Antral follicle count, anti-mullerian hormone and inhibin B: predictors of ovarian response in assisted reproductive technology?

Transcription

1 BJOG: an International Journal of Obstetrics and Gynaecology October 2005, Vol. 112, pp DOI: /j x Antral follicle count, anti-mullerian hormone and inhibin B: predictors of ovarian response in assisted reproductive technology? S. Muttukrishna, a H. McGarrigle, a R. Wakim, b I. Khadum, b D.M. Ranieri, b P. Serhal b Objective The objective of this study was to evaluate the relationship between anti-mullerian hormone (AMH), inhibin B and antral follicle count (AFC) with ovarian response. Design Retrospective study. Setting Fertility unit. Sample AFC was recorded, and a serum sample obtained on day 3 from all patients undergoing in vitro fertilisation (IVF). Patients were given 300 IU/L recombinant follicle stimulating hormone (FSH; Gonal F). The following day blood samples were collected. Methods Serum samples were assayed for FSH, AMH and inhibin B using commercial immunoassay kits and oestradiol using an in house assay. Main outcome measures Response to gonadotrophin stimulation and the number of eggs collected. Results AFC was negatively correlated to age (r ¼ 0.426, P < 0.001). Delta inhibin B (levels of inhibin B on day 4 minus day 3) had the best association to the number of eggs collected (r ¼ 0.533, P < 0.001) followed by basal AMH (r ¼ 0.51, P < 0.001) and AFC (r ¼ 0.505, P < 0.001). The number of eggs fertilised was significantly associated with basal AMH (r ¼ 0.592, P < 0.001) and inhibin B (r ¼ 0.548, P < 0.001). AMH with a cutoff of 0.2 ng/ml had the best sensitivity (87%) and specificity (64%) in predicting poor response. A cumulative score using basal FSH, basal AMH, delta E 2 (levels of oestradiol on day 4 minus day 3), delta inhibin B, AFC and age gives the best predictive statistics to identify poor responders with 87% sensitivity and 80% specificity and a positive likelihood ratio of Conclusion Delta inhibin B had the best positive association with the number of eggs collected and basal AMH is the single best predictor of poor response. AFC has a significant association with the number of eggs collected and is predictive of clinical pregnancy. It is evident that a single parameter is of limited value in predicting ovarian response. However, we have demonstrated a cumulative score using all the above markers could be useful in predicting poor response. INTRODUCTION The pool of primordial follicles in the ovary is related to the number of growing antral follicles. Antral follicles are responsive to gonadotrophin stimulation and the measure of ovarian reserve can be defined as the total number of follicles, which can be stimulated to grow under maximal stimulation. There is a gradual decline in the number of primordial follicles with increasing age. Classically, age 1 4 and follicle stimulating hormone (FSH) levels in the early follicular phase 5 8 have been used as markers of a Department of Obstetrics and Gynaecology, University College London, UK b The Assisted Conception Unit, University College Hospital, University College London, UK Correspondence: S. Muttukrishna, Department of Obstetrics and Gynaecology, University College London, Chenies Mews, London WC1E 6HX, UK. D RCOG 2005 BJOG: an International Journal of Obstetrics and Gynaecology ovarian reserve. Dynamic tests such as the clomiphene citrate challenge test (CCCT), GnRH stimulation test (G-test) and ovarian stimulation test (OST) have been used by several groups to predict the ovarian response to gonadotrophin stimulation The principle of these tests is to stimulate ovarian hormone (oestradiol) secretion by increasing the gonadotrophins in circulation. These tests give an indication of ovarian function directly compared with the measurement of FSH, which is an indirect measure of ovarian hormone feedback on the pituitary. 9,14 18 More recently, the ovarian hormones, inhibin B and anti-mullerian hormone (AMH), 19 have been used by various groups to assess the ovarian reserve. AMH and inhibin B are members of TGF-h family of proteins. AMH is a homodimer of 140 kda molecular weight. Inhibin B is a heterodimer of alpha and beta B subunits. In women, the granulosa cells of the ovarian follicles produce AMH and inhibin B. AMH production by the granulosa cells is not controlled by gonadotrophins, 19,20 while inhibin B secretion by granulosa cells is stimulated by gonadotrophins. 21 Although the cyclic pattern of AMH is not known in

