Androgen priming using aromatase inhibitor and hcg during early-follicular-phase GnRH antagonist down-regulation in modified antagonist protocols

Human Reproduction Vol.21, No.10 pp. 2593 2600, 2006 Advance Access publication June 19, 2006. doi:10.1093/humrep/del221 Androgen priming using aromatase inhibitor and hcg during early-follicular-phase GnRH antagonist down-regulation in modified antagonist protocols K.Lossl 1,3, A.N.Andersen 1, A.Loft 1, N.L.C.Freiesleben 1, S.Bangsbøll 1 and C.Yding Andersen 2 1 The Fertility Clinic 4071 and 2 Laboratory of Reproductive Biology 5712, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark 3 To whom correspondence should be addressed at: The Fertility Clinic 4071, Copenhagen University Hospital, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen, Denmark. E-mail: kristine.loessl@rh.dk BACKGROUND: Temporary exposure of follicles to increased levels of androgens may enhance their sensitivity to FSH. The aim of this study was to increase the intraovarian androgen level using aromatase inhibitors and hcg before controlled ovarian stimulation (COH) and to test this concept clinically. METHODS: In a prospective, nonrandomized study, 45 patients were treated in modified antagonist protocols including early-follicular-phase downregulation and androgen priming before COH. All patients received cetrorelix, 3 mg s.c., on cycle days 2 and 5. Group I (n = 15) received no other pretreatment. Group II (n = 15) received 1 daily tablet of aromatase inhibitor, letrozole 2.5 mg, from cycle days 2 to 8. Group III (n = 15) received letrozole as Group II and 1250 IU of hcg s.c. on cycle day 2. From cycle day 8, all patients were stimulated with highly purified menotrophin in a flexible antagonist protocol. RESULTS: Aromatase inhibitor increased the level of testosterone in follicular fluid (P < 0.002), but not in plasma. Androgen priming with aromatase inhibitor and hcg increased the number of good-quality embryos (P = 0.015) but did not increase the implantation rate. CONCLUSIONS: The use of aromatase inhibitor before COH significantly influences the local endocrine environment before and during stimulation. Androgen priming with both aromatase inhibitor and hcg may result in more good-quality embryos. Key words: androgen priming/aromatase inhibitor/follicular fluid/gnrh antagonist/ovarian stimulation Introduction Follicular health is widely recognized as an important parameter for the viability of the enclosed oocyte and its potential to undertake embryogenesis. Studies have demonstrated that the intrafollicular ratio of androgens to estradiol is a possible parameter of follicular health. Relatively high concentrations of androgens may be found in less healthy follicles, reflecting a reduced ability of the granulosa cells to synthesize estradiol (McNatty et al., 1979; Andersen, 1993). These studies suggested that androgens were a passive indirect sign of follicular health and did not suggest a direct action of androgens on the follicle. Other more recent studies suggested that androgens themselves have a direct stimulatory effect on follicular development (Hillier et al., 1997; Vendola et al., 1998; Weil et al., 1998). Androgen receptors (ARs) were shown on granulosa and theca cells of both monkey and human ovaries (Hild-Petito et al., 1991; Horie et al., 1992; Chadha et al., 1994). The most abundant expression of AR was localized to the granulosa cells of healthy pre-antral and antral follicles. AR expression was positively correlated to granulosa cell mitosis and follicular growth and negatively correlated to granulosa cell apoptosis (Hillier et al., 1997; Weil et al., 1998). The possible effects of androgens closely resembled those normally ascribed to FSH, and, surprisingly, primate follicles have shown a positive correlation between FSH receptor (FSHr) expression and AR mrna levels. Furthermore, androgen treatment significantly enhanced FSHr expression (Weil et al., 1999). These observations suggested that androgens promote follicular growth indirectly by amplifying the response to FSH. To evaluate whether these effects of androgens could be used to augment the treatment outcome of women undergoing ovarian stimulation for IVF or ICSI, we introduced a new stimulation protocol involving a period of early-follicular-phase down-regulation and androgen priming before ovarian stimulation. Two measures were employed to achieve a temporary enhanced androgen exposure to the follicles. An aromatase inhibitor was given alone to prevent androgens from being aromatized into estrogens or combined with a bolus of hcg to promote theca-cell androgen production. To arrest follicular growth during the period of androgen priming, we administered a GnRH antagonist. Furthermore, the introduction of early-follicular-phase GnRH antagonist down-regulation and The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. 2593 For Permissions, please email: journals.permissions@oxfordjournals.org

