Key words: follicle development / FSH administration/ovarian stimulation/ovulatory women

Transcription

1 Human Reproduction Vol.21, No.11 pp , 2 Advance Access publication July 27, 2. doi:.93/humrep/del259 Timing of FSH administration for ovarian stimulation in normo-ovulatory women: comparison of an early or a mid follicular phase initiation of a short-term treatment I.Cedrin-Durnerin 1,3, N.Massin 1, J.Galey-Fontaine 1, H.Bry-Gauillard 1, M.Roger 2, N.Lahlou 2 and J.N.Hugues 2 1 Centre for Reproductive Medicine Jean Verdier Hospital, Bondy Cedex, AP-HP, University Paris XIII and 2 Hormonal Biology, Saint Vincent de Paul Hospital, AP-HP, University Paris VI, Paris, France 3 To whom correspondence should be addressed at: Service de Médecine de la Reproduction, Hôpital Jean Verdier, Avenue du 14 Juillet, Bondy Cedex, Bondy, France. isabelle.cedrin-durnerin@jvr.ap-hop-paris.fr BACKGROUND: In normo-ovulatory infertile women undergoing mild ovarian stimulation out of IVF, FSH stimulation regimen must be carefully adjusted to control the number of recruited follicles and to prevent multiple pregnancies. The aim of this prospective study was to assess the effect of the timing of FSH administration (fixed dose and duration) on the number of large follicles. METHODS: Women were prospectively randomized by means of sealed envelopes to receive daily IU recombinant FSH (rfsh), either from cycle day (CD) 2 (Group A) or from CD 7 11 (Group B). Hormonal measurements and follicular ultrasound assessments were performed on CD 2, 7 and 12. RESULTS: On CD 12, the development rate of exactly two follicles ³14 mm in diameter was significantly lower in Group A than in Group B (4% of women versus 42%, P =.2). Although the pattern of serum estradiol (E 2 ) concentrations in Group A displayed a plateau from CD 7, the cancellation rate for overstimulation (more than three follicles ³14 mm in diameter) was significantly increased (P =.9). CONCLUSIONS: Preventing the closure of the FSH window by mid to late follicular phase FSH administration better fulfils the objective of obtaining a limited number of large follicles than surpassing the FSH threshold by an early administration. Key words: follicle development / FSH administration/ovarian stimulation/ovulatory women Introduction Controlling the number of growing follicles is one of the key issues of ovarian stimulation in women who do not proceed to IVF as a first line therapy, because the follicle number on the day of hcg administration is the main determinant of both pregnancy rate and risk of complications. Indeed, in women with normoovulatory cycles undergoing intrauterine insemination (IUI) for unexplained or moderate male infertility, the pregnancy rate is significantly improved when FSH stimulation leads to more than one single dominant follicle. Meanwhile, risks of multiple pregnancy and ovarian hyperstimulation syndrome (OHSS) are simultaneously increased (Stone et al., 1999; Guzick et al., 1999; Dickey et al., 21). Therefore, it is usually recognized that increasing the chance of pregnancy should not outweigh the risk of multiple birth. However, there is so far no definite consensus on the most appropriate number and size of follicles to be recruited in these situations. When a gentle stimulation strategy is applied, withholding HCG administration is currently recommended if more than three large follicles are present (Fauser et al., 25). However, triggering ovulation with exactly two large follicles could be more relevant to simultaneously increasing success rate and preventing high-order multiple pregnancies. As regards follicular sizes, several studies have demonstrated that the risk of high-order multiple pregnancy is better associated with the number of follicles mm in diameter than with the number of large follicles at the time of hcg administration. However, different threshold values, e.g. more than 3, or 7 follicles mm in diameter (Gleicher et al., 2; Tur et al., 21; Dickey et al., 25), have been proposed for withholding hcg administration. Therefore, although follicles <14 mm in diameter on the day of hcg administration are unlikely to lead to ovulation, defined