Abstract. Introduction. RBMOnline - Vol 9. No Reproductive BioMedicine Online; on web 19 August 2004

Transcription

1 RBMOnline - Vol 9. No Reproductive BioMedicine Online; on web 19 August 2004 Article Ovulation induction in women with polycystic ovary syndrome: randomized trial of clomiphene citrate versus low-dose recombinant FSH as first line therapy Dr Eugenio López was born in Orihuela (Alicante), Spain in In 1978 he obtained his medical degree at the Faculty of Medicine, University of Valencia in Spain. The PhD degree was granted to him at the Faculty of Medicine, University of Murcia in After completing his obstetrics/gynaecology residency in the Hospital Virgen de la Arrixaca of Murcia, he joined the Reproductive Medicine Unit where he is currently working. Dr López is Associate Professor of Obstetrics and Gynecology and has in excess of 60 publications to his name in national and international books and journals. Current clinical research interests include studies on ovulation induction, endometriosis and assisted reproduction. Dr Eugenio López Eugenio López 1, Joanne Gunby 2, Salim Daya 2,3, Juan J Parrilla 1, Lorenzo Abad 1, Juan Balasch 4,5 1 Department of Obstetrics and Gynecology, Hospital Universitario Virgen de la Arrixaca, Murcia, Spain; 2 Department of Obstetrics and Gynecology, and 3 Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada; 4 Institut Clinic of Gynecology, Obstetrics and Neonatology, Faculty of Medicine, University of Barcelona, Hospital Clinic-Institut d Investigacions Biomediques August Pi I Sunyer (IDIBAPS), Barcelona, Spain 5 Correspondence: Institut Clínic of Obstetrics and Gynecology, Hospital Clínic, c/casanova 143, Barcelona, Spain. Fax: ; jbalasch@ub.edu Abstract This single centre randomized controlled trial was undertaken to compare the efficacy and safety of clomiphene citrate and lowdose recombinant FSH as first line pharmacological therapy for anovulatory infertility associated with polycystic ovary syndrome (PCOS). Seventy-six infertile patients with PCOS were randomized to receive clomiphene citrate ( mg/day for 5 days) (clomiphene citrate group, n = 38) or recombinant human FSH (FSH group, n = 38) in a chronic, low-dose, step-up protocol (daily starting dose 75 IU) for up to three consecutive cycles. Ovarian response was monitored by transvaginal ultrasonography and human chorionic gonadotrophin (HCG) was given to trigger ovulation in all cycles with appropriate follicular development. The primary outcome measure was cumulative pregnancy after undergoing up to three treatment cycles. Secondary outcomes were cycle cancellation rate, ovulation rate per cycle, cumulative ovulation rate, pregnancy rate per cycle, incidence of OHSS, cumulative live birth rate, and multiple birth rate. One hundred and four clomiphene citrate cycles and 91 FSH cycles were evaluable. The relative risk and its 95% confidence interval were 1.17 ( ) for HCG cycles with ovulation, 1.78 ( ) for the pregnancy rate per woman, and 1.83 ( ) for live births per woman in favour of FSH. The cumulative pregnancy rate after three treatment cycles was 43% with FSH and 24% with clomiphene citrate (P = 0.06). By logistic regression analysis, the factors predicting ovulation included female age, serum androstenedione and use of FSH. Predictors of pregnancy were duration of infertility and use of FSH. This randomized controlled trial suggests that low-dose recombinant FSH may be an effective alternative to clomiphene citrate in first-line treatment for anovulatory PCOS patients. Thus, further studies, possibly multi-centre, in order to avoid problems with patient recruitment, are warranted to confirm these results. Keywords: clomiphene citrate, FSH, ovulation induction, PCOS 382 Introduction Approximately 20 40% of infertile women have ovulatory disorders and the most common cause of ovulatory dysfunction is polycystic ovary syndrome (PCOS), which represents approximately 70% of cases (Kousta et al., 1997; American College of Obstetricians and Gynecologists, 2002). For infertile women with PCOS, induction of ovulation with

