Full text of DOI:10.1093/humrep/15.1.72

Human Reproduction vol.15 no.1 pp.72–75, 2000
Ovarian response to repeated controlled stimulation in
in-vitro fertilization cycles in patients with ovarian
endometriosis
Majedah Al-Azemi, Andr e´ s L o´ pez Bernal, Jo Steele,
Isabelle Gramsbergen, David Barlow and
Stephen Kennedy
and associated additional causes of infertility, e.g. impaired
sperm function.
1
In-vitro fertilization (IVF) and embryo transfer is a treatment
option that is increasingly offered to couples irrespective of
the severity of disease present. However, it is unclear whether
the presence of endometriosis adversely affects pregnancy and
live birth rates, and which mechanisms are responsible if
rates are lowered; it has been suggested that women with
endometriosis have a lower ovarian response to gonadotro-
phins. Early reports suggested that women with endometriosis
undergoing ovarian stimulation in IVF–embryo transfer cycles
with clomiphene citrate alone or in combination with human
menopausal gonadotrophin (HMG) had poor ovarian response,
reduced oocyte quality, and low fertilization and pregnancy
rates (Wardle et al., 1985; Yovich et al., 1985). Subsequently,
however, outcomes following IVF–embryo transfer have been
reported to be similar in women with endometriosis compared
to those with infertility due to other causes, irrespective of the
severity of disease (Dmowski et al., 1995; Geber et al., 1995;
Olivennes et al., 1995). Other investigators have observed
differences in ovarian response in women with advanced
endometriosis, which in effect means ovarian, cystic disease
Nuffield Department of Obstetrics and Gynaecology, University of
Oxford, Oxford Radcliffe Hospital, Women’s Centre, Oxford, OX3
9
DU, UK
1To whom correspondence should be addressed
In-vitro fertilization (IVF) is an effective infertility treat-
ment for women with endometriosis, but most women need
to undergo several cycles of treatment to become pregnant.
This case-control study was designed to assess how consist-
ently women with ovarian endometriosis respond to ovarian
stimulation in consecutive treatment cycles compared to
women with tubal infertility. We compared outcome meas-
ures in 40 women with a history of surgically confirmed
ovarian endometriosis and 80 women with tubal infertility,
all of whom had at least three IVF treatment cycles. The
groups were matched for age and early follicular follicle
stimulating hormone (FSH) concentration at their first IVF
cycle. Outcome measures included number of follicles,
number of oocytes, peak oestradiol concentration and
number of FSH ampoules required per follicle. Cumulative
pregnancy and live birth rates were calculated in both
groups. The ovarian endometriosis group had a significantly
poorer ovarian response and required significantly more
ampoules of FSH per cycle, a difference that became
greater with each subsequent cycle. However, cumulative
pregnancy (63.3 versus 62.6% by fifth cycle) and live birth
(Chillik et al., 1985; O’Shea et al., 1985; Matson and Yovich,
1986; Oehninger et al., 1988; Pellicer et al., 1995). The
reduced ovarian response observed in these studies resulted in
fewer oocytes being available for fertilization as well as higher
miscarriage rates, suggesting that oocyte quality may also be
compromised by advanced disease.
Most couples who require IVF–embryo transfer need to
undergo several treatment cycles to achieve a pregnancy. They
continue having treatment despite the enormous financial and
personal costs, because they believe that the cumulative
pregnancy and live birth rates reported by most IVF centres
justify the expense. It would be important to advise women
with ovarian endometriosis if their chances of success really
are less than optimal. However, to the best of our knowledge,
there are no data on ovarian response in consecutive cycles in
women with ovarian disease.
(46.8 versus 50.9% by fifth cycle) rates were similar in
both groups. In conclusion, despite decreased ovarian
response to FSH, ovarian endometriosis does not decrease
the chances of successful IVF treatment.
Key words: endometriosis/IVF/ovarian endometriosis/ovarian
stimulation
We therefore performed a comparative analysis of a compu-
terized database to address two questions: firstly, whether
poorer ovarian responses are obtained during consecutive IVF–
embryo transfer cycles in women with ovarian endometriosis
compared to those with tubal disease and normal ovaries, and
secondly whether ovarian responsiveness affects the overall
clinical pregnancy and live birth rates in these groups.