by follicular collapse on ultrasound examination (Silverberg et al., 1991), and to yield a pregnancy (Richmond et al., 25), the presence of intermediate size follicles ( 13 mm) should be taken into account at the time of hcg administration. Accordingly, an objective of ovarian stimulation leading to the development of two follicles 14 mm in diameter with a limited number of follicles mm in diameter could be a suitable option for an optimal pregnancy rate with a limited risk of high-order multiple births in normo-ovulatory women. Today, there are no clear guidelines regarding the management of FSH stimulation in normo-ovulatory women to achieve The Author 2. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved For Permissions, please journals.permissions@oxfordjournals.org

2 I.Cedrin-Durnerin et al. this goal. In clinical practice, FSH administration is usually started from the early follicular phase with a starting dose calculated on age, BMI and history of previous ovarian response. Subsequent FSH dose adjustments are performed according to the ovarian response to modulate the number of growing follicles. Daily FSH doses ranging from 75 to 15 IU are currently used in ovarian stimulation programmes for IUI. In normoovulatory women, it has been shown (Sengoku et al., 1999) that a low dose step-up protocol (starting dose of 75 IU per day with dose increase by 37.5 IU every 7 days in the absence of recruited follicle) led to an average rate of two large follicles and two intermediate size follicles. Nevertheless, about 5% of cycles were monofollicular for large follicles. Compared with a conventional protocol (starting dose of 15 IU increased by 75 IU every 5 days), the mean number of large and intermediate size follicles was about four and three, respectively, and the rate of monofollicular development was reduced to 25%. Therefore an intermediate starting dose of IU could better fit with the objective of a high rate of bifollicular development for large follicles and a limited number of growing follicles. The duration and the timing of FSH administration are other important tools for tailoring the number of growing follicles. During the luteo-follicular transition of a normal cycle, the increase of serum FSH beyond a certain threshold initiates recruitment of several follicles. Subsequently, the closure of the FSH window from the mid follicular phase is responsible for the selection of a single dominant follicle. These two concepts of FSH threshold and FSH window were applied to anovulatory women for ovulation induction with two different protocols, namely the step-up and step-down regimen of FSH administration (Homburg and Howles, 1999). The main advantage of step-down regimen is a shorter duration of FSH administration, but its safety regarding the rate of multifollicular development compared with the step-up protocol seems lower (Christin-Maitre et al., 23). In normo-ovulatory women, several data suggest that duration and timing, rather than dose, of FSH administration are involved in the regulation of the number of growing follicles during ovarian stimulation. Indeed, both early FSH administration at the beginning of the follicular phase (Hohmann et al., 21) and extended FSH administration in the late follicular phase (Lolis et al., 1995; Hughes et al., 1998) proved to increase the percentage of cycles with more than one single dominant follicle. Moreover, it has been reported that the duration, rather than the magnitude, of FSH administration affects follicular development (Schipper et al., 1998a). Indeed, a brief elevation of FSH levels induced by a single injection of a high dose of FSH in the early follicular phase did not impair selection and dominance processes, whereas moderate but persistent elevation of FSH levels, using the same dose split in daily injection for 5 days, from the mid to the late follicular phase induced ongoing growth of multiple follicles. However, a short-term treatment from the early to the mid follicular phase to surpass the FSH threshold, followed by the discontinuation of FSH administration from the mid to the late follicular phase to close the FSH window, could be an