RBMOnline - Vol 19. No 5. 2009 737744 Reproductive BioMedicine Online; www.rbmonline.

com/Article/3992 on web 22 September 2009

Article
The effect of hysteroscopic polypectomy on the
concentrations of endometrial implantation
factors in uterine flushings
Dr Ben-Nagi has carried out research under the supervision of Mr Jurkovic in early
pregnancy complications and endometrial implantation factors in women with endometrial
polyps, submucous fibroids and with history of previous Caesarean sections. This research
has been presented in international conferences and published in peer-reviewed journals.

Dr Ben-Nagi
J Ben-Nagi1,3, J Miell2, J Yazbek1, T Holland1, D Jurkovic1
Early Pregnancy and Gynaecology Assessment Unit, Kings College Hospital, Denmark Hill, London SE5 8RX, UK;
2
Department of Endocrinology, University Hospital Lewisham, London SE13 6LH, UK;
3
Correspondence: e-mail-address: jbennagi@gmail.com

Abstract
Endometrial polyps have been associated with infertility and early pregnancy loss. The aim of this study was to investigate the eect of hysteroscopic polypectomy on the concentrations of endometrial implantation factors in uterine
ushings. Pre-menopausal women with a certain diagnosis of endometrial polyp on contrast-enhanced transvaginal
ultrasound scan were recruited into this prospective study. In all women, paired samples of uterine ushings were
obtained on the same day of the menstrual cycle prior to and post hysteroscopic polypectomy. Enzyme-linked immunoassays were performed to analyse glycodelin, interleukin-6 (IL-6), interleukin-10 (IL-10), tumour necrosis factor a
(TNFa) and osteopontin, whilst immunoradiometric assay was used to analyse insulin-like growth factor binding
protein-1 (IGFBP-1). Concentrations of IGFBP-1, TNFa and osteopontin in uterine ushings were signicantly
lower in the mid-secretory phase prior to polypectomy in comparison to the measurements obtained after complete
surgical removal of the polyp (P < 0.05). There were no dierences in the concentrations of glycodelin, IL-6 and
IL-10 in paired samples prior to and post-polypectomy. The presence of endometrial polyps is associated with
decreased mid-secretory concentrations of IGFBP-1, TNFa and osteopontin, which are reversed following surgical
polypectomy.
Keywords: cytokines, endometrial polyps, glycodelin, IGFBP-1, osteopontin

Introduction
Polyps are one of the most common endometrial abnormalities with a reported prevalence of 925% in the general
population of women (Anastasiadis et al., 2000; Sherman
et al., 2002). They are often asymptomatic but they can
sometimes cause menstrual irregularities such as intermenstrual bleeding. Some studies have suggested that endometrial polyps may be associated with infertility and early
pregnancy loss (Sanders, 2006; Tur-Kaspa et al., 2006).
However, the pathophysiological processes by which polyps
may aect womens fertility are not clear. It has been pos-

tulated that polyps aect the endometrial environment by

causing abnormal bleeding or by presenting an abnormal
site for implantation. Assessment of the uterine cavity
and removal of endometrial polyps is routinely performed
in women undergoing fertility treatment. There is some evidence to suggest that pregnancy rates in infertile women are
improved following hysteroscopic polypectomy (Varasteh
et al., 1999).
Glycodelin is one of the most abundantly secreted endometrial glycoproteins. Its secretion is limited to the mid- and
late luteal phase of the cycle, which suggests that it plays