Introduction
Endometriosis may be a factor in the infertility of up to 80%
of couples seeking medical assistance to conceive (Wardle and
Hull, 1993). The wide range of reported prevalence rates arises
primarily because the definition of what constitutes minimal
peritoneal disease has changed over time and it is this type of
disease, rather than more severe endometriosis involving the
ovaries, that is most commonly seen in infertile women.
Treatment options depend upon factors such as the severity of
endometriosis, the woman’s age, the duration of infertility,
Materials and methods
Population
Patient data have been collected prospectively in the Oxford IVF unit
since 1987 and stored in a relational database (Paradox, Borland,
72
© European Society of Human Reproduction and Embryology

Ovarian response to controlled stimulation in endometriosis
Scott Valley, CA, USA). Using this database, we identified all women
with surgically proven ovarian endometriosis, who had ജ3 IVF/
intracytoplasmic sperm injection (ICSI)–embryo transfer cycles
between 1989 and 1998. This group (n ϭ 40) was compared to 80
controls with tubal disease (all identified by laparoscopy) as the sole
cause of their infertility who had IVF/ICSI–embryo transfer in the
same time period. Controls were obtained by identifying the nearest
patient in the database with the appropriate characteristics preceding
and following each index case. Controls were matched with cases
for age (Ϯ2 years) and early follicular phase follicle stimulating
hormone (FSH) concentration (Ϯ2 IU/l). All the women in both
groups had two ovaries at the time of their first IVF cycle. Women
with polycystic ovaries were excluded from the case and control
groups. It should be noted that reliable data regarding medical
treatment between IVF cycles were not available.
Table I. Numbers of IVF cycles proceeding to oocyte recovery and cycles
cancelled either due to excessive or poor response in the control and
endometriosis groups
Cycle no.
1
2
3
4
5
Control
R
L
H
78
0
2
79
0
1
77
1
2
33
1
0
14
0
0
Treatment protocol
The standard IVF treatment protocol used in Oxford has been
described previously (Lockwood et al., 1995). In brief, all patients had
a long protocol cycle with luteal phase start using the gonadotrophin-
releasing hormone (GnRH) agonist, nafarelin (Synarel; Searle, High
Wycombe, UK; 400 µg intranasal twice a day) followed by gonadotro-
Endometriosis
R
L
H
31
7
2
37
1
2
37
2
1
23
0
0
12
0
0
P-value
Ͻ0.01
NS
NS
NS
NS
R ϭ numbers of cycles leading to oocyte recovery; L ϭ numbers of cycles
cancelled due to poor ovarian response; H ϭ numbers of cycles cancelled
due to excessive ovarian response.
2
NS ϭ not significant (χ test).
Table II. Ovarian surgery rates prior to IVF treatment in the endometriosis
and control groups
Cycle no.
1
2
3
4
5
Control
n
80
2
80
2
80
2
43
2
14
2
Ovarian surgery
Endometriosis
n
Ovarian surgery
P-value
40
10
Ͻ0.05
40
3
NS
40
2
NS
23
3
NS
12
1
NS
Figure 1. Ovarian response in terms of number of oocytes
retrieved in endometriosis (triangles) and control (squares) groups.
Results are means Ϯ SEM. The asterisks indicate significant
differences between groups. *P Ͻ 0.05; **P Ͻ 0.01; ***P Ͻ
0.001.
n ϭ number of patients.2
NS ϭ not significant (χ test).
Table III. Ovarian response in terms of number of follicles aspirated, number of oocytes retrieved, peak
oestradiol concentration (pmol/l) and number of ampoules used per follicle
Cycle no.
