You are here

Clinical implications of congenital uterine anomalies: a meta-analysis of comparative studies

Reproductive BioMedicine Online, 6, 29, pages 665 - 683

Abstract

The clinical implications of congenital uterine anomalies (CUA), and the benefits of hysteroscopic resection of a uterine septum, were evaluated. Studies comparing reproductive and obstetric outcome of patients with and without CUA and of patients who had and had not undergone hysteroscopic resection of a uterine septum, were evaluated. Meta-analysis of studies indicated that the pregnancy rate was decreased in women with CUA (RR 0.85, 95% CI 0.73 to 1.00; marginally significant finding,P= 0.05). The spontaneous abortion rate was increased in women with CUA (RR 1.68, 95% CI 1.31 to 2.15). Preterm delivery rates (RR 2.21, 95% CI 1.59 to 3.08), malpresentation at delivery (RR 4.75, 95% CI 3.29 to 6.84), low birth weight (RR 1.93, 95% CI 1.50 to 2.49) and perinatal mortality rates (RR 2.43, 95% CI 1.34 to 4.42) were significantly higher in women with CUA. Hysteroscopic removal of a septum was associated with a reduced probability of spontaneous abortion (RR 0.37, 95% CI 0.25 to 0.55) compared with untreated women. Presence of CUA might be associated with a detrimental effect on the probability of pregnancy achievement, spontaneous abortion and obstetric outcome. Hysteroscopic removal of a septum may reduce the probability of a spontaneous abortion.

Keywords: congenital uterine anomaly, miscarriage, obstetric outcome, pregnancy, treatment.

Introduction

Congenital uterine anomalies (CUA) represent a significant concern for most reproductive specialists, as it has been suggested that their presence is associated with adverse reproductive outcomes. The prevalence of CUA has been reported to vary, depending on the population and the diagnostic technique used (Chan et al, 2011b, Grimbizis et al, 2001, and Saravelos et al, 2008a); in the general population, CUA have been shown to affect 4.3–6.7% of the population, in the infertile population, the prevalence of CUA has been estimated to range between 3.4% and 8.0% and, in women characterized as ‘recurrent aborters’, this figure has been reported to be 12.6–18.2% (Chan et al, 2011b, Grimbizis et al, 2001, and Saravelos et al, 2008a).

Whether the presence of CUA affects the reproductive potential of the patient is still considered by many debatable. Different subtypes of uterine anomalies have been reported to be associated with the probability of pregnancy achievement, both in natural (Acien, 1993, Shuiqing et al, 2002, Sugiura-Ogasawara et al, 2010, and Woelfer et al, 2001) and assisted reproductive cycles (Jayaprakasan et al, 2011 and Tomazevic et al, 2009). Furthermore, certain types of CUA have been suggested to negatively affect the evolution of pregnancy, by means of an increased first- or second-trimester spontaneous abortion rate (Acien, 1993, Jayaprakasan et al, 2011, Saravelos et al, 2010, Sorensen, Trauelsen, 1987, Woelfer et al, 2001, and Zlopasa et al, 2007). Equally important is the fact that the presence of CUA has also been associated by many researchers with poor obstetric outcomes (Acien, 1993, Cooney et al, 1997, Hua et al, 2011, Woelfer et al, 2001, Zhang et al, 2010, and Zlopasa et al, 2007).

At the same time, the effectiveness of treating congenital uterine anomalies is still questionable. Even minimally invasive procedures, such as the hysteroscopic resection of a septum, are not universally accepted as effective treatments in reversing a possible adverse reproductive effect. Several studies, however, albeit not of high quality, suggest that these procedures may improve pregnancy rates in infertile patients (Colacurci et al, 1996, Lavergne et al, 1996, Mollo et al, 2009, Pabuccu, Gomel, 2004, Pabuccu et al, 1995, and Perino et al, 1987). As far as pregnancy outcome and obstetric complications are concerned, although correction of uterine septa is strongly associated with decreased abortion rates and preterm delivery rates (Fedele et al, 1993, Grimbizis et al, 1998, Grimbizis et al, 2001, Perino et al, 1987, Shuiqing et al, 2002, and Tomazevic et al, 2007), some investigators have suggested that surgical correction should be reserved only for specific cases (Acien, 1993 and Pang et al, 2011).

The first attempt to systematically summarize the existing literature in this field was by our group in 2001 ( Grimbizis et al., 2001 ). That review indicated the potential clinical implications of CUA, although a meta-analytic pooling of the data from the individual studies was not conducted. A meta-analysis was subsequently published in 2011 suggesting that the presence of certain types of CUA is associated with reduced clinical pregnancy rates, increased spontaneous abortion rates and preterm delivery rates, as well as increased probability of malpresentation at delivery in the general population ( Chan et al., 2011a ). Several clinical questions, however, still remain unanswered. The effect of CUA on infertile patients and patients experiencing recurrent spontaneous abortions has not yet been assessed and, debate is ongoing on the management of patients with this type of anomaly and whether or not these should be treated.

Considering the importance of this issue, an updated critical reappraisal of the available data is warranted to furnish the clinician with an up-to-date summary of answers to important clinical questions, and also enhance the awareness regarding the level of the existing evidence according to the principles of evidence-based medicine. At the same time, a re-evaluation of the published data can identify gaps in the relevant literature and may provide clinicians with future research objectives.

In this systematic review and meta-analysis, a holistic approach was used to gain a better insight into the clinical implications of CUA and the potential benefit of their treatment. For this purpose, the existing literature has been critically reviewed and an exhaustive updated meta-analysis has been carried out with the aim of providing answers to the following relevant clinical questions: is the presence of CUA associated with the probability of conception in a natural or an assisted reproductive technologies cycle? Is the presence of CUA associated with the evolution of pregnancy (in terms of spontaneous abortion rates)? Is the obstetric outcome different between pregnant women with CUA and women with a normal uterus? Is the hysteroscopic treatment of a septate uterus associated with a clinical benefit in terms of reproductive outcome?

Materials and methods

Sources

A computerized literature search in Pubmed, Scopus and Cochrane CENTRAL covering the period until December 2012 was independently conducted by two reviewers (CAV and SPP) either with the use of the MeSH term ‘uterus/abnormalities’ (Pubmed and Cochrane CENTRAL) or with a combination of the free-text terms ‘uter*’, ‘muller*’, ‘anomal*’, ‘abnormal*’, ‘malform*’ and ‘congenit*’. The aim of this search was to identify studies that evaluated the aforementioned research questions.

