AMH concentration is not related to effective time to pregnancy in women who conceive naturally

Reproductive BioMedicine Online (2014) 28, 216– 224

www.sciencedirect.com www.rbmonline.com

ARTICLE

AMH concentration is not related to effective time to pregnancy in women who conceive naturally ´line Paccolat a, Isabelle Streuli a,*, Jacques de Mouzon b, Ce Charles Chapron a,c, Patrick Petignat a, Olivier P Irion a, Dominique de Ziegler b a Department of Obstetrics and Gynaecology, University Hospitals of Geneva and the Faculty of Medicine of the ´ Paris Descartes, Sorbonne Paris Cite ´ – Assistance Publique Ho ˆpitaux Geneva University, Geneva, Switzerland; b Universite de Paris, CHU Cochin, Department of Obstetrics Gynaecology and Reproductive Medicine, Paris, France; c Cochin institute, CNRS UMR 8104, INSERM U1016, Paris, France

* Corresponding author. E-mail address: [email protected] (I Streuli). Dr Isabelle Streuli was trained as a specialist in obstetrics and gynaecology at the University Hospitals of Geneva, Switzerland. She then specialized in the field of reproductive medicine and gynaecological endocrinology at the University Hospitals of Geneva and the Cochin University Hospitals in Paris, France. She also completed a postgraduate MSc at the Paris Descartes University. She is now head of the unit for reproductive medicine and gynaecological endocrinology at the maternity of the Geneva University Hospitals. She has been concentrating her clinical research on the theme of ovarian reserve and endometriosis.

Abstract This study determined whether anti-Mu ¨llerian hormone (AMH) concentration influences the time necessary to conceive a

live-born child – effective time to pregnancy (eTTP) – in a population of women who conceived naturally. This is an observational study of 87 women with a planned spontaneous pregnancy resulting in a live birth. eTTP was assessed retrospectively by a questionnaire and AMH was measured in a frozen serum sample from first trimester of pregnancy. eTTP was correlated with age (r = 0.24, P = 0.02), but not with AMH (r = 0.10) or body mass index (r = 0.05). With logistic regressions, the only variable that affected the probability of pregnancy within 3 or 6 months was age, irrespective of whether an AMH concentration limit of 1.0 ng/ml or 2.0 ng/ml was chosen. In conclusion, this study suggests that there is no relationship between AMH concentration and eTTP and therefore speaks against determining AMH in women who are not infertile for the purpose of predicting their chances of pregnancy. The findings are concordant with previous reports describing AMH as a quantitative but not a qualitative marker of ovarian reserve and therefore does not reflect a woman’s ability to become pregnant. RBMOnline ª 2013, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. KEYWORDS: anti-Mu ¨llerian hormone, fecundity, Mu ¨llerian-inhibiting substance, natural conception, ovarian reserve, time to pregnancy

1472-6483/$ - see front matter ª 2013, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.rbmo.2013.10.007

AMH and time to pregnancy

Introduction Anti-Mu ¨llerian hormone (AMH), a paracrine factor implicated in the regulation of early follicular growth, is now widely used as a quantitative marker of ovarian reserve in infertility workups (Visser and Themmen, 2005; La Marca et al., 2009). In assisted reproduction treatment, AMH concentration predicts the response to ovarian stimulation and are logically used for adjusting the doses of gonadotrophins used (Al-Azemi et al., 2011; Broekmans et al., 2006; Nelson et al., 2009). The relationship between AMH concentration and oocyte quality is controversial as certain authors find an association between AMH and oocyte morphology, fertilization rates and embryo morphology, while others do not confirm these findings (Ebner et al., 2006; Irez et al., 2011; Lekamge et al., 2007; Silberstein et al., 2006; Riggs et al., 2011; Smeenk et al., 2007). Whether AMH concentration independently predicts pregnancy and live birth rates after assisted reproduction treatment is also subject to debate. The number of collected oocytes per cycle strongly correlates with live birth rates in stimulated cycles (Broekmans et al., 2008; Sunkara et al., 2011), and since AMH concentration is strongly correlated with the number of collected oocytes, it is logical to assume that AMH will also be correlated to live birth rates in assisted reproduction. This was indeed demonstrated in a study by Nelson et al. (2007) which examined whether AMH concentration predicts live birth in women undergoing their first assisted reproduction cycle with a long gonadotrophin-releasing hormone agonist protocol: their results showed an association between AMH and live birth (area under the receiver operating characteristics curve (AUC): 0.62); AMH concentration did not, however, predict live birth independently of oocyte yield. In contrast, La Marca et al. (2011) showed, in a prediction model of assisted reproduction outcome, that only age and AMH predicted live birth, and this was independent of the number of retrieved oocytes. The discriminative ability of AMH concentration in this model was low with an AUC of 0.55. When associating age and AMH, the model permitted the identification of live birth with a sensitivity of 79.2% and a specificity of 44.2% at the best cut off (La Marca et al., 2011). Furthermore, as reported by two small studies published in French (abstracts available in English), AMH concentration does not seem to foretell pregnancy chances in women whose treatment does not include mutlifollicular ovarian stimulation as in the case of intrauterine insemination or modified natural-cycle IVF (Lamazou et al., 2011a; Lamazou et al., 2011b). One propounded practical advantage of AMH over other ovarian function markers is that measurements remain relatively stable during the menstrual cycle and under hormonal treatments (Hehenkamp et al., 2006; Streuli et al., 2008; Tsepelidis et al., 2007; Mohamed et al., 2006). In contrast to the abundance of publications on AMH in infertile women, little is known about this hormone in fertile women. One recent study reports the distribution of AMH concentrations according to age in a large cohort of presumably healthy women consulting for male factor infertility (Shebl et al., 2011). Most studies on AMH in the general population have focused on the prediction of the age at

