Abstract. Introduction. RBMOnline - Vol 8. No Reproductive BioMedicine Online; on web 22 December 2003

Transcription

1 RBMOnline - Vol 8. No Reproductive BioMedicine Online; on web 22 December 2003 Article Randomized trial to compare the effect of recombinant human FSH (follitropin alfa) with or without recombinant human LH in women undergoing assisted reproduction treatment Dr Richard Marrs Richard Marrs received his medical degree and training in obstetrics and gynaecology at the University of Texas, USA, then moving into reproductive endocrinology at the University of Southern California School of Medicine. He has held several positions including Director of the Division of Reproductive Endocrinology and Infertility in the Department of Obstetrics and Gynecology at Cedars-Sinai Medical Centre in Los Angeles, Director of the Institute for Reproductive Research at the Hospital of the Good Samaritan, and Clinical Professor in the Department of Obstetrics and Gynecology at the Loma Linda University School of Medicine in California. Dr Marrs is the author of numerous papers, abstracts and books and has given many invited lectures on the use of assisted reproduction technology for infertility treatment. He was awarded the Lifetime Achievement award by the American Infertility Association and is a member of several scientific advisory and editorial boards. Currently he is the director for the Centre for Assisted Reproductive Medicine, Santa Monica/UCLA Medical Centre, Santa Monica in California and is a managing partner of California Fertility Partners there. Dr Marrs research interests include development of techniques for maximizing the success of assisted reproduction cycles. R Marrs 1, D Meldrum 2, S Muasher 3, W Schoolcraft 4, L Werlin 5, E Kelly 6,7 1 California Fertility Partners, Santa Monica, CA, USA; 2 Reproductive Partners Medical Group, Inc., Redondo Beach, CA, USA; 3 The Jones Institute for Reproductive Medicine, Fairfax, VA, USA; 4 Colorado Centre for Reproductive Medicine, Englewood, CO, USA; 5 Coastal Fertility Center, Irvine, CA, USA; 6 Serono Inc., 1 Technology Place, Rockland, MA 02370, USA 7 Correspondence: eduardo.kelly@serono.com Abstract Women undergoing intracytoplasmic sperm injection (ICSI) for male factor infertility were randomly assigned to receive ovarian stimulation in a long agonist protocol with a combination of recombinant human FSH (r-hfsh; Gonal-F ) and recombinant human LH (r-hlh; Luveris ) (n = 212) starting on day 6 of FSH stimulation until human chorionic gonadotrophin (HCG) at a daily fixed dose of 150 IU r-hlh, or with r-hfsh alone (n = 219). There was no significant difference in the number of metaphase II oocytes retrieved (10.3 versus 10.4) in patients treated with r-hfsh and r-hlh versus r-hfsh alone; however, more embryos were transferred in the LH-supplemented group (2.9 versus 2.8, P = 0.037). Overall, the implantation rates were 22.9 versus 27.0% in patients treated with r-hfsh and r-hlh versus with r-hfsh alone respectively (NS). The respective numbers of MII oocytes retrieved in patients <35 or 35 years were 11 versus 8.3 (P = 0.010) for patients treated with r-hfsh alone, and 10.7 versus 9.3 (NS) for those given supplemental r-hlh (150 IU) from day 6. Implantation rates in patients <35 years treated with r-hfsh were higher (30.7%) than those receiving r-hfsh and r-hlh, (23.5%) (P = 0.068). In patients 35 years, the implantation rates were 21.7% for those patients supplemented with 150 IU r-hlh from day 6 of stimulation versus 15.7% when treated with FSH alone (NS). Younger patients therefore do not seem to benefit from an LH-supplemented ovarian stimulation protocol, but women 35 years undergoing assisted reproduction may benefit from using r-hlh in addition to r-hfsh. Keywords: age, ovarian stimulation, recombinant FSH, recombinant LH Introduction At present, many women are experiencing age-related infertility owing to social trends that lead to deferred childbearing and to the current age of the baby boom generation (Practice Committee of the American Society for Reproductive Medicine, 2002). Although there is no strict definition of advanced reproductive age in women, infertility becomes more pronounced after 35 years of age. Thus, experts have stressed that evaluation and treatment of infertility should not be delayed in women 35 years of age or older (Speroff, 1993; Practice Committee of the American Society for Reproductive Medicine, 2002). 175

2 176 Treatment options for age-related infertility include ovarian stimulation with intrauterine insemination (IUI) and IVF. The presence of male factor, tubal disease, or endometriosis supports proceeding directly to IVF in women of advanced reproductive age (Practice Committee of the American Society for Reproductive Medicine, 2002). Pregnancy rates from IVF are generally higher than those from IUI, but these rates also decrease significantly with female age. This age-related decline in IVF success is related to both decreased ovarian responsiveness to gonadotrophins and a marked decrease in embryo implantation rates (Practice Committee of the American Society for Reproductive Medicine, 2002). The change in ovarian responsiveness with increasing age is recognized by the common practice of increasing the daily dose of gonadotrophins administered, mainly in the form of LHcontaining preparations, in association with the use of micro - dose gonadotrophin-releasing hormone (GnRH) agonist protocols. Such