Study selection

Only comparative studies were considered eligible for this review; case reports or case series were excluded. Studies had to compare women with CUA and women with a normal uterus, and provide data on the probability of pregnancy achievement, spontaneous abortion rates and obstetric outcome. Furthermore, all studies comparing women treated hysteroscopically for uterine septum with women not treated, were also considered for this review. Studies with outcome data in the same women before and after treatment were excluded. Also, studies analysing women with uterine hypoplasia or agenesis and women with T-shaped uterus were not considered for this systematic review. Studies with overlapping data were excluded, unless it was feasible to extract non-overlapping data. Selection of the studies was carried out independently by two of the reviewers (CAV and SPP). Any disagreement was resolved by discussion with a third reviewer (GFG).

Data extraction

Data extraction was carried out independently by two of the reviewers (CAV and SPP). The following data were recorded from each of the eligible studies: demographic (citation data, country, study period, number of patients included), methodological (retrospective, prospective, cohort or not), population of the study (regarding both the ‘cases’ and ‘control’ groups), uterine anomalies evaluated and mode of diagnosis. Any disagreement between the two reviewers responsible for data extraction was resolved by discussion with a third reviewer (GFG).

Outcomes

The main outcome measures used in this meta-analysis were a-priori determined, depending on the research question evaluated. More specifically, the effect of CUA on the probability of conception (either spontaneous or after assisted reproductive technology), was examined by using pregnancy rate per patient. The effect of CUA on the evolution of pregnancy was evaluated by using first-trimester spontaneous abortion rates, second-trimester spontaneous abortion rates, or both, as outcome measures. The association of the presence of CUA with the obstetric outcome was assessed by examining preterm delivery rates. For this last question, the following secondary outcome measures was evaluated: premature delivery rates; neonatal birth weight; preeclampsia rates; placental abruption rates; intrauterine growth restriction (IUGR) rates; premature rupture of membranes rates; and perinatal mortality rates. Whenever possible, data extraction was carried out per CUA subtype and overall. American Fertility Society (AFS) classification (1988) was used as the basis for the categorization of the CUA, as it is the most popular system used to date and also was the classification used in most of the studies analysed. Studies in which a septate was further classified as ‘complete’, ‘partial’, ‘large’ or ‘small’ were all included under the common AFS classification ‘septated uterus’ (Class V). The reason for this was that the AFS classification system did not have a clear definition, nor did it have a universally accepted definition of what is a ‘partial’ or ‘small’ septum. Furthermore, the potential clinical implications of arcuate uteri were also evaluated given that this is also considered a type of CUA (Class VI). Finally, the potential beneficial effect of the hysteroscopic resection of a septum was evaluated by comparing first spontaneous abortion rates (first or second trimester) and, second, the probability of pregnancy achievement and the probability of preterm labour in women treated compared with those not treated.

Quantitative data synthesis

The dichotomous data results for each of the eligible for meta-analysis studies were expressed as relative risks (RR) with 95% confidence intervals (CI). This measure was selected on the basis of its clinical interpretability. The random effects model was used to pool the data of the individual studies ( DerSimonian and Laird, 1986 ). All results were combined for meta-analysis with Revman Software (Version 5.1. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2011). Study-to-study variation (statistical heterogeneity) was assessed by using the chi-squared statistic and the I-squared index (≥50% indicating substantial heterogeneity). Furthermore, differences between subgroups (e.g. types of CUA) were also compared. Statistical significance was set atP≤ 0.05.

Quality assessment of the eligible studies

The quality of the eligible studies was assessed with the use of a Newcastle–Ottawa scale ( Wells et al., 2000 ) that was adapted by CAV and GFG for the specific research questions and the data analysed ( Supplementary Table S1 ). A sensitivity analysis that excluded studies of low quality, as assessed by this scale, was also performed (where applicable) to test the robustness of the results obtained.

Subgroup analyses

Subgroup analyses according to the type of CUA or the population analysed in each study (infertile population, women who have experienced recurrent spontansoua abortions, non-specific group of patients) were carried out in an exploratory fashion to investigate a potential differential effect of the presence of CUA in these subpopulations.

Results

Systematic review

The literature search yielded 7166 publications in total. The titles of these manuscripts were screened, resulting in 103 studies considered potentially eligible to be included in the review. Of the total of 103 potentially relevant manuscripts identified, 65 studies were excluded after the examination of the abstracts and 38 studies were further evaluated by retrieving the full text. Finally, after the exclusion of 13 studies, 25 studies were included in the present systematic review and meta-analysis (Acien, 1993, Ban-Frangez et al, 2009, Ben-Rafael et al, 1991, Cooney et al, 1997, Erez et al, 2007, Heinonen, 1997, Hua et al, 2011, Jayaprakasan et al, 2011, Lavergne et al, 1996, Liang, Hu, 2010, Lin et al, 2009, Munoz Munoz et al, 2011, Pang et al, 2011, Ravasia et al, 1999, Saravelos et al, 2010, Shuiqing et al, 2002, Sorensen, Trauelsen, 1987, Sugiura-Ogasawara et al, 2010, Tomazevic et al, 2010, Tonguc et al, 2011, Valli et al, 2004, Woelfer et al, 2001, Zhang et al, 2010, Zlopasa et al, 2007, and Zupi et al, 1996) ( Figure 1 ). Characteristics of the studies included in the systematic review are presented in Table 1 .

rbmo1230-fig-0001

Figure 1 PRISMA flow diagram of search and retrieval of studies.

Table 1 Studies comparing reproductive outcome between women with and without uterine anomalies.

Study, country, journal Design Study period Sample size cases/controls Population Exclusion criteria Cases Controls Uterine anomalies evaluated Mode of diagnosis Method of classification of anomalies
Sorensen and Trauelsen, 1987 , Am J Obstet Gynecol Retrospective 1970–1973 and 1979–1980 50/141 Infertile women with minor Müllerian anomalies. Women that could not be traced (n = 7) and women with major uterine anomalies (n = 6). Infertile women with minor Müllerian duct anomalies. Infertile women with normal uterine cavity. Arcuate uterus. HSG. According to the degree of fundal excavation, mild anomaly is when H/L ≥ 0.1

H: the distance from the nadir of the fundal indentation to the line connecting the summits of the uterine horns and L: the length of this line.
 
