Reproductive BioMedicine Online (2014) 28, 704– 713

www.sciencedirect.com
www.rbmonline.com

REVIEW

Adenomyosis and reproductive performance

after surgery for rectovaginal and
colorectal endometriosis: a systematic
review and meta-analysis
Paolo Vercellini a,b,*, Dario Consonni c, Giussy Barbara a, Laura Buggio a,
Maria Pina Frattaruolo a,b, Edgardo Somigliana b,d

a
Department of Clinical Sciences and Community Health, Università degli Studi di Milano, and Fondazione IRCCS Ca’
Granda–Ospedale Maggiore Policlinico, 20122 Milan, Italy; b Center for Research in Obstetrics and Gynaecology
(C.R.O.G.), 20122 Milan, Italy; c Epidemiology Unit, Department of Preventive Medicine, Fondazione IRCCS Ca’
Granda–Ospedale Maggiore Policlinico, 20122 Milano, Italy; d Infertility Unit, Fondazione IRCCS Ca’ Granda–Ospedale
Maggiore Policlinico, 20122 Milan, Italy
* Corresponding author. E-mail address: paolo.vercellini@unimi.it (P Vercellini).

Paolo Vercellini is associate professor of obstetrics and gynaecology at the University of Milan and chief of the
Benign Gynaecological Surgery Service of the Department of Obstetrics and Gynaecology of the ‘Luigi
Mangiagalli’ Institute of Milan, Italy. His research focuses on the pathogenesis and treatment of endometriosis,
medical and surgical therapies for uterine leiomyomas, the pathogenesis of chronic pelvic pain, diagnosis and
therapy of menorrhagia and reproductive surgery. He is a reviewer for several international journals, is author
and coauthor of over 200 peer-reviewed scientific articles and has contributed numerous book chapters.

Abstract The relationship between rectovaginal–bowel endometriosis and fertility is unclear. Nevertheless, extirpative surgery,
including colorectal resection, is being fostered as a fertility-enhancing procedure. Adenomyosis and deep endometriosis often
coexist. As the uterine condition may further impact on reproductive outcome, this work performed a systematic literature review
with the objective of identifying all English-language reports on surgical treatment for rectovaginal and colorectal endometriosis,
including bowel resection, in which participants were screened preoperatively for uterine adenomyosis. Risk ratios (RR) were then
combined in a meta-analysis. In the five selected observational studies, in women seeking pregnancy, 7/59 (11.9%) with concomitant
adenomyosis conceived, compared with 74/172 (43.0%) in those without adenomyosis. Adenomyosis was never excised. One in 10
women experienced a major surgical complication. The RR of clinical pregnancy ranged from 0.23 to 0.46, with absence of heter-
ogeneity among studies (I2 = 0.0%). Pooling of the results yielded a common RR of 0.32 (95% confidence interval 0.16 to 0.66). No
small-study effect was detected (Egger’s test). Screening for adenomyosis before suggesting difficult and risky procedures may allow

http://dx.doi.org/10.1016/j.rbmo.2014.02.006
1472-6483/ª 2014, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.
Adenomyosis, deep endometriosis and reproductive performance
identification of a subgroup of patients at particularly worse prognosis for which surgery would have a marginal effect on the like-
lihood of conception. RBMOnline
ª 2014, Reproductive Healthcare Ltd. Published by Elsevier Ltd. All rights reserved.
KEYWORDS: adenomyosis, bowel endometriosis, infertility, rectovaginal endometriosis, surgery, systematic review
Introduction
Rectovaginal and colorectal endometriosis are usually asso-
ciated with pain symptoms (Chopin et al., 2005), but the
role of these severe disease forms in causing infertility is
less clear (de Ziegler et al., 2010; Vercellini et al., 2009a).
In spite of this, radical surgery, including vaginal excision
and colorectal resection, has been proposed as a
fertility-enhancing procedure (Bianchi et al., 2009; Darai
et al., 2010a, 2011; Mohr et al., 2005; Stepniewska et al.,
2009). According to three comprehensive literature reviews
(de Ziegler et al., 2010; Meuleman et al., 2011a; Vercellini
et al., 2009a), pregnancy rates after excision of rectovagi-
nal and colorectal endometriosis vary between 42% and 44%.
However, interpretation of data is complicated by the
often-undefined baseline fertility status of the study
population. Considering only women who were infertile
before surgery and who achieved conception spontaneously
after surgery results in much less optimistic estimates,
raising doubts on the real magnitude of the effect, if any
(Vercellini et al., 2012). Unfortunately, randomized
controlled trials to clarify this issue are not available and
probably will never be.
Moreover, surgery for rectovaginal and colorectal endo-
metriosis is associated with a high incidence of complica-
tions (e.g. permanent neurological bladder dysfunction
and rectovaginal fistula formation). In particular, bowel
resection is the main determinant of major morbidity (De
Cicco et al., 2011; Douay-Hauser et al., 2011; Kondo
et al., 2011a; Vercellini et al., 2009b), and some authors
suggest less aggressive approaches in the absence of suboc-
clusive symptoms (Donnez and Squifflet, 2010; Roman
et al., 2013). Therefore, the preoperative counselling of
infertile women whose main objective is conception is diffi-
cult, as the caring gynaecologist should balance accurately
risks and benefits before suggesting the procedure. Behind
the usual unfavourable prognostic factors, such as advanced
age, a long period of infertility and coexisting dyspermia,
attention should also be paid to concomitant conditions that
could further interfere with fertility. In particular, an asso-
ciation has been repeatedly reported between deeply infil-
trating endometriosis and uterine adenomyosis (Bazot
et al., 2004; Larsen et al., 2011; Levy et al., 2013). Accord-
ing to Kunz et al. (2005) and Kissler et al. (2006), adenomy-
osis is an even more important cause of infertility than
endometriosis per se.
Thus, patients with coexistent deep endometriosis and
uterine adenomyosis may constitute a subgroup with partic-
ularly poor reproductive prognosis. Mathieu d’Argent et al.
(2010) reported that the outcome of IVF in women with
colorectal endometriosis is similar to that observed in
women with tubal or male factor infertility. However, the
same group subsequently demonstrated that the likelihood
705
of conception is significantly reduced if IVF is performed
in women with uterine adenomyosis in addition to colorectal
endometriosis compared with women with colorectal endo-
metriosis but no concomitant adenomyosis (Ballester et al.,
2012).