217 menopause (Broer et al., 2011; Freeman et al., 2012; van Disseldorp et al., 2008). La Marca et al. (2012), studying AMH concentrations in 416 healthy women between the ages of 18 and 51 years, found an inverse correlation between AMH and age but none with body mass index (BMI) or the smoking status. In a recent publication on AMH concentrations in the first trimester of pregnancy in fertile women attending an abortion centre, Masse et al. (2011) showed, as the current study group has reported before, that spontaneous pregnancies can be encountered in women with low or even undetectable AMH: 13.5% of women pregnant women in the first trimester had AMH <0.75 ng/ml and 2.8% <0.4 ng/ml (Masse et al., 2011; Fraisse et al., 2008). In a recent publication, Hagen et al. (2012) showed that women with AMH concentrations in the first quintile and even AMH <0.7 ng/ml had a monthly probability of conceiving (fertility ratio) comparable to women with AMH concentrations in the second to fourth quintiles. In contrast women with AMH concentrations in the fifth quintile had reduced fecundity. It is, however, not known to this date whether women from the general non-infertile population with very low AMH have a reduced chance of conception compared with women with normal or high AMH. There are only few reports studying AMH concentration during pregnancy. In a longitudinal study, Nelson et al. (2010) found that AMH concentration is not significantly modified in the first trimester of pregnancy, as reported in cross-sectional studies (La Marca et al., 2005); AMH concentration – unchanged in the first trimester – decreased significantly during the second and third trimesters of pregnancy, as confirmed by the current study group (unpublished data; see also discussion). Later, AMH return to concentrations comparable to those of the first trimester 17 weeks after delivery (Nelson et al., 2010). These last decades, a general trend in European societies has led to delaying childbearing until the early to mid-thirties (OFS, 2008). In turn women have become increasingly aware of fertility issues related to age, which sparked mounting interest for ‘fertility tests’. To that end, AMH and other markers of ovarian function such as FSH or antral follicle count are commonly measured in women of all reproductive ages for the purpose of counselling them about their individual medical urgency at conceiving. These markers of ovarian function have not been studied for their ability to predict natural conception in non-infertile women. Time to pregnancy (TTP; the number of months or menstrual cycles taken by a couple to conceive) is a parameter widely used in epidemiology to reflect fecundity, such as when studying the effect of environmental factors on human reproduction (Baird and Steiner, 2012; Jensen et al., 2001; Joffe et al., 2005). The term ‘effective time to pregnancy’ (eTTP) is used when referring to the time needed for achieving a pregnancy that results in a live birth (time interval between starting unprotected intercourse and the beginning of pregnancy) (De La Rochebrochard and Thonneau, 2005; Larsen and Vaupel, 1993). Time to pregnancy can be evaluated reliably using a prospective or a retrospective study design, with some limitations particular to each design (Jensen et al., 2001). One sensitive way of assessing fecundity is therefore to determine the time women take to conceive a pregnancy that will result in a live birth when they actively try to become

218 pregnant, or eTTP. Hence, one astute way of challenging fertility tests is to determine whether they actually foretell the promptness with which women will conceive. The objective of the present study was to determine whether a correlation exists between AMH concentration – first-trimester values being used as surrogate for non-pregnant reference – and fecundity reflected by eTTP in a population of women who conceived naturally.

I Streuli et al. fertility. Women were also asked to confirm that their pregnancy was planned and that they became pregnant spontaneously without any fertility treatment. Women wrote down the date at which they became exposed to the risk of pregnancy defined as the date at which they started having sexual intercourse without any form of contraception with the aim of achieving pregnancy.

Serum samples

Materials and methods The study protocol was reviewed and accepted by the ethics committee of the Geneva university hospitals (protocol no. 07-219, approved 5 October 2007) and all women gave a written informed consent for participating in the study and for measuring AMH in frozen serum aliquots.

Subjects This is an observational study in a population of fertile women aged between 23 and 41 years. French-speaking women were recruited during their postpartum hospital stay at the Maternity of the Geneva University Hospitals between December 2007 and March 2009. Recruitment was conducted by three gynaecologists who questioned women having recently delivered about their eligibility and willingness to participate in the study. Recruitment of study participants among women having delivered was therefore randomly conducted based on factors independent of the study paradigm, the recruiter’s work schedule. On recruitment days, all patients present in the postpartum unit were approached and informed about the study. If the patient agreed to participate, the recruiting gynaecologist checked whether she met the following inclusion criteria: spontaneous, desired and planned pregnancy in women over the age of 18 years. Women with unplanned, undesired pregnancies or pregnancies resulting from any form of infertility treatment (assisted reproduction and non-assisted reproduction treatment) were not eligible. The study protocol also excluded women who had been breastfeeding or taking contraception at the moment of conception. In order not to exclude women with long eTTP, women who had consulted an infertility specialist but became pregnant spontaneously without requiring treatment were considered eligible. Patients who met the inclusion criteria were given an informed consent to sign and a self-completion questionnaire to fill out during their postpartum hospital stay. The questionnaire and the signed informed consent form were returned by mail or given to the attending midwife before leaving the hospital. This study used a short version of the French translation of the time to pregnancy questionnaire used by the European Infertility Subfecundity Study Group (Juul et al., 1999). The main items of the questionnaire were: estimated delivery date (defined in Switzerland as 40 weeks of amenorrhoea), weight and height, smoking status (smoker versus non-smoker), cycle regularity (regular versus irregular), number of pregnancies, number of deliveries and previous contraception use (oral contraceptive pill and/or intrauterine device). In the questionnaire, women were also asked whether they had consulted an infertility specialist or had any medical conditions that could affect

