1 Institut Clinic of Gynecology, Obstetrics and Neonatology and 2 Hormonal Laboratory, Faculty of MedicineUniversity of Barcelona, Hospital Clinic-Institut dInvestigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain
3 To whom correspondence should be addressed at: Institut Clinic of Gynecology, Obstetrics and Neonatology, Hospital Clinic; C/Casanova 143, 08036 Barcelona, Spain. e-mail: jbalasch{at}medicina.ub.es
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: assisted reproduction/early pregnancy loss/recombinant FSH/LH/ovarian stimulation
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
From experimental and clinical evidence it seems unequivocal that ovarian follicles have development-related requirements for stimulation by LH, i.e. there is a threshold for LH requirements during folliculogenesis (Hillier, 2000; Balasch and Fábregues, 2002
; Filicori et al., 2002
; Shoham, 2002
). However, although a lower limit of LH activity associated with an appropriate ovarian response and the establishment of a pregnancy does existand it is likely to be very low since <1% of follicular LH receptors need to be occupied in order to allow normal steroidogenesis (Chappel and Howles, 1991
)the exact limit has not yet been determined. Thus, threshold values of serum LH of <3, <1.8, <1,
0.7 and <0.5 IU/l at different study points during the follicular phase of gonadotrophin ovarian stimulation cycles in down-regulated women have been proposed to diagnose the group of profoundly suppressed LH patients (Chappel and Howles, 1991
; Fleming et al., 2000
; Westergaard et al., 2000
; Esposito et al., 2001
; Lam et al., 2002
; Tesarik and Mendoza, 2002
).
In those previous studies, however, the threshold values for LH were established arbitrarily or according to the detection limit of LH assays used. In contrast, in a previous report (Balasch et al., 2001) we used a statistical method using receiver-operating characterisitc (ROC) curve analysis to test the usefulness of mid-follicular (stimulation day 7) LH concentration to discriminate between conception versus non-conception cycles and ongoing pregnancies versus early pregnancy losses. To this end, 72 consecutive patients who had become pregnant on assisted reproduction treatment were matched with the non-conception assisted reproduction treatment cycles that immediately followed. We found that mid-follicular LH concentrations were not predictive of implantation or early pregnancy loss after IVF/ICSI (Balasch et al., 2001
). It is noteworthy that in ROC curve analysis, many efficiencies of all decision levels can be calculated, resulting in an overall quantification of accuracy which is not affected by the prevalence of a condition. Also, ROC plots provide a pure index of accuracy by demonstrating the limits of a tests ability to discriminate between alternative states of health over the complete spectrum of operating conditions (Hanley and McNeil, 1982
; Zweig and Campbell, 1993
).
The aim of this prospective study, where ROC analysis was used, was to assess further the usefulness of serum LH levels as predictors of ovarian response, assisted reproduction treatment outcome, and the outcome of pregnancy when measured throughout the ovarian stimulation period in a large cohort of assisted reproduction treatment patients who received ovarian stimulation with recombinant FSH under pituitary suppression.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
All patients received standard ovarian stimulation with FSH under pituitary suppression with GnRH agonist according to a protocol previously reported (Balasch et al., 2001). In all women, pituitary desensitization was achieved by s.c. administration of triptorelin acetate (Decapeptyl 0.1 mg; Ipsen Pharma, Spain) (0.1 mg daily, which was reduced to 0.05 mg after ovarian arrest was confirmed) started in the mid-luteal phase of the previous cycle. Gonadotrophin stimulation of the ovaries was started when serum estradiol concentrations declined to <50 pg/ml and a vaginal ultrasonographic scan showed an absence of follicles >10 mm diameter. On days 1 and 2 of ovarian stimulation, 6 and 4 ampoules/day of recombinant human FSH (75 IU per ampoule) (Gonal-F; Serono S.A., Spain) respectively were administered s.c. On days 3 and 4 of ovarian stimulation, two ampoules per day of FSH were administered to each patient. From day 5 onward, FSH was administered on an individual basis according to the ovarian response as assessed by sequential transvaginal ultrasonography and serum estradiol measurements. The criteria for hCG administration (5000 IU; Profasi; Serono S.A.) were the presence of
2 follicles >18 mm in diameter with
4 follicles measuring
14 mm in association with a consistent rise in serum estradiol concentration. Oocyte aspiration was performed with vaginal ultrasonography 3536 h after hCG administration. The maturational status of the oocytes and the embryo grading were recorded according to published criteria (Veeck, 1999
); embryos of Veeck grades 1 or 2 were considered high quality. Up to three embryos per patient were replaced and the luteal phase was supported with additional doses of hCG or vaginal micronized progesterone according to ovarian response. The cycle was cancelled when there were <3 follicles with diameter
14 mm after 89 days of gonadotrophin therapy or after 45 additional treatment days without attaining, or the imminent prospect of attaining, the criteria for hCG administration.