Ben-Rafael et al., 1991 , J Reprod Med Retrospective Eight-year period (A) 27/89; (B) 40/41 (A) Infertile patients; and (B) women who have experienced recurrent spontaneous abortions. Women with arcuate or subseptate anomalies. (A) Infertile women and

(B) recurrent aborters

with Müllerian duct anomalies.
(A) Infertile women and women who have experience recurrent spontaneous abortions (B) with normal uteri. Bicornuate uterus, unicornuate uterus and uterus didelphys. HSG. Not reported.
Acien, 1993 , Hum Reprod Retrospective 1980–1991 176/28 Women with uterine anomalies. Patients with Rokitansky syndrome, hypoplastic uterus and incomplete case studies were excluded. Uterine or uterovaginal malformations. Genital, urinary anomalies, or both, but normal uterus. Arcuate, septate, bicornuate, didelphis, unicornuate. Clinical examination, ultrasound, HSG, pyelography. According to the criteria of (Jarcho, 1946) and (Buttram, Gibbons, 1979) and the AFS (1988 ).
Lavergne et al., 1996 , Eur J Obstet Gyn Reprod Biol Retrospective 1987–1992 83/5880 embryo transfers Infertile women with uterine anomalies who underwent an IVF cycle. All cases insufficiently documented. Infertile women with uterine anomalies who underwent an IVF cycle. (A) All women without CUA anomolies who underwent one IVF attempt or (B) the data of Fécondation In Vitro National (FIVNAT). Unicornuate, pseudounicornuate, bicornuate, septated (only untreated cases analysed in this meta-analysis). By at least two of the following: hysteroscopy and/or HSG and/or laparoscopy. Not reported.
Zupi et al., 1996 , Gynecol Endosc Retrospective 1989–1993 64/763 Women that underwent hysteroscopy for reasons other than infertility. Women who had undergone voluntary abortion, or women diagnosed during hysteroscopy with submucous fibroids or synechiae. Women diagnosed with septate, bicornuate and arcuate uterus. Women who underwent hysteroscopy and were diagnosed with a normal uterus. Arcuate, septate, bicornuate and unicornuate. Hysteroscopy. Septate or bicornuate uterus as defined as a double cavity, separated by a mid-cavity septum that covered at least one-third of the uterine cavity.

Arcuate uterus was defined as a uterus with a fundal notch consisting of less than one-third of the cavity.
Heinonen, 1997 , J Am Assoc Gynecol Laparosc Retrospective 1962–1995 19/19 Women with septate and subseptate uterus. Women with septate and subseptate uterus that underwent hysteroscopic metroplasty. Women with septate and subseptate with no treatment matched by age, gravidity and type of uterine anomaly. Septate and subseptate uterus. Not reported. According to Buttram and Gibbons (1979) .
Cooney et al., 1997 , J Clin Ultrasound Retrospective Not reported 22/66 Pregnant women with fetal heart activity in the first trimester undergoing ultrasound. Sonographicaly detected uterine duplication. Women with normally appearing uteri matched for gestational age. Septate uterus, bicornuate, didelphys, partial septate. Ultrasound but confirmed with surgery (n = 11), hysteroscopy (n = 1), second look ultrasound (n = 6). Not reported.
Ravasia et al.,1999 , Am J Obstet Gynecol Retrospective cohort 1992–1997 25/1788 Women undergoing trial of labour after previous caesarean section. Not reported. Women with known Müllerian anomalies opting to undergo trial of labour. Women without known Müllerian anomalies opting to undergo trial of labour. Septate, unicornuate, bicornuate, didelphys. Patient records:

laparoscopy or hysteroscopy after abnormal HSG or ultrasound (n = 21) or during the previous caesarean section.
Not reported.
Woelfer et al., 2001 , Obstet Gynecol Prospective Aug 1997 to Sep 2000 106/983 General population screened for other gynaecological conditions. Ongoing pregnancy, history of infertility, recurrent spontaneous abortion or fibroids distorting the uterine cavity. Women with uterine anomalies. Low risk for congenital malformations. Arcuate, partial septate and bicornuate. Three-dimensional ultrasound. According to the AFS (1988 ) criteria.
Shuiqing et al., 2002 , Chin Med Sci J Retrospective 1984–1998 153/27 Not reported. Not reported. Patients with uterine anomalies. Women with other renal or genital anomalies but with normal uterus (treated at the same hospital during the same period). Unicornuate uterus, bicornuate uterus, septate uterus and didelphys uterus. History, clinical examination, ultrasound, HSG, hysteroscopy According to Buttram (1983) .
Valli et al., 2004 , J Am Assoc Gynecol Laparosc Prospective 1990–2001 33/15 Women with septate uterus and who have experienced recurrent abortions. Women with bicornuate uterus. Women who underwent hysteroscopic metroplasty. Untreated women. Septate uterus. Hysteroscopy. According to the AFS (1988 ) criteria

(Fundal protrusion <15 mm was considered an arcuate uterus).
Erez et al., 2007 , Am J Obstet Gynecol Retrospective Not reported 165/5406 Women delivering after primary caesarean section. Patients with multiple pregnancies, more than one previous caesarean section and known congenital, chromosomal fetal. anomalies, or both. Women with Müllerian duct anomalies delivering after primary caesarean section. Women without Müllerian duct anomalies delivering after primary caesarean section. Arcuate uterus, bicornuate, didelphys, unicornuate, septate. Surgery (during previous caesarean section). According to the AFS (1988 ) criteria.
Zlopasa et al., 2007 , Int J Gynaecol Obstet Retrospective 1997–2000 130 /182 Pregnant women. Twin gestations, chorioamnionitis, presence of submucosal myomas, fetal chromosomopathy, maternal diabetes or IVF. Women with Müllerian duct anomalies. Women with normal uterus. Arcuate uterus, bicornuate, didelphys, unicornuate, subseptate, septate. Surgery, sonohysterography, laparoscopy, hysteroscopy. According to AFS (1988 ) criteria.
Ban-Frangez et al., 2009 , Eur J Obstet Gynecol Retrospective 1993–2004 106/212 Women with uterine septum that underwent IVF or ICSI and had a singleton pregnancy with ultrasound demonstration of fetal heart activity. Extrauterine pregnancies, multiple pregnancies and cases with an empty gestational sac. Women who had conceived afterIVF or ICSI. Women without uterine anomalies with a singleton pregnancy after IVF or ICSI matched for age, body mass index, stimulation protocol, use of IVF or ICSI and infertility indications. Septate uterus. Two-dimensional ultrasound (which confirmed the normal external uterine shape in all cases); hysteroscopy. According to AFS (1988 ) criteria.
Lin et al., 2009 , Int J Gynaecol Obstet Retrospective 1998–2007 21/15 Women with septum that expanded longitudinally from the fundal part to the vagina. Abnormalities other than utero-cervico-vaginal septum. Vaginal septum removal plus hysteroscopic resection of the uterus septum. Women that remained untreated. Complete septate uterus. Three-dimensional ultrasound and HSG. NA
Tomazevic et al., 2010 , Reprod Biomed Online Retrospective 1993–2005 289 embryo transfers before hysteroscopic resection /538 embryo transfers after hysteroscopic resection/ 1654 embryo transfers as controls Infertile patients undergoing an embryo transfer after IVF or ICSI. Not reported. Infertile patients who underwent IVF–ICSI with uterine septate divided into two groups: hysteroscopically corrected and not corrected. Infertile patients with normal uterus that underwent IVF–ICSI matched for age, body mass index, stimulation protocol, quality of embryos, use of IVF or ICSI and infertility indication. Septate uterus. Two-dimensional ultrasound and confirmed subsequently during hysteroscopic procedure. According to AFS (1988 ) criteria