protocols reduce both excessive pituitary suppression and direct effects on the ovaries by the agonist (Olivennes et al., 1996; Laml et al., 1999; Phelps et al., 1999; Surrey and Schoolcraft, 2000; Lisi et al., 2001; De Moustier et al., 2002; Meo et al., 2002; Practice Committee of the American Society for Reproductive Medicine, 2002). This clinical practice thus represents an attempt to increase FSH but also LH effects on the ovaries, acting through both exogenously administered and endogenous gonadotrophins. According to current concepts of the roles of FSH and LH in folliculogenesis, LH plays an essential role in the final stages of follicular maturation (Hillier, 2001; Zeleznik, 2001). Once an appropriate (i.e. LH-responsive) stage of follicular development has been achieved in response to treatment with FSH, granulosa cells become receptive to LH stimulation and LH becomes capable of exerting its actions on both theca cells and granulosa cells. In fact, at non-saturating concentrations of FSH and LH, the responses are additive; indeed, it has been postulated that the maturing follicle reduces its dependence on FSH by acquiring LH receptors (Hillier, 2001; Zeleznik, 2001). Thus, LH may play an essential role in determining oocyte maturity and developmental potential in IVF cycles. However, exposure of the developing follicle to inappropriately high concentrations of LH may interfere with follicular and oocyte maturation and thus adversely affect the reproductive process (Hillier, 1994; Daya, 2001; Shoham, 2002). Now that recombinant human LH (r-hlh) is available as a standalone product, the present study was undertaken to investigate the usefulness of r-hlh supplementation in women treated with recombinant human FSH (r-hfsh) under pituitary suppression for assisted reproductive techniques. The specific aim of the study was to test whether the addition of r-hlh can be useful in the subgroup of women of advanced reproductive age while being not deleterious for the remaining assisted reproduction population when administered from stimulation day 6. It is well established that the best possibility of success is in the first cycle of IVF treatment and that there is a significant negative effect with increasing number of attempts thereafter (Templeton et al., 1996). Secondly, Daya (2003) has recently advocated the enrolment of subjects into studies who are being treated for the very first time, so as to reduce any potential bias resulting from previous treatments. For these reasons this study also analysed naive patient data, as in Westergaard et al. (2001). Materials and methods Patients The study included normo-ovulatory women aged with serum/plasma FSH within the normal range (up to 11.2 miu/ml); presence of both ovaries; male partner with a diagnosis of male factor infertility; and a requirement for ICSI for assisted reproduction. Exclusion criteria included: clinically significant systemic disease; smoking more than 10 cigarettes per day; any contraindication to pregnancy; serum/plasma LH:FSH ratio >2; and more than two previous ICSI cycles in which gonadotrophin stimulation was used. All patients gave written informed consent. Protocol This was a large, randomized, open-label, multicentre study carried out at 44 centres in the USA. Patients were randomized 1:1 to undergo ovarian stimulation with r-hfsh (Gonal-F ; Serono Laboratories, Geneva, Switzerland) alone or r-hfsh + r-hlh (Luveris ; Serono Laboratories). A computer-generated randomization sequence was used. Pituitary down-regulation was carried out using leuprolide acetate (Lupron ; TAP Pharmaceuticals, Inc., IL, USA), 0.5 mg/day, starting 7 8 days after estimated ovulation. Treatment with r-hfsh (225 IU/day) was started when serum oestradiol was <75 pg/ml. After 5 days, the r-hfsh dose could be increased by IU/day every 2 3 days if necessary. The combination group began treatment with r-hlh (150 IU/day) on stimulation day 6. The dose of r-hlh remained constant throughout the treatment period. The two products could be mixed and administered in the same syringe. Oestradiol concentrations and ultrasound measurements were assessed throughout the cycle to determine response to treatment. Patients received a single intramuscular injection of human chorionic gonadotrophin (HCG) (Profasi, Serono Laboratories; 10,000 USP units) when the following criteria had been met: largest follicle 18 mm; at least two other follicles 16 mm; and oestradiol concentration within the acceptable range (approximately 150 pg/ml per mature follicle). Oocytes were retrieved h after HCG administration according to the centre s usual practice. ICSI was performed according to standard procedures at each site. The maturational status of the oocytes and embryo grading were recorded according to published criteria (Veeck, 1990). The protocol stated that up to three embryos could be replaced 2 3 days after oocyte retrieval. Luteal phase support with progesterone was provided for all patients who received HCG. A blood sample for determination of serum β-hcg concentration was collected on days after HCG administration and assessed at the local laboratory. If positive, the test was repeated 2 7 days later to confirm pregnancy. For patients who became pregnant, an ultrasound scan was performed days after HCG administration and the number of fetal sacs and fetal heart activity was recorded. If a clinical pregnancy occurred, women were followed up to determine pregnancy outcome.