Pregnancy was diagnosed by increasing serum concentrations of -hCG after embryo transfer, and the subsequent demonstration of an intrauterine gestational sac by ultrasonography. The following categorization of pregnancy outcome was made: (i) biochemical pregnancy: women showing serum
-hCG concentrations >50 IU/l 1314 days after embryo transfer and 7 days later but no intra uterine or extrauterine pregnancy could be demonstrated by ultrasonography and menses ensued; (ii) spontaneous abortion: women having increasing concentrations of
-hCG on days 1314 and 2021 after embryo transfer, and pregnancy confirmed by ultrasonic demonstration of an intrauterine gestational sac but the gestation subsequently spontaneously terminated; (iii) ongoing pregnancy: this category comprised patients achieving >20 weeks gestation.
Blood samples, hormone assays and ultrasonography
Treatment monitoring was carried out throughout gonadotrophin administration. Blood samples for hormone analyses were obtained on day S0 (the day when pituitary suppression was evidenced) and every other day from stimulation day 5 (S5) until the day of hCG injection. Ultrasonographic scans were performed the same days as hormone analyses. Each day one blood sample was drawn between 08:00 and 10:00 h, and for this study two serum aliquots were obtained. Estradiol was measured daily in one of the serum aliquots for clinical monitoring, and the second aliquot was stored at 20°C for later measurements of LH. Frozen blood samples from each patient were examined in one run.
Hormones were measured using commercially available kits as reported previously (Balasch et al., 2001). Estradiol concentrations in serum were estimated by a competitive immunoenzymatic assay (Immuno 1, Technicon; Bayer, USA). The sensitivity was 10 pg/ml and the inter-assay coefficient of variation (CV) was 5%. FSH and LH serum concentrations were measured by an immunoenzymatic assay with two monoclonal antibodies (Immuno 1, Technicon; Bayer) and data expressed in terms of IRP 78/549 and 68/40 respectively. The sensitivity of the assays was 0.1 IU/l for FSH and 0.3 IU/l for LH, and inter-assay CV were 2.7 and 3.1% respectively. The sensitivity of the hormone assays was calculated considering the mean + 2 SD of 20 replicates for samples at hormone concentration zero. The intra-assay CV of the LH assay was 1.35% at an LH of 5.88 IU/l, 1.8% at an LH of 2.98 IU/l, 2.3% at an LH of 1.13 IU/l, 3.10% at an LH of 0.53 IU/l, and 5.85% at an LH of 0.32 IU/l. There was no cross-reactivity of LH with free
subunit. Total
-hCG was measured by a solid-phase, two-site chemiluminescent enzyme immunometric assay standardized against the Third International Standard 75/537 (Immulite, Diagnostic Products Co., USA), with a detection limit of 2 IU/l. The inter-assay CV was 5.8%.
Ultrasonic scans were performed using a 5 mHZ vaginal transducer attached to an Aloka sector scanner (Model SSD-620; Aloka Co., Ltd, Japan).
Statistical methods
For statistical analysis, the Students t-test, the MannWhitney U-test and the Pearson bivariate method were used as appropriate. Results are expressed as mean ± SE. P < 0.05 was considered significant. The discrimination attained between two study groups (cancelled versus non-cancelled cycles, conception versus non-conception cycles, and ongoing pregnancy versus early pregnancy loss groups) was evaluated with ROC analysis (Hanley and McNeil, 1982; Zweig and Campbell, 1993
). ROC curves are plots of all the sensitivity and specificity pairs which are possible for all levels of a particular parameter. They are constructed by plotting the false positive rate or 100 specificity on the x-axis. The y-axis shows the true positive rate or sensitivity. The best threshold value discriminating between two conditions is the value located at the greatest distance from the diagonal.