2009 Published by Reproductive Healthcare Ltd, Duck End Farm, Dry Drayton, Cambridge CB23 8DB, UK

737

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

a role in implantation and early pregnancy development (Li

et al., 1993; Seppala et al., 2001). Studies by Richlin et al.
(2002) have shown that concentrations of glycodelin in
uterine ushings are raised in women with polyps and broids compared with the control group in the proliferative
phase of the cycle.
Insulin growth factor binding protein-1 (IGFBP-1) is
thought to have major eects on decidualization, implantation and trophoblast invasion in the female reproductive
tract. The expression of IGFBP-1 is tightly controlled and
it is predominant in the secretory phase endometrium and
decidualized stromal endometrial cells (Fowler et al., 2000).
Cytokines are produced by the human endometrium and
are important mediators between the embryo and decidua
during implantation. Piccinni et al. (1998) have shown that
there are signicantly higher concentrations of CD4+
T-helper 2 cells producing interleukin-4 (IL-4), interleukin-6 (IL-6) and interleukin-10 (IL-10) in decidua from
women with normal reproductive histories compared with
women with recurrent miscarriages. Lim et al. (1998)
reported a predominant type-2 cytokine expression prole
(IL-4, IL-6 and IL-10) and lack of type-1 cytokine expression (interferon-c, IL-2 and tumour necrosis factor-b) in
normal fertile women.
Tumour necrosis factor-a (TNFa) is known to be one of the
most versatile cytokines. It is speculated that TNFa promotes DNA synthesis in the early proliferative phase and
participates in cell dierentiation and tissue remodelling,
which is required to support embryonic attachment
(Terranova et al., 1995). TNFa also facilitates apoptosis
and therefore initiates menstrual shedding (Tabibzadeh
et al., 1995).
Osteopontin is a progesterone-regulated glycoprotein component of the extracellular matrix (Apparao et al., 2003). It
is secreted by the glandular epithelium of mammalian uteri
and later by the decidualized stroma. Osteopontin is recognized by the integrin family to facilitate cellcell attachment
and adhesion (Apparao et al., 2001).
Although endometrial polyps have been identied as a possible cause of infertility, their eect on endometrial implantation factors has been little studied so far. This prospective
study examined changes in the concentrations of endometrial markers in paired samples of uterine ushings prior
to and post hysteroscopic polypectomy. It was postulated
that removal of polyps will have a signicant impact on
the concentrations of endometrial markers, which should
help us to understand better the pathophysiological mechanisms of impaired endometrial receptivity in the presence of
endometrial polyps.

Materials and methods

738

Non-pregnant, pre-menopausal women attending the gynaecology out-patient clinic at Kings College Hospital, London, UK with a history of abnormal vaginal bleeding who
were found to have an endometrial polyp on transvaginal

ultrasound scan were invited to participate in this prospective interventional study. The inclusion criteria were: (i)
age between 18 and 45 years; (ii) regular menstrual cycles;
(iii) not using any hormonal contraception or treatment with
an eect on endometrium; (iv) presence of an endometrial
polyp on two-dimensional ultrasound scan, which was also
conrmed on saline infusion sonohysterosonography; (v)
absence of any other endometrial pathology on transvaginal
scan; (vi) ability to give written consent; and (vii) acceptance
of a second examination at follow-up, which was performed
on the same cycle day of the rst hysteroscopy. The study
was approved by Kings College Hospital ethics and research
and development committee.
All women had a two-dimensional transvaginal ultrasound
scan performed by gynaecologists with expertise in
transvaginal scanning using high-frequency transducers of
57.5 MHz (Aloka SSD-5000; Aloka, Tokyo, Japan). This
was followed by saline infusion sonohysterosonography to
conrm the presence of the endometrial polyp (Lee et al.,
2006).
Subsequently, all women included in the study had uterine
ushings, performed prior to the commencement of the hysteroscopy and polypectomy. Uterine ushings were performed in the dorsal lithotomy position. A Cuscos
bivalve speculum was inserted in the vagina in order to
expose the cervix. The cervix was cleaned with sterile saline
to remove any visible potential contaminants. An 8 F paediatric Foley balloon catheter (Schering AG, Berlin) was
then passed into the uterine cavity through the cervix and
the balloon was then inated with 2 ml of air. Under simultaneous transvaginal ultrasound scan guidance, the balloon
was withdrawn to lie above the level of the internal cervical
os. This was identied as the reection of the urinary bladder. Five aliquots of 2 ml sterile 0.154 mol/l sodium chloride solution were then sequentially injected and aspirated
from the uterine cavity over approximately 10 s. Ultrasound scan guidance was used to conrm that the uid
did not enter the Fallopian tubes or cervical canal, with
re-aspiration being done when the uid was noted to be distending the upper aspect of the uterine cavity. The ve 2-ml
aliquots were collected in ve separate universal specimen
pots, immediately frozen and stored at 20C.
Following hysteroscopic polypectomy women were asked
to attend for a follow-up appointment, which included a
transvaginal scan and repeat uterine ushings. Women were
advised to wait until they had at least two menstrual bleeds
(at least one complete normal physiological cycle) between
the hysteroscopy and the follow-up visit. The follow-up
appointments were timed to coincide with the same day
of the cycle when the hysteroscopy and pre-polypectomy
uterine ushings were performed.
Prior to the commencement of any assay, the ve 2-ml aliquots obtained per patient were thawed at room temperature and then pooled. The total pooled volume of uterine
ushing was thawed for analysis on a single occasion only.
The total volume of uid retrieved from the uterine cavity
per patient was recorded. The protein content was
RBMOnline

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

calculated in the total volume of uterine uid recovered.