alternative approach to mimic normal physiological changes and to get a limited number of growing follicles. This prospective randomized study was thus set up in normo-ovulatory women to assess the effect of the timing of 2942 FSH administration on the resulting number of follicles 14 mm and mm in diameter. Comparison was performed between women treated for 5 days with a daily dose of UI of recombinant FSH (rfsh) initiated either from the early or from the mid follicular phase. Materials and methods Subjects From May 2 to November 22, infertile women with ovulatory cycles were enrolled in this study approved by a review committee. The inclusion criteria were: age 38 years, BMI 27, cycle length days, plasma luteal progesterone values 5 ng/ml, normal basal FSH level ( 13 IU/l) and normal antral follicle count (<AFC<24 follicles for both ovaries ) at ultrasound. Exclusion criteria were previous ovarian surgery, ovarian endometriosis and endocrine and systemic disorders (eg. diabetes mellitus, hepatic, renal or cardiovascular diseases). Women with unexplained infertility or requiring IUI for cervical or moderate male factor infertility were eligible for this study. Included patients gave written informed consent. Protocol Patients were prospectively randomized by means of sealed opaque envelopes to receive daily s.c. injections of IU of rfsh (Gonal F ; Serono, Boulogne, France) either from cycle day (CD) 2 (Group A) or from CD 7 11 (Group B). Random allocation sequence was generated from a table of random numbers and was concealed to physician who enrolled and randomized patients. This study was not blind. Blood sampling for hormonal determinations and ultrasound assessments of follicular development were performed on CD 2, 7 and 12 in both groups. According to the data obtained on CD 12, another assessment of follicular development was programmed on the presumed day of hcg administration. No additional exogenous FSH administration between CD 12 and HCG administration was included in the study design. However, a few patients requiring gonadotrophin support to sustain estradiol (E 2 ) secretion or follicular growth on CD 12 did receive additional FSH to avoid cycle cancellation. When at least one follicle reached 17 mm in diameter, 5 IU urinary HCG (Gonadotrophines chorioniques Endo ; Organon, Puteaux, France) was administered. No luteal support was performed. When more than three follicles 14 mm in diameter were present, the administration of hcg was withheld and patients were informed to have no intercourse, or protected intercourse. Hormonal measurements Hormonal measurements were carried out using commercially available chemiluminescence immunoassays with automated Elecsys immunoanalyser (ECLIA; Roche diagnostic, Meylan, France). The sensitivity of the assay was.1 IU/l for FSH and LH. Intra-assay and inter-assay coefficients of variation were within 3 and % and within 3 and 4%, respectively, for FSH and LH. The sensitivity of the assay was 5 pg/ml and.3 ng/ml for E 2 and progesterone, respectively. Intra-assay and inter-assay coefficients of variation were 5 and %, respectively, for E 2 and 3 and 5%, respectively, for progesterone. Inhibin (INH) A and B were measured from frozen serum samples as previously described (Lahlou et al., 1999) using Oxford BioInnovation reagents distributed by DSL-France (Cergy-Pontoise, France). In the INH A assay, the intra-assay precision was 5.4 and 3.2% at concentrations 14 and 48 pg/ml respectively, the sensitivity was 1 pg/ml. In the INH B assay, the intra-assay precision was 7.4 and 4.2% at concentrations 44 and 225 pg/ml, respectively; the sensitivity was pg/ml.