3 Safety Adverse events were recorded at each visit according to standard procedures. All patients who received at least one dose of study medication were included in the safety evaluation. Ovarian hyperstimulation syndrome (OHSS) was defined according to the classification proposed by Golan et al. (1989). End-points The primary efficacy variable was number of metaphase II oocytes retrieved. Secondary efficacy variables included: number of fertilized oocytes; number of pregnancies (total, biochemical and clinical); implantation rates; and number of live births. Subgroup analyses Additional analyses were performed with the patient population stratified according to age ( 35 years versus <35 years) and number of previous cycles (first ICSI cycle versus those who had undergone one or more previous cycles). Sample size and statistical methods It was calculated that a sample size of 280 patients would give 80% power to detect the expected difference in the primary endpoint, the number of metaphase II oocytes retrieved (based on previous studies of r-hfsh versus HMG) of 8.2 percentage points in favour of r-hfsh alone (Imthurn et al. 1996). Allowing for withdrawals from treatment, the planned sample size was 296 patients, 148 in each group. The total number of patients recruited eventually reached 431 in total and all patients irrespective of any protocol deviation were included in this analysis (intention to treat (ITT) population). For the primary efficacy end-point, the 95% confidence interval of the difference between treatment groups was calculated using an ANOVA model that included effects for treatment group and centre. Secondary efficacy variables were analysed using ANOVA for continuous variables and logistic regression, the Cochran Mantel Haenzsel test or exact test for categorical variables. In addition, a sensitivity analysis based on a logistic regression model was employed where required. Results Patient disposition and demographics The ITT population included a total of 431 patients who were enrolled and treated; 212 were randomized to the combination treatment and 219 to r-hfsh alone. The two groups were well matched for baseline demographic characteristics (Table 1). The combined r-hfsh + r-hlh group included 60 women aged 35 years and in the r-hfsh group there were 56. Basal serum LH concentrations did not differ significantly between the two groups (Table 1). Primary end-point and stimulation characteristics The majority of patients successfully completed ovarian stimulation and received HCG (97% in the combination group and 93% in the r-hfsh group). Most patients also underwent oocyte retrieval (96 and 92% respectively). Considering the ITT population as a whole, there was no significant difference in the number of metaphase II oocytes retrieved (10.4 ± 6.3 versus 10.3 ± 5.9) for the FSH alone versus LH-supplemented groups. There were no differences between the overall treatment groups for length of treatment, ampoules of r-hfsh used, follicular development and oestradiol on the day of HCG administration (Table 2). Outcome of ICSI and pregnancy Results for the efficacy end-points of the ITT population are presented in Table 3. None of the differences between groups was statistically significant, except that the mean number of embryos transferred was higher in the r-hfsh + r-hlh group (2.9 ± 0.6 versus 2.8 ± 0.7, P = 0.037). The number of high quality (A and B grade) embryos transferred was 2.4 ± 1.0 in the FSH + LH group and 2.2 ± 1.1 in the FSH alone group. Overall implantation rates were 22.9 and 27.0% respectively for the groups treated with r-hfsh + LH and r-hfsh alone (NS). Table 1. Baseline characteristics of the study population. Values are mean ± SD unless otherwise stated. Characteristic Treatment group Total patients P-value (n = 431) r-hfsh + r-hlh r-hfsh alone (n = 212) (n = 219) Age (years) 32.4 ± ± ± 3.8 NS BMI 24.3 ± ± ± 5.1 NS Smoker (%) a 13 (6.1) 16 (7.3) 29 (6.7) NS Duration of infertility 3.9 ± ± ± 2.9 NS (years) Basal FSH (miu/ml) 5.1 ± ± ± 1.9 NS Basal LH (miu/ml) 3.8 ± ± ± 2.6 NS Basal oestradiol 80.0 ± ± ± NS (pmol/l) Previous assisted 56 (26.4) 60 (27.4) 116 (26.9) NS reproduction cycle(s) (%) Values in parentheses are percentages. a Defined as consuming less than 11 cigarettes/day; NS = not significant. 177

4 Twenty-two patients were protocol violators (15 in the r-hfsh + r-hlh and 7 in the FSH alone group), as they had more than three embryos replaced. In order to estimate the impact of this imbalance, the analysis of clinical pregnancy was repeated twice: on the set of all patients randomized (ITT data set, n = 431) adjusted with the number of embryos transferred using a logistical regression model, and on the set of patients excluding those patients who did not fulfil the protocol (n = 409). As the ITT population was the primary population of interest, and the results for the two populations were very similar, only the results for the ITT population are presented here. For the ITT population, there were no statistically significant differences between groups for the clinical pregnancy rate (Table 3). Live birth rates are not available for the whole study population because of incomplete follow-up. Subgroup analyses There were no significant differences for the primary endpoint between patient subgroups stratified by age and number of previous cycles. For the <35 and 35 age groups, the