situation in the oocyte donation model, the corpus luteum may have been another target of action in the present setting. In fact, an increase in serum concentrations of both E 2 and progesterone in the luteal phase of patients given 6 days after ICSI as compared with patients receiving placebo was observed in this study. These observations were made in both types of ovarian stimulation regimens employed in this study. The mechanism of action of on the corpus luteum remains a controversial issue. Early studies suggested that treatment in the luteal phase (one or two administrations of 500 μg of buserelin) may act as a luteolytic agent for contraceptive purposes (Lemay et al., 1982, 1983), supposedly due to GnRH receptor desensitization in gonadotroph cells. In fact, this may be true for certain doses of GnRH agonist, although the hypothesis of a contraceptive action of s has never been definitely confirmed, and attempts at inducing early abortion in women by huge doses of a superactive resulted in a failure (Skarin et al., 1982). More recently, a number of observational clinical studies 2577

7 J.Tesarik et al. reported the consequences of an inadvertent administration of in the luteal phase, and with only one exception (Herman et al., 1992), all agree in that luteal-phase administration does not compromise the continuation of pregnancy resulting from assisted reproduction attempts but rather seems to support implantation (Golan et al., 1990; Isherwood et al., 1990; Ron-El et al., 1990; Smitz et al., 1991; Jackson et al., 1992; Balasch et al., 1993; Elefant et al., 1993; Har-Toov et al., 1993; Weissman and Shoham, 1993; Wilshire et al., 1993; Young et al., 1993; Gartner et al., 1997). Moreover, a recent small randomized study suggested that low doses of (100 μg of buserelin) exert a stimulatory effect on the corpus luteum (Pirard et al., 2005). The present data strongly support this hypothesis. It is not known whether the increase in the synthesis of E 2 and progesterone in patients given in the luteal phase is mediated by pituitary hormones or results from a direct action of in the ovary. Our failure to detect a measurable increase in serum LH concentration 1 day after administration does not support a role for the pituitary in these events, but it must be reminded that a short LH peak, sufficient to produce this effect, may not be detected with the design of this study. If stimulates placental production of HCG, which is known to promote the secretion of vascular endothelial growth factor (VEGF), the main player in the development of OHSS (McClure et al., 1994; Neulen et al., 1995; Krasnow et al., 1996), a question arises about the safety of luteal-phase administration in patients at high risk of OHSS. Surprisingly, however, a recent study performed in an animal (rat) model has shown that administration during the luteal phase reduces the expression of VEGF and VEGF receptors, both at mrna and protein levels, and decreases vascular permeability in the hyperstimulated ovaries (Kitajima et al., 2004). Moreover, the continuation of GnRH agonist administration in -down-regulated IVF cycles has been shown to prevent the development of the early OHSS in at-risk patients in whom embryo transfer is not performed and all embryos are cryopreserved (Endo et al., 2002). The -induced increase in HCG production by the embryo at early stages of implantation is probably too low to affect significantly VEGF synthesis in extrauterine locations, and the decrease in the ovarian VEGF and VEGF receptor expression operated by (Kitajima et al., 2004) may be related to action on the pituitary ovarian axis, independent of its effect on the early embryo. In this study, we have not observed any case of severe OHSS requiring hospitalization. There is thus no a priori reason why GnRH agonist should be avoided as luteal-phase support in this category of patients, although caution is recommended until more complete information on the effects of luteal-phase administration at different levels of the reproductive system is available. Moreover, a GnRH receptor site was immunolocalized in murine endometrium (Murdoch, 1995), and a functional LH receptor has been detected in the human uterus (Reshef et al., 1990). A direct action of or induced LH in the uterine tissues can thus also be responsible for the observed luteal-phase effects In conclusion, the results of this prospective randomized study extend previous findings of