Content uploaded by Erhard Godehardt
Author content
All content in this area was uploaded by Erhard Godehardt on Jan 13, 2014
Content may be subject to copyright.
Relevance of anti-Mullerian hormone measurement
in a routine IVF program
C. Gnoth
1,2,5
, A.N. Schuring
3
, K. Friol
1
, J. Tigges
1
, P. Mallmann
2
and E. Godehardt
4
1
Center for Family Planning, Gynecological Endocrinology and Reproductive Medicine, Rheydter Strasse 143, 41515 Grevenbroich,
Germany;
2
Department of Obstetrics and Gynecology, University of Cologne, Cologne, Germany;
3
Department of Obstetrics and
Gynecology, University of Muenster, Muenster, Germany;
4
Biometric Research Group, Clinic for Thoracic and Cardiovascular Surgery,
Heinrich-Heine University of Duesseldorf, Duesseldorf, Germany
5
Correspondence address. Tel: þ49-2181-491513; Fax: þ49 2181-491534; E-mail: dr.christian.gnoth@rmz-nrw.de
BACKGROUND: Diminished ovarian reserve has become a major cause of infertility. Anti-Mullerian hormone
(AMH) seems to be a promising candidate to assess ovarian reserve and predict the response to controlled ovarian
hyperstimulation (COH). This prospective study was conducted to evaluate the relevance of AMH in a routine IVF
program. METHODS: Three hundred and sixteen patients were prospectively enrolled to enter their first IVF/
ICSI-cycle. Age, FSH-, inhibin B- and AMH-levels and their predictive values for ovarian response and clinical preg-
nancy rate were compared by discriminant analyses. RESULTS: A total of 132 oocyte retrievals were performed. A
calculated cut-off level
1.26 ng/ml AMH alone detected poor responders (
4 oocytes) with a sensitivity of 97%, and
there was a 98% correct prediction of normal response in COH if levels were above this threshold. With levels
<0.5 ng/ml, a correct prediction of very poor response (
2 oocytes) was possible in 88% of cases. Levels of AMH
0.5 ng/ml were not significantly correlated with clinical pregnancy rates. CONCLUSIONS: AMH is a predictor
of ovarian response and suitable for screening. Levels
1.26 ng/ml are highly predictive of reduced ovarian
reserve and should be confirmed by a second line antral follicle count. Measurement of AMH supports clinical
decisions, but alone it is not a suitable predictor of IVF success.
Introduction
Anti-Mullerian hormone (AMH) belongs to the transforming
growth factor-b superfamily. In the female, it is a product of
the granulosa cells from pre-antral and small antral follicles.
Together with two other key players (growth and differen-
tiation factor-9 and bone morphogenetic protein-15), it inhibits
the initiation of premature follicle growth and decreases the
sensitivity of follicles for the FSH-dependent selection
process (Gruijters et al., 2003 ; Visser and Themmen, 2005;
Knight and Glister, 2006; Visser et al., 2006). AMH levels
are characterized by a steady decline with age from adulthood
toward menopause reflecting the size of the ovarian follicle
pool (van Rooij et al., 2005). Therefore, AMH seems to be a
promising parameter for early detection of reduced ovarian
reserve as well as ovarian dysfunction (Visser et al., 2006).
The loss of follicles with increasing female age is variable
and the chronological age of the ovary does not always
reflect its biological and reproductive age (Fleming et al.,
2006). Nowadays, an increasing number of female patients in
their late 30s seek help for infertility problems, meaning that
diminished ovarian reserve and age-related ovarian dysfunc-
tion have become a major cause of infertility. Furthermore, epi-
demiological surveys suggest (Nikolaou and Templeton, 2003)
that approximately as many as 10% of women in their early 30s
could be approaching their perimenopause transition. At a high
risk for early ovarian exhaustion are those of our patients with a
history of repeated ovarian surgery because of functional cysts
and endometriosis. Numbers of such, mostly laparoscopic,
interventions seem to increase continuously. Most women
with reduced ovarian reserve are still displaying regular men-
strual cycle characteristics that will continue normally with
deterioration only becoming clinically apparent when ovarian
function is already severely impaired (Burger et al., 2005).
Thus ovarian aging and a reduced ovarian reserve can
become a critical factor in infertility, without any obvious clini-
cal symptoms. Furthermore, a substantial rise in FSH levels
occurs relatively late when the perimenopausal transition is
already present (Burger et al., 1999). On the other hand elev-
ated FSH levels in younger women may be related to receptor
polymorphisms and not reduced ovarian reserve (Schipper
et al., 1998). Because traditional assessments of ovarian
reserve such as early follicular phase FSH, inhibin B and the
clomifen challenge test have low sensitivity in the early
stages of reduced ovarian reserve (van Rooij et al., 2006),
there is an urgent need for a reliable and early marker for the
detection of a declining number of follicles, and prediction of
# The Author 2008. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology.
All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
1359
Human Reproduction Vol.23, No.6 pp. 1359–1365, 2008 doi:10.1093/humrep/den108
Advance Access publication on April 2, 2008
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
spontaneous pregnancy potential and assisted reproduction
technology outcome (Muttukrishna et al., 2005; van Rooij
et al., 2006; La Marca et al., 2007; Smeenk et al., 2007).