2 clomiphene citrate is commonly used as the initial treatment (Yarali and Zeyneloglu, 2004). A marked discrepancy exists, however, between the high rates of ovulation (60 85%) and the much lower rates of conception (20 40%) per cycle with clomiphene citrate usage (Dickey and Holtkamp, 1996; Kousta et al., 1997). Of clomiphene citrate-induced pregnancies, 7 13% are multiple gestations (Dickey and Holtkamp, 1996; Kousta et al., 1997; American College of Obstetricians and Gynecologists, 2002). The most common side effects experienced by women taking clomiphene citrate include vasomotor symptoms (20%), adnexal tenderness (5%), nausea (3%), headache (1%), and, rarely, blurring of vision or scotoma (American College of Obstetricians and Gynecologists, 2002). The incidence of OHSS in patients treated with clomiphene citrate is rare, but this drug can cause multiple follicular development with the risk of ovarian hyperstimulation syndrome (OHSS) and multiple pregnancy mainly in patients with PCOS (Gelety et al., 1992; Kousta et al., 1997). Additionally, in about 25% of women with PCOS, ovulation will not result (Kousta et al., 1997). Nevertheless, clomiphene citrate is a popular choice among physicians for the treatment of ovulatory dysfunction associated with PCOS, because it is easily administered, relatively safe, and inexpensive. It is important to acknowledge that diet and exercise may be first line treatments (Crosignani et al., 2003; Norman et al., 2004; Pasquali and Gambineri, 2004), and clomiphene citrate is a first line pharmacological treatment. It is also recognized that the choice of first line therapy is no longer as clear as it used to be, with insulin-sensitizing agents becoming more popular (Lord et al., 2003; Cheang and Nestler, 2004). Therefore, it is reasonable to evaluate FSH as yet another alternative. In fact, studies comparing clomiphene citrate versus urinary FSH for ovarian stimulation in intrauterine insemination cycles showed higher pregnancy rates in the FSH group (Balasch et al., 1994; Matorras et al., 2002). It is well established that the low-dose protocols for FSH currently used are safe. Over the past decade, several studies have examined the utility of a chronic, low-dose regimen of FSH in reducing the complication rate (Homburg and Howles, 1999; van Wely et al., 2003). Despite the fact that the use of FSH for ovulation induction in PCOS has traditionally been restricted to women with proven clomiphene citrate resistance, a review of results from the use of a chronic low-dose protocol indicates that mono-ovulatory cycles are observed consistently in approximately 70% of cases, pregnancy rates of 20% per cycle and 40% per woman are achieved, and the incidence of OHSS is very low (0.14%) (Homburg and Howles, 1999). For women with PCOS, ovulation induction with FSH is associated with multiple pregnancy rates of 6 20% (Homburg and Howles, 1999; American College of Obstetricians and Gynecologists, 2002). Using a low-dose FSH protocol, monitoring does not need to be as intensive as for conventional dose gonadotrophin therapy and oestradiol determinations may not be necessary. Furthermore, the high purity of new FSH preparations allows them to be given s.c. instead of i.m., thereby reducing discomfort, allowing self-injection, saving nursing time, and minimizing disruption to the woman s daily routine. The use of the more potent recombinant FSH for ovulation induction is associated with both shorter duration of treatment and lower total dose of FSH, compared with purified (Coelingh-Bennink et al., 1998) or highly purified urinary FSH (Balasch et al., 1998). In light of these observations, the requirement of clomiphene citrate resistance as a prerequisite for gonadotrophin treatment may be questionable. So far as is known, a direct head-to-head comparison of FSH and clomiphene citrate as first-line pharmacological treatment in women with PCOS has not been undertaken. The objective of this randomized controlled trial was to compare the efficacy and safety of clomiphene citrate and lowdose recombinant FSH as the first-line pharmacological treatment for anovulatory infertility associated with PCOS. Materials and methods Subjects Women aged <40 years with anovulatory infertility due to PCOS of at least 1 year duration, and attending the infertility clinic at the Hospital Virgen de la Arrixaca in Murcia (Spain), were eligible for the study. A history of amenorrhoea (no menstrual periods for three or more months) or oligomenorrhoea (fewer than nine menstrual periods in a year), monophasic basal body temperature charting, proliferative endometrium on biopsy, or serum progesterone concentrations <1 ng/ml were accepted as demonstration of anovulation. In addition, subjects were required to have the ultrasonographic appearance of polycystic ovaries (Adams et al., 1986), a positive response to the progestin challenge test (normal withdrawal bleeding after treatment with oral medroxyprogesterone acetate, 10 mg daily for 5 days), normal serum prolactin, dehydroepiandrosterone sulphate, and fasting glucose concentrations, a normal hysterosalpingogram (and laparoscopy when appropriate), and no history of pelvic surgery or pelvic inflammatory disease. When the study was started the criteria for PCOS