a beneficial effect of GnRH agonist administration in the luteal phase on embryo implantation potential, obtained in an oocyte donation programme, to ovarian stimulation cycles using both and GnRH antagonist for prevention of premature LH surge. This effect was particularly evident in ovarian stimulation cycles using GnRH antagonist. Moreover, these data corroborate the hypothesis of a direct action of on the implanting embryo by demonstrating a stimulatory effect on β-hcg secretion, but they also suggest a direct or indirect positive effect on the corpus luteum function. These observations open the idea of a broader use of as a novel lutealphase support, possibly in combination with the conventionally employed HCG, E 2 and progesterone, especially in GnRH antagonist-using ovarian stimulation regimens. However, caution is needed as to eventual late-appearing health effects of this therapy on the offspring. The present data are likely to prompt further studies into the mechanism of action of GnRH agonist in the luteal phase. Pre-clinical studies are also needed to determine the optimal dose, time and frequency of GnRH agonist administration. References Balasch J, Martinez F, Jové I, Cabré L, Coroleu B, Barri PN and Vanrell JA (1993) Inadvertent gonadotrophin-releasing hormone agonist (GnRHa) administration in the luteal phase may improve fecundity in in-vitro fertilization patients. Hum Reprod 8, Beckers ND, Macklon NS, Eijkemans MJ, Ludwig M, Felberbaum RE, Diedrich K, Bustion S, Loumaye E and Fauser BC (2003) Nonsupplemented luteal phase characteristics after the administration of recombinant human chorionic gonadotropin, recombinant luteinizing hormone, or gonadotropinreleasing hormone (GnRH) agonist to induce final oocyte maturation in in vitro fertilization patients after ovarian stimulation with recombinant follicle-stimulating hormone and GnRH antagonist cotreatment. J Clin Endocrinol Metab 88, Elefant E, Biour B, Blumberg-Tick J, Roux C and Thomas F (1993) Administration of a gonadotropin-releasing hormone agonist during pregnancy: followup of 28 pregnancies exposed to triptoreline. Fertil Steril 63, Endo T, Honnma H, Hayashi T, Chida M, Yamazaki K, Kitajima Y, Azumaguchi A, Kamiya H and Kudo R (2002) Continuation of administration for 1 week, after hcg injection, prevents ovarian hyperstimulation syndrome following elective cryopreservation of all pronucleate embryos. Hum Reprod 17, Fauser BC and Devroey P (2005) Why is the clinical acceptance of gonadotropin-releasing hormone antagonist cotreatment during ovarian hyperstimulation for in vitro fertilization so slow? Fertil Steril 83, Gartner B, Moreno C, Marinaro A, Remohi J, Simon C and Pellicer A (1997) Accidental exposure to daily long-acting gonadotrophin-releasing hormone analogue administration and pregnancy in an in-vitro fertilization cycle. Hum Reprod 12, Golan A, Ron-El R, Herman A, Weinraub Z, Soffer Y and Caspi E (1990) Fetal outcome following inadvertent administration of long-acting D-Trp 6 LH-RH microcapsules during pregnancy: a case report. Hum Reprod 5, Har-Toov J, Brenner SH, Jaffa A, Yavetz H, Peyser MR and Lessing JB (1993) Pregnancy during long-term gonadotropin-releasing hormone agonist therapy associated with clinical pseudomenopause. Fertil Steril 59, Herman A, Ron-El R, Golan A, Nachum H, Soffer Y and Caspi E (1992) Impaired corpus luteum function and other undesired results of pregnancies associated with inadvertent administration of a long-acting agonist of gonadotrophin-releasing hormone. Hum Reprod 7, Isherwood PJ, Ibrahim ZA, Matson PL, Morroll DR, Burslem RW and Lieberman BA (1990) Endocrine changes in women conceiving during treatment with a LHRH agonist. Hum Reprod 5, Iwashita M, Kudo Y, Shinozaki Y and Takeda Y (1993) Gonadotropin-releasing hormone increases serum human chorionic gonadotropin in pregnant women. Endocr J 40,