(however, a septum of 1.3–1.5 cm is considered as small and defines an arcuate uterus according to the authors).
Liang and Hu, 2010 , Int J Gynaecol Obstet Retrospective 1999–2008 876/68979 Pregnant women. Medical complications or history of caesarean delivery. Pregnant women with Müllerian duct anomalies. Pregnant women with normal uterus matched for age and parity. Uterus didelphys, bicornuate uterus, incomplete septate uterus, complete septate uterus, unicornuate uterus. Physical examination, ultrasound, HSG, hysteroscopy, laparoscopy, laparotomy before pregnancy. According to Buttram and Gibbons (1979) .
Saravelos et al., 2010 , Reprod Biomed Online Retrospective Not reported 56/107 Women who have experienced recurrent spontaneous abortion. Pregnancies in which patients received medical treatment (e.g. low molecular weight heparin, steroids, acetyl-salicylic acid) or surgery (e.g. septotomy, Strassman's metroplasty, cervical cerclage). Women with a specific CUA and no identifiable cause of recurrent abortions. Women with normal uterus and investigations for recurrent spontaneous abortions. Septate, arcuate, bicornuate, didelphys, unicornuate. Screening through two-dimensional ultrasound and HSG, and confirmation through a combined hysteroscopy/ laparoscopy. According to AFS (1988 ) criteria.
Sugiura-Ogasawara et al., 2010 , Fertil Steril Retrospective
 
1986–2007 42/1528 Patients who have had two or more consecutive spontaneous abortions. Patients with structural chromosome abnormalities. CUA and no chromosomal abnormalities. Normal uterus and no chromosomal abnormalities. Septate, unicornuate, bicornuate, didelphis. Laparoscopy, laparotomy, magnetic resonance imaging, or both. According to AFS (1988 ) criteria

(Tompkin's index was used to distinguish between arcuate uterus and mild septate or bicornuate uterus).
Zhang et al., 2010 , Chin Med J Retrospective 1998–2009 116/270 Pregnant women. Patients with multiple pregnancies and known congenital, chromosomal fetal anomalies, or both, were excluded. Pregnant women with uterine malformations. Pregnant women with normal uteri. Septate uterus, didelphys, bicornuate, arcuate, unicornuate. Surgery or sonohysterography, laparoscopy with hysterography, hysteroscopy. According to AFS (1988 ) criteria.
Hua et al., 2011 , Am J Obstet Gynecol Retrospective 1990 –2008 203/66753 Singleton pregnancies in general population. Not reported. Women with CUA and singleton pregnancies. General population with singleton pregnancies. Septate uterus, unicornuate, bicornuate uterus, didelphys. Not reported. Not reported.
Jayaprakasan et al., 2011 , Ultrasound Obstet Gynecol Prospective 2005–2009 184/1201 Infertile women undergoing assisted reproductive techniques. Patients found to have one or more uterine fibroids or polyps distorting the endometrial cavity or if the ultrasound view was unclear and not good enough to allow a definitive diagnosis to be made. Women with CUA undergoing assisted reproductive techniques. Women undergoing assisted reproductive techniques with normal uterus. Septate, unicornuate, bicornuate, didelphys. Three-dimensional ultrasound. According to AFS (1988 ) criteria as modified by Salim et al., 2003 .
Munoz Munoz et al., 2011 , Hum Reprod (abstract) Retrospective 2000–2010 64/70 Women undergoing assisted reproductive techniques with donated oocytes. Recipients with associated severe male factor. Women with Müllerian duct anomalies. Women with premature ovarian failure after oocyte donation. Bicornuate, septate, unicornuate, didelphis. Not reported. Not reported.
Pang et al., 2011 , Int J Gynecol Obstet Prospective Jan 2006 to March 2011 Recurrent spontaneous abortions: 46 cases /recurrent spontaneous abortions:32 controls

and

no history of poor reproductive outcome: 30 cases/

no history of poor reproductive outcomes: 30 controls
Women with subseptate uterus and who have experienced recurrent spontaneous abortions (two first- trimester abortions) (Group A) or no history of poor reproductive outcome (Group B). Women who had experienced only one spontaneous abortion. Women with subseptate uterus and who have experienced recurrent spontaneous abortions (Group A) or have no history of poor reproductive outcome (Group B) treated with hysteroscopic metroplasty. Women with subseptate uterus and who have experienced recurrent spontaneous abortion (Group A) or have no history of poor reproductive outcome (Group B) not treated with hysteroscopic metroplasty. Subseptated uterus. Three-dimensional ultrasound. According to AFS (1988 ) criteria.
Tonguc et al., 2011 , Int J Gynecol Obstet Retrospective Jan 2006 to Jan 2009 102/25 Patients with a uterine septum and otherwise unexplained infertility. Patients who had a history of tuberculosis or endometriosis, an endocrinologic problem, a history of abdominal surgery, a husband with mild or severe oligospermia, as revealed by the spermiogram. Patients who underwent hysteroscopic metroplasty. Patients who rejected the operation. Septate uterus. Not reported. According to AFS (1988 ) criteria.

AFS, American Fertility Society; CUA, congenital uterine analmolies; HSG, hysterosalpingogram; ICSI, intracytoplasmic sperm injection; NA, not applicable.