Indeed, the relationship between adenomyosis and infer-
tility is still unclear, and literature data are inconsistent
(Campo et al., 2012; Leyendecker et al., 2006; Maheshwari
et al., 2012; Matalliotakis et al., 2005; Sunkara and Khan,
2012; Tocci et al., 2008; Tomassetti et al., 2013; Vercellini
et al., 2006a). Adenomyosis is currently reliably detected by
both transvaginal ultrasonography (TVUS) and magnetic res-
onance imaging (MRI) without the need for histological
examination of a biopsy specimen (Dueholm and Lundorf,
2007; Meredith et al., 2009). Therefore, the present work
deemed it interesting to perform a systematic literature
review and meta-analysis with the aim of defining the preg-
nancy rate after surgery for deep endometriosis, including
colorectal resection, specifically in patients who were
screened preoperatively for the presence of uterine adeno-
myosis associated with rectovaginal endometriosis. The
study hypothesis was that patients with uterine adenomyo-
sis in addition to rectovaginal and colorectal endometriosis
have a worse reproductive prognosis than those with
endometriotic lesions only. Disentangling this issue may
have important implications for patient counselling, surgical
strategies and women’s decisions.
Materials and methods
The present literature overview was conducted according to
the PRISMA guidelines for systematic reviews (Moher et al.,
2009). As published, deidentified data were used and this
study was exempt from Institutional Review Board approval.
Sources
This review was restricted to published research articles
that compared the pregnancy rate after excisional surgery
for rectovaginal and colorectal endometriosis in women
with adenomyosis detected at TVUS and/or pelvic MRI with
that observed in women without adenomyosis. Different
strategies were adopted to identify medical papers pub-
lished in the last decade on the effect of adenomyosis on
reproductive outcome after surgery for deep endometriosis.
A Medline search was conducted for literature published
from January 2003 to August 2013 using combinations of
the medical subject heading terms ‘rectovaginal endometri-
osis’, ‘colorectal endometriosis’, ‘rectosigmoid endometri-
osis’, ‘intestinal endometriosis’, ‘bowel endometriosis’,
‘deeply infiltrating endometriosis’, ‘surgery’, ‘infertility’
and ‘adenomyosis’. Only those publications written in
English were included. All pertinent articles were retrieved
706
and the relative reference lists were systematically
reviewed in order to identify further reports that could be
included in the meta-analysis. Moreover, review articles
published on endometriosis and adenomyosis in the same
time span were consulted and their reference lists searched
for potential additional studies. No attempt was made to
identify unpublished studies.
Study selection
Two authors (GB and LB) independently performed an initial
screening of the title and abstract of all articles to exclude
citations deemed irrelevant by both observers (e.g. if only
technical details or effect on pain were reported). Abstracts
of scientific meetings were not included. Studies in which
women undergoing surgery were categorized based on pre-
operative diagnosis of uterine adenomyosis, irrespective of
fertility status, were identified. Original articles were
selected based on the definition of the total number of
women seeking pregnancy after surgery and, when possible,
on the modality for achieving conception (spontaneous
attempts versus assisted reproductive techniques). Reports
on the effect of adenomyosis on spontaneous pregnancy
rate or after IVF or intracytoplasmic sperm injection (ICSI)
carried out without preliminary surgery for deep endometri-
osis were excluded.
The quality of studies was evaluated by means of the
Newcastle–Ottawa scale, a validated modality for assessing
observational and nonrandomized studies (Wells et al.,
2000). The scale uses a score system based on three major
criteria: selection of participants, comparability of study
groups and assessment of exposure. The quality checklist
includes eight items with a score of either 0 or 1 for each
item except ‘comparability of cohorts’ where a score of 0,
1 or 2 can be awarded. Therefore, the quantitative
appraisal of overall quality of the individual studies ranged
from 0 to 9. No cut-off level was set for inclusion in the
meta-analysis.
Data extraction
Studies were categorized based on research design. The
year of publication, location, setting, number and clinical
characteristics of recruited subjects, modality for the diag-
nosis of adenomyosis, surgical approach, type of surgical
treatment and length of follow up were recorded. The num-
ber of patients seeking pregnancy and who conceived were
obtained from individual reports. The post-operative preg-
nancy rate was separately determined in women with or
without associated uterine adenomyosis. When possible,
only spontaneous conceptions were considered. If the distri-
bution of pregnancies resulting from spontaneous attempts
and from IVF/ICSI could not be discriminated between the
two study groups, this was specified and both types of con-
ception were included in the analyses. The two observers
independently evaluated all articles and abstracted data
onto standardized forms. A final extraction form was com-
piled from the two evaluation forms, with correction or res-
olution of any discrepancies between reviewers by
consensus reached after discussion or arbitration by a third
reviewer (PV).
P Vercellini et al.
Data analysis
The dichotomous data for clinical pregnancy were
expressed as risk ratios (RR) with 95% confidence intervals
(CI), and were combined in a meta-analysis using a
fixed-effect model (Mantel–Haenszel method; Egger
et al., 2001). When assessing the summary graph, the out-
come of clinical pregnancy was a positive effect, and a
higher pregnancy rate was considered a benefit. The heter-
ogeneity Cochrane Q and the I2 statistics, which describe
the proportion of the total variation of estimates across
studies due to heterogeneity rather than chance (Higgins
et al., 2003), were then calculated. Negative values of I2
are set equal to 0 so that I2 lies between 0% and 100%. A
value of 0% indicates no observed heterogeneity, whereas
I2 values of 25%, 50% and 75% indicate low, moderate and
high heterogeneity, respectively (Higgins et al., 2003).
A Begg’s funnel plot, which plots RR on a log scale
(effect) against standard error of log RR (precision), was
generated and visually inspected for asymmetry to deter-
mine if the included studies were nonrepresentative of
the body of possible studies on the subject (which could
result from small study effect or other biases, such as pub-
lication and poor-quality bias). Egger’s approach to test fun-
nel plot asymmetry was also used (Egger et al., 1997). All
analyses were performed using Stata version 13 (StataCorp,
College Station, TX, USA).
Results
Figure 1 shows the flow diagram of the literature search
results. The Medline search identified 422 records. Another
three citations were found from scanning of reference lists.