respective numbers of metaphase II oocytes retrieved were 11.0 versus 8.3 (P = 0.010) for patients treated with r-hfsh alone and 10.7 versus 9.3 (NS) for those given supplemental r-hlh from day 6. For patients aged under 35, the number of high quality embryos transferred was 2.2 ± 1.1 in the group receiving FSH alone and 2.4 ± 1.0 in those receiving FSH + LH (NS). For the 35 age group, corresponding values were 2.1 ± 1.1 and 2.4 ± 1.1, respectively. Implantation rates in patients <35 years treated with r-hfsh alone were higher than for those receiving r-hfsh + r-hlh (30.7 versus 23.5%, Table 2. Gonadotrophin treatment and ovarian response. Values are mean ± SD unless otherwise stated. Characteristic Treatment group Total patients P-value r-hfsh + r-hlh r-hfsh alone (n = 431) (n = 212) (n = 219) Days to ovarian arrest 13.3 ± ± ± 3.3 NS Days of stimulation 9.3 ± ± ± 1.8 NS Ampoules of FSH 28.8 ± ± ± 10.3 NS Ampoules of LH 8.9 ± ± ± 5.1 <0.001 Follicles >10 mm 13.0 ± ± ± 6.4 NS on HCG day Follicles >15 mm 9.2 ± ± ± 4.7 NS on HCG day Oestradiol (pmol/l) 9004 ± ± ± 5012 NS on HCG day Patients with follicular 204 (96.2) 202 (92.2) 406 (94.2) NS puncture (%) Number of oocytes 13.6 ± ± ± 7.1 NS retrieved Values in parentheses are percentages. NS = not significant. Table 3. Oocyte retrieval and outcome of ICSI. Values are mean ± SD unless otherwise stated. Characteristic Treatment group Total patients P-value r-hfsh + r-hlh r-hfsh alone (n = 431) (n = 212) (n = 219) Number of MII oocytes 10.3 ± ± ± 6.1 NS Number of 2PN oocytes 7.6 ± ± ± 4.6 NS Fertilization rate 55.4 ± ± ± 22.0 NS Embryos cryopreserved 2.1 ± ± ± 3.5 NS Patients with embryos 198 (93.4) 197 (90.0) 395 (91.6) NS transferred (%) Number of embryos 2.9 ± ± ± transferred Patients with clinical 90 (42.5) 91 (41.6) 181 (42.0) NS pregnancy (%) 178 Values in parentheses are percentages. NS = not significant.

5 P = 0.068). In patients aged 35 years, implantation rates were 21.7% for those patients supplemented with 150 IU r-hlh from day 6 of stimulation, versus 15.7% for those treated with r-hfsh alone (NS) (Figure 1). For the ITT and evaluable patient population, the clinical pregnancy rate was higher in women aged 35 years who received both r-hfsh and r-hlh, although this difference did not reach statistical significance (Table 4). However, the difference in clinical pregnancy rates was significant (P < 0.03) with an odds radio (95% CI) of 3.13 ( ) for the ITT population aged 35 years who were undergoing their first assisted reproduction cycle (Table 4). This subgroup of patients also had significantly more embryos transferred in the r-hfsh + r-hlh group (3.1 ± 0.6 versus 2.8 ± 0.7; P = 0.04). Using a sensitivity analysis on this ITT population aged 35 years based on a logistic regression model adjusted with the number of embryos transferred, the clinical pregnancy rates were 36.3% versus 19.6% (FSH + LH versus FSH, NS). Younger women, by contrast, showed a trend towards better outcomes when treated with r-hfsh alone for some of the secondary end-points, with the clinical pregnancy rate for the ITT population being 42.9 versus 45.4% (FSH + LH versus FSH, NS) and the adjusted clinical Figure 1. Implantation rates per cycle in women of different age groups who underwent ovarian stimulation using recombinant human FSH with and without supplemental recombinant human LH. pregnancy rate being 39.8 versus 45.8% (FSH + LH versus FSH, NS). The effect of age on clinical pregnancy rates was examined with and without adjustment for differences in numbers of embryos transferred. For the ITT population treated with r-hfsh alone, pregnancy rates were 74/163 (45.4%) for those aged <35 years and 17/56 (30.4%) for those aged 35 years (P = 0.051). Adjusted clinical pregnancy rates were significantly different (43.1 versus 26.8%, P = 0.038) in favour of the younger age group. For the r-hfsh + r-hlh group, there was no difference in clinical pregnancy rates between younger and older women for either the ITT (42.9 versus 41.5%, NS) or adjusted (41.9 versus 39.6%, NS) analyses. Results were similar for women undergoing their first ART cycle, except that for the r-hfsh group the difference in pregnancy rates was significant for the ITT analysis (P = 0.03) as well as the adjusted analysis. Safety results There was no difference between groups in the rate of adverse events; 45 patients in the combination group (21%) and 49 (22%) in the r-hfsh group experienced at least one adverse event. The most common adverse events were headache and OHSS, each occurring in 18 patients. Most adverse events were mild or moderate in severity. Three patients withdrew from the study because of risk of OHSS, two in the combination group and one in the r-hfsh group. Discussion It is well established that successful IVF and embryo transfer requires both stimulation of the ovary and suppression of the pituitary. Thus, exogenous gonadotrophins and GnRH analogues are the key hormones required to maximize IVF success. The long protocol of GnRH agonist administration is the most commonly adopted protocol for assisted reproduction cycles worldwide, mainly in young, normogonadotrophic women (Barbieri and Hornstein, 1999). GnRH agonists do not usually result in total elimination of LH and it is accepted that <1% of follicular LH receptors need to be occupied to elicit a maximal steroidogenic response. Accordingly, resting concentrations of Table 4. Clinical pregnancy rates per started cycle in the intention to treat (ITT) group and evaluable patient population for women aged <35 and 35 years after ovarian stimulation with r-hfsh + r-hlh or r-hfsh alone. Maternal age (years) Treatment group P-value b r-hfsh + rfsh r-hlh alone All <35 ITT (%) 63/147 (42.9) 74/163 (45.4) Adjusted a (%) All 35 ITT (%) 27/65 (41.5) 17/56 (30.4) Adjusted a (%) Aged 35 and undergoing ITT (%) 22/48 (45.8) 9/40 (22.5) first assisted reproduction Adjusted a (%) cycle Values in parentheses are percentages. a Predicted clinical pregnancy rates and P-value from a regression logistic model adjusted for the number of embryos transferred. b All P-values were not significant. 179