2 women, AMH levels in the follicular phase do not vary. Preliminary reports indicate that AMH levels decline with increasing female age 22 and AMH is associated with ovarian response in IVF patients with normal FSH levels. 23 When AMH measurement was added to inhibin B and FSH levels, the prediction of ovarian response was improved. 19 More recently antral follicle counts (AFCs) has also been reported to be useful in predicting ovarian reserve. 24 Assessment of ovarian reserve in women undergoing assisted reproduction is useful in optimising the treatment protocol and in counselling the patients. Therefore, availability of a reliable measure of ovarian reserve is essential for this purpose. While several groups have studied the use of various static and dynamic markers to assess ovarian reserve, no single parameter gives a satisfactory prediction. The objective of this study was to investigate the efficacy of AFC, AMH and inhibin B in predicting ovarian response in patients having gonadotrophin ovarian stimulation test (GOST) for the assessment of the ovary. MATERIALS AND METHODS In this study, blood samples collected from 108 patients who had OST for the assessment of their ovarian reserve were assayed. Basal ultrasound scan for AFC was performed on day 3 of the menstrual cycle, followed by a blood test for the assessment of serum FSH, oestradiol (E 2 ), inhibin B and AMH. Subsequently, 300 IU/L of recombinant FSH (Gonal F; Serono, Welland Garden City, Hertfordshire, UK) was administered. A second blood sample was taken on day 4 for the assessment of E 2 and inhibin B. The rise in E 2 after FSH stimulation (delta E 2 : day 4 E 2 minus day 3 E 2 ) and the rise in inhibin B after FSH stimulation (delta inhibin B: day 4 inhibin B minus day 3 inhibin B) were calculated. AFC, number of ampoules of gonadotrophins used for stimulation, the number of eggs collected and the number of eggs fertilised in the context of a subsequent IVF cycle were other outcomes used for analysis. Among the patients who completed an IVF cycle the causes of infertility were age (n ¼ 7), endometriosis (n ¼ 3), male factor (n ¼ 26), polycystic ovary (PCO) (n ¼ 8), tubal (n ¼ 3) and unexplained (n ¼ 22). The administration of buserelin nasal spray (Suprefact; Hoechst,Hounslow,Middlesex,UK)100Ag into each nostril 6 hourly (800 Ag/day) was started in the mid luteal phase of the same cycle of the GOST. Down-regulation was confirmed by ultrasound scan on day 3 of the following cycle and ovarian stimulation was immediately commenced with an initial daily dose of hmg (Menogon; Ferring Pharmaceuticals Ltd, Langley, Berkshire, UK) which varied according to the patients age (<30 years ¼ 150 IU, years ¼ 225 IU, years ¼ 300 IU and >38 ¼ 375 IU). Follicular growth was monitored by transvaginal scans beginning on day 6 of stimulation. The dose of gonadotrophin AFC, AMH AND INHIBIN B AS PREDICTORS OF OVARIAN RESPONSE 1385 was adjusted according to the ovarian response. Between days 12 and 14, patients who had at least two leading follicles of 17 mm received 10,000 IU of hcg (Choragon; Ferring Pharmaceuticals Ltd.). Transvaginal ultrasound controlled oocyte retrieval was performed 37 hours later. Patients who had 3 follicles of 15 mm in diameter on the day of the hcg administration and oestradiol levels <2000 pmol/l were considered to have an inadequate ovarian response for IVF and their cycles were cancelled. Fertilisation was confirmed by the presence of two pronuclei hours after insemination. Embryo transfer was performed three days after oocyte retrieval. Two embryos were routinely transferred. The patients were grouped into poor responders (4 eggs collected) and normal responders (5 eggs collected) for analysis. HORMONE ASSAYS AMH: All samples were assayed in duplicate using a commercial assay kit according to the manufacturer s sensitised assay protocol (Immunotech, Marseille, France). The sensitivity of the assay was 0.09 ng/ml. The intra- and interassay variations were <15%. Inhibin B: All samples were assayed in duplicate using a commercial ELISA kit (Oxford bio innovations-dsl, Oxford, UK) according to the manufactures protocol. The sensitivity of the assay is 10 pg/ml. The intra- and inter-assay variations were <10%. FSH: All samples were assayed in duplicate using a commercial Coat a count solid phase immuno-radiometric assay (IRMA) (DPC, Gwynedd, UK). The sensitivity of the assay is 0.06 miu/ml and the intra- and inter-assay variation is <7%. Oestradiol was measured using an in-house radioimmunoassay (following ether extraction using a tritium Fig. 1. Relationship between AFC and the number of Eggs collected. Scatterplot of the number of antral follicles and the log-transformed number of eggs collected multiplied by a constant (1000) divided by the total number of ampoules (1 ampoule ¼ 75 IU gonadotrophin) used for stimulation. Linear regression line, the regression coefficient and the P value are shown.