The amount of total protein (mg/l) measured was adjusted
per ml of uid obtained in all the samples. Glycodelin,
IGFBP-1 and osteopontin were expressed as ng per mg/
ml of protein and IL-6, IL-10 and TNFa was expressed
as pg per mg/ml of protein.

Laboratory assays
Fluid protein
The Bayer Advia method for the measurement of uid protein is a non-enzymatic assay (Siemens Medical Solutions
Diagnostics Limited, UK). Under acid conditions and in
the presence of molybdate ions, proteins form a blue-coloured complex with pyrogallol red. The absorbance of this
complex is measured at 596 nm. The sensitivity of the assay
is the lowest concentration that can be distinguished from
zero is 10 mg/l.

analyte concentration of 171.7 pg/ml and 12% at a concentration of 163 pg/ml, respectively.

Osteopontin
The osteopontin assay employed the quantitative sandwich
enzyme immunoassay technique (R and D Systems, Minneapolis, USA). Three samples were tested 20 times on one
plate to assess intra-assay sensitivity. The intra-assay coefcients of variation for samples 1, 2 and 3 were 4% at an
analyte concentration of 2.3 ng/ml, 2.6% at a concentration
of 4.9 ng/ml and 2.9% at a concentration of 9.3 ng/ml,
respectively. A further three samples were tested in 40 separate assays to assess inter-assay sensitivity. The inter-assay
coecients of variation for samples 1, 2 and 3 were 6.6% at
an analyte concentration of 2.36 ng/ml, 5.7% at an analyte
concentration of 4.81 ng/ml and 5.4% at a concentration of
9.17 ng/ml, respectively.

Statistical analysis
Glycodelin
Glycodelin was measured using a solid-phase enzymelinked immunoassay (ELISA) based on the sandwich principle (DRG Instruments, Germany). When a quality
control value diered by greater than 10% from the previous assay means, the assay was repeated. The assay was
tested for specicity, by the manufacturer against HCG
(2000 IU/l), prolactin (200 lg/l), human placental lactogen
(20 lg/l) and alpha feto protein (300 mIU/ml) with
undetectable glycodelin concentrations in all cases. The
sensitivity of this assay is 6 ng/ml. Both the intra-assay
and inter-assay coecients of variation were 9.4% at an
analyte concentration of 118.39 ng/ml and 3.9% at an
analyte concentration of 99.6 ng/ml, respectively.

All statistical analyses were carried using Medcalc version

9.2.0.2 (Medcalc Software, Mariakerke, Belgium). Paired
sample t-test was used to analyse the dierence between total
volumes retrieved in patients with and without polyps as the
data for volume of uterine ushing retrieved (ml) is normally
distributed. Wilcoxon signed rank test was used to assess the
dierence between glycodelin, IGFBP-1, IL-6, IL-10, TNFa
and osteopontin in the pre- and post-polypectomy groups as
the data for these endometrial proteins and cytokines are not
normally distributed. Spearman rank correlation was used
for non-parametric correlation to examine the relationship
between glycodelin, IGFBP-1, IL-6, IL-10, TNFa and osteopontin concentrations and protein content and volume of
uterine ushing retrieved. P < 0.05 was considered statistically signicant.

IGFBP-1
IGFBP-1 was measured by immunoradiometric assay
(IRMA). This is a non competitive assay where the analyte
is sandwiched between two antibodies (Diagnostics Systems
Laboratories, Webster, USA). The sensitivity was 0.33 ng/ml.
The intra- and inter-assay coecients of variation were 5.2%
at an analyte concentration of 5.23 ng/ml and 3.5% at an analyte concentration of 5.16 ng/ml, respectively.