A total of 396 publications were excluded because it was
clear from the abstract that they did not fulfil the selection
criteria. Full manuscripts of the remaining 29 articles were
obtained. Twenty-two publications were excluded because
no attempts were made to categorize patients based on
presence/absence of adenomyosis (Chopin et al., 2005;
Darai et al., 2010a; Donnez and Squifflet, 2010; Douay-
Hauser et al., 2011; Fleisch et al., 2005; Ghezzi et al., 2008;
Houtmeyers et al., 2006; Jelenc et al., 2012; Juhasz-Böss
et al., 2010; Kavallaris et al., 2011; Keckstein and
Wiesinger, 2005; Kondo et al., 2011b; Lyons et al., 2006;
Meuleman et al., 2009, 2011b, 2012; Minelli et al., 2009;
Mohr et al., 2005; Neme et al., 2013; Stepniewska et al.,
2009; Vercellini et al., 2006b; Vignali et al., 2005), one
was excluded because the number of women who conceived
in the groups with or without adenomyosis was not reported
(Darai et al., 2011) and one was excluded because the prev-
alence of adenomyosis among women seeking conception
was not specified (Acién et al., 2013).
Data on the effect of adenomyosis on pregnancy rate
after surgery for rectovaginal and colorectal endometriosis
were extracted from the remaining five articles, all of which
were published in full in peer-reviewed journals between
2005 and 2010 (Darai et al., 2005, 2010b; Landi et al., 2008;
Ferrero et al., 2009; Stepniewska et al., 2010). Complete
reviewers’ agreement regarding inclusion and exclusion of
studies was achieved in all instances. The characteristics
of the selected studies are shown in Table 1. All included
Adenomyosis, deep endometriosis and reproductive performance
Figure 1 Flowchart of the study selection process.
studies were observational, two of which were retrospec-
tive cohort studies (Darai et al., 2005; Landi et al., 2008)
and two were prospective cohort studies (Darai et al.,
2010b; Ferrero et al., 2009). In one article, the authors
stated that data were prospectively recorded, although it
was not formally indicated if the study was prospective or
retrospective (Stepniewska et al., 2010). Three studies
were conducted in Italy (Ferrero et al., 2009; Landi et al.,
2008; Stepniewska et al., 2010) and two in France (Darai
et al., 2005, 2010b). The number of recruited subjects varied
from 38 (Darai et al., 2005) to 83 (Darai et al., 2010b), with a
mean age at surgery between 32 and 33 years.
The diagnosis of uterine adenomyosis was based on TVUS
in one study (Stepniewska et al., 2010), on MRI in one (Landi
et al., 2008) and on both modalities in three (Darai et al.,
2005, 2010b; Ferrero et al., 2009). Discrimination between
focal and diffuse adenomyosis was never specified. In four
studies (Darai et al., 2005, 2010b; Ferrero et al., 2009;
Stepniewska et al., 2010), the prevalence of the condition
was fairly consistent, varying from 17% (Ferrero et al., 2009)
to 20% (Darai et al., 2010b; Stepniewska et al., 2010), with a
mean of 19%. In one study (Landi et al., 2008), two
707
numerically identical groups of patients with or without
adenomyosis were retrospectively selected.
Preoperative medical treatment (gonadotrophin-releasing
hormone agonists for 3 months) was used in two studies
(Darai et al., 2005, 2010b). Surgery was performed at lapa-
roscopy in three studies (Darai et al., 2005; Landi et al.,
2008; Stepniewska et al., 2010) and at laparotomy or lapa-
roscopy in two (Darai et al., 2010b; Ferrero et al., 2009).
In four studies, only patients who underwent colorectal
resection were included (Darai et al., 2005, 2010b; Ferrero
et al., 2009; Stepniewska et al., 2010), whereas in one
study, segmental bowel resection in addition to radical exci-
sion of deep lesions was performed ‘when necessary’
(Landi et al., 2008). Uterine adenomyosis was never excised
at surgery for endometriosis and in no study was post-
operative medical treatment used. Major complications
were reported in four studies, with incidence varying from
9% (Darai et al., 2005; Ferrero et al., 2009) to 12% (Darai
et al., 2010b).
The quality score of the considered studies ranged from 5
to 7 and the score for all five studies was (mean ± SD)
5.8 ± 0.8. In all studies there was adequate selection of
Table 1 Characteristics of the selected studies on the effect of adenomyosis on reproductive outcome after surgery for rectovaginal and colorectal endometriosis (2003–2013).
Year Country Study Design Modality and No. of Preop- Surgical Major Women Women with Duration Length Quality
Source
population criteria for patients erative approach compli- seeking adenomyosis of of of
diagnosis of enrolled medical cations pregnancy seeking infertility follow evidence
adenomyosis therapy pregnancy (months) up
(months)

Darai et al. 2005 France Women undergoing surgery RCS TVUS and MRI; 38 GnRH LPS 2/22 (9) 22 (58) 4 (18) >24 24 5
for colorectal endometriosis, criteria not agonists (6–42)
pain and/or infertility; mean specified for
age 31 years (range 29–42) 3 months
Landi et al. 2008 Italy Women undergoing surgery RCS MRI: junctional 80 None LPS NR 58 (73) 26 (45)a >12 20 7
for pain and infertility; zone of 12 mm, (3–33)b
radical excision of lesions ill-defined
and segmental bowel low-signal
resection when needed; intensity area of
mean age 33 ± 4 years myometrium or
spot-like
high- intensity
myometrial foci
x x x x x x x x x x x x x 21 x
(6–31)c
Ferrero et al. 2009 Italy Women undergoing PCS TVUS and MRI; 46 None LPS/LPT 4/46 (9) 46 (100) 8 (17) 25 50 ± 24 5
rectosigmoid resection for criteria not (18–36)
symptomatic endometriosis, specified
and who sought conception
post operatively; mean age
33 ± 3 years (<40).
Darai et al. 2010b France Women undergoing PCS TVUS and MRI; 83 GnRH LPS/LPT 10/83 55 (66) 11 (20) 48 34 6
segmental colorectal criteria not agonists (12) (24–120) (6–68)
resection for symptomatic specified for
endometriosis; mean age 3 months
32 years (range 21–45).
Stepniewska 2010 Italy Women undergoing NR TVUS; criteria 62 None LPS 6/62 50 (81) 10 (20) 30 20 6
et al. segmental bowel resection not specified (10) (12–168)d (6–48)
for symptomatic
endometriosis and infertility;
mean age 32 years (range
24–39).
x 20 x x
(12–72)e

Values are n (%), mean ± SD or median (range). Quality of evidence is scored according to the Newcastle–Ottawa scale (Wells et al., 2000).