3 1386 S. MUTTUKRISHNA ET AL. Table 1. Antral follicle count and pregnancy rate. No. of AFC Pregnancy rate (%) 1 4 (n ¼ 11) (n ¼ 31) (n ¼ 27) 48 The percentage of patients who succeeded in a clinical pregnancy after completing assisted reproductive technology. Based on the antral follicle count patients were put into three groups. Group 1 (1 4 AFC), group 2 (5 9 AFC) and group 3 (10 or more AFC); n ¼ total number of patients who had treatment in each group. One-way analysis of variance (ANOVA) showed a significant difference in pregnancy rate between the groups (P < 0.05). labeled tracer) as previously described. 25 The sensitivity of the assay was <80 pmol/l. The intra- and inter-assay variation was <10%. STATISTICAL ANALYSIS Regression analysis was carried out to study the relationship between the measured parameters and the number of eggs collected. In order to normalise the number of gonadotrophin units used in different patients, the number of eggs collected was multiplied by a constant (1000) and then divided by the total ampoules of gonadotrophins used for the IVF stimulation before statistical analysis. Based on the significance of the linear regression analysis a scoring system was adopted for basal FSH (10 miu/ ml ¼ 2, >10 miu/ml ¼ 1), delta E 2 (100 pmol/l ¼ 1, pmol/l ¼ 2, 300 pmol/l ¼ 3), delta inhibin B (0 40 pg/ml ¼ 1, pg/ml ¼ 2, >100 pg/ml ¼ 3), basal AMH ( ng/ml ¼ 1, ng/ml ¼ 2, >1 ng/ ml ¼ 3), AFC (0 4 ¼ 1, 5 10 ¼ 2, >10 ¼ 3) and age (<30 years ¼ 4, years ¼ 3, years ¼ 2, >40 years ¼ 1). A cumulative score was obtained by adding all individual scores for the above six parameters. Diagnostic tests were performed by calculating sensitivity and specificity using the area under the receiver operating characteristic (ROC curve) using the individual and cumulative scoring system for the prediction of poor response. ROC analysis is used to quantify the accuracy of the diagnostic tests. The area under the ROC curve commonly summarises the global performance of a diagnostic test. This area is interpreted as the probability that the result of a diagnostic test would be correct. Therefore, the greater Fig. 2. Relationship between biochemical parameters and the number of eggs collected. Scatterplot of individual concentration of (a) basal FSH, (b) delta E 2, (c) delta inhibin B and (d) basal AMH and the log transformed number of eggs collected multiplied by a constant (1000) divided by the total number of ampoules (1 ampoule ¼ 75 IU gonadotrophin) used for stimulation. Linear regression line, the regression coefficient and the P value are shown.

