GnRH antagonist-induced inhibition of the premature LH surge increases pregnancy rates in IUI-stimulated cycles. A prospective randomized trial

Transcription

1 Human Reproduction Page 1 of 8 Hum. Reprod. Advance Access published October 10, 2006 doi: /humrep/del337 GnRH antagonist-induced inhibition of the premature LH surge increases pregnancy rates in IUI-stimulated cycles. A prospective randomized trial A.Allegra 1, A.Marino, F.Coffaro, P.Scaglione, F.Sammartano, G.Rizza and A.Volpes ANDROS Day Surgery Reproductive Medicine Unit, Palermo, Italy 1 To whom correspondence should be addressed at: ANDROS Day Surgery Reproductive Medicine Unit, Via Ausonia 43 45, Palermo, I-90144, Italy. allegra@centroandros.it BACKGROUND: Our prospective randomized controlled trial was designed to assess whether the use of GnRH antagonists can improve the success rate of controlled ovarian stimulation (COS)/intrauterine insemination (IUI) treatments, via inhibition of the premature LH rise. METHODS: A total of 104 patients were randomly divided, using a randomization list, into two groups: in group A (n = 52), recombinant FSH (rfsh) was given with GnRH antagonist Cetrorelix, and in group B (n = 52), the patients received rfsh alone in a manner similar to that of group A. The primary outcome measure was clinical pregnancy rate per couple. RESULTS: The pregnancy rate per patient was 53.8% in group A and 30.8% in group B (P = 0.017). The rate of premature LH surge was 7% in group A and 35% in group B (P < ). A premature luteinization was observed in two cycles of 144 in group A (1.4%) and in 16 cycles of 154 in group B (10.4%) (P = 0.001). The mean values of LH and progesterone were significantly lower in patients receiving GnRH antagonist than in those who did not (3.3 ± 3.3 miu/ml in group A versus 9.9 ± 7.9 miu/ml in group B, P < , for LH; 1.3 ± 1.1ng/ml versus 2.1 ± 1.9ng/ml for group A and B, respectively, P < , for progesterone). CONCLUSION: The use of GnRH antagonist in COS/IUI cycles improves pregnancy rate, preventing the premature LH rise and luteinization. Key words: controlled ovarian stimulation/gnrh antagonists/infertility/intrauterine insemination/premature luteinization Introduction Intrauterine insemination (IUI) with the husband s semen is the first line technique for many conditions of infertility such as unexplained infertility, mild male factor infertility and minimal or mild endometriosis. Several studies have demonstrated that IUI with controlled ovarian stimulation (COS) is better than IUI alone (Chaffkin et al., 1991; Nulsen et al., 1993; Hannoun et al., 1998; Guzick et al., 1999; Nuojua-Huttunen et al., 1999; Dickey et al., 2002; Duran et al., 2002; Houmard et al., 2002; Kaplan et al., 2002; Steures et al., 2004). The combination of IUI with gonadotrophin treatment also resulted in higher pregnancy rates when compared with IUI cycles with clomiphene citrate stimulation (Hannoun et al., 1998; Matorras et al., 2002). Finally, the use of IUI with COS when compared with COS alone resulted in improved cycle fecundity (Serhal et al., 1988; Dodson and Haney, 1991; Zeyneloglu et al., 1998; Cohlen et al., 2000). Moreover, it has been demonstrated that COS/IUI has a similar cumulative pregnancy rate when compared with IVF (Peterson et al., 1994; Goverde et al., 2000; Hughes, 2003; Pandian et al., 2003). The success rate of IUI with ovulation induction varies widely, with pregnancy rates ranging between 8 and 18% per cycle (Dodson et al., 1991; Sengoku et al., 1994; Hannoun et al., 1998; Guzick et al., 1999; Goverde et al., 2000). These discrepancies in pregnancy rates found among the various published studies are due to the selection of patients, duration of infertility, aetiology of infertility, sperm preparation, total number of motile sperm inseminated, number of inseminations, monitoring of the cycle, timing of IUI and protocols of ovarian stimulation. However, the application of COS for IUI is associated with an increased risk of multiple pregnancies and the occurrence of ovarian hyperstimulation syndrome (OHSS). It has been reported that there is no way of reducing the rate of multiple births after ovarian stimulation without also reducing the pregnancy rate (Collins, 1994; Dickey, 2003). More recently, low-dose protocols with recombinant FSH (rfsh) have been introduced into clinical practice to avoid the risk of multiple gestation, which is the main complication of COS/ IUI treatments (Hughes et al., 1998; Ragni et al., 2004). The above-mentioned authors have shown that with low-dose treatments, it is possible to have similar clinical pregnancy rates with an important reduction of OHSS and multiple pregnancy rates. The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. 1 For Permissions, please journals.permissions@oxfordjournals.org

2 A.Allegra et al. Another challenge to optimize the COS/IUI outcomes is to prevent the occurrence of the premature LH rise and consequent luteinization which, as is well known, is a possible complication of stimulated cycles (Fleming and Coutts, 1986; Manzi et al., 1995; Cohlen et al., 1998; Cunha-Filho et al., 2003). It has been calculated that 24% of IUI cycles suffer from premature LH surge (Cohlen et al., 1998) and this can result in IUI procedure cancellation. Obviously, this represents economic and psychological stress for the patients. Manzi et al. (1995) showed that patients who underwent COS/ IUI treatment and had premature LH surge demonstrated much better pregnancy rates in the subsequent cycle when a GnRH analogue was added, thus avoiding the LH premature surge. GnRH agonists have been the standard of care for more than a decade in reducing the incidence of