3 Ovarian stimulation in normo-ovulatory women Vaginal ultrasonography Ultrasound assessments were performed with a -Mhz vaginal transducer and a Toshiba SSA-34 device. Follicle diameter was calculated as the mean diameter measured in 2Ds. Only follicles >4 mm in diameter were considered in this study. Sample size estimate This study was a feasibility study. As very few data are available so far regarding the effects of short-term FSH treatment in normo-ovulatory women, and no data exist for a IU daily dose regimen, no primary end-point was defined. However, a sample size of 25 subjects per group is able to show about 4% of variation in the development rate of follicles. This is in accordance with the results observed in previous published studies quoted in Introduction. Statistical analysis Results are expressed as mean ± SD. Statistical analysis was performed using StatView 4.5 (Abacus Concepts, Berkeley, CA, USA). Nominal or continuous variables were analysed with chi-square or Student s t-test or analysis of variance for repeated measures as required. A P value <.5 was considered as statistically significant. Results Baseline characteristics From 52 randomized normo-ovulatory women, 48 started FSH stimulation and completed the study cycle until CD 12 according Enrolment Allocation Allocated to Group A (n=2) Received FSH (n=24) Did not receive FSH (n=2) ovarian cyst personal reason to the protocol (Figure 1). Both groups were similar (Table I) as regards age, BMI, cycle length, ovarian reserve assessed by CD 3 plasma FSH and E 2 levels, and infertility factor. Follicle development Following FSH administration from CD 2 to CD in Group A, the number of follicles mm and the endometrial thickness assessed on CD 7 were significantly higher than in Group B. However, on CD 12 following similar FSH administration from CD 7 to CD 11 in Group B, this difference was no longer significant between the two groups (Table II). Nevertheless, the development rate of exactly two large follicles was significantly higher (P =.2) in Group B (42% of women) than in Group A (4%). If the development rate of two or three large follicles was grouped in one class, the difference remained significant between groups (P =.4). However, the distribution of follicles mm in diameter was similar in both groups (Table III). There was no significant association between the number of follicles 14 mm or mm in diameter on CD 12 and basal plasma FSH or INH B levels on CD 2. Serum hormone concentrations The patterns of plasma hormonal concentrations were quite different between groups (Figure 2). E 2 and INH A values increased during the early follicular phase and tended to plateau Assessed for eligibility (n=not available) Randomized (n=52) Allocated to Group B (n=2) Received FSH (n=24) Did not receive FSH (n=2) high cycle day 2 FSH personal reason Follow-up Lost to follow-up (n=) Discontinued intervention (n=) Lost to follow-up (n=) Discontinued intervention (n=) Analysis Analysed (n=24) Excluded from analysis (n=) Analysed (n=24) Excluded from analysis (n=) Figure 1. Flow chart of the patients throughout the study. 2943

4 I.Cedrin-Durnerin et al. Table I. Characteristics of patients who received daily injections of IU recombinant FSH from cycle day 2 to (Group A) or from days 7 to 11 (Group B) Group A Group B P Age (years) 3.8 ± ± 3.4 NS BMI (kg/m 2 ) 21.4 ± ± 2.2 NS Cycles length (days) 28 ± ± 1.2 NS CD 3 FSH (IU/l) 8.2 ± ± 2.7 NS CD 3 E 2 (pg/ml) 41 ± ± 17 NS CD 2 progesterone (ng/ml) 13.3 ± ± 7 NS Infertility factor NS Unexplained infertility Moderate male factor 4 Cervical factor 3 4 Results are expressed as mean + SD; E 2, estradiol; CD, cycle day. Table II. Ultrasound evaluation E2 (pg/ml) Inh A (pg/ml) FSH (UI/l) Inh B (pg/ml) NS, non-significant. between CD 7 and CD 12 in Group A, whereas they sharply increased during the late follicular phase in Group B. FSH administration in Group A prevented the decrease in FSH levels on day 7, whereas FSH values decreased from day 2 to day 7 in Group B. Late follicular phase FSH administration in Group B resulted in higher FSH levels in Group B than in Group A on day 12 (P <.9). INH B values increased on CD 7 and decreased on CD 12 in Group A, whereas they sharply increased during the late follicular phase following FSH administration in Group B. Patterns of ovarian response in Group A Follicular and hormonal responses to FSH administration were quite heterogeneous within Group A. Two subgroups were characterized according to the presence (subgroup A1, n = ) 2944 Group A Group B CD 2 Follicles < mm 8.2 ± ± 2.8 NS Endometrial thickness (mm) 4 ± ± 2 NS CD 7 Follicles mm 3.8 ± ±.8.1 Follicles < mm 7.2 ± ± 5 