Hum. Reprod.

Advance Access published March 23, 2015

Human Reproduction, Vol.0, No.0 pp. 1 10, 2015
doi:10.1093/humrep/dev060

ORIGINAL ARTICLE Reproductive endocrinology

Ovarian reserve after treatment with

alkylating agents during childhood
Cecile Thomas-Teinturier 1,2,*, Rodrigue Setcheou Allodji 2,
Ekaterina Svetlova 3, Marie-Alix Frey 3, Odile Oberlin3,
Anne-Elodie Millischer 4, Sylvie Epelboin 5, Christine Decanter 6,

Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
Helene Pacquement7, Marie-Dominique Tabone 8,
Helene Sudour-Bonnange 9, Andre Baruchel 10, Najiba Lahlou 11,
and Florent De Vathaire 2,12
1
Department of Paediatric Endocrinology, APHP, Hopital de Bicetre, rue du General Leclerc, Le Kremlin Bicetre F-94270, France 2Radiation
Epidemiology Group, Centre for Research in Epidemiology and Population Health (CESP)INSERM U1018, Institut Gustave Roussy, rue
Edouard Vaillant Villejuif F-94805, France 3Department of Paediatric Oncology, Institut Gustave Roussy, rue Edouard Vaillant Villejuif F-94805,
France 4Department of Paediatric Radiology, APHP, Hopital Necker, rue de Sevres Paris F-75015, France 5Department of Gynaecology, APHP,
Hopital Bichat, rue Henri Huchart, Paris F-75018, France 6Fertility Preservation Unit, Lille University, Hopital Jeanne de Flandre, avenue Eugene
Avinee Lille F-59000, France 7Department of Paediatric Oncology, Institut Curie, rue dUlm, Paris F-75005, France 8Department of Paediatric
Oncology, APHP, Hopital Trousseau, avenue du Dr Arnold Netter, Paris F-75012, France 9Department of Paediatric Oncology, Centre Oscar
Lambret, rue Frederic Combemale Lille F-59000, France 10Department of Paediatric Hematology, APHP, Hopital Robert Debre, boulevard
Serurier, Paris F-75019, France 11Laboratory of Hormonal Biology, APHP, Hopital Cochin, rue du faubourg Saint-Jacques, Paris F-75005, France
12
University Paris-Sud XI, Villejuif F-94800, France

*Correspondence address. Tel: +33-145-21-78-52; Fax: +33-1-45-21-78-50; E-mail: cecile.teinturier@bct.aphp.fr

Submitted on August 1, 2014; resubmitted on February 16, 2015; accepted on February 23, 2015

study question: What is the effect of different alkylating agents used without pelvic radiation to treat childhood cancer in girls on the
ovarian reserve in survivors?
summary answer: Ovarian reserve seems to be particularly reduced in survivors who received procarbazine (in most cases for Hodgkin
lymphoma) or high-dose chemotherapy; procarbazine but not cyclophosphamide dose is associated with diminished ovarian reserve.
what is known already: A few studies have demonstrated diminished ovarian reserve in survivors after various combination ther-
apies, but the individual role of each treatment is difcult to assess.
study design: Prospective cross-sectional study, involving 105 survivors and 20 controls.
participants/materials, setting, methods: One hundred and ve survivors aged 17 40 years and 20 controls investigated
on Days 25 of a menstrual cycle or Day 7 of an oral contraceptive pill-free interval. Main outcome measures: ovarian surface area (OS), total
number of antral follicles (AFC), serum levels of follicle-stimulating hormone (FSH), luteinizing hormone (LH), estradiol and anti-Mullerian
hormone (AMH).
main results and the role of chance: Survivors had a lower OS than controls: 3.5 versus 4.4 cm2 per ovary (P 0.0004), and
lower AMH levels: 10.7 versus 22 pmol/l (P 0.003). Ovarian markers (OS, AMH, AFC) were worse in patients who received high-dose com-
pared with conventional-dose alkylating agents (P 0.01 for OS, P 0.002 for AMH, P , 0.0001 for AFC). Hodgkin lymphoma survivors
seemed to have a greater reduction in ovarian reserve than survivors of leukaemia (P 0.04 for AMH, P 0.01 for AFC), sarcoma (P 0.04
for AMH, P 0.04 for AFC) and other lymphomas (P 0.04 for AFC). A multiple linear regression analysis showed that procarbazine but
not cyclophosphamide nor ifosfamide dose was associated with reduced OS (P 0.0003), AFC (P 0.0007), AMH (P , 0.0001) and higher
FSH levels (P , 0.0001).
limitations, reasons for caution: The small percentage of participating survivors (28%) from the total cohort does not allow
conclusion on fertility issues because of possible response bias. The association between procarbazine and HL makes it impossible to dissociate
their individual impacts on ovarian reserve. The number of controls is small, but ovarian volume and AMH levels in survivors were compared with
published normal values and results were unchanged.