Calculation of the area under the ROC curve (AUCROC) provides the quantitative measure of accuracy, i.e. the ability of a particular parameter (e.g. LH serum concentrations) to discriminate between two conditions (e.g. ongoing pregnancy versus early pregnancy loss). An ROC curve representing a parameter with no discrimination at all is a 45° diagonal line from the left lower corner (0% true positive rate and 0% false positive rate) to the upper right corner (100% true positive rate and 100% false positive rate) with an area under the curve of 0.5. A parameter with no overlap between the two conditions will discriminate perfectly and has a ROC curve passing along the y-axis to the upper left corner (100% true positive rate and 0% false positive rate) to end again in the upper right corner with an area under the curve of 1.0.
Data were analysed by Statistics Package for Social Sciences (SPSS version 10.0, USA) statistical software.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
There was a total of 82 gestations and 16 (19.5%) of them were early pregnancy losses (six biochemical pregnancies and 10 first-trimester spontaneous abortions) while the remaining 66 (80.5%) were ongoing pregnancies. Therefore, for the evaluation of the results, the following groups of IVF/ICSI outcome were considered: (i) cancelled cycles (n = 32); (ii) non-cancelled cycles (n =214); (iii) no conception (n = 132); (iv) conception (n = 82) which included ongoing pregnancy (n = 66) and biochemical (n = 6) and spontaneous abortions (n = 10); (v) ongoing pregnancy (n = 66); (vi) early pregnancy loss (n = 16) which included biochemical (n = 6) and first trimester spontaneous abortions (n = 10). Mean FSH treatment duration in conception (11.17 ± 0.19 days) versus non-conception (10.95 ± 0.14 days) cycles, and ongoing pregnancies (11.10 ± 0.21 days) versus early pregnancy losses (11.56 ± 0.47 days), was similar.
As shown in Figure 1, LH serum levels throughout ovarian stimulation treatment were similar for cancelled versus non-cancelled cycles, non-conception versus conception cycles, and ongoing pregnancy versus early pregnancy loss groups. In fact, LH values in these six groups of patients clearly overlapped. There was no correlation between LH serum levels on days S0, S5, S9, S11 or S13 in non-cancelled cycles and parameters of ovarian response and assisted reproduction treatment outcome such as days of ovarian stimulation, ampoules of FSH used, number of oocytes, number of metaphase II oocytes, number of embryos, and number of high quality embryos (data not shown). LH serum levels on day S7 were very poorly correlated with those parameters (Figure 2) and the same was true when the AUC for LH serum concentration during gonadotrophin treatment was considered (Figure 3).
|
|
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
According to those studies, evidence of LH deficiency during the follicular phase has been identified in a fraction of the assisted reproduction treatment population ranging between 12% (Humaidan et al., 2002) and 26% (Fleming et al., 2000
) and even 50% (Westergaard et al., 2000
) and 70% (Esposito et al., 2001
). These rather high proportions of assisted reproduction treatment women profoundly suppressed, thus having reduced serum levels of LH somewhat related to poorer reproductive performance, seem clearly disproportionate considering that the switch in stimulation regimens to a more widespread use of FSH-only preparations, without LH supplementation, has been associated with an increased rate of overall programme success (Hull et al., 1994
; FIVNAT, 1999
, 2000
; Cramer et al., 2000
). In fact, in our previous study (Balasch et al., 2001
) we reported that 31, 15 and 7% of assisted reproduction treatment patients had mid-follicular LH (day S7) serum concentration <1,
0.7 and <0.5 IU/l respectively, proportions of supposedly suppressed women much lower than those previously reported (Fleming et al., 2000
; Westergaard et al., 2000
). Corresponding figures were exactly the same in the present study, i.e. 31.3, 15.4 and 6.5% of patients had day S7 serum levels <1,
0.7 and <0.5 IU/l respectively.