6 180 LH (1 10 IU/l) should be sufficient to provide maximal stimulation of thecal cells (Chappel and Howles, 1991). This is further supported by a recent study in down-regulated young oocyte donors showing that the inclusion of exogenous LH activity (in the form of HMG) in the ovarian stimulation protocol can have beneficial effects on oocyte yield and quality in those donors with LH concentrations <1 IU/l once pituitary suppression was achieved (Tesarik and Mendoza, 2002). In this study, however, a long acting depot agonist preparation was also used to achieve pituitary suppression (Tesarik and Mendoza, 2002). This is more convenient for patient use, but it is associated with more profound, sustained down-regulation than the short acting preparations (Yim et al., 2001). The use in such a study (Tesarik and Mendoza, 2002) of oral norethisterone preceding depot agonist administration may have induced an even more profound pituitary suppression. Some studies have also reported a significant reduction in FSH efficacy versus treatment with supplemental LH and reflected in more vials of FSH required (Balasch et al., 2003) when a subcutaneous or depot GnRH agonist was employed. This was not the case here as there was no difference in the stimulation characteristics in the two treatment groups. Caution should therefore be used when considering results from studies where depot agonists are used compared with the present report, using a routine downregulation scheme that is more commonly used in assisted reproduction practice. The above notwithstanding, there seems to be a range of LH concentrations obtained in patients treated with GnRH agonists. With the use of urinary FSH preparations containing negligible LH activity it is possible that there may be a subgroup of patients with low LH concentrations in whom ovarian responses are influenced (Fleming et al., 1996, 1998; Westergaard et al., 2000). This can become especially relevant considering the following points. First, such women cannot be identified in advance by measuring LH concentrations after down-regulation (Loumaye et al., 1997; De Placido et al., 2000). Second, oocyte maturity and fertilization rate in assisted reproduction may be influenced by the particular hormonal stimulation that preceded oocyte retrieval (Pieters et al., 1991). It has recently been suggested that, in a subset of patients, a suboptimal ovarian response to the long GnRH agonist protocol associated with the use of FSH-only gonadotrophin preparations may be due to low LH activity caused by low serum concentrations of LH and/or low LH bioactivity (De Placido et al., 2001). The following facts support this possibility. First, the ovarian response to ovarian stimulation with gonadotrophins is often reduced in patients receiving long-term down-regulation (Marcus and Edwards, 1994; Fábregues et al., 1998) or in those who required prolonged GnRH agonist treatment to achieve down-regulation, with a subsequent profound suppression of endogenous gonadotrophins (Fleming et al., 1998; Ravhon et al., 2000). Although pharmacological doses of FSH alone are capable of stimulating ovarian follicular development, LH is strictly necessary to achieve final follicular maturation and oocyte fertilization (Couzinet et al., 1988; Balasch et al., 1995). Second, one alternative approach in low-responders to the prolonged down-regulation associated with the long protocol consists of the flare regimens that involve follicular phase initiation of GnRH agonist with minimal delay before commencing ovarian stimulation with gonadotrophins. In such regimens a direct effect of GnRH agonist enhancing LH secretion exists (Surrey and Schoolcraft, 2000). Finally, although immunoreactive LH and bioactive LH concentrations are related, differences are often observed (Schroor et al., 1999), and this may explain the lack of correlation reported between the immunoreactive concentrations of LH and the requirement for LH during folliculogenesis (De Placido et al., 2000). In recent years, a number of studies have provided new information about the role of LH in the process of ovulation induction. In particular, Sullivan et al. showed that r-hlh can substitute for FSH in supporting the later stages of follicular growth (Sullivan et al., 1999). Filicori and Cognigni have advocated the use of a biphasic regimen for ovarian stimulation providing mainly FSH activity in the early to mid-follicular phase and mainly LH activity in the later stages of ovarian stimulation (Filicori and Cognigni, 2001). Based on these data it might be argued that treatment with r-hlh at a higher dose or starting earlier in the cycle might have given better results