The eligible studies were published between 1987 and 2011, and only four studies were prospective in design, whereas the remaining 21 were retrospective studies. Out of the 21 retrospective studies, in only four it was reported that a matching procedure had been used to minimize imbalances between cases and controls (Ban-Frangez et al, 2009, Cooney et al, 1997, Heinonen, 1997, and Tomazevic et al, 2010). The population investigated was variable depending on the research question evaluated in each study ( Table 1 ). Therefore, certain studies evaluated the clinical implications of CUA in specific subpopulations, such as infertile patients (n= 7) or recurrent aborters (n= 3), whereas others did not (n= 11). Similarly, the potential beneficial effect of hysteroscopic septum resection was also evaluated in infertile patients (n= 2), in women who have experienced recurrent abortions (n= 2) and in unselected population (n= 2). Some studies included two subpopulations of patients, such as the study by Ben-Rafael et al. (1991) , which examined the effect on CUA in infertile patients and patients who experienced recurrent spontaneous abortions, and the study by Pang et al. (2011) , which investigated the value of hysteroscopic removal of septum in women who had experienced recurrent spontaneous abortion and women with no history of poor reproductive outcome.

Certain studies included women with a specific subtype of CUA, whereas others included a wider spectrum of CUA ( Table 1 ). The method of diagnosis of the CUA also varied between studies. Hysterosalpingography alone was used in two studies, three dimensional ultrasound alone was used in three studies, hysteroscopy alone in two studies and previous surgical records, also, in one study. In the remaining studies multiple methods to screen for, and confirm, the presence of CUA were used, whereas this information was not provided in four studies ( Table 1 ). Last but not least, the classification system or the method used in each study for the definition of the various types of CUA was also variable, although more than one-half of the eligible studies used the classification of the AFS ( American Fertility Society, 1988 ).

Most studies were of acceptable quality, although five studies (Acien, 1993, Ben-Rafael et al, 1991, Munoz Munoz et al, 2011, Shuiqing et al, 2002, and Sorensen, Trauelsen, 1987) were considered to be of low quality, based on the Newcastle–Ottawa Scale, which was specifically adapted for the current systematic review and meta-analysis ( Supplementary Table S1 ).

Meta-analysis

Achievement of pregnancy
Natural cycles

In women with arcuate (RR 0.93, 95% CI 0.70 to 1.24; two studies), didelphys (RR 0.94, 95% CI 0.83 to 1.07; two studies), unicornuate (RR 0.85, 95% CI 0.68 to 1.07; two studies) and bicornuate (RR 0.95, 95% CI 0.89 to 1.02; two studies) uterus the probability of pregnancy in natural cycles was not significantly different from the controls. In women with septated uterus, however, the probability of pregnancy was significantly decreased compared with normal controls (RR 0.86, 95% CI 0.77 to 0.96; three studies). The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When all the CUA were combined, then no significant differences in the probability of spontaneous pregnancy in women with or without CUA were detected (RR 0.96, 95% 0.89 to 1.04; heterogeneity:P= 0.02, I2= 69%) ( Figure 2a ). The exclusion of the studies of low quality did not materially change this result (RR 1.02, 95% CI 0.97 to 1.07).

rbmo1230-fig-0002

Figure 2 Forest plots for relative risk of pregnancy achievement between women with and without congenital uterine anomalies (CUA) in: (a) natural cycles (spontaneous conception); (b) assisted reproductive techniques cycles and (c) natural cycles and assisted reproductive techniques cycles combined.

Assisted reproductive techniques cycles

In women undergoing assisted reproductive techniques cycles, no statistically significant difference was observed in the probability of pregnancy achievement per embryo transfer between women with or without arcuate (RR 1.26, 95% CI 0.98 to 1.61; 1 study), septated (RR 0.89, 95% CI 0.24 to 3.29; two studies), unicornuate (RR 1.15, 95% CI 0.29 to 4.63; 1 study) and bicornuate uterus (RR 1.73, 95% CI 0.77 to 3.8; one study). The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

Moreover, when all the data were combined, regardless of type of CUA, no significant difference in the probability of pregnancy after assisted reproductive techniques was detected between women with or without CUA (RR 0.66, 95% 0.37 to 1.19; heterogeneity:P< 0.00001, I2= 92%; four studies) ( Figure 2b ). The exclusion of a study ( Munoz Munoz et al., 2011 ) that was graded as being of low quality did not materially change the results obtained (RR 0.69, 95% CI 0.30 to 1.57).

Overall

When all the studies analysing natural and assisted reproduction techniques cycles cycles were pooled, then the probability of pregnancy achievement (either spontaneously or after assisted reproduction techniques) in patients with CUA was significantly (P= 0.05) decreased compared with normal controls (RR 0.85, 95% CI 0.73 to 1.00; heterogeneity:P< 0.00001, I2= 91%) ( Figure 2c ). The exclusion of three low-quality studies (Acien, 1993, Munoz Munoz et al, 2011, and Shuiqing et al, 2002) yielded a similar effect size, which was not, however, statistically significant (RR 0.87, 95% CI 0.68 to 1.11). Furthermore, when comparing the pooled effect sizes of the two subgroups (natural cycles [non-infertile population] versus assisted reproduction techniques cycles [infertile population]), no statistically significant difference was detected.

Sponanteous abortion

First-trimester spontaneous abortion

The probability of spontaneous abortion during the first trimester was not significantly different between women with arcuate (RR 1.22, 95% CI 0.87 to 1.72; heterogeneity: I2= 50%; six studies), didelphys (RR 1.13, 95% CI 0.45 to 2.86; heterogeneity:P= 0.02, I2= 68%; four studies) and unicornuate uterus (RR 1.38, 95% CI 0.83 to 2.28; heterogeneity: I2= 0%; five studies) and the normal controls. Women, however, with septated (RR 2.65, 95% CI 1.39 to 5.06; heterogeneity:P< 0.00001, I2= 93%; six studies) and bicornuate uterus (RR 2.32, 95% CI 1.05 to 5.13; heterogeneity:P< 0.00001, I2= 87%) had a significantly increased probability of first-trimester spontaneous abortion compared with their controls. The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When all available data on the association of CUA with first-trimester spontaneous abortion rates were used (irrespective of the type of CUA), a statistically significant increase in the probability of first-trimester spontaneous abortion was present in women with CUA (RR 1.56, 95% CI 1.17 to 2.08; heterogeneityP< 0.00001, I2= 82%; nine studies) ( Figure 3a ). A sensitivity analysis, in which low-quality studies were excluded, yielded similar results (RR 1.45, 95% CI 1.07 to 1.94).

rbmo1230-fig-0003

Figure 3 Forest plots for relative risk of (a) first-trimester spontaneous abortion; (b) second-trimester spontaneous abortion; and (c) first-trimester, second-trimester spontaneous abortion, or both, when all the available data for congenital uterine anomalies are pooled. CUA, congenital uterine anomalies.