K.Lossl et al. androgen priming before COH in a modified antagonist protocol could allow a period of hypoestrogenism and thereby facilitate a complete endometrial shedding before the start of exogenous FSH stimulation. Indeed, this is what characterizes the long agonist protocol where hypoestrogenism and endometrial shedding are achieved during GnRH agonist administration before ovarian stimulation. Pioneering work with the use of aromatase inhibitors, primarily for ovulation induction before intrauterine insemination, has been performed by others, who also described the hypothesis of possible beneficial effect of accumulating androgens to improve responsiveness to FSH (Casper and Mitwally, 2005; Mitwally et al., 2005). The aim of this study was (i) to test whether aromatase inhibitor and hcg could be used for androgen priming, by monitoring plasma and follicular fluid androgen levels, and (ii) to evaluate ovarian stimulation and cycle outcome, e.g. days of stimulation, total FSH dose and the number of follicles, oocytes retrieved, embryos and good-quality embryos. Materials and methods Study design This was a prospective, non-randomized, single-centre study performed between April 2004 and April 2005. The local ethics committee approved the study [J.nr. (KF) 01-134/03]. Each subject gave written informed consent before participating in this study. Subjects A total of 47 patients fulfilling the following inclusion criteria were included: indication of IVF or ICSI, age 25 39 years, normal regular menstrual cycles within the range of 21 35 days, BMI between 18 and 30 kg/m 2, a negative urinary pregnancy test on cycle day 2 and FSH and LH levels < 12 IU/l on cycle day 2. Exclusion criterion was polycystic ovary syndrome (PCOS) (The Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group, 2004). Two patients withdrew during down-regulation for personal reasons (new job, journey). Forty-five patients thus completed the study and were treated according to three different protocols, one treatment group of 15 patients being completed before proceeding to the next. All patients were included in the analyses of follicle dynamics and endocrinology. Treatment Figure 1 gives a graphical illustration of the drug regimens. Group I (early-follicular-phase down-regulation by the use of GnRH antagonist): this group received cetrorelix 3 mg s.c. (Cetrotide, Serono Europe Limited), administered on cycle days 2 and 5. Group II (early-follicular-phase down-regulation and androgen priming by the use of GnRH antagonist and aromatase inhibitor): this group received Cetrotide 3 mg s.c. on days 2 and 5 and additionally one tablet of letrozole (Femar, Novartis Healthcare A/S, Denmark) 2.5 mg daily from cycle days 2 to 8. Group III (early-follicular-phase down-regulation and androgen priming by the use of GnRH antagonist, aromatase inhibitor and hcg): this group received Cetrotide 3 mg s.c. on days 2 and 5, one tablet of letrozole 2.5 mg daily from cycle days 2 to 8 and additionally hcg (Pregnyl, Organon AS, Denmark) 1250 IU s.c. on day 2. From cycle day 8, all patients were stimulated according to a standard flexible, antagonist protocol with highly purified menotrophin (Menopur, 2594 Figure 1. Treatment regimens including early-follicular-phase down-regulation and androgen priming before controlled ovarian stimulation in modified flexible antagonist protocols. Ferring, Denmark), 150 or 225 IU/day, as a fixed dose for 6 days. The dose could be adjusted from stimulation day 7 according to follicular development. GnRH antagonist, Cetrotide 0.25 mg s.c., was given daily when the leading follicle reached 14 mm. When three follicles had reached a diameter of 17 mm, 10 000 IU of hcg (Pregnyl) was administered followed by oocyte retrieval 36 h later. Subsequently, IVF or ICSI was performed, and a maximum of two embryos were transferred at day 2. Luteal support with vaginal progesterone (Progestan, Organon AS) 200 mg 3 daily was given for 2 weeks from the day of embryo transfer. Five patients developed less than three mature follicles. Of these, one patient with occluded tubes had oocyte retrieval performed and four patients with patent tubes had intrauterine insemination performed. Serum hcg was measured 2 weeks after embryo transfer. A live pregnancy at 7 8 weeks was considered as ongoing. Endocrine analysis and ultrasonography Plasma concentrations of FSH, LH, estradiol and testosterone were measured on cycle day 2 (just before down-regulation and androgen priming period), on cycle day 8 (just after down-regulation and androgen priming period, corresponding to stimulation day 1), on stimulation day 7 and on the day of hcg administration. Samples were analysed using electrochemiluminescence immunoassay (Roche Diagnostics, Mannheim, Germany) in Modular Analytics