NS Endometrial thickness (mm) 7.5 ± ± 1..1 CD 12 Follicles 14 mm 2. ± ±.7 NS Follicles 13 mm 2.1 ± 2 2 ± 2 NS Follicles < mm.8 ± ± 3.8 NS Endometrial thickness (mm) 8.8 ± 2 8. ± 1.8 NS Table III. Follicular cohort on CD 12 Group A Group B P Follicles 14 mm (Absent 3) (Absent 1) P =.2 One 9 11 Two 1 Three 2 Four or more 5 Follicles mm NS Three or less Four to six 5 8 Seven or more 3 P Figure 2. Serum concentrations of estradiol (E 2 ), inhibin (INH) A, FSH and INH B on cycle days 2, 7 and 12 in Group A ( ) and Group B ( ). Results are presented by mean ± 95% confidence interval. Analysis of variance for repeated measures showed significant differences between treatment groups: P <.1 for E 2, P =.9 for INH A, P <.1 for FSH and P =.2 for INH B. Size of the leading follicle (mm) E2 (pg/ml) N of follicles FSH (UI/L) < mm >mm >14 mm Figure 3. Different patterns of serum E 2 and FSH concentrations and follicular development in Group A according to the type of ovarian response. Subgroups A1 ( ) and A2 ( ) were characterized by the presence or absence of at least one follicle 14 mm in diameter on cycle day 7, respectively. Results are presented by mean ± 95% confidence interval. = values from subgroup A1 and A2 are significantly different, P <.5. or the absence (subgroup A2, n = 14) of at least one follicle 14 mm in diameter on CD 7, with a leading follicle mean size of 15.3 ± 1. mm and 11.7 ± 1. mm, respectively. As shown in Figure 3, the two subgroups differed on CD 2 only by FSH levels that were significantly lower in subgroup A1 than in subgroup A2 (7.8 ± 1.4 versus.8 ± 4.1 UI/l, P =.4), whereas the mean number of follicles < mm in diameter was not different between subgroups. Early response to FSH led to earlier monitor the ovarian response initially scheduled on D12

5 Ovarian stimulation in normo-ovulatory women Table IV. Outcome of ovarian stimulation Group A Results are expressed as mean ± SD; NS, non-significant. Group B Cancellation (% per started cycle) 8 (33.3) 1(4.2).9 On the day of hcg hcg day 12.4 ± ± 1.1 NS E 2 (pg/ml) 312 ± ± 13. Follicles >14 mm 1. ± ±.8 NS One 11 Two Three 5 4 Follicles 13 mm 1.1 ± ± 2 NS Endometrial thickness (mm) 9.1 ± ± 2 NS Luteal phase E 2 (pg/ml) 217 ± ± 448 NS Progesterone (ng/ml) 1.4 ± ±.7 NS Pregnancies (% per started cycle) 1 (4.2) 5 (2.8).8 in all patients from subgroup A1 (mean day of control 9.4 ±.7) and in 9 of 14 of patients from subgroup A2 (mean day of control 11.5 ±.8). At that time, the mean number of follicles 14 mm in diameter was significantly higher in subgroup A1 than in subgroup A2 (4.5 ± 3.1 versus 1.2 ± 1.1, P =.1) as well as the mean number of follicles mm in diameter (7.1 ± 4.5 versus 1.8 ± 1.9, P =.1). From day 7 to the assessment of ovarian response, the growth of the leading follicle was significantly higher in subgroup A1 than in subgroup A2 (2 ±.5 versus.9 ±. mm per day, P <.1). Cycle outcome Eight cycles in Group A were cancelled for excessive response (more than 3 follicles 14 mm in diameter) and only one in Group B (P =.9). Four patients from subgroup A2 received additional FSH administration to sustain follicular growth from CD 12 until hcg administration and none from Group B. On the day of hcg administration (Table IV), plasma E 2 concentrations were significantly higher in the Group B than in the Group A, whereas the mean number of follicles 14 mm in diameter was not different between groups. However, the development rate of exactly two large follicles was significantly higher in Group B. Plasma steroid concentrations during the luteal phase were similar between groups. Only one pregnancy was obtained in Group A, and the pregnancy rate in Group B was 21% per started cycle. There were five single pregnancies and one twin pregnancy (Group B) leading to delivery of seven healthy babies. Discussion These data show that the timing of FSH administration is critical for controlling the number of recruited follicles following ovarian stimulation in normo-ovulatory patients. Guidelines established in anovulatory infertility recommended an adjusted starting dose and a stepwise FSH administration (Homburg and Howles, 1999). By contrast in normo-ovulatory