IL-6, IL-10 and TNFa

IL-6, IL-10 and TNFa were analysed using commercially
available solid-phase sandwich ELISA (Diaclone,
Besancon, France). The sensitivity of the IL-6 assay was
2 pg/ml with intra- and inter-assay coecients of variation
of 4.8% at an analyte concentration of 52.2 pg/ml and 5.5%
at an analyte concentration of 53 pg/ml, respectively. For
IL-10, the sensitivity was 5 pg/ml with intra- and interassay coecients of variation of 5% at an analyte concentration 55.1 pg/ml and 8.6% at a concentration of
58.3 pg/ml, respectively.
The sensitivity of the TNFa assay was 8 pg/ml. The intraand inter-assay coecients of variation were 5.3% at an
RBMOnline

Results
A total of 20 women were recruited into the study. Of these,
8/20 (40%) failed to attend the post-operative transvaginal
ultrasound scan and uterine ushings and thus were
excluded from the nal data analysis. In the remaining 12
(60%) women who attended for post-operative follow-up,
11 (92%, 95% CI 6599) women had a single polyp and
one (8%, 95% CI 135) had two polyps on ultrasound scan.
The median polyp volume was 4.7 cm3 (range 1.3216.99).
Womens median age was 36 years (range 2945), gravidity
2 (range 03) and parity 0 (range 02). The median length
of the cycle was 28 days (range 2628). Of the 12 patients,
eight (67%, 95% CI 4094) were in the luteal phase of the
cycle (range 2021) and were included in the analysis of
the mid-secretory concentrations. The remaining four
(33%, 95% CI 660) patients were in the proliferative phase
of the cycle (range 514).
Uterine ushings were performed on the same day of the
cycle prior and post-polypectomy (median 20; range
521). The median time between the initial and follow-up
visit was 56 days (range 5284). The median volume of uid

739

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

retrieved prior to polypectomy was 4.0 ml (range 110),

which was not signicantly dierent from the median volume of 4.25 ml (range 38) following polypectomy. The
procedure for uterine ushings was well-tolerated by all
the patients and in none of the cases was the procedure
abandoned because of discomfort, pain or any other
complications.
The histologies of the hysteroscopically excised endometrial
polyps were all benign. The maturation status of the endometrium was described for ve (42%, 95% CI 1968) cases:
four proliferative and one secretory endometrium.
The protein content (mg) in the total volume (ml) of uterine
ushing retrieved was analysed in the pre-polypectomy
group. There was no signicant correlation detected between
the total protein concentration and retrieved volume
(r = 0.02,). Glycodelin showed a stronger correlation with
protein content (r = 0.76, P = 0.01) than with recovered volume of uterine ushings (r = 0.07, not signicant). Il-6 also
showed a stronger correlation with total protein content
(r = 0.73, P = 0.03) than with the volume retrieved
(r = 0.13, not signicant). However, the correlation did not
reach statistical signicance between the IGFBP-1, IL-10,
TNFa and osteopontin and total protein or total volume.
Thus, all markers were expressed as per mg of total protein
as it appeared a more reliable indicator of the eciency of
the ushing process than the volume retrieved.

Figure 1. Presence of glycodelin (ng/mg/ml  103)

throughout the cycle in the pre- and post-polypectomy
groups. Spearmans rank correlations (r) for the pre- and
post-polypectomy groups are 0.21 and 0.04, respectively.
(P = 0.002). The changes in IGFBP-1 concentrations
remained signicant when the analysis was limited to the
mid-secretory phase only (P = 0.03) (Table 1, Figure 2).

IL-6, IL-10 and TNFa

The overall median glycodelin prior to polypectomy was

169.5 ng/mg/ml  103 (range 64902), which was not signicantly dierent from 202 ng/mg/ml  103 (range 16
1886) post polypectomy. Similar results were obtained
when the analysis was limited only to the measurements,
which were performed in the mid-secretory phase (Table
1, Figure 1).

The median IL-6 and IL-10 prior to polypectomy group were

7.5 pg/mg/ml  103 (range 1141) and 3.5 pg/mg/ml  103
(range 036), respectively. In the post-polypectomy group,
the values were 13.5 pg/mg/ml  103 (238) for IL-6 and
11.5 pg/mg/ml  103 (132) for IL-10. Similar results were
also obtained when the analysis was performed in the midsecretory phase (Table 1, Figures 3 and 4). The median TNFa
increased after removal of polyps, being 5 pg/mg/ml  103
(073) pre-polypectomy versus 69.5 pg/mg/ml  103
(0289) post-polypectomy (P = 0.01)). TNFa changes
remained signicantly dierent when comparing its secretion
between the two groups in the mid-secretory phase only
(P = 0.03) (Table 1, Figure 5).