8 for luteal-phase support Jackson AE, Curtis P, Amso N and Shaw RW (1992) Exposure to LHRH agonist in early pregnancy following the commencement of mid-luteal buserelin for IVF stimulation. Hum Reprod 7, Kitajima Y, Endo T, Manase K, Nishikawa A, Shibuya M and Kudo R (2004) Gonadotropin-releasing hormone agonist administration reduced vascular endothelial growth factor (VEGF), VEGF receptors, and vascular permeability of the ovaries of hyperstimulated rats. Fertil Steril 81 (Suppl. 1), Kolibianakis EM, Bourgain C, Platteau P, Albano C, Van Steirteghem AC and Devroey P (2003) Abnormal endometrial development occurs during the luteal phase of nonsupplemented donor cycles treated with recombinant follicle-stimulating hormone and gonadotropin-releasing hormone antagonists. Fertil Steril 80, Krasnow JS, Berga SL, Guzick DS, Zeleznik AJ and Yeo KT (1996) Vascular permeability factor and vascular endothelial growth factor in ovarian hyperstimulation syndrome: a preliminary report. Fertil Steril 65, Lemay A, Faure N and Labrie F (1982) Sensitivity of pituitary and corpus luteum responses to single intranasal administration of buserelin in normal women. Fertil Steril 37, Lemay A, Faure N, Labrie F and Fazekas AT (1983) Gonadotrophin and corpus luteum responses to two successive intranasal doses of a luteinising hormone-releasing hormone agonist at different days after the mid-cycle luteinising hormone surge. Fertil Steril 39, Lin L, Roberts VJ and Yen S (1995) Expression of human gonadotropinreleasing hormone receptor gene in the placenta and its functional relationship to human chorionic gonadotropin secretion. J Clin Endocrinol Metab 80, Macklon NS and Fauser BC (2000) Impact of ovarian hyperstimulation on the luteal phase. J Reprod Fertil Suppl 55, McClure N, Healy DL, Rogers PA, Sullivan J, Beaton L, Haning RV Jr, Connolly DT and Robertson DM (1994) Vascular endothelial growth factor as capillary permeability agent in ovarian hyperstimulation syndrome. Lancet 344, Murdoch WJ (1995) Immunolocalization of a GnRH receptor site in murine endometrium that mediates apoptosis. Cell Tissue Res 282, Neulen J, Yan Z, Raczek S, Weindel K, Keck C, Weich HA, Marme D and Breckwoldt M (1995) Human chorionic gonadotropin-dependent expression of vascular endothelial growth factor/vascular permeability factor in human granulosa cells: importance in ovarian hyperstimulation syndrome. J Clin Endocrinol Metab 80, Pirard C, Donnez J and Loumaye E (2005) as novel luteal support: results of a randomized, parallel group, feasibility study using intranasal administration of buserelin. Hum Reprod 20, Pritts EA and Atwood AK (2002) Luteal phase support in infertility treatment: a meta-analysis of the randomized trials. Hum Reprod 17, Reshef E, Lei ZM, Rao CV, Pridham DD, Chegini N and Luborsky JL (1990) The presence of gonadotropin receptors in nonpregnant human uterus, human placenta, fetal membranes and deciduas. J Clin Endocrinol Metab 70, Ron-El R, Golan A, Herman A, Raziel A, Soffer Y and Caspi E (1990) Midluteal gonadotropin-releasing hormone analog administration in early pregnancy. Fertil Steril 53, Siler-Khodr TM, Smikle CB, Sorem KA, Grayson MA, Spencer DK, Yoder BK and Khodr GS (1997) Effect of excessive GnRH-binding substance on circulating maternal hcg in human pregnancy. Early Pregnancy 3, Skarin G, Nillius SJ and Wide L (1982) Failure to induce early abortion by huge doses of a superactive LHRH agonist in women. Contraception 26, Smitz J, Camus M, Devroey P, Bollen N, Tournaye H and Van Steirteghem AC (1991) The influence of inadvertent intranasal buserelin administration in early pregnancy. Hum Reprod 6, Tesarik J and Greco E (1999) The probability of abnormal preimplantation development can be predicted by a single static observation on pronuclear stage morphology. Hum Reprod 14, Tesarik J and Mendoza C (2002) Effects of exogenous LH administration during ovarian stimulation of pituitary down-regulated young oocyte donors on oocyte yield and developmental competence. Hum Reprod 17, Tesarik J, Junca AM, Hazout A, Aubriot FX, Nathan C, Cohen-Bacrie P and Dumont-Hassan M (2000) Embryos with high implantation potential after intracytoplasmic sperm injection can be recognized by a simple, non-invasive examination of pronuclear morphology. Hum Reprod 15, Tesarik J, Hazout A and Mendoza C (2004) Enhancement of embryo developmental potential by a single administration of at the time of implantation. Hum Reprod 19, Weissman A and Shoham Z (1993) Favourable pregnancy outcome after administration of a long-acting gonadotropin-releasing hormone agonist in the mid-luteal phase. Hum Reprod 8, Wilshire GB, Emmi AM, Gagliardi CC and Weiss G (1993) Gonadotropinreleasing hormone agonist administration in early human pregnancy is associated with normal outcomes. Fertil Steril 60, Young DC, Snabes MC and Poindexter AN (1993) exposure during the first trimester of pregnancy. Obstet Gynecol 81, Submitted on November 26, 2005; resubmitted on February 20, 2006, April 16, 2006; accepted on April 21,