E170 immunoassay analyser. Inter-assay coefficients of variation were 3.7% for FSH, 4.4% for LH, 7.2% for estradiol and 8.7% for testosterone. Ultrasonography was performed on the same days. The following were measured: ovarian volume (cycle days 2 and 8) and the number of antral follicles according to the follicle size categories [<5 mm, 5 7 mm, 8 10 mm (cycle days 2 and 8) and follicle size >10 mm all days]. Follicular size was determined as mean diameter. Ultrasonography was performed using a GE LOGIQ 5 Expert scanner using a transvaginal E8C probe, B-mode, 8.0 MHz. Embryo quality Embryo morphology was scored according to previously published criteria (Staessen et al., 1989) by experienced laboratory technicians blinded to which treatment had been employed. The embryos of equal

Androgen priming in modified antagonist protocols or unequal blastomer size with <10% fragmentation and no signs of multi-nucleation were considered to be of good quality (Ziebe et al., 1997). Follicular fluid analyses Two follicular fluid samples were obtained from each patient and stored. They originated from the first large aspirated follicle, where an oocyte was obtained, from each of the two ovaries. After collection of the oocytes, each separate follicular fluid sample was centrifuged (1000 g) to eliminate granulosa and blood cells. Only follicular fluid without the contamination of red blood cells was included. Samples were stored at 20 C until analysed. Estradiol, testosterone and androstenedione were measured in each separate sample using commercially available radioimmunoassay kits (DSL-43100, DSL-4000, DSL-3800, Diagnostic System Laboratories, TX, USA). Samples were diluted 1 : 500 (estradiol), undiluted (testosterone) and 1 : 50 (androstenedione) in steroid-free serum just before measurement. No samples were pooled. Study end-points The primary end-points were the levels of steroid hormones in circulation during treatment and in follicular fluid at oocyte retrieval. Secondary end-points were the duration of exogenous gonadotrophin stimulation, total FSH dose, the number of follicles, the number of oocytes retrieved and the number and quality of embryos. Statistical analysis Endocrine levels before and after down-regulation were compared using a paired t-test for continuous data. An exception was made when comparing estradiol levels where several very low and undetectable levels were recorded after the down-regulation period. Here, non-parametric analyses were used: Wilcoxon signed rank test for within-group analyses and Mann Whitney U-test and Kruskal Wallis test for between-group analyses. Analysis of variance (ANOVA) was used for the comparison of continuous data between groups. Significant differences were further analysed using independent t-test. A P-level <0.05 was considered significant. Data were analysed using the commercially available software package SPSS, version 13 (SPSS Inc., Chicago, USA). Results Subject characteristics As summarized in Table I, the three groups were similar regarding age, BMI, the number of earlier treatments, FSH, LH, estradiol and testosterone levels on cycle day 2 before initiating treatment. However, Group I happened to have smaller ovarian volume (P 0.04) and fewer antral follicles (P 0.034) at treatment start compared with the two other groups even though the FSH, LH and estradiol levels were almost identical. Early-follicular-phase down-regulation and androgen priming Table II summarizes follicular dynamics during the period of down-regulation and androgen priming. From cycle days 2 to 8, there was no follicular growth and no development of a dominant follicle, defined as a follicle 10 mm, in any of the three groups. In the two groups receiving the aromatase inhibitor, i.e. Groups II and III, a significant reduction in the number of follicles between cycle days 2 and 8 was observed in total follicle count (P = 0.005) and in the number of follicles sized 5 7 mm (P 0.034). In Group I, the number and size of follicles remained unchanged from cycle days 2 to 8. The endocrine response during treatment is shown in Figure 2. When using the plasma estradiol level on cycle day 8 (stimulation day 1) as an indicator of down-regulation, the regimens caused profound down-regulation in all three groups with mean estradiol levels of 0.009 nmol/l, <0.005 nmol/l and <0.005 nmol/l in Groups I, II and III, respectively. Comparing cycle day 8 levels to cycle day 2 levels, estradiol was significantly suppressed in all three groups (P 0.006). The use of aromatase inhibitor, i.e. Groups II and III, led to even lower estradiol levels compared to Group I (P 0.007). A total of 53% in Group II had estradiol levels below the detection limit (0.004 nmol/l) of the assay compared with 13 and 20% in Groups I and III, respectively. After down-regulation, on cycle day 8 (stimulation day 1), the stimulation was initiated