IGFBP-1

Osteopontin

The overall median IGFBP-1 prior to polypectomy

was 1 ng/mg/ml  103 (range 04), which increased to
8 ng/mg/ml  103 (range 429) post polypectomy

Before polypectomy the median osteopontin was 1 ng/mg/

ml  103 (range 021) compared with 4 ng/mg/ml  103
(range 069) after surgery but the dierence was not

There was no correlation between expression levels of the

implantation markers with age and parity.

Glycodelin

Table 1. Median concentrations of endometrial proteins and cytokines in the pre- and postpolypectomy groups in the mid-secretory phase.

Glycodelin (ng/mg/ml  10 3)
IGFBP-1 (ng/mg/ml  10 3)
IL-6 (pg/mg/ml  10 3)
IL-10 (pg/mg/ml  10 3)
TNFa (pg/mg/ml  10 3)
Osteopontin (ng/mg/ml  10 3)

740

Pre-polypectomy

Post-polypectomy

P-value

169.5 (64785)
1 (04)
6 (1141)
3.5 (027)
5 (016)
1 (021)

202.5 (601886)
7.5 (429)
13.5 (238)
13 (132)
94 (0289)
4 (069)

NS
0.03
NS
NS
0.03
0.04

Values are median (range); IGFBP-1, insulin-like growth factor binding protein-1; IL, interleukin; NS, not statistically signicant; TNFa, tumour necrosis factor-a.

RBMOnline

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

Figure 2. Presence of insulin-like growth factor binding

protein-1 (IGFBP-1) (ng/mg/ml  103) in the pre- and
post-polypectomy groups. Spearmans rank correlations (r)
for the pre-polypectomy and post-polypectomy groups are
0.26 and 0.17, respectively.

Figure 3. Presence of interleukin-6 (IL-6) (pg/mg/ml  103)

in the pre- and post-polypectomy groups. Spearmans rank
correlations (r) for the pre-polypectomy and post-polypectomy groups are 0.09 and 0.06, respectively.

statistically signicant. The dierences, however, were

signicant when the measurements obtained in the mid-secretory phase were analysed separately (P = 0.04) (Table 1,
Figure 6).

Discussion
The current study has shown that the presence of endometrial polyps within the uterine cavity has a measurable eect
on the concentrations of endometrial proteins in uterine
ushings. There was a signicant increase in IGFBP-1,
TNFa and osteopontin concentrations in uterine ushings
obtained in the mid-luteal phase following polypectomy
in comparison to the preoperative measurements.
RBMOnline

Figure 4. Presence of interleukin-10 (IL-10) (pg/mg/

ml  103) in the pre- and post-polypectomy groups.
Spearmans rank correlations (r) for the pre-polypectomy
and post-polypectomy groups are 0.20 and
0.03,
respectively.

Figure 5. Presence of tumour necrosis factor-a (TNFa) (pg/

mg/ml  103) in the pre- and post-polypectomy groups.
Spearmans rank correlations (r) for the pre-polypectomy
and post-polypectomy groups are 0.20 and 0.00,
respectively.

Endometrial polyps also aected concentrations of endometrial proteins and cytokines throughout dierent phases
of the menstrual cycle. There were no signicant changes
in glycodelin concentrations throughout the cycle prior
to polypectomy. In contrast, glycodelin concentrations
increased to reach a peak in the mid-luteal phase following
removal of the endometrial polyps. Previous studies showed
that glycodelin concentration in uterine ushings increases
rapidly 6 days following the LH peak to reach its peak in
the late luteal phase (Li et al., 1993; Seppala et al., 2001).
The fact that this increase was not observed in the
pre-polypectomy group suggests a deciency in maternal

741

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

post polypectomy was an unexpected nding. von Wol

et al. (2002) reported low concentrations in the proliferative phase but a 5- to 10-fold increase in the mid- to
late-secretory phase. The authors concluded that the high
concentration from the mid-secretory phase support the
role of IL-6 in the regulation of endometrial functions.
However, the results of this study did not conrm their
observation.