for the combined treatment group as a whole. However, at the time this study was designed, protocols involving the separate administration of r-hfsh and r-hlh were rare. The rationale for using a dose of 150 IU/day was based on the concentrations of LH measured in women down-regulated with GnRH agonists (le Cotonnec et al., 1998). This dose of r-hlh achieved a C max of 1.2 IU/l, a concentration found to be important in hypogonadotrophic women stimulated with r-hfsh alone (O Dea et al., 2000). Concentrations of LH <1.2 IU/l were associated with insufficient oestradiol concentrations and a failure to become pregnant. There appears to be a therapeutic window of LH concentrations, because high concentrations of LH have been associated with atresia of developing follicles in women with hypogonadotrophic hypogonadism or polycystic ovary disease (Shoham, 2002; Loumaye et al., 2003). Ben-Amor and colleagues have published a preliminary report of a study where infertile women down-regulated with buserelin received either FSH (150 IU/day) or FSH (150 IU/day) plus a supplementation of r-hlh (75 IU/day) started when the leading follicle had reached 14 mm (mean diameter), until the day of HCG administration (Ben-Amor et al., 2000). Mean numbers of oocytes and metaphase II oocytes retrieved were similar between the two groups, suggesting a lack of benefit of r-hlh supplementation. Compared with the present study the patients received r-hlh at a lower dose (75 IU/day) and for an average of 2 days only. The women also had a lower average age compared with those in the present study (30.4 years). In the light of current observations, their findings are consistent with the fact that r- hlh supplementation might be of benefit to older patients only. One caveat in the present study is that significantly more embryos were transferred in the group that received LH supplementation (mean 2.9 versus 2.8, P = 0.04). Although this difference was not significant in the younger patient population, there was a trend in the older subpopulation (3.1 versus 2.8, P = 0.08) receiving LH. It seems unlikely that such a small difference, even if statistically significant, could be clinically meaningful. This was confirmed by adjusting the pregnancy rates to take account of the difference in embryos transferred (Table 4), showing no difference between unadjusted and adjusted outcomes. Any effect of r-hlh supplementation on clinical outcomes may be most evident in patients of advanced reproductive age, who

7 are more prone to have a decreased response to ovarian stimulation and lower implantation rates (FIVNAT, 1999, 2000; Practice Committee of the American Society for Reproductive Medicine, 2002). The present study adds new data to the subject when showing that in patients receiving r-hfsh + r-hlh, pregnancy rates were similar in the younger and older age groups, i.e. there was no age-related fall in pregnancy rates with this regimen. This finding remained valid when the analysis was adjusted to allow for differences in the number of embryos transferred. However, in women receiving r-hfsh alone, there was a significant decline in pregnancy rates for women 35 years of age compared with those 35 years old (P = 0.05 for the ITT population and P = adjusted for differences in number of embryos transferred). Multiple novel roles of LH have been recently proposed and it has been postulated that LH may affect IVF results both by determining oocyte quality and by influencing uterine receptivity via ovarian oestradiol secretion or through direct effects on endometrium, myometrium, and uterine artery and vein (Rao, 2001; Shemesh, 2001; Tesarik and Mendoza, 2002). Therefore, it is tempting to postulate that r-hlh added to r-hfsh (from stimulation day 6 in the present investigation) improved oocyte developmental competence and/or implantation in an older patient population. Finally, it is well known that patients aged 35 years have a lower likelihood of successful pregnancy after IVF compared with younger women (Templeton et al., 1996; Templeton and Morris, 1998). Also, irrespective of the woman s age, there is a significant decline in the probability of conception with successive assisted reproduction cycles (Tan et al., 1994; de Mouzon et al., 1998; FIVNAT, 1999, 2000). Remarkably, the present study showed the greatest benefit of supplementary r- hlh compared with r-hfsh alone in older women in their first assisted reproduction cycle. This finding adds to the results reported by Lisi and colleagues (Lisi et al., 2001, 2002). In a preliminary study comparing r-hfsh alone with r-hfsh + r-hlh in patients who required >3000 IU r-hfsh on previous attempts, there were significant differences in implantation rate and ongoing pregnancy rate, favouring the r-hfsh + r-hlh group (Lisi et al., 2001). However, the number of patients (12) and cycles (17) involved was small. A larger study involving 122 cycles using r-hfsh + r-hlh and 331 using r-hfsh alone in unselected patients found no differences in endocrine, embryological or outcome measures (Lisi et al., 2002). There was a higher implantation rate in the group given added r-hlh among patients with low LH concentrations after downregulation and among those who required large amounts of r- hfsh to complete ovarian stimulation. The present results complement those of Lisi et al. by focusing attention on a subgroup of patients who may