that were used were as described and these are also in keeping with the ESHRE/ASRM guidelines for PCOS (Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group, 2003). Obesity was not an inclusion or exclusion criterion. The male partner had to have normal semen parameters according to World Health Organization criteria (Rowe et al., 2000). Women with a previous pregnancy or previous treatment with ovarian stimulation drugs were excluded. The study was approved by the Investigation and Ethics Committee of the Hospital Clinic of Barcelona. All subjects provided written informed consent to participate. Study protocol In this randomized study, the subjects received either clomiphene citrate or low-dose recombinant FSH for up to three cycles. Women were allocated to the treatment groups according to a computer-generated randomization table. Concealment of treatment allocation was achieved with the use of sealed opaque envelopes each containing a unique study number and prepared independently by a secretary. Women not conceiving after three cycles of treatment crossed over to the alternative treatment for a further three cycles, with an interval of at least 45 days between treatments. The primary outcome measure was cumulative pregnancy after undergoing up to three treatment cycles. Secondary outcomes were cycle 383

3 384 cancellation rate, ovulation rate per cycle, cumulative ovulation rate, pregnancy rate per cycle, incidence of OHSS, cumulative live birth rate, and multiple birth rate. The primary outcome of the study was pregnancy rate before the crossover, but the results of the second arm are reported separately for information purposes. Clomiphene citrate (Omifin; Laboratorios Effik, Madrid, Spain) was given at a daily dose of 50 mg for 5 days, starting on day 5 following spontaneous or induced uterine bleeding. If ovulation was documented but no pregnancy ensued, the same dose was used in the next cycle. However, if no ovulatory response occurred, the daily dose was increased by 50 mg for the subsequent cycle, up to a maximum daily dose of 150 mg in the third treatment cycle. Human chorionic gonadotrophin (HCG, Profasi; Serono S.A., Madrid, Spain), at a dose of 5000 IU i.m., was administered when the lead follicle was >17 mm in diameter on transvaginal ultrasonography. The cycle was cancelled if no growing follicle was seen by day Treatment with recombinant FSH (Gonal-F; Serono S.A.) was commenced on day 3 following spontaneous or induced menses. The chronic low-dose, step-up regimen consisted of a starting dose of 75 IU daily s.c., with dose increments of 37.5 IU daily every 7 days if there was no evidence of ovarian response by ultrasonography (i.e. no follicle >10 mm in diameter). This stepwise increase was continued until ovarian activity was seen, at which time the dose was maintained. In successive treatment cycles, the starting dose of FSH could be modified, based on the ovarian response in the previous cycle. Criteria for HCG injection (5000 IU i.m.) were the same as for clomiphene citrate treatment cycles. A treatment cycle was cancelled if there was no follicle growth after 21 days of gonadotrophin administration. The injection of HCG was withheld if four or more follicles >14 mm in diameter were present, because of the increased risks of OHSS and multiple pregnancy. Ovarian response was monitored by transvaginal ultrasonography. In the clomiphene citrate cycles, monitoring was started 4 days after the last dose of drug and scans were repeated every other day until ovarian activity was detected. In the FSH cycles, monitoring was performed according to the well-established principles of this chronic low-dose therapy (Homburg and Howles, 1999), i.e. weekly until ovarian response was observed, then every other day or daily. Ultrasonic scans were performed with a 5 mhz vaginal transducer attached to an Aloka sector scanner (model SSD- 620; Aloka, Tokyo, Japan). Couples were advised to have sexual intercourse the evening of the HCG injection and the following day. Intrauterine insemination (IUI) was not performed. Serum progesterone was measured 6 8 days after the HCG injection (or 8 10 days after the last clomiphene citrate or FSH dose in non-responsive cycles) to determine whether ovulation had occurred. A mid-luteal phase serum progesterone concentration >3 ng/ml was considered satisfactory evidence of ovulation, as recommended in the practice guidelines (American College of Obstetricians and Gynecologists, 2002). Pregnancy was diagnosed by increasing serum concentrations of β-hcg after missed menses, and the subsequent demonstration of an intrauterine gestational sac by transvaginal ultrasonography. Women