premature LH surge by reversibly blocking pituitary gonadotrophin secretion in IUIstimulated cycles (Allegra et al., 1990; Dodson et al., 1991; Gagliardi et al., 1991; Zikopoulos et al., 1993). Nevertheless, these drugs are nowadays completely abandoned in IUI cycles because of the excessive follicular simultaneous selection they determine (with consequent higher incidence of multiple pregnancy and OHSS) and because of the long pretreatment period required. More recently, as an alternative to GnRH agonists, GnRH antagonists have been proposed to prevent the premature LH surge during IVF cycles (The European and Middle East Orgalutran Study Group, 2001; Messinis et al., 2005) and COS/IUI treatments (Ragni et al., 2001, 2004; Gomez- Palomares et al., 2005; Zikopoulos et al., 2005; Lambalk et al., 2006). These drugs do not produce a flare-up effect reducing synchronous follicular pool recruitment. Moreover, the potential advantage of a GnRH antagonist is that pituitary gonadotrophin secretion is suppressed immediately after the start of therapy. Therefore, co-treatment with GnRH antagonists can be restricted to the time in the cycle where there is a risk of a premature increase in LH. Our prospective randomized controlled trial was designed to assess whether the use of GnRH antagonists can improve the success rate of COS/IUI treatments, through the inhibition of premature LH rise. We have investigated the use of Cetrorelix in a non-ivf setting in women undergoing COS/IUI cycles. The primary end-point of the trial was the clinical pregnancy rate of a group of patients treated with rfsh and GnRH antagonist compared with a group of patients treated with rfsh alone, during IUI cycles. Materials and methods This was a prospective randomized controlled trial performed between May 2002 and December 2004 to assess the efficacy of the GnRH antagonist in improving pregnancy rate in COS/IUI cycles. A total of 104 patients were recruited for the IUI programme at Reproductive Medicine Unit of ANDROS Day Surgery, Palermo. Couples included in this study suffered from unexplained infertility, mild male factor infertility and endometriosis (stage I II). The patients had a clinical primary or secondary infertility lasting for at least 18 months (but <6 years). The demographic and infertility characteristics of the women are presented in Table I. Inclusion criteria were as follows: 2 Table I. Basic characteristics of the patients Group A (n=52) Group B (n=52) P-value Age (years) a 33 ± ± 3.6 NS Duration of infertility (years) a 2.7 ± ± 1.2 NS Cause of infertility b Unexplained (%) 27 (51.9) 36 (69.2) NS Mild male factor (%) 23 (44.2) 14 (26.9) NS Endometriosis stage I II (%) 2 (3.9) 2 (3.9) NS Primary infertility (%) b 42 (80.8) 31 (59.6) Secondary infertility (%) b 10 (19.2) 21 (40.4) NS, not significant. a Values are expressed as mean ± SD. Student s t-test for unpaired data. b The frequency distribution analysis was performed by Chi-square test. Patients at least 18 years of age but not >38 years; BMI between 18 and 30 kg/m 2 ; Normal regular menstrual cycles ranging from 24 to 35 days in length, normal basal serum FSH, thyroid-stimulating hormone (TSH) and prolactin levels, normal uterine cavity and bilateral tubal patency, assessed by hysterosalpingography and/or hydrolaparoscopy with chromosalpingography and hysteroscopy. Patients with previous attempts of COS or COS/IUI were eligible at least 6 months after treatment. Couples with minimal or mild endometriosis (stage I II) were considered eligible for the study, at least 3 months after the end of any treatment. Mild male factor infertility was defined as the presence of abnormal semen variables according to the World Health Organization criteria for normality (WHO 1999) but with a normal morphology ³5% and a total number of motile spermatozoa recovered after Pellet Swim-up ³ / ml. The protocol of the study was approved by the local Ethical Committee. After informed consent was obtained, subjects who complied with all selection criteria were randomly assigned to one of two treatment groups by giving them a code number from a randomization list in order of enrolment. To guarantee the concealment of allocation, a staff member, who was not directly involved in the study, was in possession of the randomization list; in this way, after receiving information from the physician recruiting couples, the staff member followed his own randomization list when allocating each couple. Once assigned to a treatment group, the patient received up to four sequential treatment cycles. On the third day of their menstrual period, the patients were examined by a baseline ultrasound (Aloka SSD-1700, 5.0 MHz, Aloka Japan). In the first group (group A; n = 52), a dose of IU of rfsh (Gonal-F, Serono, Geneva, Switzerland) per day, according to the patient s age ( 30 or >30 years, respectively), was given to induce follicular recruitment for 5 days. Ovarian stimulation was monitored on alternate days from day 8 of the cycle (day 6 of ovarian stimulation), and the dosage of rfsh was increased or decreased by the investigator, depending on the patient s response. From the day in which at least one follicle ³14 mm in mean diameter was detected by ultrasound scan, LH, progesterone and 17b estradiol (E 2 ) serum levels were determined using the Automated Chemiluminescence Systems (ACS:180, Bayer Health Care LLC, Leverkusen, Germany) for in vitro quantitative determination. Premature LH rise was defined as LH ³10 miu/ml and premature progesterone rise as progesterone ³2 ng/ml. The combination of the above-mentioned conditions (LH ³10 miu/ml and progesterone ³2 ng/ml) was indicated as premature luteinization.