From the current literature, AMH is such a promising and
reliable marker, corresponding to the number of small antral
follicles (Pigny et al., 2006; Visser et al., 2006) with essentially
constant levels across the cycle (La Marca et al., 2006) and a
superior intercycle reproducibility compared with that of
FSH and early antral follicle count (Fanchin et al., 2005; Feyer-
eisen et al., 2006; Hehenkamp et al., 2006). Very recently some
studies reported a possible benefit of serum AMH-
measurement in assisted reproductive programs (Tremellen
et al., 2005; Ebner et al., 2006; Ficicioglu et al., 2006; La
Marca et al., 2007). However, urgently needed cut-off levels
of AMH for supporting clinical decisions are still missing
(Freour et al ., 2006).
We therefore carried out a larger prospective trial calculating
cut-off levels and evaluating the relevance of AMH-
measurement for treatment strategies and outcome in a
routine IVF program of a private institute.
Materials and Methods
Between 1 June 2005 and 30 June 2006, all women visiting our center
of reproductive medicine for the first time, because of involuntary
childlessness with at least 1 year of unprotected intercourse without
pregnancy, underwent a basic infertility investigation. On Day 3 –5
of the women’s normal cycle, we performed a routine gynecological
examination and a basic vaginal ultrasound took a basal hormone
profile (FSH, LH, Prolactin, testosterone, DHEAS, TSH and SHBG)
and measured AMH and inhibin B. Other variables such as socioeco-
nomic status, medical history, parity, duration of involuntary child-
lessness, age and BMI were recorded in a standardized manner.
Out of all examined women, 316 patients were prospectively
enrolled in our assisted-reproduction techniques (ART) program to
enter their first IVF- or ICSI-cycle. The criteria for inclusion were:
(i) early follicular phase Day 3 –5, FSH and AMH levels
available;
(ii) regular menstrual cycle pattern (cycle length .25 and ,35
days);
(iii) presence of both ovaries;
(iv) age ,45.
Main indications for ART were: male subfertility (65%), tubal path-
ology (12%), endometriosis (12%), idiopathic infertility (9%) with at
least two unsuccessful cycles of intrauterine insemination (IUI) and
repeated polyfollicular development with gonadotrophin stimulation
for IUI (2%). All included patients suffered from primary infertility
with an average of 3.4 years of involuntary childlessness.
On Day 3–5 of a spontaneous cycle within 3 months following the
start of controlled ovarian hyperstimulation (COH), patients under-
went another transvaginal ultrasound examination to exclude
ovarian cysts and to count the number of antral follicles, measuring
2– 5 mm as described before by others (Scheffer et al., 1999, 2003).
This study was conducted in accordance with the principles of the
declaration of Helsinki and women were asked for their consent for
AMH-measurement and link to their treatment outcome.
For AMH measurement, we used a sensitive ELISA (Enzyme-
Linked Immunosorbent Active MIS/AMH ELISA Kit
w
,
DSL-10-14400, Diagnostic System Laboratories, Inc./Beckman-
Coulter) according to the manufacture’s instructions. This highly
specific mono/mono two-site ELISA uses detection and capture
antibodies, with results available within 3 h. The standards cover a
range from 0.05 to 15 ng/ml (Freour et al., 2007). The sensitivity
is 0.006 ng/ml. Intra- and inter-assay coefficients of variation
were ,5% and ,8%, respectively. Inhibin B was measured using
the DSL inhibin B ELISA test using two highly specific monoclonal
antibodies (sensitivity 0.007 ng/ml with intra- and inter-assay coeffi-
cients of variation ,6% and ,8%, respectively, DSL-10-84100,
Diagnostic System Laboratories, Inc./Beckman-Coulter).
In brief, the IVF/ICSI-treatment was as follows: all patients started
with a monophasic oral contraceptive pill on Day 3–5 of their cycle.
Starting of COH was scheduled according to the patient’s possibilities
and the center’s necessities. Down-regulation with either nafarelin
acetate (0.4 –0.6 mg/day) or triptorelin acetate (0.05 mg/day) was
started with the last 3–5 pills, 10–14 days before the start of COH.