who achieved pregnancy but miscarried did not re-enter the study. Hormone assays Serum hormone concentrations were measured using commercially available kits. FSH, LH, prolactin, testosterone, oestradiol, and progesterone were measured by an electrochemiluminiscence immunoassay (Roche Diagnostics Corporation, Indianapolis, IN, USA). The sensitivity of the assays and the interassay and intra-assay coefficients of variation were as follows: 0.1 IU/l, 3.7 and 1.8% respectively for FSH; 0.1 IU/l, 2.0 and 1.1% respectively for LH; 1.25 ng/ml, 3.6 and 2.5% respectively for prolactin; 0.02 ng/ml, 1.6%, and 0.9% respectively for testosterone; 5 pg/ml, 3 and 2.6% respectively for oestradiol; 0.03 ng/ml, 4.1 and 2.7% respectively for progesterone. Androstenedione and dehydroepiandrosterone sulphate were measured by radioimmunoassay (Immunotech, Marseille, France). The sensitivity of the assay was 0.1 ng/ml for androstenedione and 0.06 μg/ml for dehydroepinandrosterone sulphate, and the interassay and intra-assay coefficients of variation were respectively 7.6 and 1.1% for androstenendione and 3.5 and 3.4% for dehydroepinandrosterone sulphate. Sample size calculation and statistical analysis The sample size calculation was based on experience with ovulation induction in women with WHO group II anovulation, in whom cumulative pregnancy rates were obtained after three treatment cycles of 25% with clomiphene citrate and 40% with FSH, an absolute difference of 15% (Balasch, 1986; Balasch et al., 1996). The sample size required to provide power of 80% to detect this magnitude of treatment effect was calculated to be 152 women per treatment group, using a two-tailed analysis with a detection limit of 5% of avoiding a type I error in hypothesis testing. Recruitment was terminated after 21 months when it became evident that it would not be possible to recruit this number of subjects from a single centre in a reasonable time period. The statistical significance of differences was determined in dichotomous variables using the chi-squared test and in continuous variables using Student s t-test (two-sided α = 0.05). Cumulative pregnancy rates were calculated and compared using Kaplan Meier survival analysis. Logistic regression analyses were performed to identify the models that best predicted each of ovulation and pregnancy as end-points. The analysis was by intention-to-treat. Analyses were carried out using the computer programmes SPSS version 10 and ARCUS. Results The study was carried out over the period April 2000 to December Seventy-six women were enrolled in the study, 38 per treatment group. Except for one woman in the FSH group who withdrew after one unsuccessful treatment cycle, all women completed the study and received the

4 allocated treatment (Figure 1). The median age of the women was 29 years, with a range from 22 to 39 years. All women presented with oligomenorrhoea or amenorrhoea and primary infertility of a median duration of 3 years (range 1 8 years). The median body mass index was 22 kg/m 2 (range 18 26); thus no woman was obese. The baseline characteristics of the participants in each group are given in Table 1; there was no significant difference between groups for any characteristic. A summary of the results of the study is presented in Table 2. First three cycles HCG was administered in 65/104 cycles (63%) in the clomiphene citrate group and 71/91 cycles (78%) in the FSH group. The cycle was cancelled for lack of follicle growth in 38% of clomiphene citrate cycles and 14% of FSH cycles. In seven FSH cycles (8%) but no clomiphene citrate cycles, HCG was withheld because of the presence of more than three midto large-size follicles. In the clomiphene citrate group, the Figure 1. Participant flow through trial. (CC = clomiphene citrate; FSH = recombinant FSH.) Table 1. Baseline characteristics in the two treatment groups. Data are expressed as median (range) or mean ± SD. No significant differences were observed between groups for any variable. Variable Clomiphene Recombinant citrate (n = 38) FSH (n = 38) Age (years) 29 (23 38) 30 (22 39) Duration of infertility (years) 3 (1 8) 3 (1 8) Body mass index (kg/m 2 ) 22.3 ± ± 1.9 FSH (IU/l) 4.7 ± ± 1.8 LH (IU/l) 9.2 ± ± 3.5 LH/FSH ratio 2.02 ± ± 1.06 Oestradiol (pg/ml) 65 ± ± 22 Testosterone (ng/ml) 1.16 ± ± 0.30 Androstendione (ng/ml) 2.21 ± ±