3 GnRH antagonists and IUI All these hormonal assessments were available on the same day of the blood test, and according to the strict criteria of our protocol, the administration of GnRH antagonist Cetrorelix (Cetrotide, Serono), at the dose of 0.25 mg per day (multiple-dose protocol), was started at the same day (follicle ³14 mm in mean diameter) only if LH level was <10 miu/ml, progesterone level <2 ng/ml and E 2 >200 pg/ml. When the leading follicle reached a mean diameter of 18 mm, IU of hcg was administered and Cetrorelix was discontinued. HCG was administered to all patients and in all cycles even if a premature LH surge was detected. Patients included in group B (n = 52) were treated with rfsh alone, at the same dosage of group A (75 or 150 IU according to patient s age), and hcg ( IU) was administered when the leading follicle reached the mean diameter of 18 mm. Also in this latter group, LH, progesterone and E 2 serum levels were determined from the day when the dominant follicle was ³14 mm in mean diameter, but, in this group, the knowledge of LH and progesterone levels did not influence clinical decision regarding ongoing cycles, which were made on the basis of ultrasound only. Two inseminations were performed 20 and 34 h after hcg, respectively. All women were provided luteal phase support with natural micronized progesterone (Progeffik, Effik, Italy) 400 mg per day, vaginally into two divided doses, starting 2 days after the second insemination. IUI was cancelled if more than five or less than two follicles with a mean diameter ³16 mm were present, in order to optimize pregnancy chance and to reduce the risk of multiple pregnancy. Furthermore, insemination was not performed when E 2 serum level was ³1500 pg/ml. Pregnancy was diagnosed by quantitative b-hcg 2 weeks after IUI. At 6 7 weeks gestation, a transvaginal ultrasound was performed to document a clinical pregnancy by observing fetal cardiac activity. High-order multiple pregnancies were defined as three or more gestational sacs visualized at ultrasound scan. The primary end-point of the study was the clinical pregnancy rate per couple. We also evaluated the incidence of premature LH rise (³10 miu/ml) with or without progesterone rise, the duration of GnRH antagonist treatment, the total amount of gonadotrophin used, the number of recruited follicles (mean diameter of ³16 mm) and E 2 serum levels on the day of hcg administration. Statistical analysis was performed by using Chi-square test for frequency distribution analysis and Student s t-test for unpaired data. P < 0.05 was considered statistically significant. A post hoc power analysis for this trial was calculated using StatXact software. In order to determine the sample size we fixed a = 0.05, and e = 0.10; we considered the latter value acceptable for our study goals. In order to estimate the sample variance and bearing in mind our previous experience, we fixed f = 0.35 (pregnancy rate per patient); our goal was the refutation of H 0 :P A = P B. A cost-effectiveness analysis was also conducted comparing costs and pregnancy rate between the two groups. Because the management of the clinical condition was the same in group A and in group B, the drugs were the only resources which may contribute to cost difference among the two groups. In fact, in our daily work, we usually perform hormonal evaluation in all patients, even when the patient is not included in any trial. This is the reason why we did not include these costs in the cost-effectiveness analysis. We strongly believe that hormonal assessment should always be carried out in all cycles when gonadotrophins are used, both in IUI and IVF treatments (obviously, excluding cycles with GnRH agonist, when only 17-b-E 2 has to be evaluated). The cost of treatment was calculated using the public price (euro currency 2004). To verify the robustness of economic result, we conducted a oneway sensitivity analysis. The cost per patient to obtain a higher chance of improving pregnancy was calculated by the incremental costeffectiveness ratio (ICER), and the incremental effectiveness with Cetrorelix was moved along the limits of 95% confidence interval (CI) of the absolute increase of pregnancy rate. Results The subject disposition is shown in Figure 1. A total of 104 patients, between May 2002 and December 2004, were randomized to participate in the study, 52 in group A and 52 in group B. The basic characteristics of the patients were similar: no statistical differences were found in the age of the patients, and duration and cause of infertility in either treatment group. However, we observed a prevalence of secondary infertility in group B (P = 0.018; Table I). The total number of started cycles was 320. The number of performed cycles was 302 (144 in group A and 158 in group B). IUI was not executed in 18 cycles, in 8 cycles because of hyper-response (more than five follicles with a mean diameter ³ 16 mm), 4 in group A and 4 in group B, and in 10 cycles because of low response (less than two follicles with a mean diameter ³16 mm), 4 in group A and 6 in group B. Thus, 8 cycles were cancelled in group A and 10 cycles in group B. In 4 cycles in group B, on the day of hcg administration, no hormonal assessment (LH, progesterone and E 2 ) was performed. The mean total (± SD) amount of rfsh administered was ± in group A and ± in group B, with a significant difference between the two groups (P = 0.009). No significant differences were seen in the duration