LPS = laparoscopy; LPT = laparotomy; MRI = magnetic resonance imaging; NR = not reported; PCS = prospective cohort study; RCS = retrospective cohort study; TVUS = transvaginal ultrasonography.
a
Not a prevalence estimate, as women were matched for presence/absence of adenomyosis.
b
Concomitant endometriosis and adenomyosis.
c
Endometriosis only.
d
Wishing to conceive but did not.
e
Wishing to conceive and achieved pregnancy.
Adenomyosis, deep endometriosis and reproductive performance
women with and without adenomyosis. Adenomyosis was
always identified preoperatively using standard nonsurgical
modalities. However, the criteria for ascertainment of pres-
ence of adenomyosis were described in detail in only one
study (Landi et al., 2008). The outcome of interest was never
present at start of the study. The overall performance on
comparability of participants was fair in three studies (Darai
et al., 2010b; Landi et al., 2008; Stepniewska et al., 2010).
Potential confounders were controlled in only one study
(Darai et al., 2010b). Assessment of outcome was presumably
based on standardized criteria, but the definition for the
determination of clinical pregnancy was formally never spec-
ified. The mean duration of follow up was adequate in all
studies, varying between 20 (Stepniewska et al., 2010) and
50 (Ferrero et al., 2009) months, but in none was it possible
to ascertain whether losses to follow up occurred.
A total of 231 women desiring pregnancy after surgery
for deeply infiltrating endometriosis were identified: 59
with and 172 without adenomyosis. The number of women
seeking conception varied from 22 (Darai et al., 2005) to
58 (Landi et al., 2008). Generally, information was available
on pre-existence and duration of infertility or concomitance
of additional infertility factors other than endometriosis,
but the distribution of these variables between the groups
of patients with or without adenomyosis was never speci-
fied. Only infertile women were included in one study
(Stepniewska et al., 2010). The number of spontaneous con-
ceptions could be abstracted in two studies (Darai et al.,
2005; Landi et al., 2008), whereas in three (Darai et al.,
2010b; Ferrero et al., 2009; Stepniewska et al., 2010),
events in women with and without adenomyosis also
included pregnancies resulting from IVF/ICSI. In three stud-
ies (Darai et al., 2005; Ferrero et al., 2009; Stepniewska
et al., 2010), no pregnancy occurred in women over 35 years
Figure 2 Forest plot showing individual and Mantel–Haenszel combined risk ratios in studies that evaluated the likelihood of
clinical pregnancy in women with or without uterine adenomyosis undergoing surgery for deep endometriosis including colorectal
resection. Dots indicate study-specific risk ratios; boxes indicate the study-specific weight; horizontal lines indicate 95% confidence
intervals; rhombus indicates combined RR and 95% CI; dashed vertical line indicates the overall estimate. NS = not statistically
significant.
709
of age, and in the study by Darai et al. (2010b), only one of
24 women who conceived was over 35 years of age.
Overall, among women seeking pregnancy, 7/59 (11.9%)
with concomitant adenomyosis conceived, compared with
74/172 (43.0%) in those without adenomyosis. The RR of
clinical pregnancy ranged from 0.23 (Ferrero et al., 2009)
to 0.46 (Landi et al., 2008), with absence of heterogeneity
among studies (I2 = 0.0%). Therefore, the fixed
random-effects model was used to compute the overall
RR. Pooling of the results derived from the five included
studies yielded a common RR of 0.32 (95% CI 0.16–0.66),
demonstrating that adenomyosis was associated with a
68% reduction in the likelihood of pregnancy in women seek-
ing conception after surgery for severe rectovaginal and
colorectal endometriosis (Figure 2). In the three studies
in which it was not possible to discriminate the number of
spontaneous conceptions in women with and without
adenomyosis (Darai et al., 2010b; Ferrero et al., 2009;
Stepniewska et al., 2010), 68 pregnancies were reported,
22 of which (32.4%) resulting from IVF/ICSI. These figures
include some women who conceived twice as reported by
Darai et al. (2010b).
Begg’s funnel plot showed no indication of asymmetry
among studies (Egger’s test; Figure 3). Considering the
absence of heterogeneity and of small-study effect, a sensi-
tivity analysis to verify if any one study unduly influenced
the pooled effect size was not conducted.
Discussion
The results of the present systematic review demonstrate
that the presence of uterine adenomyosis is strongly associ-
ated with post-operative reproductive outcome in women
710
Figure 3 Begg’s funnel plot with pseudo 95% confidence
intervals of the estimate in five studies in which the effect of
adenomyosis on the likelihood of clinical pregnancy in infertile
women undergoing surgery for deep endometriosis including
colorectal resection was evaluated. No indication of asymmetry
was observed. RR = risk ratio.
undergoing radical surgery for deep endometriosis including
colorectal resection. At meta-analysis, coexistence of
uterine adenomyosis in addition to deep endometriosis
was associated with a 68% reduction in the likelihood of
pregnancy. The RR reduction among women with character-
istics comparable to the study population would be between
34% and 84%.
These estimates should be considered with caution, as
the quality of the assessed studies was suboptimal, and con-
founding and selection bias cannot be excluded. The number
of recruited subjects was limited, and only 59 patients with
coexistent uterine adenomyosis and deep/colorectal endo-
metriosis seeking pregnancy were evaluated. Qualitative
heterogeneity was substantial, as studies included either
only infertile women (Stepniewska et al., 2010), or also
those who did not seek conception before surgery, or even
parous ones (Darai et al., 2010b; Ferrero et al., 2009). The
majority of the procedures were performed because of pain
and not because of infertility. Not all women underwent
their first operation, and it has been reported that reproduc-
tive performance in patients with endometriosis was signifi-
cantly different at first- compared with second-line surgery
(Vercellini et al., 2009c). The five considered studies had a
suboptimal design, as all were observational, and two (pos-
sibly three) of them were retrospective. Moreover, not only
infertile patients were recruited. In the single study includ-
ing only infertile patients (Stepniewska et al., 2010), not all
of them wished to conceive after surgery and, when consid-
ering their relatively young age (all the patients who con-
ceived were aged <35 years) and the fact that their
endometriosis-related symptoms might have somehow inter-
fered with their sexual life, many had a rather short (1 year)
duration of infertility. Moreover, many patients in this study
had ovarian endometriomas.