& The Author 2015. 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@oup.com
2 Thomas-Teinturier et al.
wider implications of the findings: Early detection and follow-up of compromised ovarian function after cancer therapy should
help physicians to counsel young survivors about their fertility window. However, longitudinal follow-up is required to determine the rate of pro-
gression from low ovarian reserve to premature ovarian failure.
study funding/competing interest(s): La Ligue contre le Cancer (grant no., PRAYN7497). The authors have no competing
interests to disclose.
Key words: ovarian reserve / AMH / chemotherapy / childhood cancer survivors / hodgkin lymphoma
Introduction
The overall increase in cancer prevalence and the signicant increase in
long-term survival rates have generated worldwide interest in proce-
dures that may preserve fertility in adolescents and young women
exposed to gonadotoxic chemotherapy (Gatta et al., 2014). Gonado-
toxicity is a well-known side effect of alkylating agents such as cyclophos-
phamide, procarbazine, melphalan, busulfan and thiotepa (Whitehead
et al., 1983; Teinturier et al., 1998; Couto-Silva et al., 2001). Acute
ovarian failure is rare in survivors of childhood cancer. It was estimated
as 6% in the Childhood Cancer Survivor Study (Chemaitilly et al.,
2006) and is even lower in the absence of pelvic radiation therapy and
high-dose chemotherapy regimens given before haematopoietic stem
cell transplantation. Nevertheless, it has been shown that ovarian
reserve is lower and that the risk of premature menopause is higher in
women who were treated with alkylating agents than in the age-matched
general population (Bath et al., 2003; Larsen, Muller, Rechnitzer et al.,
2003; Larsen, Muller, Schmiegelow et al., 2003; van Beek et al., 2007;
Lie Fong et al., 2009). The reported incidence of premature menopause,
i.e. before 40 years of age, in this population varies from 2% in our study
to 8% in the Childhood Cancer Survivor Study (Byrne et al., 1992;
Chiarelli et al., 1999; Sklar et al., 2006; Thomas-Teinturier et al., 2013).
However, it will probably increase in future years given the use of inten-
sive chemotherapy over the past 15 years. In addition to infertility, which
requires assisted reproductive techniques (in vitro fertilization with a
womans own oocytes or with oocyte donation), premature menopause
is associated with an increased incidence of osteoporosis, and higher
morbidity and mortality relating to cardiovascular disease (Shuster
et al., 2010). In survivors, it is difcult to predict the extent of reproduct-
ive impairment and the window of fertility for family planning. Although
assessment of ovarian reserve can predict response and pregnancy in in-
fertile women undergoing assisted reproductive technology, there are
few data on its usefulness for counselling young cancer survivors
(Larsen, Muller, Rechnitzer et al., 2003; Van Beek et al., 2007; Lie Fong
et al., 2009).
Our aim was to assess the ovarian reserve of follicles in a large homo-
geneous cohort of childhood cancer survivors treated with alkylating
agents without pelvic or total body irradiation or busulfan/thiotepa, to
determine the degree of ovarian damage and risk factors associated
with such damage.
Materials and Methods
Subjects
Approval for the study was obtained from the local medical research commit-
tee, and written informed consent was obtained from all participants. Sample
size calculation was performed for ovarian volume based on a mean (+SD)
ovarian volume of 6.55 + 1.39 ml in normal women aged 18 39 years
(Kelsey et al., 2013) and anti-Mullerian hormone (AMH) levels based on
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
our normal values, median 20.6 pmol/l (5th 95th percentile 8.5 and 59.8)
(Dewailly et al., 2010). A total of 72 patients should allow us to identify
with a power of 80%, a reduction of 1.3 ml in ovarian volume corresponding
to the observed reduction between 18 and 40 years of age in the normal
population. A total of 48 patients should allow us to identify with a power
of 80%, a difference of 10 pmol/l in AMH levels when accepting a risk
alpha of 0.05. So, the power of this study with the actual numbers recruited
(n 105) was 93.3% to identify a reduction of 1.3 ml in ovarian volume and
99.2% to identify a difference of 10 pmol/l in AMH levels with the population
averages when accepting a risk alpha of 0.05.
The les from ve centres computer databases (Institut Gustave Roussy,
Institut Curie, Hopital Trousseau, Hopital Saint-Louis, Paris and Centre
Oscar Lambret, Lille) were screened to identify all patients who met all of
the following criteria: aged 17 40 years, living in or near Paris or Lille,
without known ovarian failure, having received alkylating agents during child-
hood for a sarcoma, neuroblastoma, lymphoma, acute leukaemia or other
tumours with .3 years follow-up after the end of treatment, and none of
the following exclusion criteria: ovarian tumours, brain or pelvic radiother-
apy, or treatment with busulfan or thiotepa. A letter was sent to the 408 iden-
tied survivors fullling the above criteria. One hundred and forty-ve
women agreed to participate but only 108 performed the assessment. Of
these, three were excluded from the analysis, because their evaluation
either diagnosed a polycystic syndrome or was done during the second
week of the contraceptive pill (Fig. 1).
Controls
Twenty healthy controls without any known ovarian problem, aged 15 34
years, with regular menstrual cycle length before oral contraceptive pills
use were recruited by means of advertisement in Saint Vincent de Paul hos-
pital. They had a median age of 21.5 years and thus were slightly younger than
patients, the median age of whom was 25 years (17 40) (P 0.009), and had
a lower body mass index (BMI) (20.5 versus 21.4 kg/m2, P 0.02). Seven
controls were using oral contraceptive pills. Because our control group is
small and younger than the patient group, we also compared patients
median ovarian volume to an age-matched validated normative model of
ovarian volume (median 6.3 ml) published by Kelsey and patients
median AMH levels to normal values (median 20.6 pmol/l) published by
Dewailly (Dewailly et al., 2010; Kelsey et al., 2013).
Experimental design
Patients and controls not using an oral contraceptive pill were evaluated
during the early follicular phase of the menstrual cycle (cycle Day 2 5).
The oral contraceptive pill users were asked to stop their pill for at least
1 month before ovarian assessment. Those who did not want to stop their
pill were asked to take a 30 mg Ethinyl-Estradiol monophasic pill for at least
1 month and were evaluated on Day 7 of the pill-free interval where gonado-
trophin levels have been reported to be not different from their early follicular
Ovarian reserve in childhood cancer survivors
Figure 1 Flow chart of survivors.
phase levels (Cohen and Katz, 1981). Transvaginal or transabdominal (if ne-
cessary) ultrasound was performed to measure the volume of both ovaries,
and the size and number of antral follicles (AFC). Any nding of pelvic abnor-
mality (uterus, Fallopian tubes) was also recorded. On the same morning,
blood samples were collected for hormonal evaluation: follicle-stimulating
hormone (FSH), luteinizing hormone (LH) and estradiol serum levels were
analysed in the endocrine laboratory of Hopital Cochin, except in the case
of 11 women whose samples were analysed in the endocrine laboratory of
Lille Hospital using similar methods. All blood samples were then centrifuged
3
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
at 1300 g for 10 min at 48C and frozen (2208C) for storage and then ana-
lysed in the same run for AMH.
Hormonal assays
FSH and LH were measured as previously described using a time-resolved
uorometric assay with Dela reagents (PerkinElmer Life Sciences, Courta-
boeuf, France). In the FSH assay, intra- and inter-series coefcients of vari-
ation were at the level 3.1 IU/l 1.2 and 3.9%, respectively, and at the level
4 Thomas-Teinturier et al.