5 Table 2. Results of the study. Outcome Clomiphene citrate Recombinant P-value d (n = 38) FSH (n = 38) Number of cycles Cycles with HCG given (%) 65/104 (63) 71/91 (78) HCG cycles with one large 48/65 (74) 54/71 (76) follicle (%) Cycles with ovulation (%) 55/104 (53) 67/91 (74) HCG cycles with ovulation (%) 55/65 (85) 67/71 (94) Women who ovulated at least 30/38 (79) 35/38 (92) 0.10 once (%) a Pregnancies per cycle (%) 9/104 (9) 16/91 (18) Pregnancies per ovulatory cycle 9/55 (16) 16/67 (24) (%) Pregnancies per woman (%) b 9/38 (24) 16/38 (42) 0.09 Live births per woman (%) c 6/38 (16) 11/38 (29) 0.17 Twin births per live birth (%) 0/6 (0) 2/11 (18) a RR 1.17, 95% CI ; b RR 1.78, 95% CI ; c RR 1.83, 95% CI d As the unit of randomization was the patient but not the cycle, statistical comparisons in the univariate statistical test are presented only between patients. 386 HCG rate increased in subsequent cycles, from 50% in first cycles, to 64% in second cycles and 77% in third cycles (Figure 2). In the FSH group, the HCG rates were similar in the first and second cycles (71 and 76% respectively), but increased to 92% in third cycles. Eighty-two per cent of women in the clomiphene citrate group received HCG in at least one treatment cycle, compared with 92% of women in the FSH group (P = 0.17). The mean day of the cycle that HCG was administered was 14.9 (range 13 18) in the clomiphene citrate group and 16.5 (range 10 25) in the FSH group. In cycles in which HCG was administered, the mean total dose of FSH given was 1183 IU (range IU). On the day of HCG administration, the proportion of cycles with one follicle 17 mm was 74% in the clomiphene citrate group and 76% in the FSH group, with two large follicles 25 and 24%, respectively, and with three large follicles 1.5 and 0% respectively. The ovulation rate per cycle was 55/104 cycles (53%) with clomiphene citrate and 67/91 cycles (74%) with FSH. In the clomiphene citrate group, ovulation occurred in 40% of first cycles, 53% of second cycles, and 70% of third cycles (Figure 3). In the FSH group, ovulation occurred in 66% of first cycles, 72% of second cycles, and 88% of third cycles. Within first cycles, the ovulation rate was significantly higher with FSH than with clomiphene citrate (66 versus 40%, P = 0.02, RR 1.67, CI ). Seventy-nine per cent of women receiving clomiphene citrate ovulated in at least one treatment cycle, compared with 92% in the FSH group (P = 0.10, RR 1.17, CI ). In cycles in which HCG was administered, ovulation occurred in 84% of clomiphene citrate cycles, compared with 94% of FSH cycles. In ovulation cycles, the mean (±SD) midluteal serum progesterone concentration was 9.6 ± 6.5 ng/ml in the clomiphene citrate group and 9.9 ±4.9 ng/ml in the FSH group. By logistic regression analysis, the factors that predicted ovulation were age [odds ratio (OR) 0.82, CI ], serum androstenedione concentration (OR 0.34, CI ), and treatment group (OR 5.6, CI for FSH compared with clomiphene citrate). In the three cycles prior to crossover, there were nine pregnancies in the clomiphene citrate group and 16 pregnancies in the FSH group; thus, the pregnancy rates per woman were 24 and 42%, respectively (P = 0.09). Using FSH as the initial treatment for ovulation induction, compared with clomiphene citrate, the relative risk of pregnancy is 1.78, CI The pregnancy rate per cycle was 9% in the clomiphene citrate group and 18% in the FSH group. In the clomiphene citrate group, there were two pregnancies in first cycles, six in second cycles, and one in third cycles; in the FSH group, the numbers were eight, five, and three respectively (Figure 4). In women who achieved pregnancy, the mean time to pregnancy was 1.9 ± 0.6 cycles with clomiphene citrate and 1.7 ± 0.8 cycles with FSH (P = 0.52). In cycles in which ovulation occurred, the pregnancy rate per cycle was 16% with clomiphene citrate and 24% with FSH. By survival analysis, the cumulative probability of pregnancy after three treatment cycles was 24% with clomiphene citrate and 43% with FSH (P = 0.06) (Figure 5). By logistic regression analysis, the factors that predicted pregnancy were duration of infertility (OR 0.41, CI ) and treatment group (OR 2.9, CI for FSH compared with clomiphene citrate). There were three miscarriages in the clomiphene citrate group and five in the FSH group (P = 1.0). Thus, the live birth rates per woman were 16 and 29% respectively (P = 0.17). The relative risk of live birth after three treatment cycles is 1.83, CI , with FSH compared with clomiphene citrate. There was one multiple pregnancy in the clomiphene citrate group (11%) and three in the FSH group (19%) (P = 1.0). All multiple pregnancies were twins. One twin gestation in each group miscarried; thus, there were two twin live births (18%) in the FSH group and none in the clomiphene citrate group (P = 0.51). There were two cases of mild OHSS according to the classification proposed by Golan et al. (1989), both in women who received FSH and became pregnant.