benefit from supplementary r- hlh, and on the possible beneficial effects of supplementary r- hlh on implantation rates. In conclusion, patients of older reproductive age undergoing assisted reproduction might benefit from the addition of r-hlh. Further studies are required to examine the possible effect of LH supplementation on age-related decline in pregnancy rate and the physiological mechanisms behind this observation. Acknowledgements The study investigators were M Abuzeid, S Awadalla, D Barad, J Batzofin, L Bonaventura, S Corson, J Crain, K Doody, T Falcone, G Frishman, K Ginsburg, J Goldfarb, D Grainger, G Hill, A Jacobson, B Kaplan, M Levy, R Marrs, L Marshall, J Martin, E Marut, JB Massey, W Maxson, P McShane, D Meldrum, C Miller, V Montgomery-Rice, SJ Muasher, P Patrizio, CM Peterson, J Queenan, R Reindollar, V Sahakian, W Schoolcraft, G Sher, K Silverberg, J Stassart, M Steinkampf, E Surrey and L Werlin. The authors would also like to thank L Ferrande, H Holbrook, P Engrand and C Howles for their help in conducting the study and analysing the data. References Balasch J, Miró F, Burzaco I et al The role of luteinizing hormone in human follicle development and oocyte fertility: evidence from in-vitro fertilization in a woman with long-standing hypogonadotrohic hypogonadism and using recombinant human follicle stimulating hormone. Human Reproduction 10, Balasch J, Peñarrubia J, Fábregues F et al Ovarian responses to recombinant FSH or HMG in normogonadotropic women following pituitary desensitization by a depot GnRH agonist for assisted reproduction. Reproductive BioMedicine Online 7, Barbieri RL, Hornstein MD 1999 Assisted reproduction-in vitro fertilization success is improved by ovarian stimulation with exogenous gonadotropins and pituitary suppression with gonadotropin-releasing hormone analogues. Endocrine Reviews 20, Ben-Amor A-F, on behalf of the Study Group 2000 The effect of luteinizing hormone administered during the late follicular phase in normo-ovulatory women undergoing in-vitro fertilization. Human Reproduction 15 (Abstract Book 1), 46. Chappel SC, Howles C 1991 Re-evaluation of the roles of luteinizing hormone and follicle-stimulating hormone in the ovulatory process. Human Reproduction 6, Couzinet B, Lestrat N, Brailly S et al Stimulation of ovarian follicular maturation with pure follicle stimulating hormone in women with gonadotropin deficiency. Journal of Clinical Endocrinology and Metabolism 66, Daya S 2001 Gonadotrophin-releasing hormone agonist protocols for pituitary desensitization in in vitro fertilization and gamete intrafallopian transfer cycles (Cochrane Review). In: The Cochrane Library Issue 4. Oxford: Update Software. Daya S 2003 Pitfalls in the design and analysis of efficacy trials in subfertility: Associate editor s commentary: on the article Common statistical errors in the design and analysis of subfertility trials by A. Vail and E. Gardner. Human Reproduction 18, De Moustier B, Brinsden P, Bungum L et al The effects of combined treatment of recombinant (r) FSH and r-hlh in ratios 1:1, 2:1 and 3:1 in women aged years undergoing IVF-ICSI treatment. Human Reproduction 17 (Abstract Book 1), 55. de Mouzon J, Rossin-Amar B, Bachelot A et al FIVNAT. Influence of attempt rank in in vitro fertilization. Contraception, Fertilité, Sexualité 26, De Placido G, Alviggi C, Mollo A et al Recombinant follicle stimulating hormone is effective in poor responders to highly purified follicle stimulating hormone. Human Reproduction 15, Fábregues F, Balasch J, Creus M et al Long-term downregulation does not improve pregnancy rates in an in vitro fertilization program. Fertility and Sterility 70, Filicori M, Cognigni GE 2001 Roles and novel regimens of luteinizing hormone and follicle-stimulating hormone in ovulation induction. Journal of Clinical Endocrinology and Metabolism 86, FIVNAT 1999 Dossier FIVNAT Bilan de l année 98. FIVNAT 2000 Dossier FIVNAT Bilan de l année 99. Fleming R, Chung CC, Yates RWS et al Purified urinary follicle 181

8 182 stimulating hormone induces different hormone profiles compared with menotrophin, dependent upon the route of administration and endogenous luteinizing hormone activity. Human Reproduction 11, Fleming R, Lloyd F, Herbeert M et al Effects of profound suppression of luteinizing hormone during ovarian stimulation on follicular activity, oocytes and embryo function in cycles stimulated with purified follicle stimulating hormone. Human Reproduction 13, Golan A, Ron-El R, Herman A, Weinraub Z, Soffer Y, Caspi E 1989 Ovarian hyperstimulation syndrome: an update review. Obstetrical and Gynecological Survey 44, Hillier SG 1994 Current concepts of the roles of follicle stimulating hormone and luteinizing hormone in folliculogenesis. Human Reproduction 9, Hillier SG 2001 Gonadotropic control of ovarian follicular growth and development. Molecular and Cellular Endocrinology 179, Imthurn B, Macas E, Rosselli et al Nuclear maturity and oocyte morphology after stimulation with highly purified follicle stimulating hormone compared to human menopausal gonadotrophin. Human Reproduction 11, Laml T, Obruca A, Fischl F et al Recombinant luteinizing hormone in ovarian hyperstimulation after stimulation failure in normogonadotropic women. Gynecological Endocrinology 13, le Cotonnec JY, Loumaye E, Porchet HC et al Pharmacokinetic and pharmacodynamic interactions between recombinant human