16.6 IU/l 1.5 and 2.8%, respectively. In the LH assay, the intra- and inter-
series coefcients of variation were at the level 0.3 IU/l 1.4 and 2.6%, re-
spectively, and at the level 6.9 IU/l 1.7 and 2.0%, respectively. The sensitivity
for both assays was 0.01 IU/l. The estradiol level was measured as previously
described in a direct radioimmunoassay using Clinical Assays reagents (Dia-
sorin, Antony, France). Intra- and inter-series coefcients of variation at the
level of 96 pmol/l were 5.8 and 2.4%, respectively. The sensitivity was
7 pmol/l. Serum AMH levels were assessed using the second-generation
enzyme immunoassay AMH-EIA (Immunotech, Beckman Coulter, Marseille,
France). Intra- and inter-assay coefcients of variation were ,12.3 and
14.2%, respectively and sensitivity was 2.5 pmol/l.
Ultrasound
All transvaginal ultrasound measurements were performed by one of three
age at cancer diagnosis was 9.3 years (range: 0.04 17.7) and the median
age at evaluation was 25 years (range: 17 40.7). Five survivors (5%) had
been treated for acute lymphoblastic leukaemia (ALL), 21 (20%) for
Hodgkin lymphoma (HL), 15 (14%) for non-Hodgkin lymphoma
(NHL), 21 (20%) for bone sarcoma, 16 (15%) for soft tissue sarcoma,
23 (22%) for neuroblastoma and 4 (4%) for other tumours (three
Wilms tumours and one malignant germ cell tumour). All patients had
received alkylating agents as part of their chemotherapy, and the cumu-
lative doses of cyclophosphamide, ifosfamide and procarbazine are
shown in Table I. Among 71 patients who received cyclophosphamide,
median dose was 4.6 g/m2 and only 12 received 10 g/m2 or more.
Among 33 patients who received ifosfamide, median dose was 48 g/m2
and 20 received 40 g/m2 or more. Nineteen survivors (18%) had re-

trained investigators (A.E.M., S.E., C.D.) using a 5 9 MHz transvaginal
probe on a Voluson E8 expert, GE. Twenty-two survivors and 14 controls
who were uncomfortable with transvaginal sonography had a transabdominal
sonography. An antral follicular count (AFC) was performed by assessing the
number of follicles measuring 2 5 and 5 10 mm in both the right and left
ovaries. Furthermore, the surface area of each ovary and the ovarian
volume were estimated using the formulas 0.8 length width and d1
d2 d3 p/6 where d1, d2, d3 are the three maximal longitudinal, antero-
posterior and transverse diameters. The ovarian surface (OS) area and
volume were recorded as the mean value of the right and left ovary.
Ovarian volume was available only in 79 patients and not in controls for tech-
nical reasons.
Treatments
The medical records of all survivors were examined, and the cumulative dose
of each chemotherapeutic agent was calculated in milligrams per square
metre. Radiotherapy was recorded and the irradiation eld separated into
subdiaphragmatic, susdiaphragmatic, limb or cervical. None of the patients
underwent oophorectomy.
Statistical analysis
The results are presented as the median plus range. Correlations between
variables were tested using Spearmans correlation analysis (r correlation
coefcient). Comparisons between two independent groups were per-
formed using the Wilcoxon rank test for patients and control groups and
with the sign test for the median value observed in patients versus published
normal values. The Kruskal Wallis rank test was performed to analyse the
variance of three or more groups, and multiple comparison tests with the
Bonferroni correction were performed for the gures. Stepwise, multiple,
linear regression analysis was performed to evaluate the most important
factors inuencing the variations in AFC, OS area, AMH and FSH serum
levels. To make the variables better t the assumptions underlying regression,
the distributions of AFC, OS area, AMH and FSH serum levels were normal-
ized by log transformation. The independent variables studied were age at
diagnosis, pubertal status at diagnosis, each drug (yes or no) and its cumula-
tive dose, subdiaphragmatic irradiation (yes or no), high-dose chemotherapy
with stem cell transplant (yes or no), age at evaluation, oral contraceptive pill
use and modality of ultrasound (transvaginal versus transabdominal).
P-values of ,0.05 were considered signicant. All analyses were performed
using SAS software for Windows version 9.3.
Results
One hundred and ve female survivors of childhood cancer, aged 1740
years (median: 25 years) were assessed. The main characteristics of the
patients and cancer therapies received are shown in Table I. The median
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
ceived subdiaphragmatic radiotherapy (median dose: 20 Grays),
mostly in the lombo-aortic or splenic area for Hodgkin disease. Fourteen
survivors underwent autologous stem cell transplantation after a condi-
tioning regimen of alkylating chemotherapy without busulfan, thiotepa or
total body irradiation. Most survivors (62%) had been treated before
puberty, 13% during puberty and 25% after rst menses. Among the
70 survivors not taking the oral contraceptive pill, only one suffered
from oligomenorrhea. Ultrasound revealed an ovarian cyst in seven sur-
vivors who were subsequently withdrawn from the analysis of ovarian
size and AFC.
Thirty-ve out of 105 survivors and 7 out of 20 controls were taking an
estrogen-containing oral contraceptive. Analysis of the main characteris-
tics and results in survivors and controls taking or not taking an estrogen-
containing oral contraceptive revealed no statistical difference between
the two groups. We analysed ovarian markers stratied by oral con-
traceptive use or not, and then without this stratication, and the
results remained unchanged, so the gures show the results in the
whole population.
Comparison of ovarian markers between survivors and controls
showed a lower OS area: 3.5 versus 4.4 cm2 per ovary (P 0.0004)
and lower AMH levels: 10.7 versus 22 pmol/l (P 0.003) in survivors
(Table II). In addition, median ovarian volume in survivors (3.76 ml)
was statistically signicantly different from the median ovarian volume
of 6.3 ml reported in age-matched normal population (P , 0.0001),
and median AMH in survivors was also signicantly different from the
normal median AMH value of 20.6 pmol/l (P , 0.0001). Only eight
patients had evidence of altered ovarian function at the time of evalu-
ation, as diagnosed by FSH . 15 UI/l at least twice and AMH ,
3.6 pmol/l and amenorrhoea after stopping their pill.
In survivors, AFC was correlated with AMH (r 0.69, P , 0.0001),
the number of small follicles ,5 mm (r 0.66, P , 0.0001) and to a
lesser extent OS area (r 0.42, P , 0.0001). AMH was also correlated
with OS area (r 0.45, P , 0.0001) and negatively with FSH
(r 20.57, P , 0.0001) (Supplementary data, Table SI).
Survivors were divided into three groups based on the treatment they
receivedalkylating agents alone, alkylating agents and subdiaphrag-
matic radiotherapy or high-dose alkylating agents. In all therapeutic
groups, survivors had a lower OS area and AMH levels compared with
the controls (Fig. 2). This was conrmed by comparative analysis with
published normal values. Ovarian markers (OS, AMH, AFC) were
worse in patients given high-dose alkylating agents compared with
those treated with a conventional dose of alkylating agents (P 0.01
for OS area, P 0.002 for AMH and P , 0.0001 for AFC). The total
number of antral follicles was signicantly reduced only in these survivors
 Ovarian reserve in childhood cancer survivors
Table I Characteristics of the whole population studied.