Ovarian hyperstimulation was performed with either recombinant fol-
litrophin b (rec-FSH, Puregon
w
, Organon/Germany) or Menogon
HP
w
(Ferring Pharmaceuticals, Kiel/Germany). Starting doses for
patients under 35 years of age were 150 IE/ml; for patients 35 and
older, it was 225–300 IE/ml. After 5 days of stimulation, follicular
growth was assessed by vaginal ultrasound, estradiol measurement
and control of pituitary desensitization was assessed by LH measure-
ment. The dose of rec-FSH was adjusted according to ovarian
response. Patients received ovulation induction with 250 mg recombi-
nant human choriongonadotrophin (rec-HCG, Ovitrelle
w
, Serono/
Germany) when at least three leading follicles were 20 mm in
size. Transvaginal oocyte retrieval was performed 35 h later. Luteal
support was provided with vaginal application of micronized pro-
gesterone (Utrogest
w
, 400 mg/day, Kade-Pharma/Germany). A
maximum of two embryos were transferred in all women in their
first attempt. Supernumerous zygotes were frozen according to the
German embryo protection act.
The primary outcome measured in this study was the number of
oocytes retrieved. According to a widespread definition, a cut-off of
four or less oocytes at retrieval was regarded as poor or low ovarian
response. With a mean fertilization rate of 60%, a minimum of
four oocytes is necessary to have a possible transfer of at least two
embryos. Approximately 30% of the ART cycles were conducted as
IVF and 70% were conducted as IVF/ICSI because of male infertility.
Clinical pregnancy was considered to be the secondary outcome
measure defined as a gestational sac assessed by vaginal ultrasound
1– 2 weeks after a positive pregnancy test.
Data were analyzed using the SAS package, version 9, of statistical
procedures (SAS Institute Inc., Cary, USA). Student’s t-test was used
to compare endocrine profiles and basic characteristics of the patients.
To compare the relevance of different parameters for predicting
ovarian reserve and clinical pregnancy rate, we first performed a step-
wise discriminant analysis, followed by linear discriminant analysis to
calculate the sensitivities and specificities as well as the percentages of
positive and negative predictions. For AMH alone, cut-off values and
receiver operating characteristic (ROC) curves were calculated
(Fawcett, 2004) to minimize both the false positive and false negative
rates in order to find an optimal threshold for the discrimination
between women with poor and normal response (.4 oocytes) and
clinical pregnancy versus no pregnancy. The calculation of the ROC
curves was based on the statistical package R. Statistical significance
was considered to be reached at P-values of ,0.05.
Results
In this prospective study, 316 patients were admitted to our IVF
program with the aim of having their first IVF or IVF/ICSI
attempt between 1 June 2005 and 30 June 2006. Out of these,
Gnoth et al.
1360
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
132 patients had their first oocyte retrieval in this time interval.
There were no cycle cancelations because of impending
ovarian hyperstimulation syndrome or no ovarian response.
For 119 patients, pregnancy tests and assessments for clinical
pregnancy (completed cycles) could be performed before 1
October 2006. From these, 56 clinical pregnancies were diag-
nosed corresponding to a clinical pregnancy rate per completed
cycle of (56/119) 47%. Thirteen tests were performed after the
1st of October and not included in the further analysis. A flow
chart of the 316 patients admitted and further details are pre-
sented in Fig. 1. The remaining 184 couples did not start
their first cycle because of personal reasons (mainly: clarifica-
tion of the coverage of treatment costs and/or occupational
activity) and were treated later according to a changed protocol
(see below). There was no statistical difference in the compo-
sition of these non-starters (length of infertility, age, cause of
infertility and hormonal levels) and the 119 starters. No preg-
nancy test was missed in the 119 completed cycles. As
expected, non-pregnant women and those with less than five
oocytes were significantly older than pregnant women or
normal responders.
The mean AMH level of all treated patients was 1.7 +
2.9 ng/ml. To investigate the correlation of AMH with poor
ovarian response, we used step-wise and linear discriminant
analyses and calculated an AMH cut-off level for poor
ovarian response of 1.26 ng/ml with minimized false posi-
tive and false negative results. Figure 2 shows the typical
ROC for AMH indicating poor ovarian response and clinical
pregnancy.
Table I gives the results for the stepwise discriminant analy-
sis for poor responders (4 oocytes, cut-off level AMH
1.26 ng/ml). AMH alone showed the best performance by
detecting 32 of true 33 poor responders (sensitivity). Age,
FSH, inhibin B alone or in combinations were less precise.
As to be expected, the specificity for identifying poor respon-
ders was improved by the use of a combination of parameters.
In correctly predicting ‘normal’ ovarian response, AMH alone,
using a calculated cut-off of AMH .1.26 ng/ml, was superior
to the other parameters or their different combinations. Apply-
ing this model, only 2% of women with an AMH .1.26 ng/ml
will have four or less oocytes in IVF following COH. Using a
cut-off level of AMH 1.26 ng/ml, a correct prediction of
‘poor’ response was only possible in 36% of the cases
mainly because of the age-related high heterogeneity
between the poor and normal responder groups and probably
dose adjustment during COH (see Discussion). To address
this problem, both groups were adjusted for age in an additional
analysis with a different statistical approach. In this way, the
specificity for AMH alone increas ed to 62% and the correct
prediction rate for poor response increased to 88%. Both
sensitivity and prediction of normal response with AMH
alone were much more robust and remained unchanged with
or without adjustment for age (Table I). Using a level of
AMH ,0.5 ng/ml, a similarly robust and correct prediction
of very poor response (2 oocytes retrieved) was possible in
88% of cases regardless of adjustment for age (Table I).