6 Figure 2. Proportion of cycles in which HCG was administered. (CC = clomiphene citrate; FSH = recombinant FSH.) Figure 3. Proportion of cycles in which ovulation occurred. (CC = clomiphene citrate; FSH = recombinant FSH.) Figure 4. Pregnancy rate per cycle. (CC = clomiphene citrate; FSH = recombinant FSH.) Figure 5. Cumulative pregnancy rate. (CC = clomiphene citrate; FSH = recombinant FSH.) 387

7 388 Second arm After the first arm of the trial, 29 women in the clomiphene citrate group were not pregnant, compared with 22 in the FSH group. Twenty-seven women in the clomiphene citrate group crossed over to FSH treatment ( secondary FSH group ) and nine women (33%) became pregnant. Twenty-one women in the FSH group crossed over to clomiphene citrate treatment ( secondary clomiphene citrate group ) and one woman (5%) became pregnant. The pregnancy rate per cycle in the second arm was 9/66 FSH cycles (14%) and 1/56 clomiphene citrate cycles (2%). One of the nine pregnancies with FSH was a twin pregnancy. All the others, plus the one pregnancy with clomiphene citrate were singletons. There was no miscarriage and no case of OHSS. Discussion This study suggests that low-dose recombinant FSH may be an effective first-line treatment for anovulatory PCOS patients, although the present findings need to be further evaluated in larger studies. Recruitment in clinical trials is one of the biggest challenges in clinical research. The assumption, based on the volume of patients seen in infertility clinics, was that the required sample size could be reached in a reasonable period of time. However, once the study was commenced, the recruitment of patients did not occur at the expected rate, due to overstimation of the numbers of available patients. Because of resource limitations, it was not possible to extend the study for several more years to reach the desired sample size and it was necessary to terminate it and present the results as observed. Thus, further studies on the subject, preferably multi-centre to avoid problems with patient recruitment, are warranted. The introduction of clomiphene citrate into clinical medicine in 1967 revolutionized the treatment of infertility in general and of PCOS in particular. Traditionally, gonadotropin treatment for ovulation induction in PCOS patients has been used as a second line therapy (Yarali and Zeyneloglu, 2004). Current ideas on ovarian stimulation, however, emphasize the need for optimization rather than maximization (i.e. lower number of eggs but with a higher quality will minimize aggressive and potentially deterimental stimulation regimens) (Edwards et al., 1997; Out and Coelingh-Bennink, 1998). The use of recombinant FSH in a chronic low-dose protocol for ovulation induction in PCOS patients may contribute to this concept (Balasch et al., 1998; Coelingh-Bennink et al., 1998; Homburg and Howles, 1999) and thus it seems reasonable to evaluate it as a primary approach. The main mode of action of clomiphene citrate is to boost FSH concentrations in the early follicular phase, but unfortunately, it also raises LH concentrations at this apparently critical stage and, for patients who already have a high baseline LH concentration, the additional discharge of LH may compromise their chances of conceiving (Homburg, 1998; Out and Coelingh-Bennink, 1998). Therefore, as previously stressed (Out and Coelingh-Bennink, 1998), if the goal of clomiphene citrate administration is to increase temporarily FSH concentrations, then from a pharmacological point of view it is reasonable to test the value of administering FSH exogenously. In this regard, it should be noted that elevated serum LH and disturbed intra-ovarian regulation of FSH action are endocrine features in PCOS (Taymor, 1996; Fauser and van Heusden, 1997) and early studies, both in vitro (Erickson et al., 1979) and in vivo (Seibel et al., 1984), provided evidence that the self-perpetuating state of biochemical imbalance so characteristic of PCOS could be interrupted in a physiological way when FSH is administered in a chronic low dose. Thus, from a conceptual point of view, FSH treatment rather than clomiphene citrate could be a more appropriate approach for patients with PCOS. So far as is known, however, no other randomized controlled trial has compared clomiphene citrate and FSH as first line therapy in patients with PCOS. This randomized clinical trial is the first such study, and the results suggest that the use of recombinant human FSH may be not only as effective but even more effective than clomiphene citrate. The results support the implementation of a sufficiently powered multicentre study to test this hypothesis. The following facts support the rationale for such a study. First, the percentages of cycles receiving HCG and cycles with ovulation were higher in the FSH group. Second, the pregnancy rate per woman (P = 0.09), and the cumulative probability of pregnancy after three treatment cycles (P = 0.06), demonstrated a trend in favour of FSH. Third, logistic regression analysis showed that treatment with FSH was one of the factors predicting both ovulation and pregnancy. Finally, FSH use in the second arm was successful in patients who had failed to become pregnant with clomiphene citrate. Although the second period of treatment is biased by the drop-outs from the first arm mainly due to pregnancy, and the problems of period effect and carry-over effect are difficult to quantify, the higher pregnancy rates per woman in the secondary FSH group (33.3%) versus the secondary clomiphene citrate group (5%) support a definitive further assessment of FSH and clomiphene citrate to evaluate the comparative efficacy and safety of these treatments. Obesity is a major determinant of the response to clomiphene citrate in terms of both dose required and success rate (Kousta et al., 1997); in the present study no woman was obese. Thus, it seems unlikely that 50 mg daily of clomiphene citrate may have been too low, as the starting dose. Clomiphene citrate is given in a dose of mg per day for 5 days but it is accepted that a daily dose of 150 mg confers no benefit and only exacerbates the side-effects (Balen, 1997; Homburg and Insler, 2002); patients in the present study received a maximum daily dose of 150 mg clomiphene citrate. Clomiphene citrate is a popular drug because it is easy to administer and no close monitoring is needed. However, it has been suggested that ultrasonographic assessment of follicular growth, as done in this investigation, can improve results obtained (Balen, 1997; Homburg and Insler, 2002). HCG was administered in all clomiphene citrate cycles in patients in the present study. An ovulatory trigger in the form of HCG is not mandatory in clomiphene citrate-induced cycles but it is considered very useful for the timing of intercourse and to ensure ovulation (American College of Obstetricians and Gynecologists, 2002; Homburg and Insler, 2002). Finally, most clomiphene citrate-induced pregnancies occur within the first three menstrual cycles (American College of Obstetricians and Gynecologists, 2002); in the current investigation patients were randomized to receive three treatment cycles with clomiphene citrate or recombinant FSH.