4 AFC, AMH AND INHIBIN B AS PREDICTORS OF OVARIAN RESPONSE 1387 Table 2. Predictive statistics for poor response. Sensitivity (%) Specificity (%) Percentage of correctly classified Positive likelihood ratio Basal FSH cutoff ¼ 10 miu/ml) Delta E 2 (cutoff ¼ 100 pmol/l) Delta inhibin B (cutoff ¼ 40 pg/ml) Basal AMH (cutoff ¼ 0.2 ng/ml) AFC (cutoff ¼ 5) Cumulative score (cutoff ¼ 12) Sensitivity and specificity for the prediction of poor response to ovarian stimulation. Cutoff values, sensitivity, specificity, percentage of correctly classified and the positive likelihood ratios are given. Scoring was based on the following: basal FSH (10 miu/ml ¼ 2, >10 miu/ml ¼ 1), delta E 2 (100 pmol/l ¼ 1, pmol/l ¼ 2, 300 pmol/l ¼ 3), delta inhibin B (0 40 pg/ml ¼ 1, pg/ml ¼ 2, >100 pg/ml ¼ 3), basal AMH ( ng/ml ¼ 1, ng/ml ¼ 2, >1 ng/ml ¼ 3), AFC (0 4 ¼ 1, 5 10 ¼ 2, >10 ¼ 3) and age (<30 years ¼ 4, years ¼ 3, years ¼ 2, >40 years ¼ 1). A cumulative score was obtained by adding all individual scores for the above six parameters. Sensitivity is the fraction of positive cases that are correctly classified by the diagnostic test; in this study it is the poor responders. Specificity is the fraction of negative cases that are correctly classified; in this study it is the percentage of good responders. the area under the ROC curve the better the diagnostic test will be. STATA statistical software (Stata Corp., Texas, USA) was used for statistical analysis. RESULTS In this study GOST was carried out in 108 patients to assess their ovarian reserve before IVF treatment. Sixty-nine patients had a complete IVF cycle while, in eight patients, the treatment was converted to intrauterine insemination (IUI) (patients with less than three follicles <15 mm diameter on the day of hcg with normal sperms and patent tubes). A total of 31 patients decided not to proceed with the treatment. Linear regression analysis showed that AFC had a significant linear association with the number of eggs collected (r ¼ 0.505, P < 0.001; Fig. 1). AFC was also associated with pregnancy rates. Patients who had a complete IVF cycle were divided in to three groups based on their AFC. The percentage of patients who had clinical pregnancies in each group is shown in Table 1. There was a significant difference in pregnancy rate between the three groups of AFC (ANOVA, P < 0.05). AFCs were significantly correlated with delta E 2 (r ¼ 0.306, P ¼ 0.01), delta inhibin B (r ¼ 0.4, P ¼ 0.001) and basal AMH (r ¼ 0.49, P < 0.001). Basal FSH on day 3 (r ¼ 0.318, P ¼ 0.001), delta E 2 (r ¼ 0.453, P < 0.001), delta inhibin B (r ¼ 0.53, P < 0.001) and basal AMH on day 3 (r ¼ 0.509, P < 0.001) had the best significant association with the number of eggs collected (Fig. 2). AFC on day 3 has a significant negative association with day 3 FSH (r ¼ 0.33, P ¼ 0.001) and age ( 0.426, P < 0.001). Basal inhibin B (0.29, P < 0.01) and AMH (0.487, P < 0.001) on day 3 has a significant positive correlation with AFC. Sensitivity is the fraction of positive cases that are correctly classified by the diagnostic test; in this study it is the poor responders. Specificity is the fraction of negative cases that are correctly classified; in this study it is the percentage of good responders. Various cutoff points were analysed using ROC curves and the cutoff points that gave the best sensitivity, specificity and percentage of correctly classified were used for analysis and the scoring system. Basal FSH levels <10 miu/ml had a sensitivity of 87% and specificity of 60% in predicting a poor response. Only 76% of the population was correctly classified using this parameter. Delta E 2 with a cutoff of 100 pmol/l had a sensitivity of 90% and specificity of 49% in predicting poor response. Delta inhibin B with a cutoff of 40 pg/ml has a sensitivity of 87% and specificity of 49%. Basal AMH with a cutoff of 0.2 ng/ml has a sensitivity of 87% and specificity of 64% in predicting poor response. AFC with a Fig. 3. ROC for the cumulative score model. Receiver operating curve for the identification of cutoff score for the cumulative score model to predict poor response to ovarian stimulation. The area under the curve was ROC analysis is used to quantify the accuracy of the diagnostic tests. The area under the ROC curve commonly summarises the global performance of a diagnostic test. This area is interpreted as the probability that the result of a diagnostic test would be correct.