1
2
3
4
5
No. of patients
C
E
78
31
79
37
77
38
33
23
14
11
No. of follicles aspirated (n)
C
E
10.6 Ϯ 0.6
10.5 Ϯ 1.0
11.5 Ϯ 0.6
9.7 Ϯ 1.0*
12.1 Ϯ 0.7
10.2 Ϯ 0.9
11.6 Ϯ 0.9
7.4 Ϯ 1.0**
12.7 Ϯ 1
9 Ϯ 1.6
No. of oocytes retrieved (n)
C
E
7.1 Ϯ 0.5
6.9 Ϯ 0.7
8.1 Ϯ 0.5
6.4 Ϯ 0.7*
8.8 Ϯ 0.5
6.6 Ϯ 0.7**
8.3 Ϯ 0.7
4.1 Ϯ 0.5***
8.6 Ϯ 0.8
5.8 Ϯ 0.9*
Peak oestradiol concentration (µmol/l)
C
E
4287 Ϯ 303
4505 Ϯ 551
4144 Ϯ 246
3485 Ϯ 335
4312 Ϯ 240
4163 Ϯ 431
4022 Ϯ 342
2306 Ϯ 367** 3794 Ϯ 600
5087 Ϯ 639
Ampoules per follicle (n)
C
E
3.1 Ϯ 0.3
4.5 Ϯ 1.0
3.7 Ϯ 0.3
7.7 Ϯ 1.6*
4.4 Ϯ 0.7
12.1 Ϯ 3.5**
4.0 Ϯ 0.6
16 Ϯ 4.0**
3.6 Ϯ 0.7
9.4 Ϯ 2.3
Results are means Ϯ SEM; C ϭ controls; E ϭ endometriosis group.
Differences between control and endometriosis groups (t-test): *P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.001.
73

M.Al-Azemi et al.
Figure 4. Cumulative live birth rates in the endometriosis
triangles) and control (squares) groups.
(
Figure 2. Number of human menopausal gonadotrophin (HMG)
ampoules required per follicle aspirated in the endometriosis
defined as an oestradiol concentration Ͼ8000 pmol/l with ജ15
follicles that were ജ12 mm in diameter (Forman et al., 1990).
(
triangles) and control (squares) groups. Results are means Ϯ SEM.
The asterisks indicate significant differences between groups. *P Ͻ
.05; **P Ͻ 0.01.
0
Outcome measures
The main aim of the study was to compare the ovarian responses
over consecutive treatment cycles in the endometriosis and control
groups. The following parameters of ovarian response were analysed:
(i) number of follicles aspirated at oocyte recovery, (ii) number of
oocytes retrieved, (iii) oestradiol concentration on the day of HCG
administration and (iv) number of gonadotrophin ampoules used per
follicle aspirated. In addition, cumulative pregnancy and live birth
rates were calculated in both groups. The definition of a clinical
pregnancy was the presence of an intrauterine fetal heartbeat on
transvaginal ultrasound scanning 4 weeks post-embryo transfer.
Statistical analysis
The data were tabulated and analysed using Prism and Instat
(GraphPad software, San Diego, CA, USA). Contingency tables,
analysis of variance (ANOVA), t-tests and regression analysis were
used as appropriate.
Results
At the time of the first IVF treatment cycle, there were no
significant differences in age (33 Ϯ 4.2 versus 33.3 Ϯ
Figure 3. Cumulative pregnancy rates in the endometriosis
triangles) and control (squares) groups.
(
4.1 years; means Ϯ SEM) or early follicular phase FSH
concentration (5.1 Ϯ 1.9 versus 5.1 Ϯ 2 IU/l) between the
endometriosis and control groups respectively. The mean
numbers of treatment cycles undergone were 3.9 Ϯ 0.9 and
3.5 Ϯ 0.7 in the endometriosis and control groups respectively.
The mean numbers of months between treatments were: cycles
phin stimulation with highly purified human urinary FSH (Metrodin
HP; Serono, Welwyn Garden City, UK). The dose in the initial
treatment cycle was determined on an individual basis according to
age, body mass index, and follicular phase FSH concentration. Upon
ultrasonographic (Pie Medical Equipment; Netherlands) detection of
at least three follicles ജ16 mm in diameter, 10 000 IU of human
chorionic gonadotrophin (HCG) (Profasi; Serono) was administered.