AMH was measured in serum samples stored frozen in the first trimester of pregnancy. The study was initially planned in 2008 under the assumption than AMH remained stable during pregnancy and after delivery (La Marca et al., 2005). Subsequently, in 2010 Nelson et al. (2010) showed a decrease of AMH concentration in the second and third trimesters of pregnancy. AMH concentrations in the first trimester of pregnancy and 17 weeks after delivery were not different. This decrease in the second and third trimesters of pregnancy was confirmed in 42 women in this study who had serum samples from the three trimesters of pregnancy (first trimester 2.16 ± 1.4 ng/ml, second trimester 1.75 ± 1.1 ng/ml, third trimester 1.1 ± 1.0 ng/ml; P < 0.0001). These findings led us to exclude serum samples collected after the first trimester. In Geneva, nearly all first-trimester blood analyses are performed in three laboratories, two of which (Dianalabs Geneva and the University Hospitals of Geneva Biochemistry laboratory) routinely freeze an aliquot of each serum sample for quality and research purposes after obtaining the patient’s written consent. Both laboratories were contacted in order to retrieve all frozen serum aliquots from women who had signed the informed consent and completed the study questionnaire. The serum aliquots were then centralized at the Geneva University Hospitals laboratory for analysis.

Measurement of AMH Serum AMH was retrospectively measured by enzyme-linked immunosorbent assay (ELISA) using the AMH/MIS ELISA kit (Generation I, A16507; Immunotech-Beckman, Marseilles, France) on stored frozen specimens. The same trained laboratory technician performed all measurements according to the manufacturers’ instructions. According to the references in this study’s laboratory, the detection limit of the assay is 0.1 ng/ml; the inter-assay coefficient of variation (CV) is 9.6% for serum AMH 3.5 ng/ml and 13.7% for serum AMH 9.0 ng/ml. The intra-assay CV is 4.1% for serum AMH >0.4 ng/ml.

Determination of effective time to pregnancy For each woman, the effective time to pregnancy (eTTP) was calculated as the time interval in months between the start of the exposure to the risk of pregnancy (absence of contraception and unprotected sexual intercourse) and the start of pregnancy, which was defined as the calculated date of the last monthly period. In the self-completion questionnaire, women reported the date of start of exposure and the estimated delivery date (defined in Switzerland

AMH and time to pregnancy as 40 completed weeks since the last monthly period) of their pregnancy. The estimated last monthly period was calculated by subtracting 280 days from the estimated delivery date.

Statistical analysis All statistical data were anonymized and collected in a computerized database. Statistical analysis was performed using the Statistical Package for Social Sciences version 19.0 for Macintosh (IBM SPSS, USA) and SAS version 9.1.3 (SAS, USA). Continuous data is presented as mean ± standard deviation. Student’s t-test and ANOVA were carried out when appropriate. Correlation analysis (Pearson’s correlation coefficient r) was performed between AMH, age and eTTP. eTTP was also analysed as a dichotomous variable with categories of more or less than 3 and 6 months. Two types of multivariate analysis were conducted: (i) variance–covariance analysis (generalized linear model) to explain the variations of eTTP according to covariables; and (ii) logistic regressions to determine the risk of having an eTTP below 3 and 6 months according to several potential confounding factors. A P-value <0.05 was considered statistically significant. When planning the study, it was estimated using Medcalc version 12.3.0 (MedCalc Software, Belgium) that a sample of 84 women would be sufficient to show a coefficient of correlation of 0.3 between AMH and eTTP with a power of 80% and a type-I error of 0.05. It was retrospectively calculated using PS version 3.0 (2009; William D Dupont and W Dale Plummer, Department of Biostatistics, Vanderbilt University, Nashville, TN, USA) that the sample size of 51 subjects with AMH <2.0 ng/ml (eTTP 4.3 ± 5.1 months) and 36 subjects with AMH 2.0 ng/ml (eTTP 6.6 ± 9.2 months) would show a difference in eTTP between the groups of >4.5 months with a power of 84.5% and a type I error of 0.05.

Results

219 Table 1

Characteristics of the study population.

Characteristic

Study population (n = 87)

Age (years; n = 87) <30 30–34 35

31.0 ± 4.5 37 (42.5) 29 (33.3) 21 (24.1) xxxx 2.1 ± 1.5 16 (18.4) 15 (17.2) 20 (23.0) 36 (41.4) xxxx 22.9 ± 3.9 3 (3.5) 60 (69.8) 19 (22.1) 4 (4.7) xxxx

x AMH (ng/ml; n = 87) <1.0 1.0–1.4 1.5–1.9 2.0

x BMI (kg/m2; n = 86) <18.5 18.5–24.9 25.0–29.9 30.0

x Smoking status Non smoker Smoker

x Previous contraception No contraception Oral contraceptive pill Oral contraceptive pill and IUD

x Menstrual cycles Regular Irregular

x Sexual intercourse 1 /week <1 week Undetermined

x A total of 279 women were approached for study participation, 252 of of whom met the inclusion criteria and received the questionnaire and the study consent form. Of them, 186 women returned a completed questionnaire and signed the study consent form and 130 had at least one serum sample frozen during her pregnancy available for AMH measurements. Of these, 102 confirmed in the questionnaire that their pregnancy was actively planned while 28, who stated in the questionnaire that their pregnancy was desired but not planned, were excluded. Finally, since AMH concentration was shown to decrease in the second and third trimesters of pregnancy, the study population was restricted to the 87 women with a planned pregnancy in whom the frozen serum sample was strictly obtained before 14 weeks of amenorrhoea (12 weeks of gestation; Nelson et al., 2010). In the study population, 35.6% had a university degree, 23.0% a higher education other than the university and 24.0% had completed an apprenticeship; 82.7% were professionally active. As shown in Table 1, the women’s mean age was 31.0 ± 4.5 years; 57.5% of them were aged 30 years, 24.1% were aged 35 years and only one woman was aged

No. of pregnancies 1 2 3

x 66 (75.9) 21 (24.1) xxxx

x 6 (6.9) 70 (80.5) 11 (12.6) xxxx

x 67 (80.7) 16 (19.3) xxxx

x 69 (80.2) 11 (12.8) 6 (7.0) xxxx

x

x

28 (32.2) 30 (34.5) 29 (33.3) xxxx

No. of deliveries 1 2 3

46 (53.5) 31 (36.0) 9 (10.5)

x

Values are mean ± SD or n (%). AMH = anti-Mu ¨llerian hormone; BMI = body mass index; IUD = intrauterine device.