from a normogonadotrophic state in all three groups. The FSH levels in Groups I and III were unsuppressed on cycle day 8, measured 3 days after the second GnRH antagonist injection. In Group II, there was a significant increase in FSH level when cycle day 8 level was compared to cycle day 2 level (P < 0.001). The LH level following downregulation was 2.8, 4.0 and 2.6 IU/l on cycle day 8 in Groups I, II and III, respectively. These levels were significantly reduced when compared to cycle day 2 levels (P 0.004). Plasma testosterone level was not increased by the androgen priming period in any of the three groups when cycle day 8 level was compared to cycle day 2 level. On the contrary, the testosterone level was significantly decreased in Groups I and III (P 0.034). There was no significant difference in plasma testosterone levels between the three groups. After down-regulation, on stimulation day 1, we observed a very thin and completely shedded endometrium in all patients. Follicular fluid endocrine levels Follicular fluid steroid levels are summarized in Table III. The testosterone level was significantly (P 0.003) higher in patients treated with aromatase inhibitor during the period of earlyfollicular-phase down-regulation and androgen priming, i.e. Groups II and III, compared to Group I that only received the GnRH antagonist. The level of androstenedione was significantly (P = 0.008) higher in Group II compared to Group I. Also in Group III, the level of androstenedione was higher than that in Group I, but this difference did not reach statistical significance. The follicular fluid estradiol level did not differ between the three groups, whereas the ratios of estradiol to testosterone and estradiol to androstenedione were significantly lower in Group II, where the aromatase inhibitor was used alone for androgen priming (P 0.003). Characteristics of ovarian stimulation and cycle outcome The endocrine response during treatment is shown in Figure 2. There were no significant differences between the groups in the levels of FSH, LH, testosterone and estradiol on stimulation day 7 and on the day of hcg administration. 2595

K.Lossl et al. Table I. Patient characteristics at cycle day 2 before initiation of early-follicular-phase down-regulation and androgen priming Group I Group II Group III P-value Age 32.4 ± 1.0 32.6 ± 0.9 32.8 ± 1.0 0.956 BMI 23.9 ± 1.1 22.7 ± 0.7 21.6 ± 0.5 0.138 Infertility diagnosis [n (%)] Male factor 11 (73) 10 (67) 12 (80) Tubal factor 1 (7) 2 (13) 1 (7) Endometriosis 0 (0) 1 (7) 0 (0) Idiopathic 3 (20) 2 (13) 2 (13) Treatment [n (%)] IVF 5 (33) 5 (33) 5 (33) ICSI 10 (67) 10 (67) 10 (67) Number of earlier treatments 0.7 ± 0.3 0.9 ± 0.3 0.8 ± 0.3 0.881 Endocrinology, cycle day 2 FSH (IU/l) 7.9 ± 0.6 7.3 ± 0.4 7.9 ± 0.4 0.661 LH (IU/l) 5.1 ± 0.5 5.2 ± 0.5 5.5 ± 0.5 0.828 Estradiol (nmol/l) 0.14 ± 0.01 0.13 ± 0.01 0.14 ± 0.01 0.916 Testosterone (nmol/l) 1.2 ± 0.1 1.3 ± 0.1 1.3 ± 0.1 0.755 Ovarian volume (cm 3 ), cycle day 2 4.2 ± 0.2 a,b 5.3 ± 0.3 a 5.4 ± 0.5 b 0.004 a 0.040 b Total number of antral follicles, cycle day 2 17.5 ± 2.0 a,b 24.5 ± 2.5 a 24.3 ± 1.9 b 0.034 a 0.018 b Values are given as mean ± SEM. Groups were compared with one-way analysis of variance (ANOVA) followed by independent t-test when the former showed a significant difference. a Respective. b Indicates which two groups have been compared in that row to obtain the P-value given to the right. Table II. Total number of follicles according to size before and after early-follicular-phase down-regulation and androgen priming Follicular size (mm) Group I Group II Group III Cycle day 2 Cycle day 8 P-value Cycle day 2 Cycle day 8 P-value Cycle day 2 Cycle day 8 P-value <5 14.1 ± 2.0 14.3 ± 1.7 0.908 18.3 ± 3.0 15.7 ± 2.2 0.179 18.7 ± 1.5 17.8 ± 1.8 0.462 5 7 2.9 ± 0.6 3.3 ± 0.6 0.517 5.7 ± 1.1 3.1 ± 0.6 0.010 5.3 ± 0.9 2.7 ± 0.7 0.034 8 10 0.5 ± 0.3 0.3 ± 0.2 0.334 0.5 ± 0.3 0.8 ± 0.2 0.469 0.4 ± 0.3 0.1 ± 0.1 0.265 >10 0 ± 0 0 ± 0 0 ± 0 0 ± 0 0 ± 0 0.1 ± 0.1 0.334 Total 17.5 ± 2.0 17.8 ± 1.8 0.722 24.5 ± 2.5 19.5 ± 2.1 0.005 24.3 ± 1.9 20.7 ± 1.5 0.005 Values are given as mean ± SEM. Paired-sample t-tests were used for statistical analysis. Stimulation characteristics and cycle outcome are summarized in Table IV. Most of the characteristics of ovarian stimulation and cycle outcome were similar between the three groups. The number of oocytes retrieved was 5.7, 8.4 and 8.4 in Groups I, II and III, respectively, not being statistically different using ANOVA (P = 0.102). The number of embryos (P 0.028) and especially the number of good-quality embryos (P 0.015) were increased in the group of patients receiving both aromatase inhibitor and hcg during the androgen priming period. The ongoing pregnancy rate per initiated cycle was 27, 27 and 33% in Groups I, II and III, respectively. Four, three and five patients had elective single-embryo