The diagnosis of uterine adenomyosis was based on valid
and reliable modalities (Dueholm and Lundorf, 2007;
Meredith et al., 2009), but the depth and extension of
lesions could not be defined, impeding the evaluation of
any dose–response relation (Gordts et al., 2008; Vercellini
et al., 2006a). Exclusion from the analysis of subjects lost
P Vercellini et al.
to follow up may have inflated the purported benefit of
surgery, as dropouts generally have a worse prognosis. In
general, the finding that the prevalence of adenomyosis in
women with severe deep endometriosis was not substan-
tially higher than that observed in parous women undergo-
ing hysterectomy for various reasons (Naftalin et al.,
2012; Vercellini et al., 2006a) is not consistent with the
much higher estimates reported previously (Kissler et al.,
2006; Kunz et al., 2005; Leyendecker et al., 2006). If the
prevalence of adenomyosis in patients with endometriosis
is similar to that observed in the population of patients
without endometriosis, a causal relationship between endo-
metriosis and adenomyosis seems unlikely.
The best hypothetical study cohort would include only
women who are infertile before surgery, who do not have
other infertility factors in addition to adenomyosis and
endometriosis and who are undergoing their first operation
for endometriosis. Moreover, only spontaneous post-
operative conceptions should be considered, as the effect
of surgery with colorectal resection for deep endometriosis
on the likelihood of pregnancy at IVF/ICSI is undefined.
Ascribing IVF/ICSI pregnancies to this kind of surgery seems
currently unfounded. The type of adenomyosis should be
described in detail in order to discriminate between the
nodular and the diffuse form and, in the latter case,
between limited, sub-basalis lesions and those infiltrating
most of the uterine wall (Gordts et al., 2008; Vercellini
et al., 2006a). Results should be managed according to an
intention-to-treat principle, allowing for women lost to fol-
low up. Finally, a multivariate approach should be adopted
when analysing data, as studies in this specific area are
observational. Therefore, several variables that could be
associated with the outcome should be adequately con-
trolled for. Factors such as patients’ age, ovarian reserve,
previous surgery, presence of endometriomas and other
infertility factors must be taken into account, especially
considering that some already-observed associations
between the aforementioned factors and adenomyosis or
deep endometriosis may enhance confounding. For
instance, it is not possible to exclude that the effect of ade-
nomyosis here described is actually due to patient age,
adhesions or impaired ovarian function.
In spite of the above considerations, quantitative heter-
ogeneity was absent and both the direction and the magni-
tude of the effect of adenomyosis were exceedingly
consistent among studies. A small study effect or other
biases, such as publication and poor-quality bias, were not
demonstrated at Begg’s funnel plot inspection, and no sig-
nificant asymmetry was detected. Thus, the included stud-
ies were likely representative of the body of possible
studies on the subject. Moreover, the data included in this
analysis constitute the only available evidence on which to
base clinical understanding and treatment decision making.
This work performed a thorough literature review adopt-
ing an accepted modality for article search. To avoid major
bias in data gathering, these were extracted from the
reports of two independent observers who were not
blinded. Rejected studies and the reason for their exclusion
have been described. A formal evaluation of the quality of
the selected studies was performed, heterogeneity was
explored in a funnel plot and potential sources of confound-
ing and bias have been addressed (Egger et al., 1997).
Adenomyosis, deep endometriosis and reproductive performance
The finding of a strong association between concomitant
uterine adenomyosis and reproductive outcome after com-
plex surgery for severe deep endometriosis is of concern
and supports the performance of systematic preoperative
screening for adenomyosis using pelvic MRI scan or TVUS.
Independently of the purported pathogenic mechanisms
involved (Kissler et al., 2006; Kunz et al., 2005; Tomassetti
et al., 2013), great caution should be adopted in these
patients before suggesting major surgery when the main
objective is attainment of a conception. As the mean age
of women seeking their first pregnancy is progressively
increasing, presumably adenomyosis will be encountered
more and more frequently in the infertility clinic (Devlieger
et al., 2003).
If the results of this systematic review are confirmed,
women with rectovaginal/colorectal endometriosis and
concomitant uterine adenomyosis should be informed that
surgery may not have an appreciable effect on the likeli-
hood of conception and that the same or possibly better
chances of pregnancy could be obtained through IVF/ICSI
(Ballester et al., 2012; Mijatovic et al., 2010). Indeed, reli-
able evidence supporting radical endometriosis surgery as a
fertility-enhancing procedure is scarce and weak, regardless
of the presence of adenomyosis. In this context, the results
of this meta-analysis may be important for counselling
patients undergoing surgery for their symptoms regarding
their post-operative chances to conceive. However, the
choice of surgery for patients, with or without adenomyosis,
with the main goal of increasing their fertility is currently
not based on robust evidence. In these cases, surgery should
be considered as a last option, when no other less-invasive
treatment is available. This seems particularly important
in infertile women aged >35 years of age, whose reproduc-
tive potential would not benefit from the complex and risky
surgery needed in these difficult conditions.
The overall major complication rate observed in the five
selected studies is around 10%, which is consistent with find-
ings of other literature reviews and with the experience of
several authors (De Cicco et al., 2011; Kondo et al., 2011a;
Vercellini et al., 2009b). Moreover, it may not be excluded
that the complication rate in the hands of most gynaecolo-
gists is higher than that reported by particularly skilful and
expert surgeons. In addition, a publication bias is likely, as
surgeons with unfavourable outcomes may be less willing
to try to publish their data.
Unbearable pain in women wanting a spontaneous preg-
nancy still constitutes an indication for excisional treatment
of rectovaginal endometriosis in infertile subjects. However,
when adenomyosis is also present, patients should be
informed in detail of the very low probability of post-
operative conception and the possibility of treating symp-
toms with medical therapy (Vercellini et al., 2009d) and
trying to achieve a pregnancy with IVF/ICSI. If pain is toler-
able, symptomatic bowel stenosis is absent and large
endometriomas or cysts with alarming ultrasonographic
characteristics are not observed, the latter alternative
should be favoured, especially in women aged >35 years of
age. Bowel resection in women with concomitant adenomy-
osis should be reserved for cases with subocclusive symptoms
and should never be performed with the exclusive objective
of increasing the chances of conception. In general, it has
been repeatedly demonstrated that, when feasible, bowel
711
lesion shaving or disk excision is equally effective but less
risky than colorectal resection for rectosigmoid endometri-
osis (Donnez and Squifflet, 2010; Roman et al., 2013).