luteinizing hormone and recombinant human follicle-stimulating hormone. Fertility and Sterility 69, Lisi F, Rinaldi L, Fishel S et al Use of recombinant FSH and recombinant LH in multiple follicular stimulation for IVF: a preliminary study. Reproductive BioMedicine Online 3, Lisi F, Rinaldi L, Fishel S et al Use of recombinant LH in a group of unselected IVF patients. Reproductive BioMedicine Online 5, Loumaye E, Engrand P, Howles CM et al Assessment of the role of serum luteinizing hormone and estradiol response to follicle-stimulating hormone on in vitro fertilization outcome. Fertility and Sterility 67, Loumaye E, Engrand P, Shoham Z et al Clinical evidence for an LH ceiling effect induced by administration of recombinant human LH during the late follicular phase of stimulated cycles in World Health Organization type I and type II anovulation. Human Reproduction 18, Marcus SF, Edwards RG 1994 High rates of pregnancy after longterm down-regulation of women with severe endometriosis. American Journal of Obstetrics and Gynecology 171, Meo F, Rainieri DM, Khadum I et al Ovarian response and in vitro fertilization outcome in patients with reduced ovarian reserve who were stimulated with recombinant folliclestimulating hormone or human menopausal gonadotropin. Fertility and Sterility 77, O Dea L St L and The US Recombinant LH Study Group 2000 Recombinant LH in support of recombinant FSH in female hypogonadotropic hypogonadism evidence of threshold effects (abstract). Fertility and Sterility 74 (suppl. 3), S36. Olivennes F, Righini C, Fanchin R et al A protocol using a low dose of gonadotropin-releasing hormone agonist might be the best protocol for patients with high follicle-stimulating hormone concentrations on day 3. Human Reproduction 11, Phelps JY, Figueira-Armada L, Levine AS et al Exogenous luteinizing hormone (LH) increases estradiol response patterns in poor responders with low serum LH concentrations. Journal of Assisted Reproduction and Genetics 16, Pieters MHEC, Dumoulin JCM, Engelhart CM et al Immaturity and aneuploidy in human oocytes after different stimulation protocols. Fertility and Sterility 56, Practice Committee of the American Society for Reproductive Medicine 2002 Aging and infertility in women: a committee opinion. Fertility and Sterility 78, Rao CV 2001 Multiple novel roles of luteinizing hormone. Fertility and Sterility 76, Ravhon A, Aurell R, Lawrie H et al The significance of delayed suppression using buserelin acetate and recombinant follicle-stimulating hormone in a long protocol in vitro fertilisation program. Fertility and Sterility 73, Schroor DC, van Weissenbruch M, Engelbert M et al Bioactivity of luteinizing hormone during normal puberty in girls and boys. Hormone Research 51, Shemesh M 2001 Actions of gonadotrophins on the uterus. Reproduction 121, Shoham Z 2002 The clinical therapeutic window for luteinizing hormone in controlled ovarian stimulation. Fertility and Sterility 77, Speroff L 1993 Aging and reproduction. Postgraduate Obstetrics and Gynecology 13, 1 5. Sullivan MW, Stewart-Akers A, Krasnow JS et al Ovarian responses in women to recombinant follicle-stimulating hormone and luteinizing hormone (LH): a role for LH in the final stages of follicular maturation. Journal of Clinical Endocrinology and Metabolism 84, Surrey ES, Schoolcraft WB 2000 Evaluating strategies for improving ovarian response of the poor responder undergoing assisted reproductive techniques. Fertility and Sterility 73, Tan SL, Doyle P, Maconochie N et al Pregnancy and birth rates of live infants after in vitro fertilization in women with and without previous in vitro fertilization pregnancies: a study of eight thousand cycles at one center. American Journal of Obstetrics and Gynecology 170, Templeton A, Morris JK 1998 Reducing the risk of multiple births by transfer of two embryos after in vitro fertilization. New England Journal of Medicine 339, Templeton A, Morris JK, Parslow W 1996 Factors that affect outcome of in vitro fertilisation treatment. Lancet 348, Tesarik J, Mendoza C 2002 Effects of exogenous LH administration during ovarian stimulation of pituitary down-regulated young oocyte donors on oocyte yield and developmental competence. Human Reproduction 17, Veeck L 1990 The morphological assessment of human oocytes and early concepti. In: Keel BA, Webster BW (eds) Handbook of the Laboratory Diagnosis and Treatment of Infertility. CRC Press, Boca Ratón, pp Westergaard LG, Laursen SB, Andersen YC 2000 Increased risk of early pregnancy loss by profound suppression of luteinizing hormone during ovarian stimulation in normogonadotrophic women undergoing assisted reproduction. Human Reproduction 15, Westergaard LG, Erb K, Laursen SB, Rex S, Rasmussen PE 2001 Human menopausal gonadotropin versus recombinant folliclestimulating hormone in normogonadotropic women downregulated with a gonadotropin-releasing hormone agonist who were undergoing in vitro fertilization and intracytoplasmic sperm injection: a prospective randomized study. Fertility and Sterility 76, Yim SF, Lok IH, Cheung, LP et al Dose-finding study for the use of long-acting gonadotrophin-releasing hormone analogues prior to ovarian stimulation for IVF. Human Reproduction 16, Zeleznik AJ 2001 Follicle selection in primates: Many are called but few are chosen. Biology of Reproduction 65, Received 1 May 2003; refereed 6 June 2003; accepted 7 November 2003.