Table II gives the results of the stepwise discriminant analy-
sis for age, FSH and AMH for ‘not’ achieving pregnancy. The
cut-off level minimizing for false positive and false negative
failures was calculated as 1.8 ng/ml (Fig. 2). Neither AMH
alone nor one of the other analyzed parameters alone or their
combinations were suitable for predicting the success of
ART cycles.
Table III shows the results of the logistic regression analysis
for low and normal ovarian response in ART cycles and for
AMH with a cut-off level of 1.26 ng/ml. As expected, low
responders were significantly older and had a sign ificantly
lower clinical pregnancy rate with a significantly lower
AMH, than normal responders. In the poor responder group,
FSH was elevated but still within the normal range and was
weakly achieving statistical significance (P ¼ 0.034) in discri-
minating between low and normal responders. The overall
clinical pregnancy rate in the low-AMH group (AMH
1.26 ng
/ml) was not significantly different from the group
with levels above the cut-off.
Discussion
This prospective study was conducted to evaluate the relevance
of routine AMH measurements before IVF. Therefore, 316
patients were prospectively enrolled to enter their first IVF or
IVF/ICSI-cycle. Age, FSH-, inhibin B- and AMH-levels and
their predictive values for ovarian response and clinical preg-
nancy rate were compared by discriminant analyses. A total
of 132 oocyte retrievals were performed. Hundred and eighty-
four couples did not start their first cycle in the scheduled
observation period because of individual reasons and were,
after a preliminary analysis of the data already collected,
treated later following a changed protocol (see below), facili-
tating our clinical management. Although the number of
patients of the study group was relatively small, the results
seemed to be promising and, according to power calculations,
would not be significantly changed with increasing size of the
treatment group.
A significant correlation of AMH levels with ovarian
response as expressed by the number of oocytes retrieved
during the first IVF or IVF /ICSI treatment cycle has been
observed. Cut-off levels for clinical decisions were calculated.
The ovarian response to high levels of gonadotrophins reflects
the size of the cohort of selectable antral follicles. Because the
number of antral follicles is related to the size of the primordial
follicle pool (Gougeon et al., 1994), the ovarian response to
stimulation reflects the ovarian reserve (Lawson et al ., 2003).
Our resu lts show that age and AMH levels are the superior par-
ameters predicting ovarian response. However, even in women
of comparable age, there was a wide variation in the individual
ovarian reserve (te Velde and Pearson, 2002; Gnoth et al.,
2003; van Rooij et al., 2005). For them, AMH is a very prom-
ising and possibly the best actual candidate to evaluate their
individual ovarian response to gonadotrophin stimulation and
to detect poor responders with levels of AMH 1.26 ng/ml
(97% sensitivity). With an AMH .1.26 ng/ml, a correct
prediction of normal response (.4 oocytes) to gonadotrophin
stimulation of 98% can be achieved. Combinations of age
and serum parameters FSH, inhibin B and AMH are less
useful in predicting ovarian response in a subsequent
AMH and ART
1361
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
ART-cycle than is AMH alone. As a result of the low
specificity for AMH alone, the age-related high heterogeneity
between the normal and poor responder group, and probably
an effect of dose adjustment during COH in this study, the
correct overall prediction rate for poor response during COH
only reaches 36%. If groups are of comparable age, specificity
increases to 62% and a correct prediction of poor response will
be possible in 88% of the COHs. Sensitivity and prediction of
normal response with AMH alone are much more robust and
remain unchanged even if there is a high age-related heterogen-
eity. Levels of AMH ,0.5 ng/ml robustly predicted very poor
response (2 oocytes retrieved) with impaired IVF success.
The antral follicle count is still regarded as a relatively good
marker to predict poor ovarian response in ART programs pro-
viding better information than the patient’s age alone or
several endocrine markers (Scheffer et al., 2003). However, in
our routine clinical setting, we found a high inter-observer varia-
bility of antral follicle count due to the duration of the vaginal
ultrasound examination, observers experience and, notably,
expected ovarian reserve due to patient’s age. We therefore
excluded the ‘first line antral follicle count’ from further analysis
and focused on objectively assessed endocrine markers avoiding
observer related variability. Although we found a strong corre-
lation of AMH levels and subsequent ovarian response in ART-
cycles by using the cut-off level of AMH 1.26 ng/ml, we only
have a correct prediction rate of 36%. By using groups of com-
parable age, the specificity and ‘correct prediction rate of poor
response’ will increase significantly, although sensitivity and
‘correct prediction rate of normal response’ are much more
robust and will remain unchanged whether or not a high
age-related heterogeneity is given. Therefore, AMH measure-
ment and a subsequent antral follicle count should be combined
in unselected groups of patients, typical for a clinical setting, to
minimize false positive results, as has also been proposed by
others previously (Muttukrishna et al., 2005; Tremellen et al.,
2005). But the antral follicle count should be performed accu-
rately as a confirmatory test only in the second line and not as
a screening method.