transfer performed in Groups I, II and III, respectively. Two, three and no single-embryo transfers were performed in Groups I, II and III, respectively. The ongoing implantation rate was 28, 22 and 21% in Groups I, II and III, respectively. Premature rise of LH One patient in each group experienced a premature LH rise during ovarian stimulation, all three continued treatment as planned. In Group I, the patient had an LH level of 14.6 IU/L the day before hcg administration. She achieved an ongoing pregnancy. In Group II, one patient had an LH rise to 25.3 IU/l on stimulation day 7, just before the onset of daily antagonist treatment. Treatment resulted in a biochemical pregnancy. In Group III, one patient had an LH rise to 30.5 IU/l on the day of hcg administration. The patient had an ongoing pregnancy. Discussion This study introduced a new concept of ovarian stimulation involving a 1-week period of androgen priming in combination with early-follicular-phase GnRH antagonist down-regulation before stimulation with exogenous gonadotrophins. It has earlier been hypothesized that if the objective of androgen priming is to enhance the sensitivity to FSH, the optimal timing may be before rather than when the follicles are stimulated by FSH (de Ziegler, 2003). The results showed a significantly higher level of testosterone in follicular fluid at oocyte retrieval, almost 2 weeks after the androgen priming period, in the two groups that received the aromatase inhibitor compared to the group without (i.e. Groups II and III versus Group I). This demonstrates that the aromatase inhibitor effect persists throughout the entire follicular phase and that follicles are likely to have experienced enhanced levels of androgens during the whole course of ovarian stimulation. Furthermore, it shows that the use of aromatase inhibitor actually does enhance the local, intrafollicular concentration of 2596

Androgen priming in modified antagonist protocols Figure 2. Mean plasma levels of FSH, LH, estradiol and testosterone during treatment. Bars represent standard error of mean. During the period of down-regulation and androgen priming (comparing stimulation day 1 to cycle day 2), the FSH level was significantly increased in Group II (P < 0.001). The LH level (P = 0.004) and the estradiol level (P = 0.006) were suppressed in all three groups. The testosterone level was significantly suppressed in Groups I and III (P = 0.034). On stimulation day 1, the estradiol level was significantly lower in Groups II and III, where the aromatase inhibitor was applied during down-regulation and androgen priming, compared to Group I (P = 0.007). Also, on stimulation day 1, the FSH level in Group II was significantly higher than that in Group I (P = 0.002). Otherwise, there were no significant differences in the endocrine levels between the three groups. Cd2, cycle day 2; stim1, stimulation day 1 (corresponding to cycle day 8); stim7, stimulation day 7. 2597

K.Lossl et al. Table III. Steroid levels in follicular fluid from mature follicles Group I Group II Group III P-value Patients with oocyte retrieval (n) 13 15 13 Testosterone (nmol/l) 30.0 ± 3.2 a,b 51.0 ± 2.8 a 46.0 ± 3.7 b <0.001 a 0.003 b Androstenedione (nmol/l) 317 ± 35 a 414 ± 21 a 379 ± 21 0.008 a Estradiol (nmol/l) 1902 ± 194 1855 ± 133 2316 ± 239 0.170 Estradiol/testosterone 65.0 ± 3.9 a,b 37.0 ± 2.3 a,c 52.0 ± 4.0 b,c <0.001 a 0.020 b 0.003 c Estradiol/androstenedione 6.2 ± 0.5 a 4.5 ± 0.3 a,b 6.0 ± 0.4 b 0.001 a 0.003 b Values are given as mean ± SEM. Groups were compared with one-way analysis of variance (ANOVA) followed by independent t-test when the former showed a significant difference. a Respective. b Respective. c Indicates which two groups have been compared in that row to obtain the P-value to the right. Table IV. Stimulation characteristics and cycle outcome in modified antagonist protocols with early-follicular-phase down-regulation and androgen priming before ovarian stimulation testosterone without any detectable change in circulating levels, as the plasma levels of testosterone remained similar between the three groups at all measured time points. This is in contrast to the systemic administration of androgens and suggests that possible side effects with a systemic administration are likely to be avoided by the use of aromatase inhibitors. Recently, another group reported an increase in intrafollicular, but not in circulating, androgen level when letrozole, 2.5 mg daily, was added during the first 6 days of gonadotrophin stimulation in a short IVF/ICSI protocol for low responder patients (Garcia-Velasco et al., 2005). They used the aromatase inhibitor when rather than before ovarian stimulation. It is surprising that the follicular fluid androgen level remained almost twice as high in women who received the aromatase inhibitor as compared with those who did not, because 