Median age Pubertal status at cancer Median Age at evaluation by BMI at Subdiaphragmatic cyclo dose ifo dose Pcb dose HDCT
at cancer diagnosis (n) age at year class (n) evaluation radiotherapy (n) median median median (n)
diagnosis
...................................... evaluation
............................... (range) (range) (range) (range)
Before During After 18 25 30 35
(range) (range) (kg/m2) (g/m2) (g/m2) (g/m2)
puberty puberty rst 25 30 35 40
(year) (year)
menses
..........................................................................................................................................................................................................................................................
Whole 9.3 (0.0417.7) 65 14 26 25 (17 40.7) 53 36 6 10 21.4 (1642.7) 19 4.6 (1 22) 48 (3 104) 3 (0.3 9.1) 14
population, n 71 n 33 n 23
n 105
By cancer type
ALL, n 5 8.0 (3.7 15.5) 4 0 1 24.3 (21 26) 4 1 0 0 21.1 (18.632.8) 0 0.9 (0.9 5.6) 0 0 0
n5
HL, n 21 13.4 (8.617.7) 6 5 10 24.4 (19 38) 12 7 0 2 21.2 (18 26.7) 12 2.7 (1 4.1) 14.9 (14.5 15.3) 3 2.2 9.1) 2
n 11 n2 n 20
NHL, n 15 9.7 (0.9 17.4) 9 2 4 25.5 (19 33) 7 4 4 0 23.7 (18.428.8) 0 4.8 (3 10.3) 0 2.4 1
n 15 n1
Bone sarcoma, 11.5 (3.816.7) 7 6 8 24 (18 37) 13 7 0 1 20.5 (17 42.7) 1 10.1 (3.111.8) 58.6 (3.1 104) 0 0
n 21 n 11 n 16
Soft tissue 6.3 (0.1 16.9) 12 1 3 27.5 (19 40) 4 7 1 4 22.3 (19 27.7) 0 9.3 (2.1 22) 48.2 (26.5 60.7) 3.8 (0.37.3) 1
sarcoma, n 16 n5 n 12 n2
Neuroblastoma, 1.3 (0.04 8.2) 23 0 0 25 (17 37) 10 9 1 3 21.1 (16 34.7) 5 4.7 (1.2 19.7) 0 0 6
n 23 n 23
Other, n 4 4.4 (1.5 7.8) 4 0 0 20 (17 25) 3 1 0 0 21.1 (20.123.4) 1 9.2 12 (12 12) 0 4
n1 n3

cyclo, cyclophosphamide; ifo, ifosfamide; pcb, procarbazine; HDCT, high-dose chemotherapy.

5
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
6 Thomas-Teinturier et al.