According to the current literature, existing assessments of
ovarian reserve have limited predictive properties especially
for the occurrence of pregnancy (Broekmans et al., 2006).
Newest investigations see AMH as a superior predictor of
Figure 1: Flowchart of 316 ART patients with 132 oocyte retrievals and 119 completed cycles (means and standard deviations).
Gnoth et al.
1362
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
IVF success (live birth rate) (Nelson et al., 2007) which cannot
be supported by our findings. In our study, AMH and the other
examined markers were not predictive for the success of the
subsequent IVF treatment, confirming the results of others
(van Rooij et al., 2006). However, a case wise analysis of all
IVF treatments in this study revealed that in all cases with an
AMH 1.26 ng/ml, a dose adjustment after 5 days of gonado-
trophin stimulation was necessary. In this way, we could
obviously increase the number of retrieved oocytes and
improve treatment outcome. This bias may also explain the
low correct prediction rates for poor response in the ‘AMH
only’ analysis (Table I).
Recent studies have shown that a low response to gonado-
trophin stimulation is associated with early menopause (Nikolaou
et al., 2002; Lawson et al., 2003). AMH also seems to be a prom-
ising candidate for the occurrence of menopausal transition
(van Rooij et al., 2004). Our results show that AMH levels
1.26 ng/ml are strongly indicative of significantly diminished
ovarian reserve and AMH levels ,0.5 ng/ml robustly and
correctly predict a very low response (2 oocytes) in 88% of
cases, supporting the value of AMH as an early warning test for
exhausted ovarian function when FSH levels are still within the
normal range (Feyereisen et al., 2006). The low correlation of
inhibin B with poor ovarian response as well as with the
success of a subsequent IVF treatment may be explained by the
fact that inhibin B not only represents the number of antral fol-
licles, but is closely related to functional changes of the follicles
during the cycle and with advancing age (Knight and Glister,
2006).
Concluding from our results, we now use AMH as a reliable
pre-warning and screening marker for reduced ovarian reserve
in women as previously suggested by others (van Rooij et al.,
2005). In considera tion of our results, we generally adapted our
treatment strategies for all patients from 30 June 2006 onwards.
Regardless of age, women with an AMH .1.26 ng/ml receive
a starting dose of 150 mU/ml recombinant FSH or HMG.
Patients with an AMH ,0.5 ng/ml are started with 375 mU/
ml recombinant FSH or HMG and counseled about their
impaired prognosis, independently of age. In cases of AMH
levels 0.5 ng/ml and 1.26 ng/ml, the start dosage is fixed
Figure 2: Receiver operating characteristic (ROC)-curves for AMH as an indicator for poor ovarian response and clinical pregnancy.
Table I. Detection and prediction of poor responders (n ¼ 33 with 4 oocytes, calculated cut-off AMH 1.26 ng/ml) in 132 oocyte retrievals.
Observations Age, FSH, inhibin B, AMH (n ¼ 60) Age, FSH, AMH (n ¼ 120) Age, AMH (n ¼ 132) AMH (n ¼ 132)
Sensitivity (true positive rate) 0.79 0.69 0.76 0.97*
Specifity (true negative rate) 0.83 0.66 0.65 0.41*
Correct prediction ‘poor response’ 0.58 0.42 0.42 0.36*
Correct prediction ‘normal response’ 0.93 0.85 0.88 0.98*
Stepwise discriminant analysis of age, FSH, inhibin B and AMH.
*If, in a further analysis with a calculated cut-off of AMH 1.26 ng/ml, the poor and normal responder groups are adjusted for age, specificity increases to
62% and the correct prediction rate ‘poor response’ to 88%. Sensitivity and correct prediction rate of ‘normal response’ remain unchanged. If levels of AMH
are , 0.5, there is 88% correct prediction of 2 oocytes whether or not groups are adjusted for age.
AMH and ART
1363
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
to 300 mU/ml recombinant FSH or HMG if an antral follicle
count showed ,5 follicles/ovary (Kline et al., 2005).