The present report confirms and expands our previous study (Balasch et al., 2001) showing that serum LH measurements during ovarian stimulation with recombinant FSH under pituitary suppression, in normogonadotrophic women undergoing assisted reproduction treatment, cannot predict ovarian response, IVF/ICSI outcome, implantation, and the outcome of pregnancy. There may be several reasons for discrepancies with previous studies suggesting that low circulating levels in the follicular phase may have a significant impact on ovarian response and the outcome of assisted reproduction treatment cycles (Fleming et al., 2000
; Westergaard et al., 2000
; Esposito et al., 2001
; Humaidan et al., 2002
; Tesarik and Mendoza, 2002
).
First, it is notable that in those previous studies the threshold values for LH were established arbitrarily or according to the detection limit of the LH assay used and individual studies differ as to the definition of the threshold serum LH level. In contrast, we analysed the diagnostic accuracy of LH concentrations measured throughout ovarian stimulation using ROC curve analysis. The clinical performance of a laboratory test can be described in terms of diagnostic accuracy, or the ability to correctly classify subjects into clinically relevant subgroups. The ROC graph is a plot of all of the sensitivity/specificity pairs resulting from continuously varying the decision threshold over the entire range of results observed. Thus, the ROC plot, representing the fundamental ability of a test to discriminate between two states of health, is an index of pure accuracy (Zweig and Campbell, 1993).
Second, in those previous reports a potent (buserelin) or a long-acting GnRH agonist preparation (even preceded by oral norethisterone administration which further increases pituitary suppression) were used. Long-acting depot preparations produce more profound pituitary desensitization than do the short-acting preparations (Yim et al., 2001). In contrast we use daily doses of s.c. leuprolide acetate (Balasch et al., 2001
) or triptorelin acetate (present investigation). GnRH agonist formulations differ from the native decapeptide by one amino acid at the sixth position, and some also by another amino acid at the tenth position (Conn and Crowley, 1991
). The substitution of an amino acid at these positions is crucial because it increases the drugs resistance to peptidases, prolongs the drugs half-life, and increases its potency. Thus, the relative potency of buserelin is higher than that of triptorelin and leuprolide which are equally potent (Conn and Crowley, 1991
; Cheung et al., 2000
). The most potent agonists have a greater hydrophobicity and increased hydrophobicity may have non-specific effects: (i) decrease the rate of clearance of the compound from the circulation, and (ii) alter the apparent binding constant to membrane-bound receptors by increasing the affinity for the hydrophobic plasma membrane compartment (Nestor et al., 1984
). Thus, structural differences may be responsible for GnRH agonist pharmacological properties and their ability to penetrate into follicular fluid and affect granulosa cell functions. In fact, intact buserelin acetate after intranasal administration has been detected in follicular fluid of women treated with this agonist and gonadotrophins (Loumaye et al., 1989
) at concentrations that may be sufficient to affect granulosa cell function (Parinaud et al., 1988
). On the contrary, it has been shown that concentrations of leuprolide acetate in serum and follicular fluid during adjunctive therapy are within a range that does not affect progesterone accumulation of cultured human granulosa-lutein cells (Dodson et al., 1988
; Frederick et al., 1991
). Accordingly, in-vivo studies have reported that the inhibitory effect on ovarian steroidogenesis and follicular development is more evident with the more potent buserelin than with triptorelin or leuprolide acetate (Parinaud et al., 1992
; Balasch et al., 1992
). Thus, the agonist seems to be the major effect modifier and most of the present controversy in the literature about the value of LH levels in controlled ovarian hyperstimulation for assisted reproduction treatment could be explained on the basis of the type and formulation of the GnRH agonist used rather than the gonadotrophin preparation.
Finally, we use a step-down method of gonadotrophin administration where the highest dose of recombinant FSH is given on stimulation days 1 and 2. This implies that the starting dose of FSH used by us is 2-fold the dose used in those previous studies. This may be important with respect to ovarian paracrine mechanisms. It has been shown that FSH, acting on granulosa cells, activates paracrine mechanisms that up-regulate theca cell response to LH (Hillier, 2000). Thus, it can be postulated that higher doses of FSH used at a critical period of ovarian stimulation during the early follicular phase can overcome too low residual LH concentrations existing in some women once pituitaryovarian suppression has been achieved (Hillier, 2000
).