Table II Main results in survivors versus controls and normal population.

n Patients Controls P Normal values P

105 20
.............................................................................................................................................................................................
Median age (years) and range 25 (1740.7) 21.5 (15 34) 0.009 (18 39)
2
Median BMI (kg/m ) 21.4 (1642.7) 20.5 (17.2 22.5) 0.02
Median ovarian volume (ml) 3.8 (0.79.9) NA 6.3 ,0.0001
Median OS per ovary (cm2) 3.5 (1.17.1) 4.4 (2.1 10.3) 0.0004
Median total antral follicular count 12 (140) 11 (8 24) 0.8
Median FSH (IU/l) 6.2 (2.152.6) 5.8 (3.5 11) 0.1
Median AMH (pmol/l) 10.7 (098) 22 (3.3 47) 0.003 20.6 ,0.0001
5th 95th pc (8.3 59.8)

Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
Median LH (IU/l) 3.7 (0.530.3) 3.3 (0.5 6.9) 0.4

Ovarian volume was compared with median ovarian volume reported by Kelsey et al. (2013) for women aged 1839 years and AMH levels with normal values reported by Dewailly et al. (2010).

Figure 2 Distribution of ovarian markers: (a) OS area, (b) antral follicle count, (c) AMH level and (d) FSH level, according to treatment received. Central
line indicates the median of the data, the circle the mean, the bottom and top of the box the 25th and 75th percentiles. The whiskers represent the outliers.
Alkyl, alkylating agents alone; Alkyl + RT, alkylating agents and subdiaphragmatic radiotherapy; BMT, high-dose chemotherapy.

(P 0.0004). An increase in FSH levels over controls was observed only We then studied ovarian markers according to the ve types of
in patients who had received alkylating agents and subdiaphragmatic cancerALL, HL, NHL, sarcoma and neuroblastoma with miscellan-
radiotherapy (12/19 have been treated for HL) (P 0.0009). eous tumours (Fig. 3). Although AMH levels were signicantly lower in
Ovarian reserve in childhood cancer survivors
Figure 3 Distribution of ovarian markers: (a) OS area, (b) antral follicle count, (c) AMH level and (d) FSH level, according to cancer diagnosis. Central line
indicates the median of the data, the circle the mean, the bottom and top of the box the 25th and 75th percentiles. The whiskers represent the outliers. ALL,
acute lymphoblastic leukaemia; HL, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma; NRB, neuroblastoma.
survivors of all types of cancer other than ALL and sarcoma compared
with the controls, OS area was signicantly reduced only in HL, neuro-
blastoma and sarcoma survivors. This was conrmed by comparative
analysis with published normal values. The somewhat reduced AFC in
HL survivors did not reach signicance compared with the controls
(P 0.07). HL survivors had a signicantly lower ovarian reserve than
other survivors as demonstrated by the lower AMH levels as well as
AFC and OS area, and higher FSH levels, compared with survivors of
ALL (P 0.04 for AMH, P 0.02 for FSH, P 0.01 for AFC),
sarcoma (P 0.04 for AMH, P 0.001 for FSH, P 0.04 for AFC,
P 0.02 for OS area), NHL (P 0.004 for FSH, P 0.04 for AFC,
P 0.04 for OS area). Neuroblastoma survivors had a reduced OS
area and antral follicular count compared with ALL (P 0.05 and
0.009) and sarcoma survivors (P 0.006 and 0.04).
To identify factors inuencing the variations in AFC, OS area, AMH and
FSH serum levels, a linear regression analysis was performed (Supplemen-
tary data, Table SII). A stepwise, multiple, linear regression analysis was
performed controlling for the use of oral contraceptive pill and modalities
of ovarian sonography for OS area and AFC. In the nal model explaining
2442% of the variance depending on the tested variable, predictors were
age at evaluation, procarbazine dose and a history of high-dose alkylating
agents (Table III). The use of an oral contraceptive pill as a factor of low
AMH serum levels was on the borderline of statistical signicance (P
0.05). Age or pubertal status at treatment, subdiaphragmatic irradiation
7
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
and the total cumulative dose of cyclophosphamide or ifofasmide were
not predictive factors for markers of ovarian reserve. As procarbazine
therapy was strongly associated with HL (20/23 patients who received
procarbazine were treated for HL and 20/21 patients with HL received
procarbazine (Table I)), it seems impossible to separate the effects of
this drug and the type of cancer.
Discussion
This is the largest study with a detailed analysis of markers of ovarian
reserve in a population of survivors of childhood cancer who all received
alkylating agents without direct pelvic radiotherapy and with no known
ovarian failure. Only 8/105 (8%) patients were diagnosed with
unknown altered ovarian function at the time of evaluation (FSH .
15 UI/l at least twice, AMH , 3.6 pmol/l and amenorrhoea after stop-
ping oral contraceptive use). However, only 16 women aged .30 years
agreed to participate, 2 of which had premature ovarian failure (13%).
Their limited number is not sufcient to draw conclusions regarding
the incidence of premature ovarian failure in this population. Reasons
for non-participation in this study in the case of women aged . 30
may bias the results, for example, they may already have children or con-
versely may have known ovarian failure.
Comparison of markers of ovarian reserve between controls and sur-
vivors showed signicant impairment in the size of the ovaries and AMH
8 Thomas-Teinturier et al.

arising mostly in adolescents. This could be potential confounder in

Table III Stepwise multivariate linear regression these studies, because an older age at diagnosis of cancer may be
analysis of predictive factors of AMH and FSH levels, total
linked with procarbazine treatment.
antral follicular count and mean OS area.
Diminished ovarian reserve in survivors of childhood cancer has
Variables Coefcient b SE P-value already been described in smaller series (Bath et al., 2003; Larsen,
........................................................................................ Muller, Rechnitzer et al., 2003; Larsen, Muller, Schmiegelow et al.,
AMH levels (R 2 0.33) 2003; Van Beek et al., 2007; Gracia et al., 2012; El-Shalakany et al.,
Age at evaluation 20.063 0.017 0.0004 2013; Krawczuk-Rybak et al., 2013; Lunsford et al., 2014). These
Oral contraceptive use 20.373 0.189 0.05 series have several limitations: small sample size in some cases, inclusion
Procarbazine dose (g/m2) 20.019 0.004 ,0.0001 of subjects with various diagnoses and heterogeneous therapies making it
High-dose chemotherapy 21.143 0.276 ,0.0001 impossible to assess the effect of each chemotherapeutic agent and
FSH levels (R 2 0.42) dosage, and inclusion of subjects with known ovarian failure or non-
Age at evaluation 0.018 0.007 0.01 extensive analysis of ovarian markers. Larsen reported that survivors

Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
had a smaller ovarian volume (4.8 versus 6.8 ml) and a lower number
Oral contraceptive use 0.140 0.081 0.09
2 of antral follicles per ovary (7.5 versus 11) than controls. Ovarian irradi-
Procarbazine dose (g/m ) 0.012 0.002 ,0.0001
ation, alkylating agents, older age at diagnosis and longer time off treat-
High-dose chemotherapy 0.197 0.114 0.09
ment were associated with a reduced AFC (Larsen, Muller,
Total antral follicular count (R 2 0.39) Schmiegelow et al., 2003). Gracia reported an impaired ovarian
Age at evaluation 20.041 0.010 0.0001 reserve in a heterogeneous prospective cohort of survivors (n 71),
Oral contraceptive use 20.085 0.108 0.4 with AMH levels and AFC seeming to be the most sensitive measures
Procarbazine dose (g/m2) 20.007 0.002 0.0007 for quantifying damage to the ovaries (Gracia et al., 2012). Lie Fong
High-dose chemotherapy 20.777 0.165 ,0.0001 et al. (2009) reported only AMH serum levels in 185 female survivors
Ovarian sonography 0.340 0.126 0.008 of childhood cancer, but the serum samples were taken randomly
procedures (transvaginal during the menstrual cycle, and data were studied without taking into
versus transabdominal) account the use of oral contraceptive pills and were compared with con-
Mean OS area (R 2 0.24) trols not taking contraceptive pills. However, there is controversy sur-
Age at evaluation 0.002 0.005 0.7 rounding the possible reduction in AMH levels due to oral
Oral contraceptive use 20.080 0.056 0.1 contraceptive pill use in healthy volunteers and the same trend was on
Procarbazine dose (g/m2) 20.004 0.001 0.0003 the borderline of statistical signicance in our patients (P 0.05) (Van
High-dose chemotherapy 20.266 0.081 0.002 den Berg et al., 2010; Bentzen et al., 2012; Kristensen et al., 2012). Fur-
Ovarian sonography 20.046 0.065 0.5 thermore, in cancer survivors, Charpentier et al. (2014) found that use of
procedures (transvaginal versus oral contraceptive pills remained signicantly associated with lower
transabdominal) AMH in their multivariate model. Therefore, this factor could have
increased the difference in AMH levels between survivors and controls
in this study. Moreover, AMH levels were not measured using the
same assay in survivors and controls and even if results were adjusted
levels. No signicant difference was detected in AFC between survivors for comparison this could have biased the global results.
and controls, although there was a signicant difference between survi- Early detection and follow-up of compromised ovarian function after
vors of ALL and HL. This could be due to the high percentage of trans- cancer therapy should help physicians to counsel young survivors about
abdominal ultrasound in our controls (70%), which is associated with a their fertility window and guide fertility preservation strategies when ne-
lower AFC compared with transvaginal ultrasound, because of its cessary. Nevertheless, the question that remains unanswered is whether
lower resolution. We were able to compare the effect of specic che- a reduction in ovarian reserve truly reects future fertility impairment and
motherapeutic regimens on ovarian reserve in this study. This revealed predicts earlier age at menopause in this population. Some ndings and
that only procarbazine treatment (and not cyclophosphamide or ifofas- our own experience suggest that predictions about future fertility in sur-
mide treatment) was associated with a diminished ovarian reserve. But it vivors of childhood cancer based only on one measurement of ovarian
is worth noting that most patients received ,10 g/m2 cyclophospha- reserve are not adequate. Indeed, Dillon et al. (2013) reported that
mide. In the absence of high-dose alkylating agents therapy, HL survivors despite diminished ovarian reserve markers, pregnancies were achieved
seemed to have a greater reduction in ovarian reserve than other survi- in some survivors with a mean time to conception that seemed longer
vors connected to procarbazine dose. Nowadays most HL patients than in controls, although signicance was not reached because of the
receive protocols without procarbazine with less ovarian damage (Van small numbers involved: 8.6 months (range 028) compared with 3.1
Beek et al., 2007). months (07) (P 0.3). It is likely that measures of ovarian reserve
Age and pubertal status at the time of treatment were not associated provide more information about the quantity of ovarian follicles available
with diminished ovarian reserve. These ndings are consistent with those rather than their quality in young women. Additional studies are needed
of Lie Fong et al. (2009) but not with those of Charpentier et al. (2014) to better understand the predictive value of these measures for preg-
and Larsen, Muller, Schmiegelow et al. (2003). However, procarbazine nancy rate and time to menopause in this population of young survivors
dose has not been tested as a covariate in the multivariate model in (Ledger, 2010; Loh and Maheshwari, 2011; Nelson et al., 2012). Longi-
these last two studies, and procarbazine is used in the treatment of HL tudinal measurement of ovarian reserve in these patients will be required
Ovarian reserve in childhood cancer survivors
to study the rate of ovarian reserve decline, pregnancy rate, time taken to