A subgroup analysis according to the type of population is presented in Table 2 . According to this analysis, the effect sizes were significantly different (P= 0.006) in the various subgroups with the larger ones deriving from studies carried out in non-specific populations, and the most modest one in the study carried out among women who had experienced recurrent spontaneous abortions.

Table 2 Subgroup analysis according to the type of population analysed in each study.

  Non-specific population Infertile patients Women who have experienced recurrent spontaneous abortions Difference between subgroups

P
  Studies RR (95% CI) Studies RR (95% CI) Studies RR (95% CI)
Achievement of pregnancy 4 0.66 (0.37 to 1.19) 4 0.96 (0.89 to 1.04) 0 N/A NS
Obstetric outcome              
 First-trimester spontaneous abortion 5 1.88 (1.33 to 2.66) 3 1.24 (0.82 to 1.88) 1 1.07 (0.97 to 1.17) 0.006
 Second-trimester spontaneous abortion 4 2.21 (1.40 to 3.48) 1 2.78 (0.83 to 9.32) 1 2.39 (1.32 to 4.33) NS
 First and second-trimester spontaneous abortion combined 6 2.28 (1.57 to 3.32) 5 1.46 (0.83 to 2.57) 3 1.13 (1.06 to 1.22) 0.001
 Preterm delivery <37 weeks 8 2.25 (1.51 to 3.34) 1 2.49 (1.32 to 4.69) 1 1.73 (1.03 to 2.90) NS
 Preterm delivery <34 weeks 2 4.90 (3.52 to 6.84) 1 1.19 (0.32 to 4.48) 0 - 0.04
 Preterm delivery <28 weeks 2 3.15 (0.94 to 10.54) 1 2.53 (0.58 to 11.02) 1 2.63 (1.04 to 6.62) NS
 Malpresentation at delivery 7 4.83 (3.20 to 7.29) 2 6.19 (0.91 to 42.05) 1 2.98 (1.20 to 7.38) NS
 Birth weight <2500 g 2 1.93 (1.50 to 2.49) 0 N/A 0 N/A N/A
 Birth weight <1500 g 2 2.07 (1.10 to 3.88) 0 N/A 0 N/A N/A
 Intrauterine growth restriction 2 3.17 (0.89 to 11.24) 0 N/A 0 N/A N/A
 Placental abruption 4 2.66 (1.36 to 5.23) 1 0.55 (0.03 to 11.34) 0 N/A NS
 Premature rupture of membranes 4 1.47 (0.69 to 3.15) 1 2.89 (1.01 to 8.31) 1 4.20 (0.47 to 37.24) NS
 Perinatal mortality 4 2.50 (1.33 to 4.69) 1 1.39 (0.13 to 15.09) 0 N/A NS
The effect of hysteroscopic resection of the septum              
 Achievement of pregnancy 1 0.96 (0.76 to 1.22) 1 2.16 (0.95 to 4.87) 2 1.09 (0.60 to 1.98) NS
 Spontaneous abortion 3 0.51 (0.27 to 1.00) 1 0.19 (0.06 to 0.56) 2 0.35 (0.20 to 0.61) NS
 Preterm labour <37 weeks 3 0.92 (0.27 to 3.14) 1 0.91 (0.14 to 5.85) 2 0.36 (0.09 to 1.42) NS

N/A, not applicable; NS, not statistically significant.

Second-trimester spontaneous abortion

The probability of spontaneous abortion during the second trimester was not significantly different between women with didelphys (RR 1.71, 95% CI 0.63 to 4.59; heterogeneity I2= 0%; four studies) and unicornuate uterus (RR 2.27, 95% CI 0.64 to 7.96; heterogeneity I2= 0%; four studies) and their controls. On the other hand, women with arcuate (RR 1.98, 95% CI 1.06 to 3.69; heterogeneity: I2= 10%; five studies), septated (RR 2.95, 95% CI 1.51 to 5.77; heterogeneity I2= 39%; five studies) and bicornuate uterus (RR 2.90, 95% CI 1.56 to 5.41; heterogeneity I2= 0%; four studies) had a statistically significant increased probability for second-trimester spontaneous abortion compared with their controls. The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When all available data on the association of CUA with second-trimester spontaneous abortion rates were used (irrespective of the type of CUA), a statistically significant increase in the probability of second-trimester spontaneous abortion was present in women with CUA (RR 2.31, 95% CI 1.63 to 3.27; heterogeneity I2= 0%; six studies) ( Figure 3b ). The exclusion of low-quality studies did not substantially alter these results (RR 2.26, 95% CI 1.55 to 3.28).

A subgroup analysis according to the type of population is presented in Table 2 ,and suggests that the effect sizes were not significantly different.

Combined spontaneous abortion risk

When all the data for spontaeous abortions were combined (first trimester, second trimester, or both), it was shown that the risk for spontaneous abortion was not significantly different between women with arcuate (RR 2.03, 95% CI 0.96 to 4.29; heterogeneityP= 0.0003, I2= 81%; five studies) and didelphys uterus (RR 1.17, 95% CI 0.38 to 3.66; heterogeneityP= 0.04, I2= 68%; three studies). On the other hand, women with septated (RR 2.81, 95% CI 2.13 to 3.71; heterogeneity:P= 0.33, I2= 14%; six studies), bicornuate (RR 2.40, 95% CI 1.42 to 4.08; heterogeneityP= 0.02, I2= 67%; five studies) and unicornuate uterus (RR 2.10, 95% CI 1.12 to 3.94; heterogeneity:P= 0.82, I2= 0%; four studies) were shown to have an increased risk of spontaneous abortion compared with their controls. The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When the available data on all the CUA were pooled, then the probability of spontaneous abortion in women with CUA was confirmed to be higher compared with the controls (RR 1.68, 95% CI 1.31 to 2.15; heterogeneityP< 0.00001, I2= 89%; 14 studies) ( Figure 3c ). This was also the case when low-quality studies were excluded (RR 1.81, 95% CI 1.29 to 2.52).

A subgroup analysis according to the type of population is presented in Table 2 . According to this analysis, the effect sizes were significantly different (P= 0.001) in the various subgroups, with the larger ones deriving from studies performed in non-specific populations and the most modest one in the study performed in recurrent aborters.