8 Although all these considerations were taken into account when planning this study in order to maximize the success rate with clomiphene citrate, the use of recombinant FSH may be preferable from an efficiency of achieving success perspective. Even in this small trial, some aspects of short-term safety were demonstrated for the treatment groups: (i) the proportion of cycles with two or three large follicles was similar; (ii) multiple pregnancy rates were low in both groups and all multiple pregnancies were twins; and (iii) mid-luteal serum progesterone concentrations were similar. The last point is important when considering that the haemodynamic changes characterizing the OHSS are associated with the luteinization process (Manau et al., 2002). Thus, apart from proposed cytokines (Rizk et al., 1997), a substance or substances produced by the corpora lutea plays a major role in the deterioriation of circulatory function in patients developing OHSS (Manau et al., 2002). In fact, the numbers of cases of OHSS observed in the current study were low; both cases of OHSS recorded in the FSH group were mild in nature. Therefore, the increased incidences of multiple gestation and OHSS that burdened the early days of gonadotrophin treatment now appear surmountable by employing recombinant FSH in a low-dose regimen that is associated with minimal side effects. This notwithstanding, OHSS is a serious iatrogenic complication of ovulation induction and PCOS is a condition clearly predisposing to the development of the syndrome (Golan et al., 1989; Gelety et al., 1992). Thus, this major complication as well as the other side effects both of clomiphene citrate and FSH discussed above should be considered if a larger trial is contemplated. In conclusion, this study strongly suggests that a chronic, lowdose, step-up regimen of recombinant FSH may be not only as effective but even more effective, and as safe as clomiphene citrate when used as first line of therapy for ovulation induction in patients with PCOS. Further studies, possibly multi-centre, are warranted to confirm these results. References Adams J, Polson DW, Franks S 1986 Prevalence of polycystic ovaries in women with anovulation or idiopathic hirsutism. Brithis Medical Journal 293, American College of Obstetricians and Gynecologists 2002 ACOG Practice Bulletin. Management of infertility caused by ovulatory dysfunction. Obstetrics and Gynecology 99, Balasch J 1986 Empleo del citrato de clomifeno y del epimestrol para la inducción de la ovulación. Clinica Ginecológica 10, Balasch J, Ballescá JL, Pimentel C et al Late low-dose pure follicle stimulating hormone for ovarian stimulation in intrauterine insemination cycles. Human Reproduction 9, Balasch J, Tur R, Peinado JA 1996 The safety and effectiveness of stepwise and low-dose administration of follicle stimulating hormone in WHO group II anovulatory infertile women: evidence from a large multicenter study in Spain. Journal of Assisted Reproduction and Genetics 13, Balasch J, Fábregues F, Peñarrubia J et al Follicular development and hormonal levels following highly purified or recombinant follicle-stimulating hormone administration in ovulatory women and WHO Group II anovulatory infertile patients. Journal of Assisted Reproduction and Genetics 15, Balen A 1997 Anovulatory infertility and ovulation induction. Journal of the British Fertility Society 2, Cheang KI, Nestler JE 2004 Should insulin-sensitizing drugs be used in the treatment of polycystic ovary syndrome? Reproductive BioMedicine Online 8, Coelingh-Bennink HJT, Fauser BCJM, Out HJ 1998 Recombinant follicle-stimulating hormone (FSH; Puregon) is more efficient than urinary FSH (Metrodin) in women with clomiphene citrateresistant, normogonadotropic, chronic anovulation: a prospective, multicenter, assessor-blind, randomized, clinical trial. Fertility and Sterility 69, Crosignani PG, Colombo M, Vegetti W et al Overweight and obese anovulatory patients