achieve pregnancy and age at menopause to answer this question.
The results of this study bear out that the patients most at risk of pre-
mature ovarian failure are those who are going to receive high-dose alkyl-
ating agents therapy and the majority of the others will probably have an
opportunity for natural fertility in the absence of direct pelvic radiation.
This could help oncologists to offer fertility preservation, in particular
the experimental option of ovarian tissue cryopreservation in young girls.
This study has some limitations. First, our control group had a younger
median age than the survivors (21.5 versus 25 years), and this could have
increased the difference in ovarian markers between the two groups.
Nevertheless, when adjusting for age at evaluation, the results remained
unchanged. In addition, the number of controls is small, so to strengthen
the analysis ovarian volume and AMH levels in survivors were compared
with published normal values, and results were absolutely the same. The
small percentage of participating survivors (28%) from the total cohort
does not allow conclusion on fertility issues, because this cohort may
not be representative of the population of survivors, as participating
women could be more concerned about their fertility than non-
participants. The strong association between procarbazine and HL
(20/23 patients who received procarbazine were treated for HL)
makes it impossible to dissociate their individual impacts on ovarian
reserve. Oral contraceptive use may modify hormonal results but as
the same percentage of survivors and controls were taking an oral
contraceptive, this factor could not have modied the global results.
Moreover, multivariate analysis was adjusted according to the use of
an oral contraceptive.
This study has several strengths. An extensive evaluation of ovarian
reserve was performed and hormone variability was minimized by
obtaining early follicular phase measures and by centralizing dosages in
only two laboratories. The studied population was selected according
to the use of alkylating agents without pelvic or brain radiotherapy to min-
imize the heterogeneity of treatments received on the basis of known
ovarian toxicity.
In conclusion, this study shows that ovarian reserve seems to be
reduced particularly in survivors who received procarbazine (most of
them for HL) or high-dose alkylating agents before stem cell transplant-
ation. Procarbazine but not cyclophosphamide or ifosfamide dose is
associated with diminished ovarian reserve. Currently, counselling
young patients with a diminished ovarian reserve after cancer cure is
like looking into the crystal ball. Longitudinal follow-up is required to de-
termine the rate of progression from low ovarian reserve to premature
ovarian failure.
Supplementary data
Supplementary data are available at http://humrep.oxfordjournals.org/.
Authors roles
C.T.-T. and R.S.A. had full access to the data and take the responsibility of
the integrity of the data and the accuracy of the data analysis. Study
concept and design: C.T.-T., R.S.A., M.-A.F., O.O. and F.D.V.. Execu-
tion: C.T.-T., M.-A.F., E.S., A.-E.M., S.E., C.D., O.O., H.P., M.-D.T.,
H.S.-B., A.B. and N.L. Validation of data: C.T.-T., M.-A.F., E.S. and
H.S.-B. Analysis and interpretation of the data: C.T.-T., R.S.A. and
O.O. Manuscript writing: C.T.-T. Critical review of the manuscript for
9
important intellectual content: R.S.A., E.S., M.-A.F., A.-E.M., S.E.,
C.D., O.O., H.P., M.-D.T., H.S.-B., A.B., N.L. and F.D.V. Statistical ana-
lysis: C.T.-T., R.S.A. and F.D.V. Obtaining funding: C.T.-T. and O.O.
Funding
This study was nancially supported by La Ligue contre le Cancer (grant
no. PRAYN7497). The funding agency had no role in the design and
conduct of the study; nor in the collection, management, analysis and in-
terpretation of the data; nor in the preparation, review and approval of
the manuscript.
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
Conict of interest
None declared.
References
Bath LE, Wallace WHB, Shaw MP, Fitzpatrick C, Anderson RA. Depletion of
ovarian reserve in young women after treatment for cancer in childhood:
detection by anti-Mullerian hormone, inhibin B and ovarian ultrasound.
Hum Reprod 2003;18:2368 2374.
Bentzen JG, Forman JL, Pinborg A, Lidegaard O, Larsen EC, Friis-Hansen L,
Johannsen TH, Nyboe Andersen A. Ovarian reserve parameters: a
comparison between users and non-users of hormonal contraception.
Reprod Biomed Online 2012;25:612 619.
Byrne J, Fears TR, Gail MH, Pee D, Connelly RC, Austin DF, Holmes GF,
Holmes FF, Latourette HB, Meigs JW et al. Early menopause in
long-term survivors of cancer during adolescence. Am J Obstet Gynecol
1992;166:788 793.
Charpentier AM, Lee Chong A, Gingras-hill G, Ahmed S, Cigsar C,
Gupta AA, Greenblatt E, Hodgson DC. Anti-Mullerian hormone
screening to assess ovarian reserve among female survivors of childhood
cancer. J Cancer Surviv 2014;8:548 554.
Chemaitilly W, Mertens AC, Mitby P, Whitton J, Stovall M, Yasui Y,
Robison LL, Sklar CA. Acute ovarian failure in the Childhood Cancer
Survivor Study. J Clin Endocrinol Metab 2006;91:1723 1728.
Chiarelli AM, Marrett LD, Darlington G. Early menopause and infertility in
females after treatment for childhood cancer diagnosed in 1964 1988