patients, no recommendation has been published so far for the management of ovarian stimulation, where exogenous FSH administration interferes with endogenous FSH dynamics, required for follicle P selection. FSH dose adjustments are commonly used, whereas modulation of the timing of FSH administration is rarely performed to tailor the number of follicles. However, surpassing the FSH threshold or preventing the closure of the FSH window may affect the number of growing follicles in different ways. Whereas daily FSH administration of 75 IU from CD 7 overrules the process of single dominant follicle selection in a small proportion of cycles (Balasch et al., 1994), starting administration early on CD 5 or 3 (Hohmann et al., 21) or extending administration in the late follicular phase (Lolis et al., 1995; Hughes et al., 1998) constantly increase the rate of multifollicular development. However in these studies, timing, duration and dose of FSH administration are linked and vary together. In contrast, comparison of the effects of a single high dose of FSH in the early follicular phase and the same amount split into 5-day administration in the late follicular phase (Schipper et al., 1998a) led to the suggestion that duration and timing rather than dose of FSH administration are involved in the control of the number of growing follicles during ovarian stimulation. Therefore, our data obtained from cycles treated with a fixed FSH dose and duration are in agreement with the concept that changing only the timing of FSH administration actually impacts on the number of recruited follicles. The objective to obtain a limited number of large follicles was more successfully achieved in patients whose FSH support was given in the late follicular phase than in patients FSH-supplemented only in the early follicular phase. This study, thus, suggests that more attention should be paid to adjusting the timing of exogenous FSH administration in normo-ovulatory patients. Although FSH supplementation restricted to the early follicular phase mimics closely the physiological rise of FSH levels observed in normo-ovulatory cycles, this regimen of FSH administration was associated with a high risk of multifollicular recruitment, although the ovarian response was quite heterogeneous. Indeed, about half of patients presented an early and excessive ovarian response. The low basal FSH levels observed in these patients could reflect a high ovarian sensitivity to FSH (Schipper et al., 1998b). It may thus be presumed that the daily FSH dose used in our study was too high and that these patients should benefit from lower starting doses (Papageorgiou et al., 24). The other half of patients displayed an initial follicular response consistent with the objective of a mild stimulation followed by a plateau or a drop of E 2 secretion concomitant with a slow follicular growth in the late follicular phase. Thus, discontinuation of FSH administration in conjunction with the estrogen-induced negative feedback on endogenous gonadotrophin secretion seems to be detrimental for follicular maturation as long as the leading follicle has not fully reached dominance. Although this study was not designed, for sample size, to look at pregnancy rate, the marked reduction of pregnancies in this group shows that a regimen of FSH administration restricted to the early follicular phase is inadequate for normo-ovulatory women and suggests that FSH supplementation is required in the late follicular phase. FSH administration from the mid to the late follicular phase resulted in a more homogeneous ovarian response with a high proportion of patients who achieve the objective of only two follicles 14 mm in diameter. Furthermore, this regimen of late FSH administration was associated with a limited number of 2945