of rfsh stimulation and the number of mature follicles (³16 mm in mean diameter). The duration of GnRH antagonist treatment was 3.5 ± 0.8 days (Table II). A total of 64 cycles with a premature LH rise were observed on the day of hcg administration, 10 of 144 (7%) in group A and 54 out of 154 (35%) in group B, with a significant difference between the two groups (P < ). Seventy-four cases of progesterone rise were seen on the day of hcg administration, 18 of 144 (12.5%) in group A and 56 of 154 in group B (36.4%); also in this case, a high statistical difference was found (P < ). A premature luteinization was observed in 2 cycles of 144 in group A (1.4%) and in 16 cycles of 154 in group B (10.4%) with a significant difference between the two groups (P = 0.001). As expected, the mean value of LH, on the day of hcg administration, was significantly lower in patients receiving GnRH antagonist than in those who did not (3.3 ± 3.3 miu/ml in group A versus 9.9 ± 7.9 miu/ml in group B, P < ). In the same manner, progesterone mean values, on the day of hcg administration, were significantly different between the two groups (1.3 ± 1.1ng/ml versus 2.1 ± 1.9 ng/ml for group A and B, respectively, P < ). In both groups, no spontaneous ovulation was detected. On the contrary, E 2 serum levels on the day of hcg administration were found not to be statistically different between the two groups (828.9 ± pg/ml for group A; ± pg/ml for group B). 3

4 A.Allegra et al. Assessed for eligibility (n=113) Excluded (n=9) Not meeting inclusion criteria (n=6) Refused to participate (n=3) ENROLLMENT Randomized (n=104) ALLOCATION ANALYSIS Figure 1. Trial profile. A 52 patients allocated to receive rfsh + Cetrorelix A 52 patients completed 144 cycles Table II. Characteristics of the cycles B 52 patients allocated to receive rfsh alone B 52 patients completed 158 cycles Group A (n=144) Group B (n=158) P-value Total dose of rfsh (IU) a ± ± Duration of rfsh stimulation (days) a 9.2 ± ± 2.7 NS Follicles ³16 mm a 3.08 ± ± 1.2 NS Estradiol a,b (pg/ml) ± ± NS LH a,b (miu/ml) 3.3 ± ± 7.9 < Progesterone a,b (ng/ml) 1.3 ± ± 1.9 < Cycles with premature LH rise (%) b,c 10/144 (7) 54/154 (35) d < Cycles with premature progesterone rise (%) b,c 18/144 (12.5) 56/154 (36.4) d < Cycles with premature luteinization (%) b,c 2/144 (1.4) 16/154 (10.4) d Duration of GnRH antagonist injection a days 3.5 ± NS, not significant. a Values are expressed as mean ± SD. Student s t-test for unpaired data. b The hormonal values are referred to the day of hcg administration. c The frequency distribution analysis was performed by Chi-square test. d In four cycles, the hormonal assessments on the day of hcg administration (E 2, LH and progesterone) were not available. 4

5 GnRH antagonists and IUI Moreover, in both groups, to assess the role of LH and progesterone to predict the outcome of IUI cycles, we compared the mean serum value of LH and progesterone, on the day of hcg administration, in pregnant and non-pregnant cycles. In group A, we did not find any statistical difference for both LH and progesterone mean values in pregnant and in non-pregnant cycles (2.33 ± 1.7 miu/ml for LH and 1.06 ± 0.72 ng/ml for progesterone, in pregnant cycles; 3.53 ± 3.61 miu/ml for LH and 1.32 ± 1.16 ng/ml for progesterone, in non-pregnant cycles). In group B, LH and progesterone were significantly lower in pregnant cycles (5.21 ± 3.08 miu/ml for LH and 1.06 ± 0.89 ng/ml for progesterone) compared with non-pregnant cycles (10.4 ± 8.03 miu/ml for LH and 2.20 ± 1.98 ng/ml for progesterone; P = and P = 0.02, respectively). The total number of clinical pregnancies achieved was 44 (28 obtained in group A and 16 in group B). Of these 44 pregnancies, 5 were twins, 4 observed in group A and 1 in group B. No high-order multiple pregnancies were observed. No pregnancy was achieved in the four cycles in which no hormonal assessment was performed. Thus, the clinical pregnancy rate per patient was 53.8% in group A (28 of 52) and 30.8% in group B (16 of 52), with a significant difference between the two groups (P = 0.017) (Table III). Pregnancy rate per cycle and cumulative pregnancy rate are summarized as life table analysis (Table IV). The number of dropouts, for each cycle, is reported in the same table. With regard to power analysis, correlated to our results, H 0 : P A = P B can be refused (a = 0.05) and the corresponding power of test (1 b) is Table III. Outcome of the treatments Group A Group B P-value Pregnancies Clinical Pregnancy rate (%) a 28/52 (53.8) 16/52 (30.8) Abortions (%) a 7/28 (25.0) 4/16 (25.0) NS Multiple pregnancies (%) a 4/28 (14.3) 1/16 (6.3) NS Twins 4 1 NS High-order 0 0 OHSS 0 0 OHSS, ovarian hyperstimulation syndrome; NS, not significant. a The frequency distribution analysis was performed by using Chi-square test. Table IV. Life table analysis Cycle Number of Couples Pregnancies Drop out Pregnancy rate per cycle % Group A Group B Cumulative Pregnancy rate % Concerning the abortion rate, no difference was observed between the two groups (7 of 28, 25.0% in group A and 4 of 16, 25.0% in group B). No case of OHSS was observed. Concerning the cost-effectiveness analysis, the total costs of stimulation were and for group A and B, respectively. The management of the 52 patients per group generated a small difference between the cost per cycle ( 825 and 742 for group A and B, respectively). Consequently, the incremental cost per patient was 30 in group A, where the antagonist was