with polycystic ovaries: parallel improvements in anthropometric indices, ovarian physiology and fertility rate induced by diet. Human Reproduction 18, Dickey RP, Holtkamp DE 1996 Development, pharmacology and clinical experience with clomiphene citrate. Human Reproduction Update 2, Edwards RG, Lobo RA, Bouchard P 1997 Why delay the obvious need for milder forms of ovarian stimulation? Human Reproduction 12, Erickson GF, Hsueh AJW, Quigley ME et al Functional studies of aromatase activity in human granulosa cells from normal and polycystic ovaries. Journal of Clinical Endocrinology and Metabolism 49, Fauser BCJM, van Heusden AM 1997 Manipulation of human ovarian function: physiological concepts and clinical consequences. Endocrine Reviews 18, Gelety TJ, Kerin JF, Surrey ES 1992 Ovarian hyperstimulation syndrome. Infertility and Reproductive Medicine Clinics of North America 3, Golan A, Ron-El R, Herman A et al Ovarian hyperstimulation syndrome: an update review. Obstetrics and Gynecological Survey 44, Homburg R 1998 Adverse effects of luteinizing hormone on fertility: fact or fantasy. Baillière s Clinical Obstetrics and Gynaecology 12, Homburg R, Howles CM 1999 Low-dose FSH therapy for anovulatory infertility associated with polycystic ovary syndrome: rationale, results, reflections and refinements. Human Reproduction Update 5, Homburg R, Insler V 2002 Ovulation induction in perspective. Human Reproduction Update 8, Kousta E, White DM, Franks S 1997 Modern use of clomiphene citrate in induction of ovulation. Human Reproduction Update 3, Lord JM, Flight IH, Norman RJ 2003 Metformin in polycystic ovary syndrome: systematic review and meta-analysis. British Medical Journal 327, Manau D, Arroyo V, Jiménez W et al Chronology of hemodynamic changes in asymptomatic in vitro fertilization patients and relationship with ovarian steroids and cytokines. Fertility and Sterility 77, Matorras R, Diaz T, Corcostegui B et al Ovarian stimulation in intrauterine insemination with donor sperm: a randomized study comparing clomiphene citrate in fixed protocol versus highly purified urinary FSH. Human Reproduction 17, Norman RJ, Noakes M, Wu R et al Improving reproductive performance in overweight/obese women with effective weight management. Human Reproduction Update 10, Out HJ, Coelingh-Bennink HJT 1998 Clomiphene citrate or gonadotrophins for ovulation induction? Human Reproduction 13, Pasquali R, Gambineri A 2004 Role of changes in dietary habits in polycystic ovary syndrome. Reproductive BioMedicine Online 8, Rizk B, Aboulghar M, Smitz J, Ron-El R 1997 The role of vascular endothelial growth factor and interleukins in the pathogenesis of severe ovarian hyperstimulation syndrome. Human Reproduction Update 3, Rotterdam ESHRE/ASRM-sponsored PCOS consensus workshop group 2004 Revised 2003 consensus on diagnostic criteria and 389

9 long-term health risks related to polycystic ovary syndrome (PCOS). Human Reproduction 19, Rowe PJ, Comhaire FH, Hargreave TB, Mahmoud AMA 2000 WHO Manual for the Standardized Investigation, Diagnosis and Management of the Infertile Male. Cambridge University Press, Cambridge, UK. Seibel MM, Kamreva MM, MacArdle C, Taymor ML 1984 Treatment of polycystic ovary disease with chronic low-dose follicle stimulating hormone: biochemical changes and ultrasound correlation. International Journal of Fertility 29, Taymor ML 1996 The regulation of follicle growth: some clinical implications in reproductive endocrinology. Fertility and Sterility 65, van Wely M, Bayram N, van der Veen F 2003 Recombinant FSH in alternative doses or versus urinary gonadotrophins for ovulation induction in subfertility associated with polycystic ovary syndrome: a systematic review based on a Cochrane review. Human Reproduction 28, Yarali H, Zeyneloglu HB 2004 Gonadotrophin treatment in patients with polycystic ovary syndrome. Reproductive BioMedicine Online 8, Received 2 July 2004; refereed 19 July 2004; accepted 30 July