6 I.Cedrin-Durnerin et al. follicles mm in diameter. These data are clinically relevant because the risk of high-order multiple pregnancies is associated with the number of follicles mm in diameter on the day of hcg administration in patients younger than 32 (Tur et al., 21) or 38 years (Dickey et al., 25). Withholding hcg when more than three (Tur et al., 25) or six (Dickey et al., 25) follicles mm in diameter are present could significantly reduce the number of high-order multiple pregnancies, with only a slight reduction in the overall pregnancy rate. However, the issue of the contribution of medium-sized follicles to pregnancy rate is still a matter of debate. Indeed, in a series of conception cycles performed in patients downregulated with GnRH agonist, no multiple pregnancies were observed in the presence of only one follicle 14 mm in diameter, and no high order multiple pregnancies when the tertiary follicle measured <14 mm in diameter (Richmond et al., 25). Nevertheless, the authors postulated but could not prove that pregnancies came from the largest follicles. Furthermore, for stimulation regimens without GnRH agonist, the possibility of 13 mm follicles reaching ovulation after hcg administration following a spontaneous LH surge cannot be excluded. These data suggest that the ideal regimen of FSH administration in normo-ovulatory women could combine FSH supplementation from the mid follicular phase with subsequent reduction in FSH doses when two follicles have reached 14 mm in diameter to decrease the number of intermediate size follicles, as previously shown in anovulatory patients (Hugues et al., 199). An alternative to the reduction of FSH supply could be a timely administration of a GnRH antagonist in order to decrease the endogenous FSH secretion. The use of GnRH antagonist concomitantly with daily low-dose FSH administration proved to be effective and safe in preventing multiple pregnancies (Ragni et al., 24). Although the usefulness of mild ovarian stimulation has been questioned compared with natural cycles for IUI (Goverde et al., 25; Van Rumste et al., 2), we do believe that future refinements of the FSH administration regimen could improve the pregnancy rate and minimize the risk of multiple pregnancy, as has been demonstrated over the last decade for anovulatory infertility. In conclusion, our study shows that a timely FSH administration in normo-ovulatory women is an important factor in tailoring the number of large preovulatory follicles. Initiation of FSH administration from the mid follicular phase using a steady dose for a fixed duration better achieves the objective of obtaining two large follicles ( 14 mm in diameter) than an early administration. These results suggest that preventing the closure of the FSH window is more successful than surpassing the FSH threshold in creating mild overstimulation in normoovulatory patients. However, others studies are needed to compare efficacy and safety of the late follicular phase regimen with those of the most commonly used FSH administration, during the whole follicular phase. Acknowledgements This work was supported for hormonal measurement of INH B and INH A by Institut de Recherche Endocrinienne et Métabolique, Paris (France). 294 References Balasch J, Ballesca JL, Pimentel C, Creus M, Fabregues F and Vanrell JA (1994) Late low-dose pure follicle stimulating hormone for ovarian stimulation in intra-uterine insemination cycles. Hum Reprod 9, Christin-Maitre S, Hugues JN and Recombinant FSH Study Group (23) A comparative randomized multicentric study comparing the step-up versus step-down protocol in polycystic ovary syndrome. Hum Reprod 18, Dickey RP, Taylor SN, Lu PY, Sartor BM, Rye PH and Pyrzak R (21) Relationship of follicle numbers, estradiol levels to multiple implantation in 38 intrauterine insemination cycles. Fertil Steril 75,9 78. Dickey RP, Taylor SN, Lu PY, Sartor BM, Rye PH and Pyrzak R (25) Risk factors for high-order multiple pregnancy and multiple birth after controlled ovarian hyperstimulation: results of 42 intrauterine insemination cycles. Fertil Steril 83, Fauser BCJM, Devroey P and Macklon NS (25) Multiple birth resulting from ovarian stimulation for subfertility treatment. Lancet 35, Gleicher N, Oleske DM, Tur-Kaspa I, Vidali A and Karande V (2) Reducing the risk of high-order multiple pregnancy after ovarian stimulation with gonadotropins. N Engl J Med 343,2 7. Goverde AJ, Lambalk CB, McDonnell J, Schats R, Homburg R and Vermeiden JPW (25) Further considerations on natural