40 years. Mean AMH was 2.1 ± 1.5 ng/ml; 18.4% of women had AMH <1.0 ng/ml, 35.6% had <1.5 ng/ml and 58.6% had <2.0 ng/ml. Most women had normal weight, with mean BMI 22.9 ± 3.9 kg/m2, 22.1% were overweight (BMI 25.0– 29.9 kg/m2) and 4.6% were obese (30.0 kg/m2). Moreover, 24.1% women were active smokers at the time of conception. Concerning previous contraception use, 80.5% had used oral contraceptives only, 12.6% women had a history

220 of oral contraception and intrauterine device use and 6.9% had never used any contraception. Most women had regular menstrual cycles (80.7%) and 80.2% had at least one sexual intercourse per week. In this study population, there were 32.2% for whom the pregnancy was their first and 53.5% for whom the child was their first-born. A total of 55.9% of women conceived within a period of 3 months, 82.0% within 6 months and 88.1% within a year. Only few women reported a medical condition than could have impacted fertility (two with past treated salpingitis, two with thyroid function alteration and three with diabetes). There were no women reporting endometriosis, polycystic ovarian syndrome or with a history of radiotherapy or chemotherapy. Ten women had undergone infertility investigations (six had had hysterosalpingography, five had had sperm evaluation of their partner and three had had hormonal measurements). AMH was measured at a mean of 11 ± 2.3 weeks of amenorrhoea: 35.1% of women had AMH measured before 11 weeks of amenorrhoea and 64.9% between 11 and 14 weeks of amenorrhoea. AMH concentration did not differ between the early first trimester and the late first trimester (1.8 ± 1.2 ng/ml and 2.1 ± 1.5 ng/ml, respectively). A correlation analysis confirmed the absence of change of AMH concentration throughout the first trimester (r = 0.10). This study first conducted a correlation analysis, which showed a decrease of AMH concentration (r = 0.24, P = 0.02) with increasing age (Figure 1A). There was no significant correlation between AMH and eTTP (r = 0.10) (Figure 1B) and between AMH and BMI (r = 0.05). The correlation between age and eTTP did not reach statistical significance (r = 0.19) (Figure 1C). As shown in Table 2, eTTP increased with increasing age, however not significantly. AMH did not influence eTTP significantly, irrespectively of the AMH concentration boundary chosen. eTTP was not influenced by other variables such as previous contraception, cycle regularity (irregular versus regular cycles), smoking status (smoker versus non smoker) and number of pregnancies and deliveries. Overweight women had a significantly longer eTTP compared with women with a weight within the normal range (4.5 ± 5.3 versus 9.1 ± 11.8 months; P = 0.02). The study population was then dichotomized according to AMH > or <2.0 ng/ml. There was no significant difference in the eTTP between women with AMH <2.0 (n = 51, 4.3 ± 5.1 months) and above 2.0 ng/ml (n = 36, 6.6 ± 9.2 months). A univariate analysis choosing eTTP boundaries of 3 and 6 months showed a significant association of eTTP only with age (P < 0.01) but not with BMI. In a variance–covariance analysis (generalized linear model), eTTP was analysed according to the following covariables: AMH (<2.0 versus 2.0 ng/ml), age (35 versus <35 years), BMI (25.0 versus <25.0 kg/m2), smoking (smoker versus non-smoker), cycle regularity (irregular versus regular cycles), frequency of sexual intercourse (<1/week versus 1/week) and number of deliveries (none versus 1). In this multivariate analysis, eTTP significantly increased with increasing age (P = 0.01) and in overweight women compared with women with normal weight (P = 0.03). Logistic regressions were conducted to determine the cofactors related to an eTTP of more than 3 and 6 months

I Streuli et al.

Fig. 1 Correlation between age and AMH (A), AMH and effective time to pregnancy (B) and age and effective time to pregnancy (C).

using AMH thresholds 1.0 ng/ml and 2.0 ng/ml. As shown in Table 3, the only variable that affected the probability of pregnancy within 3 or 6 months was age. AMH concentration did not influence the probability of pregnancy irrespectively of whether a limit of 1.0 ng/ml or 2.0 ng/ml was chosen.

Discussion This shows that eTTP – a marker of fecundity – and AMH are not related in women of reproductive age who

AMH and time to pregnancy

221

Table 2 Effective time to pregnancy according to the characteristics of subjects.

x Age (years) <30 30–34 35

x AMH (ng/ml) <1.0 1.0–1.4 1.5–1.9 2.0

x BMI (kg/m2) <25 25

x Smoking Non smoker Smoker

x Previous contraception None Oral contraceptive pill Oral contraceptive pill and IUD

x Menstrual cycles Regular Irregular

x Sexual intercourse 1/week <1 /week or undetermined

x

n

Study population

x

x

37 29 21 xxxx

5.4 ± 8.8 4.2 ± 3.3 8.6 ± 9.7

x

x

16 15 20 36 xxxx

7.9 5.9 6.2 4.5

± ± ± ±

10.5 10.3 7.4 5.0

x

x

63 23 xxxx

4.5 ± 5.3 9.1 ± 11.8a

x

x

66 21 xxxx

5.6 ± 8.0 6.0 ± 7.2

x

x

6 70 11 xxxx

8.2 ± 8.3 5.8 ± 8.3 3.8 ± 2.0

x

x

67 16 xxxx

5.6 ± 7.4 7.4 ± 9.8

x

x

69 17 xxxx

5.5 ± 6.7 6.9 ± 11.4

x

x 5.8 ± 8.0 6.2 ± 8.9 5.2 ± 6.3

x

28 30 29 xxxx

No. of deliveries 1 2 3

x

x

46 31 9

6.7 ± 8.6 4.6 ± 7.1 4.8 ± 5.5

No. of pregnancies 1 2 3

Values are mean ± standard deviation. AMH = anti-Mu ¨llerian hormone; BMI = body mass index; IUD = intrauterine device. a P = 0.02.