Obstetric outcome

Preterm deliveries less than 37 weeks

All types of CUA individually and combined seemed to be associated with statistically increased risk of a delivery before the completion of the 37th gestational week compared with the controls ( Table 3 ). In the case of women with arcuate uterus, the difference was marginally not significant (RR 2.04, 95% CI 0.99 to 4.19). The difference in the pooled effect sizes between the various types of CUA was not statistically significant.

Table 3 Pooled relative risks (95% confidence interval) for specific obstetric outcomes in women with congenital uterine anomolies compared with their controls.

  Arcuate Septated Didelphys Bicornuate Unicornuate Combined/undefined
Preterm delivery <37 weeks 2.04 (0.99 to 4.19) 2.11 (1.51 to 2.94) 3.39 (2.06 to 5.58) 2.16 (1.55 to 3.02) 3.14 (1.90 to 5.18) 2.21 (1.59 to 3.08)
Premature delivery <34 weeks 1.19 (0.32 to 4.48) N/A N/A N/A N/A 3.81 (1.48 to 9.83)
Premature delivery <28 weeks 2.72 (0.53 to 14.06) 2.40 (0.46 to 12.42) 3.45 (0.40 to 29.83) 5.37 (1.53 to 18.80) 6.30 (0.29 to 137.42) 2.84 (1.52 to 5.31)
Malpresentation at delivery 2.53 (1.45 to 4.43) 4.35 (2.52 to 7.50) 2.62 (2.25 to 3.06) 4.65 (3.43 to 6.32) 3.12 (2.39 to 4.08) 4.75 (3.29 to 6.84)
Birth weight <2500 g 1.42 (0.74 to 2.72) 1.60 (0.94 to 2.71) 2.40 (1.40 to 4.11) 1.74 (1.13 to 2.69) 3.54 (2.22 to 5.64) 1.93 (1.50 to 2.49)
Birth weight <1500 g 1.95 (0.47 to 8.00) 1.97 (0.58 to 6.64) 1.15 (0.16 to 8.28) 2.78 (1.16 to 6.69) 3.18 (0.79 to 12.82) 2.07 (1.10 to 3.88)
Intrauterine growth restriction 5.03 (0.33 to 76.19) 2.54 (1.04 to 6.23) 4.94 (2.20 to 11.09) 2.80 (1.06 to 7.34) 3.54 (0.21 to 61.01) 3.17 (0.89 to 11.24)
Placental abruption 6.60 (2.35 to 18.53) 4.37 (1.12 to 17.08) 2.04 (0.28 to 15.10) 1.48 (0.29 to 7.55) 4.31 (0.77 to 23.97) 2.47 (1.28 to 4.77)
Premature rupture of membranes 1.67 (0.84 to 3.34) 0.61 (0.28 to 1.37) 1.30 (0.56 to 3.02) 1.12 (0.51 to 2.43) 0.54 (0.20 to 1.50) 1.74 (0.92 to 3.29)
Perinatal mortality 2.66 (0.89 to 7.93) 2.43 (1.10 to 5.36) 1.97 (0.69 to 5.66) 3.32 (1.61 to 6.86) 2.20 (0.76 to 6.39) 2.43 (1.34 to 4.42)

By pooling the data from all CUA, it was shown that the probability of preterm labour (<37 weeks) was significantly higher in women with CUA compared with women without CUA (RR 2.21, 95% CI 1.59 to 3.08). Excluding studies of low quality did not materially alter these results (RR 2.12, 95% CI 1.37 to 3.29). A subgroup analysis according to the type of population, indicating that the effect sizes were not significantly different, is presented in Table 2 .

Premature delivery less than 34 weeks

Data for preterm deliveries below the 34th gestational week were limited ( Table 3 ). Women with an arcuate uterus presented no statistically significant increase in the probability of premature delivery at less than 34 gestational weeks. When the data that were available for all types of CUA were combined (including data from studies that did not provide an analysis per type of CUA), a significantly higher risk for premature delivery at less than 34 gestational weeks was present (RR 3.81, 95% CI 1.48 to 9.83) in women with CUA compared with the controls. The exclusion of one low-quality study ( Sorensen and Trauelsen, 1987 ) produced an even more pronounced effect (RR 4.90, 95% CI 3.52 to 6.84). A subgroup analysis according to the type of population is presented in Table 2 .

Premature delivery less than 28 weeks

An increased risk for premature delivery at less than 28 weeks of gestation was suggested for all types of CUA ( Table 3 ). This effect, however, was significant only for women with bicornuate uterus (RR 5.37, 95% CI 1.53 to 18.80). The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

Furthermore, when all available data were combined, women with CUA were shown to be at an increased risk for premature delivery at less than 28 gestational weeks compared with the control group (RR 2.84, 95% CI 1.52 to 5.31). A sensitivity analysis conducted with the exclusion of two low-quality studies, resulted in an even larger effect size, with wider confidence intervals (RR 4.84, 95% CI 1.44 to 16.21).

Malpresentation at delivery

All types of CUA individually and combined seemed to be associated with statistically increased risk of malpresentation (e.g. breech, transverse, or both) at delivery compared with women without CUA, with the largest effect sizes being present in women with septated (RR 4.35, 95% CI 2.52 to 7.50) and bicornuate uterus (RR 4.65, 95% CI 3.43 to 6.32) ( Table 3 ). The difference in the observed pooled effect sizes between the various types of CUA was statistically significant (P= 0.01). Further analysis suggested that the effect sizes seemed to be similar between the arcuate, unicornuate and didelphys groups and also between septated and bicornuate uterus.

Not unexpectedly, when all available data were combined, the association of the presence of CUA with an increased risk of malpresentation at delivery was confirmed (RR 4.75, 95% CI 3.29 to 6.84) ( Table 3 ). The exclusion of four low-quality studies resulted in similar results (RR 5.60, 95% CI 3.55 to 8.84). A subgroup analysis according to the type of population, indicated that the effect sizes were not significantly different ( Table 2 ).

Birth weight less than 2500 g

Women with an arcuate or septated uterus did not seem to have a statistically increased risk for a neonatal birth weight of less than 2500 g ( Table 3 ). In women with didelphys, however (RR 2.40, 95% CI 1.40 to 4.11), bicornuate (RR 1.74, 95% CI 1.13 to 2.69) and unicornuate uterus (RR 3.54, 95% CI 2.22 to 5.64), a significantly higher probability of giving birth to neonates with birth weight less than 2500 g was present compared with the controls. The difference in the observed pooled effect sizes between the various types of CUA was marginally not statistically significant. Further secondary analysis suggested similar effect sizes in women with didelphys uterus and unicornuate uterus and in women with arcuate, septated and bicornuate uterus.