2598 Group I Group II Group III P-value Patients included/initiated cycles (n) 15 15 15 Days of stimulation 9.4 ± 0.4 9.6 ± 0.5 10.3 ± 0.5 0.400 Total units of FSH (and LH activity) 2030 ± 130 1917 ± 162 2110 ± 205 0.718 Endometrium day of hcg (mm) 9.9 ± 0.3 9.1 ± 0.7 8.9 ± 0.5 0.288 Day of hcg Follicles = 10 mm 7.2 ± 0.9 10.9 ± 1.6 9.2 ± 1.3 0.153 Follicles = 12 mm 6.1 ± 0.7 9.5 ± 1.5 8.4 ± 1.3 0.163 Follicles = 14 mm 5.3 ± 0.7 6.1 ± 1.1 6.1 ± 1.0 0.780 Patients with oocyte retrieval (n) 13 15 13 Oocytes retrieved (per retrieval) 5.7 ± 0.6 8.4 ± 1.2 8.4 ± 1.1 0.102 Frequency of fertilization (%) 60.4 ± 6.2 a 40.0 ± 5.2 a,b 60.4 ± 7.4 b 0.017 a 0.030 b Embryos (n) 3.3 ± 0.4 a 2.7 ± 0.6 b 5.2 ± 0.7 a,b 0.028 a 0.012 b Good-quality embryos (n) 2.0 ± 0.5 a 1.9 ± 0.5 b 4.2 ± 0.7 a,b 0.015 a 0.012 b Patients with embryo transfer (n) 12 12 12 Embryos transferred e 1.5 ± 0.15 1.5 ± 0.15 1.6 ± 0.15 0.903 Embryos cryopreserved 1.1 ± 0.5 1.3 ± 0.6 2.8 ± 0.7 0.082 Positive hcg [n (% per initiated cycle)] 7 (47) 6 (40) 7 (47) Ongoing pregnancies[n (%per initiated cycle)] 4 (27) 4 (27) 5 (33) d Implantation rate 7/18 (41%) c 4/18 (22%) 4/19 (21%) d Ongoing implantation rate 5/18 (28%) 4/18 (22%) 4/19 (21%) d Values are given as mean ± SEM. Groups were compared with one-way analysis of variance (ANOVA) followed by independent t-test when the former showed a significant difference. a Respective. b Indicates which two groups have been compared in that row to obtain the P-value given to the right. c In Group I, there was a high implantation rate because of two gemelli gestations, the one being a missed abortion. d In Group III, one patient became pregnant following conversion to homologous intrauterine insemination (IUI-H); this case was excluded from the calculation of implantation rate (number of gestational sacs/number of transferred embryos). e Four, three and five patients had elective single-embryo transfer performed in Groups I, II and III, respectively. Two, three and no single-embryo transfers were performed in Groups I, II and III, respectively. the termination of aromatase inhibitor administration occurred almost 2 weeks earlier. In the light of a reported half-life of 48 h, the effect of letrozole would not be expected to persist for so long, especially in an environment where estrogen production occurs at a full swing. It may reflect that the aromatase inhibitor is accumulated in the follicular fluid and therefore escapes clearance. This will be important to evaluate in future studies because the increased risk of locomotor and cardiac malformations has recently been reported following the use of letrozole from days 3 to 7 for ovulation induction (Biljan et al., 2005). The interpretation of these results must be done with care because the study design had several methodological problems; for example, it is a retrospective study with relatively few children in the study group and mostly low-risk pregnancies in the control group; the median age of mothers was 35.2 years in

Androgen priming in modified antagonist protocols the study group, probably being younger in the control group (no information is given); and a high proportion in the study group had gestational diabetes (20/130), a condition with a higher incidence of malformations, especially cardiac anomalies. In the present study, the follicular fluid testosterone level was independent of whether or not a bolus of hcg was given during androgen priming. However, the ratio of estradiol to testosterone and that of estradiol to androstenedione in follicular fluid were significantly higher when hcg and aromatase inhibitor were combined compared to the use of aromatase inhibitor alone. This suggests that follicles aromatize androgens more efficiently when hcg and aromatase inhibitor are combined, possibly explained by an enhanced follicular FSH sensitivity. Also the higher follicular fluid estradiol/androgen ratio could explain the higher fertilization rate seen when hcg and aromatase inhibitor were combined compared to the use of aromatase inhibitor alone. These results support and extend an earlier study, which suggested that the estradiol/androgen ratio was higher in healthy follicles containing an oocyte with better developmental potential (Andersen, 1993). The lack of randomization and the limited sample size in the present study surely prevent any firm conclusions to be drawn regarding reproductive outcome. However, the results showed significantly more embryos, and good-quality embryos, when aromatase inhibitor and hcg were combined during the androgen priming period. The number of oocytes retrieved was not significantly different when analysed with the ANOVA test. However, it could be argued that the control group could be compared individually to either of the two other groups. When Group I (the control group) was compared to either Group II (P=0.048) or Group