conceived naturally within short delays (88% of studied women conceived in the first year). In the variance– covariance analyses, increasing age and overweight were the only two factors associated with longer eTTP. This therefore implies that AMH, a marker of ovarian function that quantitatively reflects the number of ovarian follicles, does not predict eTTP. The strength of this study lies in the fact that it assessed the relationship between AMH and the time needed for

conceiving a live birth naturally. This contrasts with most studies available on AMH concentrations that were conducted in infertile women. Remarkably, women actively seeking to conceive are precisely those who might be interested in ‘fertility tests’ such as measuring their AMH. The primary limitations of this study are the following: (i) Participating women were recruited during their postpartum hospital stay following the delivery of a live-born child, which could lead to the following shortcomings: (a) determining the exact moment when women attempted to conceive might have suffered recall biases; Joffe et al. (2005), however, reported, in an extensive review on the subject, that TTP could be accurately assessed retrospectively; (b) the retrospective assessment of eTTP diminishes the proportion of women with subfertility and excludes those with frank infertility; the retrospective assessment of eTTP also leads to an overrepresentation of women with short eTTP and a reduced number of women with long eTTP; however, the latter is not a true concern in this study as the research aim was precisely to assess the relationship of AMH and eTTP in fertile women and the exclusion of frankly infertile women is therefore not problematic; and (c) all women had serum samples stored during the first trimester of pregnancy, which limited the study population; there are, however, no reasons to believe that this caused any sort of recruitment bias, as the existence or absence of blood samples solely depended on the laboratory in which the first-trimester serum sample had been collected; (ii) this study had a sufficient sample size to show a difference of >4.5 months in the time to pregnancy between women with AMH > and <2.0 ng/ml; however, it was underpowered to show smaller differences; furthermore, few women had AMH in the extremes (five with <0.5 ng/ml and five with >5 ng/ml) and eTTP could not therefore be analysed according to low and high AMH thresholds; and (iii) AMH measurements performed on first-trimester blood samples served as surrogate markers of pre-pregnancy AMH; while a decline in AMH has been reported in the second and third trimesters of pregnancy (Lutterodt et al., 2009), the assumption that first-trimester AMH concentration reflects pre-pregnancy values holds true today. Furthermore, Lutterodt et al. showed no relationship between AMH concentration and human chorionic gonadotrophin, oestradiol and progesterone concentrations during the first trimester of pregnancy, with AMH also showing no gender differences, despite the fact that AMH is secreted by male fetuses as early as 40 days after conception. Taken together, these reports continue to support the assumption that AMH determined in the first trimester of pregnancy are comparable to pre-pregnancy concentrations for the purpose of this study. Yet, to this date, there is no study reporting a longitudinal assessment of AMH in the preconception period and during the first trimester of pregnancy. Until this is available, it is impossible to rule out the existence of slight differences in pre-pregnancy and first-trimester AMH concentration. The medical literature was reviewed and only two studies assessing the relationship between AMH concentration and fecundity were found. Hagen et al. (2012) studied in a prospective cohort 430 couples from the stop of contraception until the occurrence of pregnancy or for six menstrual cycles. AMH was determined using the generation

222 Table 3

I Streuli et al. Logistic regression of factors predicting an effective time to pregnancy of more than 3 or 6 months.

x Effective time to pregnancy >3 months AMH (ng/ml) Age (years) BMI (kg/m2) Smoking Menstrual cycles Sexual intercourse No. of deliveries

x Effective time to pregnancy >6 months AMH (ng/ml) Age (years) BMI (kg/m2) Smoking Menstrual cycles Sexual intercourse No. of deliveries

AMH concentration boundary of 2 ng/ml

AMH concentration boundary of 1 ng/ml

Comparison

OR (95% CI)

Comparison

OR (95% CI)

<2.0 versus 2.0 35 versus <35 25.0 versus <25.0 Smoker versus non-smoker Irregular versus regular <1/week versus 1 None versus 1 xxxx xxx

0.85 0.21 0.59 2.44 0.52 1.75 0.72

(0.32–2.80) (0.06–0.72)a (0.18–1.89) (0.73–8.17) (0.13–2.04) (0.36–8.52) (0.26–2.03)

<1.0 versus 1.0 35 versus <35 25.0 versus <25.0 Smoker versus non-smoker Irregular versus regular <1/week versus 1 None versus 1

0.92 (0.34–3.60) 0.25 (0.08–0.75)a 0.43 (0.13–1.39) 2.64 (0.78–8.98) 0.76 (0.18–3.28) 2.11 (0.45–9.82) 0.67 (0.24–1.88)

<2.0 versus 2.0 35 versus<35 25.0 versus <25.0 Smoker versus non-smoker Irregular versus regular <1/week versus 1 None versus 1

0.95 (0.19–4.62) 0.13 (0.03–0.56)a 0.53 (0.11–2.43) 2.38 (0.48–11.89) 0.30 (0.06–1.60) 0.41 (0.06–2.94) 0.28 (0.06–1.35)

<1.0 versus 1.0 35 versus <35 25.0 versus <25.0 Smoker versus non-smoker Irregular versus regular <1/week versus 1 None versus 1

0.28 (0.05–1.69) 0.11 (0.03–0.46)a 0.53 (0.11–2.40) 1.85 (0.34–9.90) 0.59 (0.09–3.89) 0.41 (0.06–2.85) 0.27 (0.05–1.36)

xxx

AMH = anti-Mu ¨llerian hormone; BMI = body mass index; CI = confidence interval; OR = odds ratio. a P = 0.01.