administered. Regarding the outcome, group A was associated with a statistically significant improvement of pregnancy rate, with a difference of 23% (95%CI: %). The cost per patient to obtain a higher chance of improving pregnancy was calculated by the ICER. Thus, the current costeffectiveness analysis showed that the patient should pay 130 for each pregnancy gained using GnRH antagonist instead of rfsh alone. The one-way sensitivity analysis showed that the ICER of Cetrorelix-based protocol could vary between a cost per pregnancy achieved of 76 and 735, in the best and worst case, respectively. The protocol based on rfsh and Cetrorelix, thus, was more costly but also more effective than the protocol without the antagonist. All these results are reported in Table V. Discussion The current investigation demonstrates that treatment with Cetrorelix in COS/IUI cycles increases the pregnancy rate per patient (53.8%) compared with treatment without GnRH antagonist (30.8%; P = 0.017). Furthermore, we observed a significant reduction in the number of premature LH rises, progesterone rises and premature luteinization in patients treated with GnRH antagonist. These differences are reflected in the mean serum LH and progesterone levels which were significantly higher in patients who used rfsh alone. Then, we compared the serum mean values of LH and progesterone, in the two groups, between pregnant and nonpregnant cycles. In group A (antagonist group), as expected, we did not find any statistical difference (between pregnant and non-pregnant cycles) for both LH and progesterone mean values; in group B (not receiving GnRH antagonist), LH and progesterone were significantly lower in pregnant cycles than in non-pregnant (P = and P = 0.02, for LH and progesterone, respectively). According to these data, we found that no pregnancy was obtained when the LH serum level, on the day of hcg administration, was >10 miu/ml, in both groups. Our data demonstrated with strong evidence that a LH serum level >10 miu/ml has a dramatic impact on achieving pregnancy; it can be thus considered a negative predictive factor. In fact, as well known, the multifollicular recruitment during ovulation induction may determine an increase in E 2 serum levels that induce an LH surge when the follicles are still growing. Few studies have documented the efficacy of GnRH antagonist in preventing LH rise in COS/IUI cycles for the treatment of subfertile couples with contradictory results. In 2004, a randomized study comparing different gonadotrophin dosages 5

6 A.Allegra et al. Table V. Cost-effectiveness analysis Patients Cycles Total cost of stimulation ( ) Cost per patient ( ) Cost per cycle ( ) Pregnancy rate per patient (95% CI) (%) Group A a 825 ± ( ) g Group B b 742 ± ( ) d Difference ( ) ICER=130.6 a, cost per patient in group A; b, cost per patient in group B; g, pregnancy rate per patient in group A; d, pregnancy rate per patient in group B; CI, Confidence interval; ICER, Incremental cost-effectiveness ratio [(a b/g d) 100]. showed that a protocol of rfsh 50 IU daily and GnRH antagonist represented an effective and safe treatment in COS/IUI cycles (Ragni et al., 2004). Furthermore, another prospective randomized study (Zikopoulos et al., 2005) compared the efficacy of GnRH antagonist versus GnRH agonist (long protocol) administration in patients undergoing COS and IUI. This study showed that there were no differences in the fecundity per cycle using either protocol even if the use of GnRH antagonists facilitated short and simple treatment. More recently, a large double-blind, placebo-controlled, multicentre trial (Lambalk et al., 2006) designed to evaluate the role of the GnRH antagonist ganirelix in the treatment of subfertile couples did not document any significant differences for clinical pregnancy rates (12.6 and 12% for the ganirelix and placebo groups, respectively); conversely, the incidence of LH rise was lower in the ganirelix group compared with the placebo group (2.9 versus 23.4%; P = 0.003). The above-mentioned investigation confirmed that treatment with a GnRH antagonist prevents premature LH rises and luteinization in patients using superovulatory drugs for IUI, but the use of GnRH antagonist did not improve pregnancy outcome. On the contrary, Gomez-Palomares et al. (2005) revealed more mature follicles and a significant improvement in pregnancy outcome (pregnancy rate 38 versus 14%), when the GnRH antagonist was added to the COS protocol. This multicentre, prospective, randomized research showed for the first time that the enhancement in overall pregnancy rate observed in the study group obtained statistical significance. Nevertheless, the authors of this study suggest that the increased pregnancy rate is not related to the use of the antagonist itself but to the fact that this drug gives the opportunity to continue the administration of rfsh until a greater number of ovulatory follicles is recruited. The rationale for this hypothesis is based on the consideration that the number of mature follicles was significantly higher in the GnRH antagonist-treated group. In contrast, in our work, the number of ovulatory follicles is similar between the two groups (3.08 ± 1.2 for group A and 3.2 ± 1.2 for group B). We therefore believe that the higher pregnancy rate observed in group A is related to the lower premature luteinization. Furthermore, as previously stated, no pregnancy was obtained when the LH serum level was >10 miu/ml. This last observation confirms that premature luteinization is an important limiting factor in achieving pregnancy. 