Further data are needed to specify the extent of the con-
nection between adenomyosis and deep endometriosis and
to clarify if resection of uterine adenomyotic lesions (that
were never treated in the selected studies) in addition to
excision of deep endometriosis is associated with a better
reproductive outcome compared with radical treatment of
pelvic endometriotic lesions only. If further data suggest
that adenomyosis could be the major infertility factor, an
alternative option to be studied would be excision of adeno-
myosis only and not endometriosis. Overall, the findings of
the present systematic review suggest an association
between adenomyosis and infertility and call for further
investigation in this elusive and little explored area of
research. The possible role of adenomyosis in reproductive
failure should be taken into account in women with severe
endometriosis, in order to avoid addressing only part of
the problem.
References
Acién, P., Núñez, C., Quereda, F., Velasco, I., Valiente, M., Vidal,
V., 2013. Is a bowel resection necessary for deep endometriosis
with rectovaginal or colorectal involvement? Int. J. Womens
Health 29, 449–455.
Ballester, M., d’Argent, E.M., Morcel, K., Belaisch-Allart, J.,
Nisolle, M., Daraı̈, E., 2012. Cumulative pregnancy rate after
ICSI-IVF in patients with colorectal endometriosis: results of a
multicentre study. Hum. Reprod. 27, 1043–1049.
Bazot, M., Darai, E., Hourani, R., Thomassin, I., Cortez, A., Uzan,
S., Buy, J.N., 2004. Deep pelvic endometriosis: MR imaging for
diagnosis and prediction of extension of disease. Radiology 232,
379–389.
Bianchi, P.H., Pereira, R.M., Zanatta, A., Alegretti, J.R., Motta,
E.L., Serafini, P.C., 2009. Extensive excision of deep infiltrative
endometriosis before in vitro fertilization significantly improves
pregnancy rates. J. Minim. Invasive Gynecol. 16, 174–180.
Campo, S., Campo, V., Benagiano, G., 2012. Adenomyosis and
infertility. Reprod. Biomed. Online 24, 35–46.
Chopin, N., Vieira, M., Borghese, B., Foulot, H., Dousset, B., Coste,
J., Mignon, A., Fauconnier, A., Chapron, C., 2005. Operative
management of deeply infiltrating endometriosis: results on
pelvic pain symptoms according to a surgical classification. J.
Minim. Invasive Gynecol. 12, 312–317.
Darai, E., Marpeau, O., Thomassin, I., Dubernard, G., Barranger, E.,
Bazot, M., 2005. Fertility after laporascopic colorectal resection
for endometriosis: preliminary results. Fertil. Steril. 84, 945–950.
Darai, E., Dubernard, G., Coutant, C., Frey, C., Rouzier, R.,
Ballester, M., 2010a. Randomized trial of laparoscopically
assisted versus open colorectal resection for endometriosis:
morbidity, symptoms, quality of life, and fertility. Ann. Surg.
251, 1018–1023.
Darai, E., Carbonnel, M., Dubernard, G., Lavoué, V., Coutant, C.,
Bazot, M., Ballester, M., 2010b. Determinant factors of fertility
outcomes after laparoscopic colorectal resection for endome-
triosis. Eur. J. Obstet. Gynecol. Reprod. Biol. 149, 210–214.
Darai, E., Lesieur, B., Dubernard, G., Rouzier, R., Bazot, M.,
Ballester, M., 2011. Fertility after colorectal resection for
endometriosis: results of a prospective study comparing lapa-
roscopy with open surgery. Fertil. Steril. 95, 1903–1908.
De Cicco, C., Corona, R., Schonman, R., Mailova, K., Ussia, A.,
Koninckx, P., 2011. Bowel resection for deep endometriosis: a
systematic review. BJOG 118, 285–291.
712
Devlieger, R., D’Hooghe, T., Timmermann, D., 2003. Uterine
adenomyosis in the infertility clinic. Hum. Reprod. Update 9,
139–147.
de Ziegler, D., Borghese, B., Chapron, C., 2010. Endometriosis and
infertility: pathophysiology and management. Lancet 376,
730–738.
Donnez, J., Squifflet, J., 2010. Complications, pregnancy and
recurrence in a prospective series of 500 patients operated on by
the shaving technique for deep rectovaginal endometriotic
nodules. Hum. Reprod. 25, 1949–1958.
Douay-Hauser, N., Yazbeck, C., Walker, F., Luton, D., Madelenat,
P., Koskas, M., 2011. Infertile women with deep and intraperi-
toneal endometriosis: comparison of fertility outcome according
to the extent of surgery. J. Minim. Invasive Gynecol. 18,
622–628.
Dueholm, M., Lundorf, E., 2007. Transvaginal ultrasound or MRI for
diagnosis of adenomyosis. Curr. Opin. Obstet. Gynecol. 19,
505–512.
Egger, M., Smith, G.D., Phillips, A.N., 1997. Meta-analysis: princi-
ples and procedures. BMJ 315, 1533–1537.
Egger, M., Davey, S.G., Altman, D.G., 2001. Systematic Reviews in
Health Care: Meta-Analysis in Context, Part IV, second ed. BMJ
Publishing Group, pp. 285–312 (Chapters 15–16).
Ferrero, S., Anserini, P., Abbamonte, L.H., Ragni, N., Camerini, G.,
Remorgida, V., 2009. Fertility after bowel resection for endo-
metriosis. Fertil. Steril. 92, 41–46.
Fleisch, M.C., Xafis, D., De Bruyne, F., Hucke, J., Bender, H.G.,
Dall, P., 2005. Radical resection of invasive endometriosis with
bowel or bladder involvement––long-term results. Eur. J.
Obstet. Gynecol. Reprod. Biol. 123, 224–229.
Ghezzi, F., Cromi, A., Ciravolo, G., Rampinelli, F., Braga, M., Boni,
L., 2008. A new laparoscopic-transvaginal technique for recto-
sigmoid resection in patients with endometriosis. Fertil. Steril.
90, 1964–1968.
Gordts, S., Brosens, J.J., Fusi, L., Benagiano, G., Brosens, I., 2008.
Uterine adenomyosis: a need for uniform terminology and
consensus classification. Reprod. Biomed. Online 17, 244–248.