Figure 6. Presence of osteopontin (ng/mg/ml  103) in the

pre- and post-polypectomy groups. Spearmans rank correlations (r) for the pre-polypectomy and post-polypectomy
groups are 0.04 and 0.01, respectively.

immunosuppression during the putative window of

implantation. This may play a role in women with endometrial polyps resulting in rejection of the pre-implanting blastocyst by the maternal immune system. Previous studies
have shown that glycodelin has immunosuppressive properties through its suppression of natural killer cell activity
(Okamoto et al., 1991). The elevated glycodelin concentrations at the time of implantation may protect the embryo
from natural killer cell activity (Richlin et al., 2002). The
observed dierences in glycodelin concentrations pre- and
post-polypectomy did not reach statistical signicance,
which may be attributed to a small sample size and the lack
of use of LH assay to predict more accurately the midsecretory phase.
IGFBP-1 concentrations were signicantly dierent
between the pre- and post-polypectomy groups. Women
with endometrial polyps produced lower concentrations
of IGFBP-1 than those without polyps. IGFBP-1 is a major
protein product of the non-pregnant endometrium during
the middle- to late-secretory phase (Giudice et al., 1998).
IGFBP-1 has been shown to inhibit trophoblast invasion
into the decidualized endometrial stromal cultures, suggesting that IGFBP-1 is a maternal restraint on trophoblast
invasion. The presence of a polypoid lesion within the
endometrium may alter the secretion of IGFBP-1 from
the glandular epithelium resulting in signicantly lower
concentrations in women with polyps. Low concentrations
of IGFBP-1 may result in inadequate endometrial decidualization and failure of implantation in women with endometrial polypoid lesions.

742

The median concentrations of IL-6 and IL-10 were higher

in the post-polypectomy than in the pre-polypectomy
group, although this was not statistically signicant. Furthermore, the removal of polyps did alter the pattern of
secretion of IL-6 and IL-10 throughout the menstrual
cycle. IL-6 remained static whilst IL-10 increased post-surgical polypectomy. In this study, a static secretion of IL-6

A sustained increase of IL-10 was observed once the polyps were removed compared with the pre-polypectomy
group. IL-10 is up-regulated in the secretory phase and
in early pregnancy (Vigano et al., 2002). IL-10 is a paracrine mediator produced at the maternalfetal interface
to control the fetal immune responses. Lower concentrations have been detected in the mid-secretory phase of
the cycle in women with recurrent miscarriage in comparison with normal controls (von Wol et al., 2000). Furthermore, the endometrium in women with a history of
normal pregnancies is associated with higher concentrations of IL-6 and IL-10 (Krasnow et al., 1996). It is therefore possible that low concentrations of IL-10 indicate an
aberrant response to the implanting blastocyst in women
with endometrial polyps.
TNFa concentration was signicantly higher following surgical polypectomy. The secretion of TNFa also increased
throughout the menstrual cycle after removal of polyps
reaching its peak in the secretory phase. This observation
is concordant with previous ndings (Philippeaux and
Piguet, 1993; Tabibzadeh et al., 1995; von Wol et al.,
2000). Hunt et al. (1992) demonstrated highest ribonucleic
acid expression in human endometrial epithelium and
stroma in the late proliferative phase and the mid- to latesecretory phase. TNFa is a multifunctional cytokine and
its expression at dierent phases of the menstrual cycle suggests its complex function on the endometrium and the preimplantation embryo leading to a successful implantation
(von Wol et al., 2000). Therefore, signicantly increased
TNFa in the post-polypectomy group promotes DNA synthesis, enhances endometrial tissue dierentiation and
remodelling, which are fundamental for implantation. Garcia et al. (1994) showed benign endometrial polyps contained little TNFa mRNA or protein but were abundant
in high-grade endometrial tumours. Abnormal TNFa
expression may be a contributory factor in the role of endometrial polyps in infertility and miscarriage.
This study showed that endometrial polyps signicantly
aect the concentration of osteopontin in the mid-secretory phase. von Wol et al. (2001) demonstrated that
osteopontin concentrations in uterine secretion were low
in the proliferative phase and increased 3- to 4-fold in
the secretory phases of the menstrual cycle. Osteopontin
plays a role in the attachment of trophoblast to the endometrial epithelium and in the late stages of implantation.
The lower concentrations and lack of increase in the
secretory phase found in the pre-polypectomy group suggest the negative eect of polyps on osteopontin secretion, which may cause failure of the pre-implanting
blastocyst to attach to the decidua in women with endometrial polyps.
RBMOnline