III (P = 0.038), using an independent t-test, significantly more oocytes were found in the two groups that received androgen priming. This leads to our preliminary hypothesis: Androgen priming by use of aromatase inhibitor increases the intraovarian androgen level, which results in an increased follicular sensitivity to FSH and may result in more retrieved oocytes. When the aromatase inhibitor is used alone for androgen priming, the follicular fluid estradiol/androgen ratio is decreased to an extent that negatively affects the developmental potential of the enclosed oocyte resulting in a lower fertilization rate. But when hcg and aromatase inhibitor are combined, the follicles may experience an active stimulation of androgen production in the early follicular phase and a higher local androgen concentration during the priming period, and hence they may develop a better FSH responsiveness that translates into a more favourable estradiol/androgen ratio securing healthy oocytes, an appropriate fertilization rate and a high number of good-quality embryos. Collectively, a specific positive effect seems to be exerted by hcg and suggests that the combined use of hcg and aromatase inhibitor is a valid method of inducing androgen priming that warrants further investigation. Women with PCOS (The Rotterdam ESHRE/ASRM- Sponsored PCOS Consensus Workshop Group, 2004) often show an aggravated response to ovarian stimulation with the development of multiple follicles, but also a good reproductive outcome is often observed when ovulation is achieved. Follicles of women with PCOS have usually been exposed to enhanced androgen levels for an extended period of time. As a consequence, their follicles may have developed a high sensitivity to FSH that may induce their hyper-responsiveness. We have chosen a period of 1 week for androgen priming. Even of short duration, the androgen priming period may still be described as a temporary reversible PCOS-like endocrine condition. The length of the androgen priming period obviously needs to be determined in more detail in future studies. The length of the androgen priming period may also depend on the dose of hcg used. We have chosen a dose of 1250 IU as a bolus injection to stimulate androgen synthesis substantially but without creating conditions that may have a negative impact on follicular health. However, the hcg dose obviously also may be subjected to alterations that need to be determined in future studies. The half-life of hcg (Pregnyl) is 32 33 h (Mannaerts et al., 1998), which means that hcg should be cleared from circulation within 7 days, and concentrations may not be maintained at sufficient levels during the entire androgen priming period. The administration of GnRH antagonist in the early follicular phase, before exogenous gonadotrophin stimulation, arrested both the follicular growth and the development of a dominant follicle in all three groups. This has perspectives not only for creating a period for androgen priming but also for a flexible start of stimulation in the short protocol and perhaps for a more complete endometrial shedding that subsequently might lead to a more receptive endometrium. After 6 days of early-follicular-phase down-regulation and androgen priming in the modified antagonist protocols, we noted a very thin endometrium in all patients similar to what is seen in the long GnRH agonist protocol. In the agonist cycle, the down-regulation results in a considerable number of days with hypoestrogenism. Also in the natural menstrual cycle, there are low estrogen levels during the period of menstruation (Hohmann et al., 2005), and the rise in endogenous estrogens is really first measured around the time of the dominant follicle at day 8. However, using the normal short antagonist protocol with exogenous FSH stimulation from cycle day 2 or 3, estradiol level shows a marked rise after the start of exogenous FSH (Hohmann et al., 2005; Iwase et al., 2005). The immediate rise of estrogens, secreted by the growing follicles, might stimulate the endometrium before it has been appropriately shedded, and one could speculate whether this could have a negative impact on the endometrial receptivity. This could explain why metaanalyses comparing the short antagonist versus the long agonist protocol have shown that the former is inferior in relation to pregnancy rates despite the comparable number and quality of transferred embryos (Ludwig et al., 2001; Al-Inany and Aboulghar, 2002). Further studies are obviously needed to clarify any possible benefits on the endometrial receptivity of early-follicular-phase down-regulation in short antagonist protocols. In conclusion, the use of aromatase inhibitor in combination with hcg significantly influences the local endocrine environment, not just during the period of androgen priming, but as far as 14 days after the cessation of treatment. Androgen priming with hcg and aromatase inhibitor may result in an increased number of good-quality embryos. This new protocol may be 2599