6 It is well known that premature luteinization is associated with a less favourable outcome because of poor oocyte quality and decreased fertilization and implantation rate (Loumaye, 1990; Manzi et al., 1995). Regarding the role of premature luteinization on the outcome of COS/IUI cycles, the above-mentioned important multicentre trial of Lambalk et al. (2006) shows that the incidence of LH rise is lower in the antagonist treated group compared with the placebo group, but pregnancy rates are similar. Also, in our data, the number of LH rises is lower in the antagonist group, but, in contrast to the above-mentioned study, we obtained (as expected) a significant improvement in pregnancy rate for patients receiving GnRH antagonist when compared with patients treated with rfsh alone. A possible explanation for this observed difference may be due to the immediate availability in our protocol of the hormonal assessments (LH, progesterone and E 2 ) before the beginning of the administration of the antagonist. As a consequence of this availability, when the leading follicle reached the mean diameter of 14 mm, in contrast with Lambalk (who started antagonist administration without considering other parameters), we delayed the administration of the antagonist until the moment in which the E 2 serum level was at least 200 pg/ml. In this way, the endogenous LH can contribute to better follicular development for a longer period of time. In fact, it is well known that GnRH antagonist immediately determines an inhibition of pituitary release of LH and FSH. However, while exogenous administration of rfsh substitutes endogenous secretion, LH secretion, which has been blocked, is not replaced. In keeping with this hypothesis, in our study, the mean number of hours of stimulation with rfsh before antagonist administration was 16.8 h more than in Lambalk s study. Our hypothesis has been confirmed by the above-mentioned work of Gomez-Palomares et al., In fact, in this study, the authors obtained better clinical pregnancy rate per cycle than our clinical pregnancy rate in first cycle (38 versus 21.2%). This result probably depended on the fact that the antagonist administration was delayed even further until the leading follicle was ³16 mm in diameter and the E 2 serum level >300 pg/ml. Furthermore, in a multicentre trial (unpublished data) in which we participated (with couples who were different from those of the present study), we did not find a difference in pregnancy rate in patients treated with GnRH antagonist and rfsh compared with patients treated with rfsh alone.

7 GnRH antagonists and IUI The observed difference between this multicentre trial data and the results of the present article is probably related to starting the GnRH antagonist (in the multicentre trial) on the day in which the leading follicle was mm in diameter, without considering the hormonal parameters above discussed. We believe that to improve the outcome of COS/IUI cycles, the antagonist administration should be started when the leading follicle is at least 16 mm in mean diameter, the E 2 serum level is >200 pg/ml, LH <10 mui/ml and progesterone <2 ng/ml. Moreover, in clinical practice, when it is not possible to know the hormonal measurements on the same day as the blood test and antagonist is to be used, it is possible to administer synchronously a small amount (e.g miu/ml) of LH. Concerning IUI success rate, there are only two papers which report a very high pregnancy rate (Ragni et al., 2004; Gomez Palomares et al., 2005). The difference between our pregnancy rate in first cycle (21.2%) and Ragni s data (34.4%) is probably due to different basic characteristics of the patients enrolled in the two studies. In fact, in our study, there is an incidence of male factor infertility (42.3%) higher than in Ragni s experience (28.1%). According to this observation, a recent and larger retrospective study performed by the same author did not confirm his previous data regarding pregnancy rate (Ragni et al., 2006). The pregnancy rate per initiated cycle was only 9.2%, but in this study mild male factor was present in 59.3% of the indications. Gomez-Palomares has also an improved pregnancy rate per cycle (38%) compared with our pregnancy rate in first cycle (21.2%) with a rfsh dosage similar to our protocol. As well as for the above-mentioned reasons (the antagonist was started later than in our study), also in this case, the baseline characteristics of the patients were different from ours. In fact, no couple suffering from mild male factor infertility was enrolled in this study. As is well known, the success rate of IUI is better in unexplained infertility than in male factor infertility (Guzick et al., 1999; Dickey et al., 2005). Moreover, in contrast with previous studies, in which the premature luteinization seems to be associated with a higher abortion rate (Manzi et al., 1995), we did not find any increased incidence of spontaneous abortion and biochemical pregnancies in group B, where the premature luteinization rate was 10.4%. Furthermore, we did not observe an increase of miscarriages in patients treated with the GnRH antagonist. With regard to stimulation protocol, it has been documented that low-dose rfsh treatment with the GnRH antagonist may ensure an important reduction in twins and the absence of highorder multiple pregnancies and OHSS (Ragni et al., 