Oocyte retrieval was performed transvaginally using intravenous
pethidine and diazepam for sedation and analgesia. The practice in
the unit is to aspirate all follicles ജ12 mm diameter. Embryo transfer
was performed ~48 h after the oocyte retrieval. An IVF cycle was
defined as the start of gonadotrophin stimulation; in most cases the
cycle proceeded to oocyte retrieval, but in some cases the cycle was
cancelled because of poor or excessive ovarian response. Poor
response was defined by the presence of less than two 16 mm follicles
after at least 10 days ovarian stimulation. Excessive response was
1
(
–2 (endometriosis ϭ 9.5; controls ϭ 11.8); cycles 2–3
endometriosis ϭ 14.0; controls ϭ 10.2); cycles 3–4 (endomet-
riosis ϭ 10.5; controls ϭ 8.7) and cycles 4–5 (endometriosis ϭ
1
2
0.4; controls ϭ 11.9). There were two cancellations (2.5%,
/80) in the first cycle for excessive response in the control
group and nine cancellations (11.3%, 9/40) (two for excessive,
and seven for poor, responses) in the endometriosis group (P
Ͻ 0.01). In subsequent cycles (cycles 2–5) there were no
statistically significant differences in cancellation rates between
the endometriosis and control groups. The numbers of IVF
74

Ovarian response to controlled stimulation in endometriosis
cycles proceeding to oocyte recovery, and the numbers can-
celled due to excessive or poor response in the endometriosis
and control groups are shown in Table I. The cancelled
cycles were excluded from the ovarian response calculations.
Significantly more endometriosis patients (10/40; 25%) had
had ovarian surgery before the first cycle compared to controls
poor response to HMG is a consequence of the loss of healthy
tissue occupied by the disease, or the result of deleterious
paracrine effects of endometriotic tissue on the intra-ovarian
mechanisms of follicle selection and oocyte maturation. Fur-
thermore, significantly more women in the endometriosis group
had undergone ovarian surgery prior to their first IVF cycle
and it is possible that surgical damage to the ovaries comprom-
ised their subsequent response to stimulation. Another reason
for only including women with ovarian disease in this paper
is that, in conducting a case-control study, it is important to
have clear objective differentiation between the study and
control groups. This is easily achieved at surgery and with
ultrasound in cases with ovarian disease but would be difficult
with endometriotic disease confined to the peritoneum.
In conclusion, the outcome of IVF treatment in patients
with ovarian endometriosis is as good as in women with tubal
disease alone. However, these patients require higher doses of
HMG for ovarian stimulation and the cost of treatment to
achieve pregnancy is higher.
(2/80; 2.5%) (P Ͻ 0.05), but there were no differences between
the groups in the ovarian surgery rates in subsequent cycles
(Table II).
The data relating to the ovarian response in terms of number
of follicles, number of oocytes, oestradiol concentration on
the day of HCG administration and the number of ampoules
required per follicle aspirated are presented in Table III. In
the control group, the ovarian response and gonadotrophin
requirements remained relatively constant and there were no
significant differences for any of the above parameters from
cycles 1–5. However, the ovarian response in the endometriosis
group in terms of number of follicles aspirated and number of
oocytes retrieved was poorer and was significantly decreased
by cycle number 4 (P Ͻ 0.05 and P Ͻ 0.01 respectively).
The difference in the number of oocytes obtained between
the endometriosis and control groups became increasingly
significant with subsequent cycles (Figure 1). The total number
of HMG ampoules required per follicle was significantly higher
in the endometriosis compared to the control group (Table III).
The requirement for HMG increased with subsequent cycles
in the endometriosis (P Ͻ 0.01) but not in the control group
References
Ahmed-Ebbiary, N.A., Morgan, C., Martin, K. et al. (1995) Ovarian response
in consecutive cycles of ovarian stimulation in normally ovulating women.
Hum. Reprod., 10, 536–543.
Chillik, C.F., Acosta, A.A., Garcia, J.E. et al. (1985) The role of in vitro
fertilization in infertile patients with endometriosis. Fertil. Steril., 44, 56–61.
Dmowski, W.P., Rana, N., Michalowska, J. et al. (1995) The effect of
endometriosis, its stage and activity, and of autoantibodies on in vitro
fertilization and embryo transfer success rates. Fertil. Steril., 63, 555–562.