I Immunotech assay from Beckman-Coulter in a subgroup of 186 women (age 19–35 years), 110 of whom (59%) achieved pregnancy within 6 months. Compared with the reference group of women with AMH in the second to fourth quintiles (1.69–5.46 ng/ml), fecundity did not differ significantly in women with AMH in the first quintile (<1.96 ng/ml) with a fecundity ratio of 0.81 (95% CI 0.44–1.4). In a subanalysis of the 11% of women with very low AMH (<0.7 ng/ml), fecundity was not reduced with a fecundity ratio of 0.88 (95% CI 0.48–1.61). Interestingly, women with AMH in the fifth quintile (>5.46 ng/ml) had a reduced fecundity, with a fecundity ratio of 0.62 (95% CI 0.39–0.99). This reduction persisted even after exclusion of women with irregular menstrual cycles. The authors postulated that this reduced fecundity could be related to phenotypes related to polycystic ovary syndrome. Since evaluation of hirsutism and antral follicle count were not performed in that study, this hypothesis could, however, not be confirmed. Steiner et al. (2011) recruited 100 women between the ages of 30 and 42 years who had attempted to conceive for less than 3 months and without history of infertility. They collected data prospectively on bleeding patterns and sexual intercourse and measured markers of ovarian function in serum samples (FSH, oestradiol, AMH and inhibin B concentrations) collected on days 2–4 of the menstrual cycle. The occurrence of a spontaneous pregnancy was assessed during a 6-month period. Their results show reduced day-specific probabilities of pregnancy in women aged >35 years (n = 29, fecundity ratio 0.42, 95% CI 0.15–0.85) as well as in women with AMH <0.7 ng/ml (n = 18, fecundity ratio 0.36, 95% CI 0.01–0.84 and 0.38,

95% CI 0.01–0.84 when age adjusted). This study presents interesting results but suffers from several limitations: (i) the study only included women over the age of 30 years, their results can therefore not be generalized; (ii) 100 women were included with a total of 224 menstrual cycles; hence only a few cycles in the period of assessment were included; (iii) AMH was measured by a two-site enzyme-linked immunosorbent assay. It is, however, not specified which assay was actually used. The cut-off concentration of 0.7 ng/ml is, therefore, not interpretable since values given by different AMH assays at different time periods have varied considerably since commercial AMH assays have become available (Nelson and La Marca, 2011; Streuli et al., 2009); and (iv) there is no description of the relationship between fecundity and other AMH cut-off concentrations. In both Steiner et al. (2011) and Hagen et al. (2012), fecundity was evaluated prospectively and the occurrence of a spontaneous pregnancy was evaluated for a relatively short period of 6 months; these studies therefore selected the most fertile women. Furthermore, the study results only report clinical pregnancies but give no further information about the evolution of the pregnancy (miscarriages, live birth rate). In this study study, three women had undetectable AMH and 16 had AMH <1.0 ng/ml, confirming prior reports (Fraisse et al., 2008; Masse et al., 2011) that low and undetectable AMH do not exclude the possibility of a pregnancy. It can, however, not be excluded that low AMH could, in a subpopulation of women, be related to a process affecting the ovarian reserve and leading to a reduction of the long-term

AMH and time to pregnancy pregnancy chances and to infertility. Low AMH has indeed been associated with an early age at menopause and, therefore could be related to a reduced reproductive life span. (Broer et al., 2011; Tehrani et al., 2011; Freeman et al., 2012). The results of this study and Hagen et al. (2012) suggest that there is no relationship between AMH concentration and time to pregnancy and therefore speak against determining AMH in women who are not infertile for the purpose of predicting their chances of pregnancy. There is currently insufficient data to predict the long-term reproductive outcome in women with low AMH in terms of spontaneous pregnancies and risk of infertility, ovarian insufficiency or premature menopause. In this study group’s opinion, measuring AMH concentrations in women without infertility might lead to unjustified anxieties about the reproductive potential of women with low AMH values. Further studies on AMH concentration in the general and the non-infertile population should be conducted in order to determine whether AMH measurements could be useful in certain conditions. In an era in which fertility-preserving measures such as cryopreservation of oocytes have become available and are widely offered to non-infertile women who are not yet ready to conceive, there is an urgent need for ways to assess future fertility. In conclusion, this study shows no relationship between AMH concentration and the time needed to conceive a live-born child. The findings support the current view that AMH is a quantitative, not a qualitative, marker of ovarian follicles. To this date, there is insufficient evidence to recommend the measurement of AMH in women who are not infertile for the purpose of predicting their live birth chances or to detect women of lower fertility. Further studies should be conducted to assess the relationship of AMH concentration and the fertility potential in the general population.

Acknowledgements Beckman Coulter Switzerland donated the AMH measurement kits used in this research. The authors thank the midwives from the postpartum unit of the Maternity Hospital of the Geneva University Hospitals for their help in retrieving the completed questionnaires and signed consents. They also thank Prof P Bischof and Prof M Boulvain for their advice in the conception of the study and Dr T Fraisse for his help in recruiting the study subjects. The authors also thankfully acknowledge the assistance of Dr R Stricker from Dianalabs who helped us retrieve the serum samples for AMH measurement and Mrs C Souverain who performed the AMH measurements.