Our studies have important clinical implications. Thus some authors advocate the routine addition of hormone preparations containing LH activity to ovarian stimulation protocols with GnRH agonists (Filicori et al., 2002) but this is not supported by our studies. On the other hand, two recent reports (Humaidan et al., 2002
; Tesarik and Mendoza, 2002
), one of them in young oocyte donors, have indicated that the inclusion of exogenous LH to the ovarian stimulation protocol can have detrimental effects on ovarian response and assisted reproduction treatment and pregnancy outcomes depending on the level of endogenous LH. Both studies concluded that LH should neither be too high nor too low and support the concept of a window for LH requirement in ovarian stimulation (Hillier, 2000
; Shoham, 2002
). Remarkably, an LH window cannot be detected using ROC analysis.
In conclusion, using daily doses of an appropriate GnRH agonist (leuprolide or triptorelin having lower potency than buserelin) and a step-down regimen of recombinant FSH administration, circulating levels of LH during ovarian stimulation for assisted reproduction treatment have no significant impact on ovarian response and IVF/ICSI outcome, and thus there is little underlying physiological support for the addition of LH in stimulation protocols.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Balasch, J., Jové, I., Moreno, V., Civico, S., Puerto, B. and Vaniell, J.A. (1992) The comparison of two gonadotropin-releasing hormone agonists in an in vitro fertilization program. Fertil. Steril., 58, 991994.[ISI][Medline]
Balasch, J., Vidal, E., Peñarrubia, J., Casamitjana, R., Carmona, F., Creus, M., Fábregues, F. and Vanrell, J.A. (2001) Suppression of LH during ovarian stimulation: analysing threshold values and effects on ovarian response and the outcome of assisted reproduction in down-regulated women stimulated with recombinant FSH. Hum. Reprod., 16, 16361643.
Chappel, S.C. and Howles, C. (1991) Reevaluation of the roles of luteinizing hormone and follicle-stimulating hormone in the ovulatory process. Hum. Reprod., 6, 12061212.[Abstract]
Cheung, T., Lo, K.W., Lam, C.W., Lau, W. and Lam, P.K. (2000) A crossover study of triptorelin and leuprolide acetate. Fertil. Steril., 74, 299305.[CrossRef][ISI][Medline]
Conn, P.M. and Crowley, W.F. (1991) Gonadotropin-releasing hormone and its analogues. N. Engl. J. Med., 324, 93103.[ISI][Medline]
Cramer, D.W., Liberman, R.F., Powers, D., Hornstein, M.D., Mcshane, P. and Barbieri, R.L. (2000) Recent trends in assisted reproductive techniques and associated outcomes. Obstet. Gynecol., 95, 6166.
Dodson, W.C., Myers, T., Morton, P.C. and Conn, P.M. (1988) Leuprolide acetate: serum and follicular fluid concentrations and effects on human fertilization, embryo growth, and granulosa-lutein cell progesterone accumulation in vitro. Fertil. Steril., 50, 612617.[ISI][Medline]
Esposito, M.A., Barhnart, K.T., Coutifaris, C. and Patrizio, P. (2001) Role of periovulatory luteinizing hormone concentrations during assisted reproductive technology cycles stimulated exclusively with recombinant follicle-stimulating hormone. Fertil. Steril., 75, 519524.[CrossRef][ISI][Medline]
Filicori, M., Cognigni, G.E., Samara, A., Melappioni, S., Perri, T., Cantelli, B., Parmegiani, L., Pelusi, G. and DeAloysio, D. (2002) The use of LH activity to drive folliculogenesis: exploring uncharted territories in ovulation induction. Hum. Reprod. Update, 8, 543557.
FIVNAT (1999) Dossier FIV-NAT-99. Bilan de lannée 98.
FIVNAT (2000) Dossier FIV-NAT-2000. Bilan de lannée 99.
Fleming, R., Rehka, P., Deshpande, N., Jamieson, M.E., Yates, R.W.S. and Lyall, H. (2000) Suppression of LH during ovarian stimulation: effects differ in cycles stimulated with purified urinary FSH and recombinant FSH. Hum. Reprod., 15, 14401445.
Frederick, J.L., Hickey, M.J., Francis, M.M., Sauer, M.V. and Paulson, R.J. (1991) The effect of leuprolide acetate on steroidogenesis by granulosa and theca cells in vitro. J. In Vitro Fertil. Embryo Transfer, 8, 230234.[ISI][Medline]
Hanley, J.A. and McNeil, B.J. (1982) The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology, 143, 2936.[Abstract]
Hillier, S.G. (2000) Controlled ovarian stimulation in women. J. Reprod. Fertil., 120, 201210.