5 1388 S. MUTTUKRISHNA ET AL. cutoff of five follicles had a sensitivity of 89% and specificity of 39% in predicting poor response (Table 2). A cumulative score cutoff of 12 gives 87% sensitivity, 80% specificity in predicting poor response. The area under the ROC curve for the cumulative score is 0.91 (Fig. 3). Although a cumulative score 9 is 100% predictive of a poor response and a score 15 is 100% predictive of a good response. However, it could only correctly classify 76% and 66% of the population, respectively. That is all patients with 9 have a poor response but, 24% of patients with a score >9 also have a poor response while all patients with a score 15 have a good response, whereas 34% with a score <15 also have a good response. DISCUSSION Diminished ovarian reserve and maternal age have a direct and profound impact on the success of assisted reproductive technology (ART). The success of in vitro fertilisation (IVF) and other ARTs is critically dependent on optimal ovarian stimulation protocols that provide adequate numbers of good quality oocytes and embryos. Therefore, it is important to individualise stimulation protocols according to age, ovarian reserve and endocrine status. Over the years, several groups have investigated the use of various basal markers such as FSH, age, inhibin B, oestradiol and dynamic tests such as CCCT, OST and G-test. 13,26 Some have reported that inhibin B is useful in predicting ovarian response, 10,11,14,15,27,28 whereas others have reported inhibin B does not add value to FSH, age or oestradiol. 9,17,29 In the menstrual cycle, inhibin B levels progressively rise and are at peak concentration around days 5 6 of the cycle. 30 The levels gradually fall and there is a sharp rise postovulation followed by a steep decline to undetectable levels in the luteal phase suggesting that inhibin B is from the small antral follicles and that the levels rise in the early follicular phase as these follicles grow in the presence of rising FSH. Then as the dominant follicle emerges and the small follicles undergo atresia inhibin B levels fall in late follicular phase while inhibin A levels rise. Almost all groups that have investigated basal inhibin B levels to assess ovarian reserve have looked at samples collected on days 2 3 of the cycle. At this time of the cycle, the levels of inhibin B are lower and are measured on the flat part of the standard curve making the results unreliable with high variability. This probably is the reason for the discordant results found in the basal inhibin B measurements. However, some investigators have looked at the increase in inhibin B to a dynamic test 10,11,31 and found that the increase in inhibin B to ovarian stimulation is useful in predicting ovarian response. More recently, some groups have reported the use of AFC 24,31 and AMH 19,23 in predicting the ovarian reserve. They conclude that AMH is a novel marker of ovarian reserve and AFC has the best predictive value. We have shown in this study that AFC is significantly correlated with the number of eggs collected and is indicative of possible pregnancy rate in patients who undergo ART. However, AFC is subjective and also might vary between cycles in patients. AMH levels are reported to be similar at different stages of the cycle 32 although a complete AMH profile throughout the menstrual cycle has not yet been published. In this study, for the first time we have evaluated the use of AMH in the GOST and also assessed the value of AFC, inhibin B, AMH, FSH and E 2 in combination to predict ovarian response. Linear regression analysis shows a significant association between AFC, FSH, delta E 2, delta inhibin B and AMH and the number of eggs collected. Among these parameters delta inhibin B had the best association (r ¼ 0.53, P < 0.001) followed by AMH (r ¼ 0.509, P < 0.001) and AFC (r ¼ 0.505, P < 0.001). Combination of AFC, basal FSH, delta E 2, delta inhibin B and basal AMH had a better association (r ¼ 0.647, P < 0.001) with the number of eggs collected. Addition of age to these combined markers did not have a marked effect on the association (r ¼ 0.652, P < 0.001). Delta inhibin B (r ¼ 0.548, P < 0.001) and basal AMH (r ¼ 0.592, P < 0.001) had the best association with the number of eggs fertilised among the parameters measured suggesting a relationship between inhibin B, AMH and the quality of oocytes. A combination of basal FSH, delta E 2, delta inhibin B, basal AMH, AFC and age had a modestly greater association with the