Forman, R.G., Frydman, R., Egan, D. et al. (1990) Severe ovarian
hyperstimulation syndrome using agonists of gonadotrophin-releasing
hormone for in vitro fertilization: a European series and a proposal for
prevention. Fertil. Steril., 53, 502–509.
(Figure 2). There were no significant differences in oestradiol
concentrations between the two groups except in cycle 4 when
the concentrations were lower in the endometriosis group.
Despite the differences observed in ovarian stimulation and
response, the cumulative pregnancy and live birth rates were
not significantly different in the two groups (Figures 3 and 4).
Geber, S., Paraschos, T., Atkinson, G. et al. (1995) Results of IVF in patients
with endometriosis: the severity of the disease does not affect outcome, or
the incidence of miscarriage. Hum. Reprod., 10, 1507–1511.
Lockwood, G.M., Pinkerton, S.M. and Barlow, D.H. (1995) A prospective
randomized single-blind comparative trial of nafarelin acetate with buserelin
in long-protocol gonadotrophin-releasing hormone analogue controlled in-
vitro fertilization cycles. Hum. Reprod., 10, 293–298.
Matson, P.L. and Yovich, J.L. (1986) The treatment of infertility associated
with endometriosis by in vitro fertilization. Fertil. Steril., 46, 432–434.
O’Shea, R.T., Chen, C., Weiss, T. and Jones, W.R. (1985) Endometriosis and
in-vitro fertilisation. Lancet, ii, 723
Oehninger, S., Acosta, A.A., Kreiner, D. et al. (1988) In vitro fertilization
and embryo transfer (IVF/embryo transfer): an established and successful
therapy for endometriosis. J. In Vitro Fert. Embryo. Transf., 5, 249–256.
Olivennes, F., Feldberg, D., Liu, H.C. et al. (1995) Endometriosis: a stage by
stage analysis – the role of in vitro fertilization. Fertil. Steril., 64, 392–398.
Pellicer, A., Oliveira, N., Ruiz, A. et al. (1995) Exploring the mechanism(s)
of endometriosis-related infertility: an analysis of embryo development and
implantation in assisted reproduction. Hum. Reprod., 10, 91–97.
Ron-El, R., Raziel, A., Herman, A. et al. (1990) Ovarian response in repetitive
cycles induced by menotrophin alone or combined with gonadotropin
releasing hormone analogue. Hum. Reprod., 5, 427–430.
Discussion
The extent to which endometriosis affects IVF outcome has
been a controversial matter for some time. This paper, however,
indicates that the outcome in women with ovarian endomet-
riosis over three or more consecutive cycles is comparable
to that in women with tubal disease and normal ovaries.
Nevertheless, women with ovarian endometriosis required
significantly higher doses of HMG to achieve adequate stimula-
tion resulting in significantly increased costs per treatment
cycle. Moreover, it is clear that women with ovarian endomet-
riosis have a progressive decline in ovarian reserve over time.
This was shown by the poorer response with each subsequent
cycle despite the use of significantly higher HMG doses. In
contrast, women in the control group maintained a constant
ovarian response over the five cycles studied. The data in the
control group confirm previous findings (Ron-El et al., 1990;
Ahmed-Ebbiary et al., 1995) that women with normal ovulatory
function are likely to have similar ovarian responses in
consecutive cycles of ovarian stimulation.
Wardle, P.G. and Hull, M.G. (1993) Is endometriosis a disease? Clin. Obstet.
Gynaecol., 7, 673–685.
Wardle, P.G., Mitchell, J.D., McLaughlin, E.A. et al. (1985) Endometriosis
and ovulatory disorder: reduced fertilisation in vitro compared with tubal
and unexplained infertility. Lancet, ii, 236–239.
Yovich, J.L., Yovich, J.M., Tuvik, A.I. et al. (1985) In-vitro fertilisation for
endometriosis. Lancet, ii, 552.
In this paper, we only included women with confirmed
ovarian endometriosis because this form of the disease is the
most likely to affect the outcome of ovarian stimulation.
Further research is required to determine whether the relatively
Received on June 21, 1999; accepted on October 6, 1999
75