Hull, M.G.R., Armatage, R.J. and McDermott, A. (1994) Use of follicle-stimulating hormone alone (urofollitropin) to stimulate the ovaries for assisted conception after pituitary desensitization. Fertil. Steril., 62, 9971003.[ISI][Medline]
Humaidan, P., Bungum, L., Bungum, M. and Yding Andersen, C. (2002) Ovarian response and pregnancy outcome related to mid-follicular LH levels in women undergoing assisted reproduction with GnRH agonist down-regulation and recombinant FSH stimulation. Hum. Reprod., 17, 20162021.
Lam, P.M., Cheung, L.P., Choy, C.M.Y., Lau, Y.P. and Haines, C. (2002) Effects of the intensity of downregulation on outcome of in vitro fertilization and embryo transfer. Gynecol. Endocrinol., 16, 143150.[ISI][Medline]
Lévy, D.P., Navarro, J.M., Schattman, G.L., Davis, O.K. and Rosenwaks, Z. (2000) The role of LH in ovarian stimulation. Exogenous LH: lets design the future. Hum. Reprod., 15, 22582265.
Loumaye, E., Coen, G., Pampfer, S., Vankrieken, L. and Thomas, K. (1989) Use of a gonadotropin-releasing hormone agonist during ovarian stimulation leads to significant concentrations of peptide in follicular fluids. Fertil. Steril., 52, 256263.[ISI][Medline]
Loumaye, E., Engrand, P., Howles, C.M. and ODea, L. (1997) Assessment of the role of serum luteinizing hormone and estradiol response to follicle-stimulating hormone on in vitro fertilization outcome. Fertil. Steril., 67, 889899.[CrossRef][ISI][Medline]
Nestor, J.J., Ho, T.L., Tahilramani, R., McRae, G.I. and Vickery, B.H. (1984) Long acting LHRH agonists and antagonists. In Labrie, F., Bélanger, A., Dupont, A. (eds), LHRH and its Analogues: Basic and Clinical Aspects. Excerpta Medica, Amsterdam, pp. 2435.
Parinaud, J., Beaur, A., Burreau, E., Vieitez, G. and Pontonnier, G. (1988) Effect of a luteinizing hormone-releasing hormone agonist (buserelin) on steroidogenesis of cultured human preovulatory granulosa cells. Fertil. Steril., 50, 597602.[ISI][Medline]
Parinaud, J., Oustry, P., Perineau, M., Rème, J.M., Monroziès, X and Pontonnier, G. (1992) Randomized trial of three luteinizing hormone-releasing hormone analogues used for ovarian stimulation in an in vitro fertilization program. Fertil. Steril., 57, 12651268.[ISI][Medline]
Shoham, Z. (2002) The clinical therapeutic window for luteinizing hormone in controlled ovarian stimulation. Fertil. Steril., 77, 11701177.[CrossRef][ISI][Medline]
Tesarik, J. and Mendoza, C. (2002) Effects of exogenous LH administration during ovarian stimulation of pituitary down-regulated yound oocyte donors on oocyte yield and developmental competence. Hum. Reprod., 17, 31293137.
Veeck, L.L. (1999) An Atlas of Human Gametes and Conceptuses. Parthenon, New York.
Westergaard, L.G., Laursen, S.B. and Andersen, C.Y. (2000) Increased risk of early pregnancy loss by profound suppression of luteinizing hormone during ovarian stimulation in normogonadotropic women undergoind assisted reproduction. Hum. Reprod., 15, 10031008.
Yim, S.F., Lok, I.H., Cheung, L.P., Briton-Jones, C.M., Chiu, T.T.Y. and Haines, C.J. (2001) Dose-finding study for the use of long-acting gonadotrophin-releasing hormone analogues prior to ovarian stimulation for IVF. Hum. Reprod., 16, 492494.
Zweig, M.H. and Campbell, G. (1993) Receiver-operating characterisitc (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin. Chem., 39, 561577.
Submitted on June 27, 2003; accepted on September 11, 2003.