number of eggs fertilised (r ¼ 0.684, P < 0.001). AFC was significantly associated with basal AMH levels (r ¼ 0.49, P < 0.001) suggesting that as a single biochemical marker, AMH is best associated with AFC on day 3 than basal inhibin B or basal E 2. Age is also negatively correlated to AFC (r ¼ 0.426, P < 0.001), number of eggs collected (r ¼ 0.34, P ¼ 0.001) and number of eggs fertilised (r ¼ 0.33, P ¼ 0.001). Diagnostic tests were carried out to assess the predictive value of each parameter independently and using a cumulative score to predict ovarian response. A good marker should have a high sensitivity (in identifying the true poor responders) and high specificity (in identifying the true good responders) to precisely identify the good and poor responders. Interestingly, all parameters independently had a sensitivity >85% in predicting poor response. However, the specificity was variable. We have chosen a cutoff that provides higher sensitivity rather than higher specificity to identify the true poor responders. Among all the markers delta oestradiol has the highest specificity of 90% (Table 2) suggesting that it could be used for determining the stimulation protocol. However, AMH was the best single marker in predicting response, identifying 87% of the poor responders (sensitivity of 87%) and 64% of the good responders (64% specificity). We devised a cumulative scoring system using age, AFC, basal FSH, basal AMH, delta E 2 and delta inhibin B to use all these markers in combination to predict ovarian response from GOST. A

6 AFC, AMH AND INHIBIN B AS PREDICTORS OF OVARIAN RESPONSE 1389 cumulative score of 12 can correctly identify 87% of poor responders and 80% of the good responders. This cutoff of 12 was obtained using the ROC curve for the cumulative score (Fig. 3). It is the best cutoff score to use as it correctly classifies 84%. However, caution should be used in interpreting these data, as complete IVF treatment was not carried out on all patients in the population studied. Ideally, if all patients were stimulated for IVF irrespective of the GOST results the resulting data would not be biased. Although the hormone measurements and the AFC were done at the time of GOST, the regression analysis to evaluate the relationship between the measured parameters and number of eggs collected or number of eggs fertilised was carried out only in the population that underwent a complete IVF cycle. Diagnostic tests were also carried out only in these patients. However, this is probably the best population available for such studies as the cost and nature of the treatment does not allow the whole test population to complete a treatment cycle. In conclusion, this study shows that AFC has a significant association with number of eggs collected and is also associated with the likelihood of clinical pregnancy. A single parameter is of limited value in predicting ovarian response to gonadotrophin stimulation in IVF patients. We have shown a panel of markers such as age, AFC, basal FSH and AMH, delta E 2 and inhibin B that could be useful in predicting ovarian response and optimising the gonadotrophin stimulation protocol for each patient. References 1. Hughes EG, King C, Wood EC. A prospective study of prognostic factors in in vitro fertilization and embryo transfer. Fertil Steril 1989; 51: Meldrum DR. Female reproductive aging ovarian and uterine factors. Fertil Steril 1993;59: Navot D, Drews MR, Berg PA, et al. Age related decline in female fertility is not due to diminished capacity of the uterus to sustain embryo implantation. Fertile Steril 1994;61: Scott RT, Opsahl MS, Leonardi MR, Neall GS, Illions EH, Navot D. Life table analysis of pregnancy rates in a general infertility population relative to ovarian reserve and patient age. Hum Reprod 1995;10: Pearlstone AC, Fournet N, Gambone JC, Pang SC, Buyalos RP. Ovulation induction in women age 40 and older: the importance of basal follicle stimulating hormone level and chronological age. Fertil Steril 1992;58: Toner JP. The significance of elevated FSH for reproductive function. Baillieres Clin Obstet Gynaecol 1993;7: Cahill DJ, Prosser CJ, Wardle PG, Ford WCL, Hull MGR. Relative influence of serum follicle stimulating hormone, age, and other factors on ovarian response to gonadotrophin stimulation. Br J Obstet Gynaecol 1994;101: Hansen LM, Batzer FR, Gutmann JN, Corson SL, Kelly MP, Gocial B. Evaluating ovarian reserve: follicle stimulating hormone and oestradiol variability during cycle days 