Higgins, J.P., Thompson, S.G., Deeks, J.J., Altman, D.G., 2003.
Measuring inconsistency in meta-analyses. BMJ 327, 557–560.
Houtmeyers, P., Ceelen, W., Gilardin, J.M., Dhondt, M., Pattyn, P.,
2006. Surgery for gastrointestinal endometriosis: indications and
results. Acta Chir. Belg. 106, 413–416.
Jelenc, F., Ribič-Pucelj, M., Juvan, R., Kobal, B., Sinkovec, J.,
Salamun, V., 2012. Laparoscopic rectal resection of deep
infiltrating endometriosis. J. Laparoendosc. Adv. Surg. Tech. A
22, 66–69.
Juhasz-Böss, I., Lattrich, C., Fürst, A., Malik, E., Ortmann, O.,
2010. Severe endometriosis: laparoscopic rectum resection.
Arch. Gynecol. Obstet. 281, 657–662.
Kavallaris, A., Chalvatzas, N., Hornemann, A., Banz, C., Diedrich,
K., Agic, A., 2011. 94 months follow-up after laparoscopic
assisted vaginal resection of septum rectovaginale and recto-
sigmoid in women with deep infiltrating endometriosis. Arch.
Gynecol. Obstet. 283, 1059–1064.
Keckstein, J., Wiesinger, H., 2005. Deep endometriosis, including
intestinal involvement––the interdisciplinary approach. Minim.
Invasive Ther. Allied Technol. 14, 160–166.
Kissler, S., Hamscho, N., Zangos, S., Wiegratz, I., Schlichter, S.,
Menzel, C., Doebert, N., Gruenwald, F., Vogl, T.J., Gaetjie, R.,
Rody, A., Siebzehnruebl, E., Kunz, G., Leyendecker, G., Kauf-
mann, M., 2006. Uterotubal transport disorder in adenomyosis
and endometriosis-a cause for infertility. BJOG 113, 902–908.
Kondo, W., Bourdel, N., Tamburro, S., Cavoli, D., Jardon, K.,
Rabischong, B., Botchorishvili, R., Pouly, J., Mage, G., Canis, M.,
2011a. Complications after surgery for deeply infiltrating pelvic
endometriosis. BJOG 118, 292–299.
Kondo, W., Darai, E., Yazbeck, C., Panel, P., Tamburro, S.,
Dubuisson, J., Jardon, K., Mage, G., Madelanat, P., Canis, M.,
P Vercellini et al.
2011b. Do patients manage to achieve pregnancy after a major
complication of deeply infiltrating endometriosis resection? Eur.
J. Obstet. Gynecol. Reprod. Biol. 154, 196–199.
Kunz, G., Beil, D., Huppert, P., Noe, M., Kissler, S., Leyendecker,
G., 2005. Adenomyosis in endometriosis––prevalence and
impact on fertility. Evidence from magnetic resonance imaging.
Hum. Reprod. 20, 2309–2316.
Landi, S., Mereu, L., Pontrelli, G., Stepniewska, A., Romano, L.,
Tateo, S., Dorizzi, C., Minelli, L., 2008. The influence of
adenomyosis in patients laparoscopically treated for deep
endometriosis. J. Minim. Invasive Gynecol. 15, 566–570.
Larsen, S.B., Lundorf, E., Forman, A., Dueholm, M., 2011. Adeno-
myosis and junctional zone changes in patients with endometri-
osis. Eur. J. Obstet. Gynecol. Reprod. Biol. 157, 206–211.
Levy, G., Dehaene, A., Laurent, N., Lernout, M., Collinet, P.,
Lucot, J.P., Lions, C., Poncelet, E., 2013. An update on
adenomyosis. Diagn. Interv. Imaging 94, 3–25.
Leyendecker, G., Kunz, G., Kissler, S., Wildt, L., 2006. Adenomyosis
and reproduction. Best Pract. Res. Clin. Obstet. Gynaecol. 20,
523–546.
Lyons, S.D., Chew, S.S., Thomson, A.J., Lenart, M., Camaris, C.,
Vancaillie, T.G., Abbott, J.A., 2006. Clinical and quality-of-life
outcomes after fertility-sparing laparoscopic surgery with bowel
resection for severe endometriosis. J. Minim. Invasive Gynecol.
13, 436–441.
Maheshwari, A., Gurunath, S., Fatima, F., Bhattacharya, S., 2012.
Adenomyosis and subfertility: a systematic review of preva-
lence, diagnosis, treatment and fertility outcomes. Hum.
Reprod. Update 18, 374–392.
Matalliotakis, I.M., Katsikis, I.K., Panidis, D.K., 2005. Adenomyosis:
what is the impact on fertility? Curr. Opin. Obstet. Gynecol. 17,
261–264.
Mathieu d’Argent, E., Coutant, C., Ballester, M., Dessolle, L.,
Bazot, M., Antoine, J.M., Daraı̈, E., 2010. Results of first in vitro
fertilization cycle in women with colorectal endometriosis
compared with those with tubal or male factor infertility.
Fertil. Steril. 94, 2441–2443.
Meredith, S.M., Sanchez-Ramos, L., Kaunitz, A.M., 2009. Diagnostic
accuracy of transvaginal sonography for the diagnosis of adeno-
myosis: systematic review and meta-analysis. Am. J. Obstet.
Gynecol. 201, e1–E6.
Meuleman, C., D’Hoore, A., Van Cleynenbreugel, B., Beks, N.,
D’Hooghe, T., 2009. Outcome after multidisciplinary CO2 laser
laparoscopic excision of deep infiltrating colorectal endometri-
osis. Reprod. Biomed. Online 18, 282–289.
Meuleman, C., Tomassetti, C., D’Hoore, A., Van Cleynenbreugel,
B., Pennickx, F., Vergote, I., D’Hooghe, T., 2011a. Surgical
treatment of deeply infiltrating endometriosis with colorectal
involvement. Hum. Reprod. Update 17, 311–326.
Meuleman, C., Tomassetti, C., D’Hoore, A., Buyens, A., Van
Cleynenbreugel, B., Fieuws, S., Pennickx, F., Vergote, I.,
D’Hooghe, T., 2011b. Clinical outcome after CO2 laser radical
excision of endometriosis with colorectal wall invasion com-
bined with laparoscopic segmental bowel resection and reanas-
tomosis. Hum. Reprod. 26, 2336–2343.