in Ontario, Canada. Am J Epidemiol 1999;150:245 254.
Cohen BL, Katz M. Further studies on pituitary and ovarian function in women
receiving hormonal contraception. Contraception 1981;24:159 172.
Couto-Silva AC, Trivin C, Thibaud E, Esperou H, Michon J, Brauner R. Factors
affecting gonadal function after bone marrow transplantation during
childhood. Bone Marrow Transplant 2001;28:67 75.
Dewailly D, Pigny P, Soudan B, Catteau-Jonard S, Decanter C, Poncelet E,
Duhamel A. Reconciling the denitions of polycystic ovary syndrome:
the ovarian follicule number and serum anti-Mullerian hormone
concentrations aggregate with markers of hyperandrogenism. J Clin
Endocrinol Metab 2010;95:4399 4405.
Dillon KE, Sammel MD, Ginsberg JP, Lechtenberg L, Prewitt M, Gracia CR.
Pregnancy after cancer: results from a prospective cohort study of
cancer survivors. Pediatr Blood Cancer 2013;60:2001 2006.
El-Shalakany AH, Ali MS, Abdelmaksoud AA, El-Ghany SA, Hasan EA.
Ovarian function in female survivors of childhood malignancies. Pediatric
Hematol Oncol 2013;30:328 335.
Gatta G, Botta L, Rossi S, Aareleid T, Bielska-Lasota M, Clavel J, Dimitrova N,
Jakab Z, Kaatsch P, Lacour B et al. Childhood cancer survival in Europe
1999 2007: results of EUROCARE-5-a population based study. Lancet
Oncol 2014;15:35 47.
10
Gracia CR, Sammel MD, Freeman E, Prewitt M, Carlson C, Ray A, Vance A,
Ginsberg JP. Impact of cancer therapies on ovarian reserve. Fertil Steril
2012;97:134 140.
Kelsey TW, Dodwell SK, Wilkinson AG, Greve T, Andersen CY,
Anderson RA, Wallace WHB. Ovarian volume throughout life: a
validated normative model. PLoS ONE 2013;8:e71465.
Krawczuk-Rybak M, Leszczynska E, Poznanska M, Zelazowska-Rutkowska B,
Wysocka J. Anti-Mullerian hormone as a sensitive marker of ovarian
function in young cancer survivors. Int J Endocrinol 2013;2013:125080.
Kristensen SL, Ramlau-Hansen CH, Andersen CY, Ernst E, Olsen SF, Bone JP,
Vested A, Toft G. The association between circulating levels of
antimullerian hormone and follicular number, androgens, and menstrual
cycle characteristics in young women. Fertil Steril 2012;97:779 785.
Larsen EC, Muller J, Rechnitzer C, Schmiegelow K, Andersen AN.
Diminished ovarian reserve in female childhood cancer survivors with
regular menstrual cycles and basal FSH,10 IU/l. Hum Reprod 2003;
18:417 422a.
Larsen EC, Muller J, Schmiegelow K, Rechnitzer C, Andersen AN. Reduced
ovarian function in long-term survivors of radiation- and chemotherapy-
treated childhood cancer. J Clin Endocrinol Metab 2003;88:5307 5314b.
Ledger WL. Clinical utility of measurement of anti-Mullerian hormone in
reproductive endocrinology. J Clin Endocrinol Metab 2010;95:5144 5154.
Lie Fong S, Laven JSE, Hakvoort-Cammel FGAJ, Schipper I, Visser JA,
Themmen APN, de Jong FH, van den Heuvel-Eibrink MM. Assessment
of ovarian reserve in adult childhood cancer survivors using
anti-Mullerian hormone. Hum Reprod 2009;24:982 990.
Loh JS, Maheshwari A. Anti-Mullerian hormone is it a crystal ball for
predicting ovarian ageing? Human Reprod 2011;26:2925 2932.
Lunsford AJ, Whelan K, McCormick K, McLaren JF. Antimullerian hormone
as a measure of reproductive function in female childhood cancer
survivors. Fertil Steril 2014;101:227 231.
Thomas-Teinturier et al.
Nelson SM, Anderson RA, Broekmans FJ, Raine-Fenning N, Fleming R, La
Marca A. Anti-Mullerian hormone: clairvoyance or crystal clear? Human
Reprod 2012;27:631 636.
Shuster LT, Rhodes DJ, Gostout BS, Grossardt BR, Rocca WA. Premature
menopause or early menopause: long term health consequences.
Maturitas 2010;65:161 166.
Sklar CA, Mertens AC, Mitby P, Whitton J, Stovall M, Kasper C, Mulder J,
Green D, Nicholson HS, Yasui Y. Premature menopause in survivors of
childhood cancer: a report from the childhood cancer survivor study.
J Natl Cancer Institute 2006;98:890 896.
Teinturier C, Hartmann O, Valteau-Couanet D, Benhamou E, Bougneres PF.
Ovarian function after autologous bone marrow transplantation in
childhood: high-dose busulfan is a major cause of ovarian failure. Bone
Marrow Transplant 1998;22:989 994.
Downloaded from http://humrep.oxfordjournals.org/ at Medical Center Library, Duke University on March 31, 2015
Thomas-Teinturier C, El Fayech C, Oberlin O, Pacquement H, Haddy N,
Labbe M, Veres C, Guibout C, Diallo I, De Vathaire F. Age at menopause
and its inuencing factors in a cohort of survivors of childhood cancer:
earlier but rarely premature. Hum Reprod 2013;28:488495.
Van Beek RD, Van den Heuvel-Eibrink MM, Laven JSE, de Jong FH,
Themmen APN, Hakvoort-Cammel FG, van den Bos C, van den Berg H,
Pieters R, de Muinck Keizer-Schrama SMPF. Anti-Mullerian hormone is a
sensitive serum marker for gonadal function in women treated for Hodgkins
lymphoma during childhood. J Clin Endocrinol Metab 2007;92:38693874.
Van den Berg MH, van Dulmen-den Broeder E, Overbeek A, Twisk JWR,
Schats R, van Leeuwen FE, Kaspers GJ, Lambalk CB. Comparison of
ovarian function markers in users of hormonal contraceptives during the
hormone-free interval and subsequent natural early follicular phases.
Human Reprod 2010;25:1520 1527.
Whitehead E, Shalet SM, Blackledge G, Todd I, Crowther D, Beardwell CG.
The effect of combination chemotherapy on ovarian function in women
treated for Hodgkins disease. Cancer 1983;52:988 993.