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

The pathophysiological mechanisms by which endometrial

polyps cause signicant reduction of IGFBP-1, TNFa and
osteopontin concentrations in uterine ushings are unclear.
Endometrial polyps dier from the surrounding endometrium and they contain extensive brotic stroma and dilated
thick-walled blood vessels (Nogueira et al., 2007). This
abnormal endometrial architecture may be the cause of
the impaired secretion of implantation factors. Furthermore, endometrial polyps also have decreased expression
of hormone receptors in the stromal component. The glands
and stroma of the polyps are unresponsive to progesterone
stimulation (Fox, 1992; Nogueira et al., 2007). This may
cause abnormalities in the secretion of progesterone-regulated endometrial markers in the presence of endometrial
polyps. A polyp may also cause mechanical blockage of
the lumen of the glandular epithelium, which could also lead
to decreased concentrations of implantation factors.
However, it is also possible that there are other factors
other than the presence of endometrial polyps that may
account for the dierence in the concentrations of endometrial proteins in the pre- and post-polypectomy groups. The
alteration in the concentration of implantation factors postsurgical polypectomy may also be caused by the mechanical
damage and repair of endometrial tissue. However, this is
unlikely to be the case as all women had at least two menstrual periods before attending for post-polypectomy ushing. Further studies comparing post-polypectomy ushing
to samples obtained from women with normal endometrium who underwent hysteroscopy and biopsy would be
required to resolve this issue.
In conclusion, this prospective study showed signicant
dierences in concentrations of IGFBP-1, TNFa and osteopontin following surgical removal of endometrial polyps.
These ndings provide a possible explanation of pathophysiological mechanisms leading to the increased risk of
infertility and early pregnancy loss in women with endometrial polyps. Limitations of this study include small
sample size and lack of timed measurements using LH
assay. Further work involving a larger number of subjects
is necessary to conrm the reproducibility of these ndings
and further explore the eect of endometrial polyps on
factors regulating implantation and early placental
development.

References
Anastasiadis P, Koutlaki N, Skaphida P et al. 2000 Endometrial
polyps: prevalence, detection, and malignant potential in
women with abnormal uterine bleeding. European Journal of
Gynaecological Oncology 21, 180183.
Apparao K, Illera M, Beyler S et al. 2003 Regulated expression of
osteopontin in the peri-implantation rabbit uterus. Biology of
Reproduction 68, 14841490.
Apparao K, Murray M, Fritz M et al. 2001 Osteopontin and its
receptor alphavbeta(3) integrin are coexpressed in the human
endometrium during the menstrual cycle but regulated
dierentially. The Journal of Clinical Endocrinology and
Metabolism 86, 49915000.
Fowler D, Nicolaides K, Miell J 2000 Insulin-like growth factor
binding protein-1 (IGFBP-1): a multifunctional role in the