Women with an AMH 0.1 ng/ml are treated only in excep-
tional cases (trial stimulation) and seriously informed about
their bad prognosis (Meduri et al., 2007) and the probably cost-
ineffective stimulation. Women with an AMH .6ng/ml are at
risk of polyfollicular development and ovarian hyperstimula-
tion (Pigny et al., 2003, 2006) and are therefore started pru-
dently with 100 mU/ml recombinant FSH. Future studies
will be analyzed to see if a further decrease in cycle cancella-
tion rate because of impending hyperstimulation or insufficient
ovarian response to gonadotrophin stimulation is possible
although in the latter case, past studies have shown disappoint-
ing results (Klinkert et al., 2005).
In summary, our results confirm the relevance of AMH
measurement as an important screening test for reduced
ovarian reserve in women. Using AMH alone with a cut-off
level 1.26 ng/ml, we will identify 97% of all women with
a reduced ovarian reserve and correctly predict low response
to gonadotrophin stimulation in 88% of cases in groups of com-
parable age. Furthermore, a level of AMH , 0.5 ng/ml cor-
rectly predicts very poor response with 2 oocytes in 88%
of cases and poor outcome even in groups of high age-related
heterogeneity. If AMH levels are 0.5 ng/ml and 1.26 ng/
ml, an antral follicle count should be added to exclude false
positive results and increase specificity especially in a routine
clinical setting with heterogeneous groups of patients.
Chronological age fairly well predicts ovarian response,
upcoming perimenopausal and menopausal transition. Import-
ant studies have shown that AMH is a predictor for the occur-
rence of perimenopausal transition (cycle length irregularity)
within 3–5 years (de Vet et al., 2002; van Rooij et al., 2004)
at levels ,0.92 ng/ml. Our results show that by early detection
of women with a reduced ovarian response, using a cut-off
level 1.26 ng/ml, treatment options are still available
(high-dose stimulation in IVF, polyovulation) with favorable
pregnancy rates, not statistically different from those women
with ‘normal’ ovarian reserve. Therefore, females should be
advised to get AMH levels determined if they do not achieve
pregnancy after six cycles of regular intercourse (Gnoth
et al., 2005). AMH is suitable as a screening test and may
replace FSH which gains relevance only in the late reproduc-
tive phase.
Acknowledgements
The authors wish to thank Mrs Silvia Heil for recording and preparing
the row data and Mrs Andreja Daniels, Mrs Claudia Hammelstein and
Mrs Claudia Monissen for all her efforts performing the laboratory
essays.
Conflict of interest: The first author was supported by small grant for
statistical analysis (software adaptation) by Diagnostic System Lab-
oratories, Inc./Beckman-Coulter, Germany.
Funding
Small grant for statistical analysis (software adaption) by Diag-
nostic System Laboratories, Inc./Beckman-Coulter, Germany.
References
Broekmans FJ, Kwee J, Hendriks DJ, Mol BW, Lambalk CB. A systematic
review of tests predicting ovarian reserve and IVF outcome. Hum Reprod
Update 2006;12:685– 718.
Burger HG, Dudley EC, Hopper JL, Groome N, Guthrie JR, Green A,
Dennerstein L. Prospectively measured levels of serum follicle-stimulating
hormone, estradiol, and the dimeric inhibins during the menopausal
transition in a population-based cohort of women. J Clin Endocrinol
Metab 1999;84:4025– 4030.
Burger HG, Robertson DM, Baksheev L, Collins A, Csemiczky G, Landgren
BM. The relationship between the endocrine characteristics and the
regularity of menstrual cycles in the approach to menopause. Menopause
2005;12:267–274.
Table III. Logistic regression analysis for low and normal ovarian response in ART cycles and for AMH using 1.26 ng/ml as a cut-off level.
Observations Groups
low responder
(n ¼ 33)
normal responder
(n ¼ 99)
Significance low
versus normal
AMH 1.26 ng/ml
(n ¼ 91)
AMH.1.26 ng/ml
(n¼ 41)
Significance AMH 1.26
versus AMH . 1.26
Ages 39.2 + 3.1 36.5 + 3.6 P , 0.05 38.0 + 3.0 36.0 + 4.0 P , 0.05
Clin. Preg. rate 12% (n ¼ 24) 42% (n ¼ 94) P , 0.05 35% (n ¼ 79) 41% (n ¼ 39 ns
AHM 0.32 + 0.33 2.20 + 3.1 P , 0.05
FSH 9.94 + 6.39 7.34 + 3.27 P , 0.05 8.67 + 4.9 6.36 + 2.0 P , 0.05
Retrieved
oocytes
2.0 + 1.0 10.0 + 5.91 P , 0.05 6.0 +
4.0 14.1 + 6.7 P , 0.05
V
alues are given in mean + standard deviations.
Table II. Prediction of ‘no pregnancy’ in 119 complete ART cycles (cut-off: AMH 1.8 ng/ml, n ¼ 63 unsuccessful cycles).