2006). In our study, we documented the possibility of achieving a remarkably high clinical pregnancy rate for patients with an acceptable number of twin gestations (14.3% in group A and 6.3% in group B with no high-order multiple pregnancies in either group) and no case of OHSS, without using a low-dose protocol. We found that the patients in group A were treated with a lower dose of rfsh compared with those in group B. The explanation of this observation is difficult; also Lambalk found a lower total dose of rfsh in Ganirelix-treated patients (550 IU) compared with the placebo group (600 IU), even if this difference was not statistically significant (Lambalk et al., 2006). With regard to the cost-effectiveness analysis, the use of GnRH antagonist with rfsh proved to be a more costly but more effective option than the use of rfsh alone. Because of the negligible value of incremental cost, and the notable increase of the effectiveness after the addition of GnRH antagonist to standard treatment, the ICER was only 130. This value indicates that Cetrorelix-based protocol was cost-effective. In conclusion, our data indicate that the use of the GnRH antagonist in COS/IUI cycles may improve pregnancy rate, preventing the occurrence of premature LH rise and luteinization. Moreover, we demonstrated the predictive significance of premature LH rise on the day of hcg administration for the outcome of cycles. In fact, no pregnancy was achieved, in either group, when the LH value was ³10 miu/ml. However, we believe that our results must be supported in larger multicentre studies. References Allegra A, Volpes A, Coffaro F, Guida S and Francofonte R (1990) Superovulation with buserelin and gonadotropins dramatically improves the success rate of intrauterine insemination with husband s washed semen. Acta Eur Fertil 21, Chaffkin LM, Nulsen JC, Luciano AA and Metzger DA (1991) A comparative analysis of the cycle fecundity rates associated with combined human menopausal gonadotropin (HMG) and intrauterine insemination (IUI) versus either HMG or IUI alone. Fertil Steril 55, Cohlen BJ, te Velde ER, van Kooij RJ, Looman CW and Habbema JD (1998) Controlled ovarian hyperstimulation and intrauterine insemination for treating male subfertility: a controlled study. Hum Reprod 13, Cohlen BJ, Vandekerckhove P, te Velde ER and Habbema JD (2000) Timed intercourse versus intra-uterine insemination with or without ovarian hyperstimulation for subfertility in men. Cochrane Database Syst Rev CD Collins JA (1994) Reproductive technology-the price of progress. N Engl J Med 331, Cunha-Filho JS, Kadoch J, Righini C, Fanchin R, Frydman R and Olivennes F (2003) Premature LH and progesterone rise in intrauterine insemination cycles: analysis of related factors. Reprod Biomed Online 7, Dickey RP (2003) A year of inaction on high-order multiple pregnancies due to ovulation induction. Fertil Steril 79, Dickey RP, Taylor SN, Lu PY, Sartor BM, Rye PH and Pyrzak R (2002) Effect of diagnosis, age, sperm quality, and number of preovulatory follicles on the outcome of multiple cycles of clomiphene citrate-intrauterine insemination. Fertil Steril 78, Dickey RP, Taylor SN, Lu PY, Sartor BM, Rye PH and Pyrzak R (2005) Risk factors for high-order multiple pregnancy and multiple birth after controlled ovarian hyperstimulation: results of 4,062 intrauterine insemination cycles. Fertil Steril 83, Dodson WC and Haney AF (1991) Controlled ovarian hyperstimulation and intrauterine insemination for treatment of infertility. Fertil Steril 55, Dodson WC, Walmer DK, Hughes CL Jr, Yancy SE and Haney AF (1991) Adjunctive leuprolide therapy does not improve cycle fecundity in controlled ovarian hyperstimulation and intrauterine insemination of subfertile women. Obstet Gynecol 78, Duran HE, Morshedi M, Kruger T and Oehninger S (2002) Intrauterine insemination: a systematic review on determinants of success. Hum Reprod Update 8, Fleming R and Coutts JR (1986) Induction of multiple follicular growth in normally menstruating women with endogenous gonadotropin suppression. Fertil Steril 45, Gagliardi CL, Emmi AM, Weiss G and Schmidt CL (1991) Gonadotropinreleasing hormone agonist improves the efficiency of controlled ovarian hyperstimulation/intrauterine insemination. Fertil Steril 55,

8 A.Allegra et al. Gomez-Palomares JL, Julia B, Acevedo-Martin B, Martinez-Burgos M, Hernandez ER and Ricciarelli E (2005) Timing ovulation for intrauterine insemination with a GnRH antagonist. Hum Reprod 20, Goverde AJ, McDonnell J, Vermeiden JP, Schats R, Rutten FF and Schoemaker J (2000) Intrauterine insemination or in-vitro fertilisation in idiopathic subfertility and male subfertility: a randomised trial and costeffectiveness analysis. Lancet 355, Guzick DS, Carson SA, Coutifaris C, Overstreet JW, Factor-Litvak P, Steinkampf MP, Hill JA, Mastroianni L, Buster JE, Nakajima ST et al. (1999) Efficacy of superovulation and intrauterine insemination in the treatment of infertility. National Cooperative Reproductive Medicine Network. N Engl J Med 340, Hannoun A, Abu-Musa A, Kaspar H and Khalil A (1998) Intrauterine insemination IUI: the effect of ovarian stimulation and infertility diagnosis on pregnancy outcome. Clin Exp Obstet Gynecol 25, Houmard BS, Juang MP, Soules MR and Fujimoto VY (2002) Factors influencing pregnancy rates with a combined clomiphene citrate/gonadotropin protocol for non-assisted reproductive technology fertility treatment. Fertil Steril 77, Hughes EG (2003) Stimulated intra-uterine