References Al-Azemi, M., Killick, S.R., Duffy, S., Pye, C., Refaat, B., Hill, N., Ledger, W., 2011. Multi-marker assessment of ovarian reserve predicts oocyte yield after ovulation induction. Hum. Reprod. 26, 414–422. Baird, D.D., Steiner, A.Z., 2012. Anti-Mullerian hormone: a potential new tool in epidemiologic studies of female fecundability. Am. J. Epidemiol. 175, 245–249.

223 Broekmans, F.J., Kwee, J., Hendriks, D.J., Mol, B.W., Lambalk, C.B., 2006. A systematic review of tests predicting ovarian reserve and IVF outcome. Hum. Reprod. Update 12, 685–718. Broekmans, F.J., Visser, J.A., Laven, J.S., Broer, S.L., Themmen, A.P., Fauser, B.C., 2008. Anti-Mullerian hormone and ovarian dysfunction. Trends Endocrinol. Metab. 19, 340–347. Broer, S.L., Eijkemans, M.J., Scheffer, G.J., van Rooij, I.A., de Vet, A., Themmen, A.P., Laven, J.S., de Jong, F.H., te Velde, E.R., Fauser, B.C., Broekmans, F.J., 2011. Anti-mullerian hormone predicts menopause: a long-term follow-up study in normoovulatory women. J. Clin. Endocrinol. Metab. 96, 2532–2539. De La Rochebrochard, E., Thonneau, P., 2005. Paternal age: are the risks of infecundity and miscarriage higher when the man is aged 40 years or over? Epidemiol. Sante Publique 53, 2S47–2S55. Ebner, T., Sommergruber, M., Moser, M., Shebl, O., Schreier-Lechner, E., Tews, G., 2006. Basal level of anti-Mullerian hormone is associated with oocyte quality in stimulated cycles. Hum. Reprod. 21, 2022–2026. Fraisse, T., Ibecheole, V., Streuli, I., Bischof, P., de Ziegler, D., 2008. Undetectable serum anti-Mullerian hormone levels and occurrence of ongoing pregnancy. Fertil. Steril. 89, e9–11. Freeman, E.W., Sammel, M.D., Lin, H., Gracia, C.R., 2012. Anti-mullerian hormone as a predictor of time to menopause in late reproductive age women. J. Clin. Endocrinol. Metab. 97, 1673–1680. Hagen, C.P., Vestergaard, S., Juul, A., Skakkebaek, N.E., Andersson, A.M., Main, K.M., Hjollund, N.H., Ernst, E., Bonde, J.P., Anderson, R.A., Jensen, T.K., 2012. Low concentration of circulating antimullerian hormone is not predictive of reduced fecundability in young healthy women: a prospective cohort study. Fertil. Steril. 98, e2. Hehenkamp, W.J., Looman, C.W., Themmen, A.P., de Jong, F.H., te Velde, E.R., Broekmans, F.J., 2006. Anti-Mullerian hormone levels in the spontaneous menstrual cycle do not show substantial fluctuation. J. Clin. Endocrinol. Metab. 91, 4057–4063. Irez, T., Ocal, P., Guralp, O., Cetin, M., Aydogan, B., Sahmay, S., 2011. Different serum anti-Mullerian hormone concentrations are associated with oocyte quality, embryo development parameters and IVF-ICSI outcomes. Arch. Gynecol. Obstet. 284, 1295–1301. Jensen, T.K., Slama, R., Ducot, B., Suominen, J., Cawood, E.H., Andersen, A.G., Eustache, F., Irvine, S., Auger, S., Jouannet, P., Vierula, M., Jorgensen, N., Toppari, J., Skakkebaek, N.E., Keiding, N., Spira, A., 2001. Regional differences in waiting time to pregnancy among fertile couples from four European cities. Hum. Reprod. 16, 2697–2704. Joffe, M., Key, J., Best, N., Keiding, N., Scheike, T., Jensen, T.K., 2005. Studying time to pregnancy by use of a retrospective design. Am. J. Epidemiol. 162, 115–124. Juul, S., Karmaus, W., Olsen, J., 1999. Regional differences in waiting time to pregnancy: pregnancy-based surveys from Denmark, France, Germany, Italy and Sweden. The European infertility and subfecundity study group. Hum. Reprod. 14, 1250–1254. la Marca, A., Giulini, S., Orvieto, R., de Leo, V., Volpe, A., 2005. Anti-Mullerian hormone concentrations in maternal serum during pregnancy. Hum. Reprod. 20, 1569–1572. la Marca, A., Broekmans, F.J., Volpe, A., Fauser, B.C., Macklon, N.S., 2009. Anti-Mullerian hormone (AMH): what do we still need to know? Hum. Reprod. 24, 2264–2275. la Marca, A., Nelson, S.M., Sighinolfi, G., Manno, M., Baraldi, E., Roli, L., Xella, S., Marsella, T., Tagliasacchi, D., D’Amico, R., Volpe, A., 2011. Anti-Mullerian hormone-based prediction model for a live birth in assisted reproduction. Reprod. Biomed. Online 22, 341–349. la Marca, A., Spada, E., Grisendi, V., Argento, C., Papaleo, E., Milani, S., Volpe, A., 2012. Normal serum anti-Mullerian