Observations Age, FSH, inhibin B, AMH (n ¼ 58) Age, FSH (n ¼ 108) Age, AMH (n ¼ 119) AMH (n ¼ 119)
Sensitivity (true positive rate) 0.50 0.52 0.60 0.83
Specifity (true negative rate) 0.54 0.60 0.64 0.34
Correct prediction ‘not pregnant’ 0.60 0.60 0.60 0.58
Correct prediction ‘pregnant’ 0.43 0.52 0.56 0.63
Stepwise discriminant analysis of age, FSH, inhibin B and AMH.
Gnoth et al.
1364
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
de Vet A, Laven JS, de Jong FH, Themmen AP, Fauser BC. Antimullerian
hormone serum levels: a putative marker for ovarian aging. Fertil Steril
2002;77:357–362.
Ebner T, Sommergruber M, Moser M, Shebl O, Schreier-Lechner E, Tews G.
Basal level of anti-Mullerian hormone is associated with oocyte quality in
stimulated cycles. Hum Reprod 2006;21:2022–2026.
Fanchin R, Taieb J, Lozano DH, Ducot B, Frydman R, Bouyer J. High
reproducibility of serum anti-Mullerian hormone measurements suggests a
multi-staged follicular secretion and strengthens its role in the assessment
of ovarian follicular status. Hum Reprod 2005;20:923– 927.
Fawcett T. ROC Graphs: Notes and Practical Considerations for
Researchers, 2004. http://home.comcast.net/~tom.fawcett/public_html/
papers/ROC101.pdf.
Feyereisen E, Mendez Lozano DH, Taieb J, Hesters L, Frydman R, Fanchin R.
Anti-Mullerian hormone: clinical insights into a promising biomarker of
ovarian follicular status. Reprod Biomed Online 2006;12:695– 703.
Ficicioglu C, Kutlu T, Baglam E, Bakacak Z. Early follicular antimullerian
hormone as an indicator of ovarian reserve. Fertil Steril 2006;85:592–596.
Fleming R, Deshpande N, Traynor I, Yates RW. Dynamics of FSH-induced
follicular growth in subfertile women: relationship with age, insulin
resistance, oocyte yield and anti-Mullerian hormone. Hum Reprod
2006;21:1436–1441.
Freour T, Mirallie S, Colombel A, Bach-Ngohou K, Masson D, Barriere P.
Anti-mullerian hormone: clinical relevance in assisted reproductive
therapy. Ann Endocrinol (Paris) 2006;67:567–574.
Freour T, Mirallie S, Bach-Ngohou K, Denis M, Barriere P, Masson D.
Measurement of serum anti-Mullerian hormone by Beckman Coulter
ELISA and DSL ELISA: comparison and relevance in assisted
reproduction technology (ART). Clin Chim Acta 2007;375:162–164.
Gnoth C, Frank-Herrmann P, Freundl G, Godehardt D, Godehardt E. Time to
pregnancy: results of the German prospective study and impact on the
management of infertility. Hum Reprod 2003;18:1959– 1966.
Gnoth C, Godehardt E, Frank-Herrmann P, Friol K, Tigges J, Freundl G.
Definition and prevalence of subfertility and infertility. Hum Reprod
2005;20:1144–1147.
Gougeon A, Ecochard R, Thalabard JC. Age-related changes of the population
of human ovarian follicles: increase in the disappearance rate of
non-growing and early-growing follicles in aging women. Biol Reprod
1994;50:653–663.
Gruijters MJ, Visser JA, Durlinger AL, Themmen AP. Anti-Mullerian hormone
and its role in ovarian function. Mol Cell Endocrinol 2003;211:85–90.
Hehenkamp WJ, Looman CW, Themmen AP, de Jong FH, te Velde ER,
Broekmans FJ. Anti-Mullerian hormone levels in the spontaneous
menstrual cycle do not show substantial fluctuation. J Clin Endocrinol
Metab 2006;91:4057 –4063.
Kline J, Kinney A, Kelly A, Reuss ML, Levin B. Predictors of antral
follicle count during the reproductive years.
Hum Reprod 2005;20:
2179
–2189.
Klinkert ER, Broekmans FJ, Looman CW, Habbema JD, te Velde ER.
Expected poor responders on the basis of an antral follicle count do not
benefit from a higher starting dose of gonadotrophins in IVF treatment: a
randomized controlled trial. Hum Reprod 2005;20:611– 615.
Knight PG, Glister C. TGF-beta superfamily members and ovarian follicle
development. Reproduction 2006;132:191–206.
La Marca A, Stabile G, Artenisio AC, Volpe A. Serum anti-Mullerian hormone
throughout the human menstrual cycle. Hum Reprod 2006;21:
3103–3107.
La Marca A, Giulini S, Tirelli A, Bertucci E, Marsella T, Xella S, Volpe A.
Anti-Mullerian hormone measurement on any day of the menstrual cycle
strongly predicts ovarian response in assisted reproductive technology.
Hum Reprod 2007;22:766– 771.