K.Lossl et al. used to provide flexibility to the short protocol. We find that the present results warrant further investigation and a randomized controlled study to evaluate the reproductive outcome of this protocol as compared with the conventional ones. References Al-Inany H and Aboulghar M (2002) GnRH antagonist in assisted reproduction: a Cochrane review. Hum Reprod 17,874 885. Andersen CY (1993) Characteristics of human follicular fluid associated with successful conception after in vitro fertilization. J Clin Endocrinol Metab 77,1227 1234. Biljan MM, Hemmings R and Brassard N (2005) The outcome of 150 babies following the treatment with letrozole or letrozole and gonadotropins. Fertil Steril 84(Suppl. 1), S64 S65. Casper RF and Mitwally MF (2005) Aromatase inhibitors for ovulation induction. Reprod Biol Endocrinal 3,54. Chadha S, Pache TD, Huikeshoven JM, Brinkmann AO and van der Kwast TH (1994) Androgen receptor expression in human ovarian and uterine tissue of long-term androgen-treated transsexual women. Hum Pathol 25,1198 1204. Garcia-Velasco JA, Moreno L, Pacheco A, Guillen A, Duque L, Requena A and Pellicer A (2005) The aromatase inhibitor letrozole increases the concentration of intraovarian androgens and improves in vitro fertilization outcome in low responder patients: a pilot study. Fertil Steril 84,82 87. Hild-Petito S, West NB, Brenner RM and Stouffer RL (1991) Localization of androgen receptor in the follicle and corpus luteum of the primate ovary during the menstrual cycle. Biol Reprod 44,561 568. Hillier SG, Tetsuka M and Fraser HM (1997) Location and developmental regulation of androgen receptor in primate ovary. Hum Reprod 12,107 111. Hohmann FP, Laven JS, de Jong FH and Fauser BC (2005) Relationship between inhibin A and B, estradiol and follicle growth dynamics during ovarian stimulation in normo-ovulatory women. Eur J Endocrinol 152,395 401. Horie K, Takakura K, Fujiwara H, Suginami H, Liao S and Mori T (1992) Immunohistochemical localization of androgen receptor in the human ovary throughout the menstrual cycle in relation to oestrogen and progesterone receptor expression. Hum Reprod 7,184 190. Iwase A, Ando H, Kuno K and Mizutani S (2005) Use of follicle-stimulating hormone test to predict poor response in in vitro fertilization. Obstet Gynecol 105,645 652. Ludwig M, Katalinic A and Diedrich K (2001) Use of GnRH antagonists in ovarian stimulation for assisted reproductive technologies compared to the long protocol. Meta-analysis. Arch Gynecol Obstet 265,175 182. Mannaerts BM, Geurts TB and Odink J (1998) A randomized three-way crossover study in healthy pituitary-suppressed women to compare the bioavailability of human chorionic gonadotrophin (Pregnyl) after intramuscular and subcutaneous administration. Hum Reprod 13,1461 1464. McNatty KP, Smith DM, Makris A, Osathanondh R and Ryan KJ (1979) The microenvironment of the human antral follicle: interrelationships among the steroid levels in antral fluid, the population of granulosa cells, and the status of the oocyte in vivo and in vitro. J Clin Endocrinol Metab 49,851 860. Mitwally MF, Casper RF and Diamond MP (2005) The role of aromatase inhibitors in ameliorating deleterious effects of ovarian stimulation on outcome of infertility treatment. Reprod Biol Endocrinol 3,54. The Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group (2004) Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 81,19 25. Staessen C, Camus M, Khan I, Van Smitz JWL, Wisanto A, Devroey P and Van Steirteghem AC (1989) An 18-month survey of infertility treatment by in vitro fertilization, gamete and zygote intrafallopian transfer, and replacement of frozen-thawed embryos. J In Vitro Fert Embryo Transf 6,22 29. Vendola KA, Zhou J, Adesanya OO, Weil SJ and Bondy CA (1998) Androgens stimulate early stages of follicular growth in the primate ovary. J Clin Invest 101,2622 2629. Weil SJ, Vendola K, Zhou J, Adesanya OO, Wang J, Okafor J and Bondy CA (1998) Androgen receptor gene expression in the primate ovary: cellular localization, regulation, and functional correlations. J Clin Endocrinol Metab 83,2479 2485. Weil S, Vendola K, Zhou J and Bondy CA (1999) Androgen and follicle stimulating hormone interactions in primate ovarian follicle development. J Clin Endocrinol Metab 84,2951 2956. Ziebe S, Petersen K, Lindenberg S, Andersen AG, Gabrielsen A and Andersen AN (1997) Embryo morphology or cleavage stage: how to select the best embryos for transfer after in-vitro fertilization. Hum Reprod 12,1545 1549. de Ziegler D (2003) The dawning of the non-cancer uses of aromatase inhibitors in gynaecology. Hum Reprod 18,1598 1602. Submitted on December 22, 2005; resubmitted on February 28, 2006, April 21, 2006; accepted on May 15, 2006 2600