2 5. Hum Reprod 1996;11: Corson SL, Gutmann J, Batzer FR, et al. Inhibin B as a test of ovarian reserve for infertile women. Hum Reprod 1999;14: Kwee J, Elting MW, Schats R, Bezemer PD, Lambalk CB, Schoemaker J. Comparison of endocrine tests with respect to their predictive value on the outcome of ovarian hyperstimulation in IVF treatment: results of a prospective randomised study. Hum Reprod 2003;18: Dzik A, Lambert-Messerlian G, Izzo VM, Soares JB, Pinotti JA, Seifer DB. Inhibin B response to EFFORT is associated with the outcome of oocyte retrieval in the subsequent in vitro fertilization cycle. Fertil Steril 2000;74: Buckman A, Heineman MJ. Ovarian reserve testing and the use of prognostic models in patients with sub fertility. Hum Reprod Update 2001;7: Ranieri DM, Phophong P, Khadum I, Meo F, Davis C, Serhal P. Simultaneous evaluation of basal FSH and oestradiol response to GnRH analogue (F G-test) allows effective drug regimen selection for IVF. Hum Reprod 2001;16: Danforth DR, Arbogast LK, Mroueh J, et al. Dimeric inhibin: a direct marker of ovarian aging. Fertil Steril 1998;70: Seifer DB, Scott RT, Bergh PA, et al. Women with declining ovarian reserve may demonstrate a decrease in day 3 serum inhibin B before a rise in day 3 follicle stimulating hormone. Fertil Steril 1999;72: Welt CK, McNicholl DJ, Taylor AE, Hall JE. Female reproductive aging is marked by decreased secretion of dimeric inhibin. J Clin Endocrinol Metab 1999;84: Creus M, Penarrubia J, Fabregues F, et al. Day 3 serum inhibin B and FSH and age as predictors of assisted reproduction treatment outcome. Hum Reprod 2000;15: Millot F, Antoine JM, Merviel P, Mathieu E, Carpeau J, Uzan S. Comparison of predictive values of inhibins A and B, and plasma estradiol in IVF patients treated with GnRH agonists and recombinant FSH. Gynecol Obstet Fertil 2002;30: Van Rooij IAJ, Broekmans FJM, te Velde ER, et al. Serum anti mullerian hormone levels: a novel measure of ovarian reserve. Hum Reprod 2002;17: Vontilainen R, Miller WL. Human Mullerian inhibitory factor messenger ribonucleic acid is hormonally regulated in the fetal testis and in adult granulose cells. Mol Endocrinol 1987;1: Muttukrishna S, Groome NP, Ledger WL. Gonadotrophic control of secretion of dimeric inhibins and activin A by human granulosa-luteal cells in vitro. J Assist Reprod Genet 1997;14: De Vet A, Laven JS, de Jong FH, Themmen AP, Fauser BC. Anti mullerian hormone serum levels: a putative marker for ovarian aging. Fertil Steril 2002;77: Seifer DB, MacLaughlin DT, Christian BP, Feng B, Shelden RM. Early follicular serum mullerian-inhibiting substance levels are associated with ovarian response during assisted reproductive technology cycle. Fertil Steril 2002;77: Banci L, Broekmans F, Eijkemans M, de Jong F, Habbema D, Velde E. Predictors in poor ovarian response in in vitro fertilisation: a prospective study comparing basal markers of ovarian reserve. Fertil Steril 2002;77: Darne FJ, McGarrigle H, Lachelin GCL. Diurnal variation of plasma and saliva oestrogen, progesterone, cortisol and plasma dehydroepiandrosterone sulphate in late pregnancy. Eur J Obs Gyn Reprod Biol 1989;32: Kligman I, Rosenwaks Z. Differentiating clinical profiles: predicting good responders, poor responders and hyper responders. Fertil Steril 2001;76: Ravhon A, Lavery R, Michael S, et al. Dynamic assays of inhibin B and oestradiol following buserelin acetate administration as predictors of ovarian response in IVF. Hum Reprod 2000;15: Tinkanen H, Blauer M, Laippala P, Tuohimaa P, Kujansuu E. Correlation between serum inhibin B and other indicators of ovarian function. Eur J Obstet Gynecol Reprod Biol 2001;94: Hall JE, Welt CK, Cramer DW. Inhibin A and inhibin B reflect

7 1390 S. MUTTUKRISHNA ET AL. ovarian function in assisted reproduction but are less useful at predicting outcome. Hum Reprod 1999;14: Groome NP, Illingworth PJ, O Brien M, et al. Measurement of dimeric inhibin B throughout the human menstrual cycle. J Clin Endocrinol Metab 1996;81: Yong PY, Baird DT, Thong KJ, McNeilly AS, Anderson RA. Prospective analysis of the relationships between the ovarian follicle cohort and basal FSH concentration, the inhibin response to exogenous FSH and ovarian follicle number at different stages of the normal menstrual cycle and after pituitary down-regulation. Hum Reprod 2003;18: Cook CL, Siow Y, Taylor S, Fallat ME. Serum Mullerian inhibiting substance levels during normal menstrual cycles. Fertil Steril 2000; 73: Accepted 11 February 2005