RBMOnline

human female reproductive tract. Human Reproduction Update

6, 495504.
Fox H 1992 The female reproductive tract and the breast. In:
MacSween R, Whaley K, editors Muirs Textbook of
Pathology. Kent, UK: Edward Arnold; 1992. p. 10201021.
Garcia F, Chen H, Yang Y et al. 1994 Tumour necrosis factoralpha mRNA and protein in endometrial tumors: analysis by insitu hybridization and immunohistochemistry. Human
Pathology 25, 13241331.
Giudice L, Mark S, Irwin J 1998 Paracrine actions of insulin-like
growth factors and IGF binding protein-1 in non-pregnant
human endometrium and at the decidualtrophoblast interface.
Journal of Reproductive Immunology 39, 133148.
Hunt J, Chen H, Hu X, Tabibzadeh S 1992 Tumour necrosis
factor-alpha messenger ribonucleic acid and protein in human
endometrium. Biology of Reproduction 47, 141147.
Krasnow J, Tollerud D, Naus G et al. 1996 Endometrial Th2
cytokine expression throughout the menstrual cycle and early
pregnancy. Human Reproduction 11, 17471754.
Lee C, Ben-Nagi J, Oli-Yebovi D et al. 2006 A new method of
transvaginal ultrasound-guided polypectomy: a feasibility
study. Ultrasound in Obstetrics and Gynecology 27, 198201.
Li T, Ling E, Dalton C et al. 1993 Concentration of endometrial
protein PP14 in uterine ushings throughout the menstrual
cycle in normal, fertile women. British Journal of Obstetrics and
Gynaecology 100, 460464.
Lim K, Odukoya O, Ajjan R et al. 1998 Prole of cytokine mRNA
expression in peri-implantation human endometrium.
Molecular Human Reproduction 4, 7781.
Nogueira A, Reis F, Silva J et al. 2007 Endometrial polyps: a
review. Journal of Gynecologic Surgery 23, 111116.
Okamoto N, Uchida A, Takakura K et al. 1991 Suppression by
human placental protein 14 of natural killer cell activity.
American Journal of Reproductive Immunology 26, 137142.
Philippeaux M, Piguet P 1993 Expression of tumour necrosis
factor-alpha and its mRNA in the endometrial mucosa during
the menstrual cycle. American Journal of Pathology 143,
480486.
Piccinni M, Beloni L, Livi C et al. 1998 Defective production of
both leukaemia inhibitory factor and type 2 T-helper cytokines
by decidual T cells in unexplained recurrent abortions. Nature
Medicine 4, 10201024.
Richlin S, Ramachandran S, Shanti A et al. 2002 Glycodelin levels
in uterine ushings and in plasma of patients with leiomyomas
and polyps: implications for implantation. Human Reproduction
17, 27422747.
Sanders B 2006 Uterine factors and infertility. Journal of
Reproductive Medicine 51, 169176.
Seppala M, Koistinen H, Koistinen R 2001 Glycodelins. Trends in
Endocrinology and Metabolism 12, 111117.
Sherman M, Maur M, Kurman R 2002 Benign diseases of the
endometrium. In: Kurman RJ, editor Blaunsteins pathology of
the female genital tract. New York, USA: Springer; 2002. p.
421466.
Tabibzadeh S, Zupi E, Babaknia A et al. 1995 Site and menstrual
cycle-dependent expression of proteins of the tumour necrosis
factor (TNF) receptor family, and BCL-2 oncoprotein and
phase specic production of TNF alpha in human
endometrium. Human Reproduction 10, 277286.
Terranova P, Hunter V, Roby K et al. 1995 Tumour necrosis
factor-alpha in the female reproductive tract. Proceedings of the
Society for Experimental Biology and Medicine 209, 325342.
Tur-Kaspa I, Gal M, Hartman M et al. 2006 A prospective
evaluation of uterine abnormalities by saline infusion
sonohysterography in 1,009 women with infertility or abnormal
uterine bleeding. Fertility and Sterility 86, 17311735.
Varasteh N, Neuwirth R, Levin B et al. 1999 Pregnancy rates after
hysteroscopic polypectomy and myomectomy in infertile
women. Obstetrics and Gynecology 94, 168171.
Vigano P, Somigliana E, Mangioni S et al. 2002 Expression of
interleukin-10 and its receptor is up-regulated in early pregnant

743

Article - Hysteroscopic polypectomy and endometrial implantation factors - J Ben-Nagi et al.

versus cycling human endometrium. Journal of Clinical

Endocrinology and Metabolism 87, 57305736.
von Wol M, Thaler C, Zepf C et al. 2002 Endometrial expression
and secretion of interleukin-6 throughout the menstrual cycle.
Gynecological Endocrinology 16, 121129.
von Wol M, Strowitzki T, Becker V et al. 2001 Endometrial
osteopontin, a ligand of beta3-integrin, is maximally expressed
around the time of the implantation window. Fertility and
Sterility 76, 775781.
von Wol M, Thaler C, Strowitzki T et al. 2000 Regulated
expression of cytokines in human endometrium throughout the

menstrual cycle: dysregulation in habitual abortion. Molecular

Human Reproduction 6, 627634.

Declaration: The authors report no financial or commercial

conflicts of interest.
Received 14 October 2008; refereed 24 February 2009; accepted 19
June 2009.

744
RBMOnline

Copyright of Reproductive BioMedicine Online is the property of Reproductive Healthcare Limited and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.