Meuleman, C., Tomassetti, C., D’Hooghe, T.M., 2012. Clinical
outcome after laparoscopic radical excision of endometriosis
and laparoscopic segmental bowel resection. Curr. Opin. Obstet.
Gynecol. 24, 245–252.
Mijatovic, V., Florijn, E., Halim, N., Schats, R., Hompes, P., 2010.
Adenomyosis has no adverse effects on IVF/ICSI outcomes in
women with endometriosis treated with long-term pituitary
down-regulation before IVF/ICSI. Eur. J. Obstet. Gynecol.
Reprod. Biol. 151, 62–65.
Minelli, L., Fanfani, F., Fagotti, A., Ruffo, G., Ceccaroni, M.,
Mereu, L., Landi, S., Pomini, P., Scambia, G., 2009. Laparo-
scopic colorectal resection for bowel endometriosis: feasibility,
complications, and clinical outcome. Arch. Surg. 144, 234–239.
Adenomyosis, deep endometriosis and reproductive performance
Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G.PRISMA Group, 2009.
Preferred reporting items for systematic reviews and meta-anal-
yses: the PRISMA statement. Ann. Int. Med. 151, 264–269.
Mohr, C., Nezhat, F.R., Nezhat, C.H., Seidman, D.S., Nezhat, C.R.,
2005. Fertility considerations in laparoscopic treatment of
infiltrative bowel endometriosis. JSLS 9, 16–24.
Naftalin, J., Hoo, W., Pateman, K., Mavrelos, D., Holland, T.,
Jurkovic, D., 2012. How common is adenomyosis? A prospective
study of prevalence using transvaginal ultrasound in a gynaecol-
ogy clinic. Hum. Reprod. 27, 3432–3439.
Neme, R.M., Schraibman, V., Okazaki, S., Maccapani, G., Chen,
W.J., Domit, C.D., Kaufmann, O.G., Advincula, A.P., 2013. Deep
infiltrating colorectal endometriosis treated with
robotic-assisted rectosigmoidectomy. JSLS 17, 227–234.
Roman, H., Bridoux, V., Tuech, J.J., Marpeau, L., da Costa, C.,
Savoye, G., Puscasiu, L., 2013. Bowel dysfunction before and
after surgery for endometriosis. Am. J. Obstet. Gynecol. 13,
00364–513.
StataCorp, 2013. Commands metan, metafunnel, metabias. Stata:
Release 13. Statistical Software. StataCorp LP, College Station,
TX.
Stepniewska, A., Pomini, P., Bruni, F., Mereu, L., Ruffo, G.,
Ceccaroni, M., Scioscia, M., Guerriero, M., Minelli, L., 2009.
Laparoscopic treatment of bowel endometriosis in infertile
women. Hum. Reprod. 24, 1619–1625.
Stepniewska, A., Pomini, P., Scioscia, M., Mereu, L., Ruffo, G.,
Minelli, L., 2010. Fertility and clinical outcome after bowel
resection in infertile women with endometriosis. Reprod. Bio-
med. Online 20, 602–609.
Sunkara, S.K., Khan, K.S., 2012. Adenomyosis and female fertility: a
critical review of the evidence. J. Obstet. Gynaecol. 32, 113–116.
Tocci, A., Greco, E., Ubaldi, F.M., 2008. Adenomyosis and ‘endo-
metrial-subendometrial myometrium unit disruption disease’
are two different entities. Reprod. Biomed. Online 17, 281–291.
Tomassetti, C., Meuleman, C., Timmerman, D., D’Hooghe, T.,
2013. Adenomyosis and subfertility: evidence of association and
causation. Semin. Reprod. Med. 31, 101–108.
Vercellini, P., Viganò, P., Somigliana, E., Daguati, R., Abbiati, A.,
Fedele, L., 2006a. Adenomyosis: epidemiological factors. Best.
Pract. Res. Clin. Obstet. Gynaecol. 20, 465–477.
713
Vercellini, P., Pietropaolo, G., De Giorgi, O., Daguati, R., Pasin, R.,
Crosignani, P.G., 2006b. Reproductive performance in infertile
women with rectovaginal endometriosis: is surgery worthwhile?
Am. J. Obstet. Gynecol. 195, 1303–1310.
Vercellini, P., Somigliana, E., Viganò, P., Abbiati, A., Barbara, G.,
Crosignani, P.G., 2009a. Surgery for endometriosis-associated
infertility: a pragmatic approach. Hum. Reprod. 24, 254–269.
Vercellini, P., Carmignani, L., Rubino, T., Barbara, G., Abbiati, A.,
Fedele, L., 2009b. Surgery for deep endometriosis: a pathogen-
esis-oriented approach. Gynecol. Obstet. Invest. 68,
88–103.
Vercellini, P., Somigliana, E., Viganò, P., De Matteis, S., Barbara,
G., Fedele, L., 2009c. The effect of second-line surgery on
reproductive performance of women with recurrent endometri-
osis: a systematic review. Acta Obstet. Gynecol. Scand. 88,
1074–1082.
Vercellini, P., Crosignani, P.G., Somigliana, E., Berlanda, N.,
Barbara, G., Fedele, L., 2009d. Medical treatment for recto-
vaginal endometriosis: what is the evidence? Hum. Reprod. 24,
2504–2514.
Vercellini, P., Barbara, G., Buggio, L., Frattaruolo, M.P., Somigli-
ana, E., Fedele, L., 2012. Effect of patient selection on estimate
of reproductive success after surgery for rectovaginal endome-
triosis: literature review. Reprod. Biomed. Online 24, 389–395.
Vignali, M., Bianchi, S., Candiani, M., Spadaccini, G., Oggioni, G.,
Busacca, M., 2005. Surgical treatment of deep endometriosis
and risk of recurrence. J. Minim. Invasive Gynecol. 12, 508–513.
Wells, G.A., Shea, B., O’Connell, D., Peterson, J., Welch, V.,
Tugwell, P., 2000. The Newcastle-Ottawa Scale (NOS) for
assessing the quality of nonrandomized studies in meta-analy-
ses. 3rd Symposium on Systematic Reviews: Beyond the Basics,
Oxford, U.K. Available from: <http://www.ohri.ca/programs/
clinical_epidemiology/oxford.htm>.
Declaration: The authors report no financial or commercial
conflicts of interest.
Received 20 September 2013; refereed 5 February 2014; accepted 5
February 2014.