Lawson R, El Toukhy T, Kassab A, Taylor A, Braude P, Parsons J, Seed P. Poor
response to ovulation induction is a stronger predictor of early menopause
than elevated basal FSH: a life table analysis. Hum Reprod 2003;18:
527–533.
Meduri G, Massin N, Guibourdenche J, Bachelot A, Fiori O, Kuttenn F, Misrahi
M, Touraine P. Serum anti-Mullerian hormone expression in women with
premature ovarian failure. Hum Reprod 2007;22:117– 123.
Muttukrishna S, McGarrigle H, Wakim R, Khadum I, Ranieri DM, Serhal P.
Antral follicle count, anti-Mullerian hormone and inhibin B: predictors of
ovarian response in assisted reproductive technology? BJOG
2005;112:1384–1390.
Nelson SM, Yates RW, Fleming R. Serum anti-Mullerian hormone and FSH:
prediction of live birth and extremes of response in stimulated cycles
implications for individualization of therapy. Hum Reprod 2007;22:2414–
2421.
Nikolaou D, Lavery S, Turner C, Margara R, Trew G. Is there a link between an
extremely poor response to ovarian hyperstimulation and early ovarian
failure? Hum Reprod 2002;17:1106–1111.
Nikolaou D, Templeton A. Early ovarian ageing: a hypothesis: detection and
clinical relevance. Hum Reprod 2003;18:1137–1139.
Pigny P, Merlen E, Robert Y, Cortet-Rudelli C, Decanter C, Jonard S, Dewailly
D. Elevated serum level of anti-mullerian hormone in patients with
polycystic ovary syndrome: relationship to the ovarian follicle excess and
to the follicular arrest. J Clin Endocrinol Metab 2003;88:5957–5962.
Pigny P, Jonard S, Robert Y, Dewailly D. Serum anti-Mullerian hormone as a
surrogate for antral follicle count for definition of the polycystic ovary
syndrome. J Clin Endocrinol Metab 2006;91:941– 945.
Scheffer GJ, Broekmans FJ, Dorland M, Habbema JD, Looman CW, te Velde
ER. Antral follicle counts by transvaginal ultrasonography are related to age
in women with proven natural fertility. Fertil Steril 1999;72:845–851.
Scheffer GJ, Broekmans FJM, Looman CWN, Blankenstein M, Fauser BCJM,
de Jong FH, te Velde ER. The number of antral follicles in normal women
with proven fertility is the best reflection of reproductive age. Hum
Reprod 2003;18:700– 706.
Schipper I, de Jong FH, Fauser BC. Lack of correlation between maximum
early follicular phase serum follicle stimulating hormone concentrations
and menstrual cycle characteristics in women under the age of 35 years.
Hum Reprod 1998;
13:1442
–1448.
Smeenk JM, Sweep FC, Zielhuis GA, Kremer JA, Thomas CM, Braat DD.
Antimullerian hormone predicts ovarian responsiveness, but not embryo
quality or pregnancy, after in vitro fertilization or intracyoplasmic sperm
injection. Fertil Steril 2007;87:223–226.
te Velde ER, Pearson PL. The variability of female reproductive ageing. Hum
Reprod Update 2002;8:141– 154.
Tremellen KP, Kolo M, Gilmore A, Lekamge DN. Anti-mullerian hormone as a
marker of ovarian reserve. Aust N Z J Obstet Gynaecol 2005;45:20 –24.
van Rooij IA, Tonkelaar I, Broekmans FJ, Looman CW, Scheffer GJ, de Jong
FH, Themmen AP, te Velde ER. Anti-mullerian hormone is a promising
predictor for the occurrence of the menopausal transition. Menopause
2004;11:601–606.
van Rooij IA, Broekmans FJ, Scheffer GJ, Looman CW, Habbema JD, de Jong
FH, Fauser BJ, Themmen AP, te Velde ER. Serum antimullerian hormone
levels best reflect the reproductive decline with age in normal women with
proven fertility: a longitudinal study. Fertil Steril 2005;83:979–987.
van Rooij IA, Broekmans FJ, Hunault CC, Scheffer GJ, Eijkemans MJ, de Jong
FH, Themmen AP, te Velde ER. Use of ovarian reserve tests for the
prediction of ongoing pregnancy in couples with unexplained or mild male
infertility. Reprod Biomed Online 2006;12:182– 190.
Visser JA, Themmen AP. Anti-Mullerian hormone and folliculogenesis. Mol
Cell Endocrinol 2005;234:81– 86.
Visser JA, de Jong FH, Laven JS, Themmen AP. Anti-Mullerian hormone: a
new marker for ovarian function. Reproduction 2006;131:1–9.
Submitted on August 21, 2007; resubmitted on March 3, 2008; accepted on
March 12, 2008
AMH and ART
1365
by guest on May 30, 2013http://humrep.oxfordjournals.org/Downloaded from