or mild hyperstimulation cycles for intrauterine insemination treatment: effects on pregnancy and multiple pregnancy rates. Hum Reprod 2, 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 34, Hohmann FP, Laven JSE, de Jong FH, Eijkemans MJC and Fauser BCJM (21) Low dose exogenous FSH initiated during the early, mild or late follicular phase can induce multiple dominant follicle development. Hum Reprod 1, Homburg R and Howles CM (1999) Low-dose FSH therapy for anovulatory infertility associated with polycystic ovary syndrome: rationale, results, reflections and refinements. Hum Reprod Update 5, Hughes EG, Collins JA and Gunby J (1998) A randomised controlled trial of three low-dose gonadotrophin protocols for unexplained infertility. Hum Reprod, Hugues JN, Cedrin-Durnerin I, Avril C, Bulwa S, Hervé F and Uzan M (199) Sequential step-up and step-down dose regimen: an alternative method for ovulation induction with follicle stimulating hormone in polycystic ovary syndrome. Hum Reprod 11, Lahlou N, Chabbert-Buffet N, Christin-Maitre S, Le Nestour E, Roger M and Bouchard P (1999) Main inhibitor of follicle stimulating hormone in the luteal-follicular transition: inhibin A, oestradiol, or inhibin B? Hum Reprod 14, Lolis DE, Tsolas O and Messinis I (1995) The follicle-stimulating hormone threshold level for follicle maturation in superovulated cycles. Fertil Steril 3, Papageorgiou TC, Guibert J, Savale M, Goffinet F, Fournier C, Merlet F, Janssens Y and Zorn JR (24) Low dose recombinant FSH treatment may reduce multiple gestations caused by controlled ovarian hyperstimulation and intrauterine insemination. BJOG 111, Ragni G, Alagna F, Brigante C, Riccaboni A, Colombo M, Somigliana E and Crosignani PG (24) GnRH antagonists and mild ovarian stimulation for intrauterine insemination: a randomized study comparing different gonadotrophin dosages. Hum Reprod 19, Richmond JR, Deshpande N, Lyall H, Yates RWS and Fleming R (25) Follicular diameters in conception cycles with and without multiple pregnancy after stimulated ovulation induction. Hum Reprod 2,75 7. Schipper I, Hop WCJ and Fauser BCJM (1998a) The follicle-stimulating hormone (FSH) threshold/window concept examined by different interventions with exogenous FSH during the follicular phase of the normal menstrual cycle: duration rather than magnitude of FSH increase affects follicle development. J Clin Endocrinol Metab 83, Schipper I, de Jong FH and Fauser BCJM (1998b) Lack of correlation between maximum early follicular phase serum follicle stimulating hormone concentrations and menstrual cycle characteristics in women under the age of 35 years. Hum Reprod 13, Sengoku K, Tamate K, Takaoka Y, Horikawa M, Goishi K, Komori H, Okada R, Tsuchiya K and Ishikawa M (1999) The clinical efficacy of low-dose stepup follicle stimulating hormone administration for treatment of unexplained infertility. Hum Reprod 14,

7 Ovarian stimulation in normo-ovulatory women Silverberg KM, Olive DL, Burns WN, Rye PH, Lu PY and Pyrzark R (1991) Follicular size at the time of human chorionic gonadotropin administration predicts ovulation outcome in human menopausal gonadotropin-stimulated cycles. Fertil Steril 5,29 3. Stone BA, Vargyas JM, Ringler GE, Stein AL and Marss RP (1999) Determinants of the outcome of intrauterine insemination: analysis of outcomes of 993 consecutive cycles. Am J Obstet Gynecol 18, Tur R, Barri PN, Coroleu B, Buxaderas R, Martinez F and Balasch J (21) Risk factors for high-order multiple implantation after ovarian stimulation with gonadotrophins: evidence from a large series of 1878 consecutive pregnancies in a single centre. Hum Reprod 1, Tur R, Barri PN, Coroleu B, Buxaderas R, Parera N and Balasch J (25) Use of a prediction model for high-order multiple implantation after ovarian stimulation with gonadotrophins. Fertil Steril 83, Van Rumste MME, den Hartog JE, Dumoulin JCM, Evers JHL and Land JA (2) Is controlled ovarian stimulation in intrauterine insemination an acceptable therapy in couples with unexplained non-conception in the perspective of multiple pregnancies? Hum Reprod 21, Submitted on February 14, 2; resubmitted on April 28, 2; accepted on May 25,