224 hormone levels in the general female population and the relationship with reproductive history. Eur. J. Obstet. Gynecol. Reprod. Biol. Lamazou, F., Fuchs, F., Genro, V., Malagrida, L., Torre, A., Albert, M., Wainer, R., 2011a. Intra-uterine insemination outcomes according to the serum AMH level on day 3. J. Gynecol. Obstet. Biol. Reprod. (Paris) 41, 122–127. Lamazou, F., Genro, V., Fuchs, F., Grynberg, M., Gallot, V., Achour-Frydman, N., Fanchin, R., Frydman, R., 2011b. Serum AMH level is not a predictive value for IVF in modified natural cycle: analysis of 342 cycles. J. Gynecol. Obstet. Biol. Reprod. (Paris) 40, 205–210. Larsen, U., Vaupel, J.W., 1993. Hutterite fecundability by age and parity: strategies for frailty modeling of event histories. Demography 30, 81–102. Lekamge, D.N., Barry, M., Kolo, M., Lane, M., Gilchrist, R.B., Tremellen, K.P., 2007. Anti-Mullerian hormone as a predictor of IVF outcome. Reprod. Biomed. Online 14, 602–610. Lutterodt, M., Byskov, A.G., Skouby, S.O., Tabor, A., Yding Andersen, C., 2009. Anti-Mullerian hormone in pregnant women in relation to other hormones, fetal sex and in circulation of second trimester fetuses. Reprod. Biomed. Online 18, 694–699. Masse, V., Ferrari, P., Boucoiran, I., Delotte, J., Isnard, V., Bongain, A., 2011. Normal serum concentrations of anti-Mullerian hormone in a population of fertile women in their first trimester of pregnancy. Hum. Reprod. 26, 3431–3436. Mohamed, K.A., Davies, W.A., Lashen, H., 2006. Antimullerian hormone and pituitary gland activity after prolonged down-regulation with goserelin acetate. Fertil. Steril. 86, 1515–1517. Nelson, S.M., la Marca, A., 2011. The journey from the old to the new AMH assay: how to avoid getting lost in the values. Reprod. Biomed. Online 23, 411–420. Nelson, S.M., Yates, R.W., Fleming, R., 2007. Serum anti-Mullerian hormone and FSH: prediction of live birth and extremes of response in stimulated cycles – implications for individualization of therapy. Hum. Reprod. 22, 2414–2421. Nelson, S.M., Yates, R.W., Lyall, H., Jamieson, M., Traynor, I., Gaudoin, M., Mitchell, P., Ambrose, P., Fleming, R., 2009. Anti-Mullerian hormone-based approach to controlled ovarian stimulation for assisted conception. Hum. Reprod. 24, 867–875. Nelson, S.M., Stewart, F., Fleming, R., Freeman, D.J., 2010. Longitudinal assessment of antimullerian hormone during pregnancy-relationship with maternal adiposity, insulin, and adiponectin. Fertil. Steril. 93, 1356–1358. OFS, 2008. Office fe ´de ´ral de la statistique Suisse. Available from: . Riggs, R., Kimble, T., Oehninger, S., Bocca, S., Zhao, Y., Leader, B., Stadtmauer, L., 2011. Anti-Mullerian hormone serum levels predict response to controlled ovarian hyperstimulation but not embryo quality or pregnancy outcome in oocyte donation. Fertil. Steril. 95, 410–412.

I Streuli et al. Shebl, O., Ebner, T., Sir, A., Schreier-Lechner, E., Mayer, R.B., Tews, G., Sommergruber, M., 2011. Age-related distribution of basal serum AMH level in women of reproductive age and a presumably healthy cohort. Fertil. Steril. 95, 832–834. Silberstein, T., Maclaughlin, D.T., Shai, I., Trimarchi, J.R., Lambert-Messerlian, G., Seifer, D.B., Keefe, D.L., Blazar, A.S., 2006. Mullerian inhibiting substance levels at the time of HCG administration in IVF cycles predict both ovarian reserve and embryo morphology. Hum. Reprod. 21, 159–163. Smeenk, J.M., Sweep, F.C., Zielhuis, G.A., Kremer, J.A., Thomas, C.M., Braat, D.D., 2007. Antimullerian hormone predicts ovarian responsiveness, but not embryo quality or pregnancy, after in vitro fertilization or intracyoplasmic sperm injection. Fertil. Steril. 87, 223–226. Steiner, A.Z., Herring, A.H., Kesner, J.S., Meadows, J.W., Stanczyk, F.Z., Hoberman, S., Baird, D.D., 2011. Antimullerian hormone as a predictor of natural fecundability in women aged 30–42 years. Obstet. Gynecol. 117, 798–804. Streuli, I., Fraisse, T., Pillet, C., Ibecheole, V., Bischof, P., de Ziegler, D., 2008. Serum antimullerian hormone levels remain stable throughout the menstrual cycle and after oral or vaginal administration of synthetic sex steroids. Fertil. Steril. 90, 395–400. Streuli, I., Fraisse, T., Chapron, C., Bijaoui, G., Bischof, P., de Ziegler, D., 2009. Clinical uses of anti-Mullerian hormone assays: pitfalls and promises. Fertil. Steril. 91, 226–230. Sunkara, S.K., Rittenberg, V., Raine-Fenning, N., Bhattacharya, S., Zamora, J., Coomarasamy, A., 2011. Association between the number of eggs and live birth in IVF treatment: an analysis of 400 135 treatment cycles. Hum. Reprod. 26, 1768–1774. Tehrani, F.R., Shakeri, N., Solaymani-Dodaran, M., Azizi, F., 2011. Predicting age at menopause from serum antimullerian hormone concentration. Menopause 18, 766–770. Tsepelidis, S., Devreker, F., Demeestere, I., Flahaut, A., Gervy, C., Englert, Y., 2007. Stable serum levels of anti-Mullerian hormone during the menstrual cycle: a prospective study in normo-ovulatory women. Hum. Reprod. 22, 1837–1840. van Disseldorp, J., Faddy, M.J., Themmen, A.P., de Jong, F.H., Peeters, P.H., van der Schouw, Y.T., Broekmans, F.J., 2008. Relationship of serum antimullerian hormone concentration to age at menopause. J. Clin. Endocrinol. Metab. 93, 2129–2134. Visser, J.A., Themmen, A.P., 2005. Anti-Mullerian hormone and folliculogenesis. Mol. Cell. Endocrinol. 234, 81–86. Declaration: The authors report no financial or commercial conflicts of interest. Received 17 September 2012; refereed 3 October 2013; accepted 8 October 2013.