insemination is not a natural choice for the treatment of unexplained subfertility. Effective treatment or not a natural choice? Hum Reprod 18, Hughes EG, Collins JA and Gunby J (1998) A randomised controlled trial of three low-dose gonadotrophin protocols for unexplained infertility. Hum Reprod 13, Kaplan PF, Katz SL, Thompson AK and Freund RD (2002) Cycle fecundity in controlled ovarian hyperstimulation and intrauterine insemination. Influence of the number of mature follicles at hcg administration. J Reprod Med 47, Lambalk CB, Leader A, Olivennes F, Fluker MR, Nyboe Andersen A, Ingerslev J, Khalaf Y, Avril C, Belaisch-Allart J, Roulier R et al. (2006) Treatment with the GnRH antagonist ganirelix prevents premature LH rises and luteinization in stimulated intrauterine insemination: results of a double-blind, placebo-controlled, multicentre trial. Hum Reprod 21, Loumaye E (1990) The control of endogenous secretion of LH by gonadotrophin-releasing hormone agonists during ovarian hyperstimulation for in-vitro fertilization and embryo transfer. Hum Reprod 5, Manzi D, Dumez S, Scott LB and Nulsen JC (1995) Selective use of leuprolide acetate in women undergoing superovulation with intrauterine insemination results in significant improvement in pregnancy outcome. Fertil Steril 63, Matorras R, Diaz T, Corcostegui B, Ramon O, Pijoan JI and Rodriguez- Escudero FJ (2002) Ovarian stimulation in intrauterine insemination with donor sperm: a randomized study comparing clomiphene citrate in fixed protocol versus highly purified urinary FSH. Hum Reprod 17, Messinis IE, Loutradis D, Domali E, Kotsovassilis CP, Papastergiopoulou L, Kallitsaris A, Drakakis P, Dafopoulos K and Milingos S (2005) Alternate day and daily administration of GnRH antagonist may prevent premature luteinization to a similar extent during FSH treatment. Hum Reprod 20, Nulsen JC, Walsh S, Dumez S and Metzger DA (1993) A randomised and longitudinal study of human menopausal gonadotropin with intrauterine insemination in the treatment of infertility. Obstet Gynecol 82, Nuojua-Huttunen S, Tomas C, Bloigu R, Tuomivaara L and Martikainen H (1999) Intrauterine insemination treatment in subfertility: an analysis of factors affecting outcome. Hum Reprod 14, Pandian Z, Bhattacharya S, Nikolaou D, Vale L and Templeton A (2003) The effectiveness of IVF in unexplained infertility: a systematic Cochrane review Hum Reprod 18, Peterson CM, Hatasaka HH, Jones KP, Poulson AM Jr, Carrell DT and Urry RL (1994) Ovulation induction with gonadotroins and intrauterine insemination compared with in vitro fertilization and no therapy: a prospective, non randomized, cohort study and meta-analysis. Fertil Steril 62, Ragni G, Vegetti W, Baroni E, Colombo M, Arnoldi M, Lombroso G and Crosignani PG (2001) Comparison of luteal phase profile in gonadotrophin stimulated cycles with or without a gonadotrophin-releasing hormone antagonist. Hum Reprod 16, Ragni G, Alagna F, Brigante C, Riccaboni A, Colombo M, Somigliana E and Crosignani PG (2004) GnRH antagonists and mild ovarian stimulation for intrauterine insemination: a randomized study comparing different gonadotrophin dosages. Hum Reprod 19, Ragni G, Caliari I, Nicolosi AE, Arnoldi M, Somigliana E and Crosignani PG (2006) Preventing high-order multiple pregnancies during controlled ovarian hyperstimulation and intrauterine insemination: 3 years experience using low-dose recombinant follicle-stimulating hormone and gonadotropinreleasing hormone antagonists. Fertil Steril 85, Sengoku K, Tamate K, Takaoka Y, Morishita N and Ishikawa M (1994) A randomised prospective study of gonadotrophin with or without gonadotrophinreleasing hormone agonist for treatment of unexplained infertility. Hum Reprod 9, Serhal PF, Katz M, Little V and Woronowski H (1988) Unexplained infertility the value of Pergonal superovulation combined with intrauterine insemination. Fertil Steril 49, Steures P, van der Steeg JW, Verhoeve HR, van Dop PA, Hompes PG, Bossuyt PM, van der Veen F, Habbema JD, Eijkemans MJ and Mol BW (2004) Does ovarian hyperstimulation in intrauterine insemination for cervical factor subfertility improve pregnancy rates? Hum Reprod 19, The European and Middle East Orgalutran Study Group (2001) Comparable clinical outcome using the GnRH antagonist ganirelix or a long protocol of the GnRH agonist triptorelin for the prevention of premature LH surges in women undergoing ovarian stimulation. Hum Reprod 16, WHO (World Health Organization) (1999) WHO Laboratory Manual for the Examination of Human Semen and Sperm-Cervical Mucus Interation, 4th edn. Cambridge University Press, Cambridge, UK. Zeyneloglu HB, Arici A, Olive DL and Duleba AJ (1998) Comparison of intrauterine insemination with timed intercourse in superovulated cycles with gonadotropins: a meta-analysis. Fertil Steril 69, Zikopoulos K, West CP, Thong PW, Kacser EM, Morrison J and Wu FC (1993) Homologous intra-uterine insemination has no advantage over timed natural intercourse when used in combination with ovulation induction for the treatment of unexplained infertility. Hum Reprod 8, Zikopoulos K, Kaponis A, Adonakis G, Sotiriadis A, Kalantaridou S, Georgiou I and Paraskevaidis E (2005) A prospective randomized study comparing gonadotropin-releasing hormone agonists or gonadotropin-releasing hormone antagonists in couples with unexplained infertility and/or mild oligozoospermia. Fertil Steril 83, Submitted on March 19, 2006; resubmitted on June 1, 2006, July 8, 2006; accepted on July 25,