When all the CUA were combined, then a statistically increased risk for low (<2500 g) birth weight neonates was shown for women with, compared with those without CUA (RR 1.93, 95% CI 1.50 to 2.49). A subgroup analysis per type of population analysed was not feasible, as all the studies were carried out at the same type ( Table 2 ).

Birth weight less than 1500 g

In women giving birth, no statistically significant association was detected between the presence of an arcuate, septated, didelphys or unicornuate uterus and the probability of very low neonatal birth weight (<1500 g) ( Table 3 ). Women with a bicornuate uterus were shown to be at an increased risk for very low neonatal birth weight compared with their controls (RR 2.78, 95% CI 1.16 to 6.69). The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When all the available data were combined, it was suggested that women with CUA are at an increased risk of giving birth to a baby with very low birth weight (RR: 2.07, 95% CI 1.10 to 3.88). A subgroup analysis per type of population analysed was not feasible, as all the studies were performed in the same type ( Table 2 ).

Intrauterine growth restriction

An association between the presence of CUA and the probability of intrauterine growth restriction was suggested in all types of CUA ( Table 3 ), yet this difference was statistically significant only in women with septated (RR 2.54, 95% CI 1.04 to 6.23), bicornuate (RR 2.80, 95% CI 1.06 to 7.34) and didelphys uterus (RR 4.94, 95% CI 2.20 to 11.09). The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When all the CUA were combined, then the observed difference regarding the probability of IUGR was not statistically significant (RR 3.17, 95% CI 0.89 to 11.24). A sensitivity analysis in which the study by Acien (1993) was excluded, led to the analysis of relevant data from a single study ( Zlopasa et al., 2007 ) which suggested a statistically significant increase in the probability of IUGR in women with CUA as compared with women without CUA (RR 5.63, 95% CI 2.65 to 11.99). A subgroup analysis per type of population analysed was not feasible, as all the studies were conducted in the same type ( Table 2 ).

Placental abruption

Women with an arcuate (RR 6.60, 95% CI 2.35 to 18.53) and septated (RR 4.37, 95% CI 1.12 to 17.08) uterus presented a statistically significant difference in the probability of placental abruption compared with the controls. On the other hand women with didelphys, bicornuate and unicornuate uterus had a non-significantly increased risk of placenta abruption during pregnancy ( Table 3 ). The difference in the observed pooled effect sizes between the various types of CUA was not statistically significant.

When all the CUA were combined, then significant differences in the probability of placental abruption in women with CUA were detected (RR 2.47, 95% CI 1.28 to 4.77). The exclusion of a study of low quality did not materially change the results obtained (RR 2.66, 95% CI 1.36 to 5.23). A subgroup analysis according to the type of population is presented in Table 2 .

Premature rupture of membranes

Premature rupture of membranes was not significantly associated with the presence of a certain type of CUA ( Table 3 ) and no significant difference was observed between the pooled effect sizes of the various types of CUA. Similarly, when all the available data on this outcome measure were combined, no association was detected with the presence of CUAs. A sensitivity analysis conducted with the exclusion of low-quality studies yielded similar results (RR 1.89, 95% CI 0.79 to 4.49).

Perinatal mortality

Women with a septated uterus (RR 2.43, 95% CI 1.10 to 5.36) or a bicornuate uterus (RR 3.32, 95% CI 1.61 to 6.86) were shown to have a significantly increased risk for perinatal mortality compared with the controls ( Table 3 ). In the groups of arcuate, didelphys and unicornuate uterus, the relative risk for perinatal mortality was increased; however, this increase was not statistically significant. No significant difference was observed in the pooled effect sizes between the various types of CUA.

When all the CUA were combined, the overall risk for perinatal mortality in women with CUA was higher than in women without CUA (RR 2.43, 95% CI 1.34 to 4.42). When only the two studies of acceptable quality were pooled, however (Zhang et al, 2010 and Zlopasa et al, 2007), the observed difference was no longer statistically significant (RR 1.98, 95% CI 0.44 to 9.02). A subgroup analysis according to the type of population suggesting that the effect sizes were not significantly different is presented in Table 2 .

The effect of hysteroscopic resection of the septum

Achievement of pregnancy

Data on the effect of hysteroscopic treatment of a uterine septum on the probability of pregnancy achievement were limited. Three studies offered relevant data, two of which in women with recurrent abortions (Pang et al, 2011 and Valli et al, 2004) and one in women with unexplained infertility ( Tonguc et al., 2011 ). In the study of Pang et al. (2011) , data were also provided in women with no history of poor reproductive outcome. When these data were combined, no significant association between hysteroscopic treatment of a uterine septum and the probability of pregnancy achievement was detected (RR 1.14, 95% CI 0.79 to 1.65; heterogeneityP= 0.002, I2= 80%; four datasets) ( Figure 4a ). The subgroup analysis according to the type of population did not detect a significant difference among subgroups ( Table 2 ).

rbmo1230-fig-0004

Figure 4 Forest plots for the effect of hysteroscopic resection of a septum on the probability of (a) pregnancy achievement; (b) spontaneous abortion; and (c) preterm labour (<37 weeks).

First-trimester or second-trimester spontaneous abortion

None of the eligible studies provided data on the effect of hysteroscopic treatment of septa on the probability of first- or second-trimester spontaneous abortion.

Spontaneous first-trimester or second-trimester spontaneous abortion or both combined)

Five studies offered data about the association of hysteroscopic septotomy and the probability of spontaneous abortion in general (Heinonen, 1997, Lin et al, 2009, Pang et al, 2011, Tonguc et al, 2011, and Valli et al, 2004). On the basis of the meta-analytic pooling of these studies, women who underwent hysteroscopic septotomy had a significantly decreased probability of spontaneous abortion compared with women who did not undergo treatment (RR 0.37, 95% CI 0.25 to 0.55; heterogeneity I2= 0%; six datasets) ( Figure 4b ). A subgroup analysis according to the type of population, indicating that the effect sizes were not significantly different, is presented in Table 2 .

Preterm labour less than 37 weeks

Five studies (Heinonen, 1997, Lin et al, 2009, Pang et al, 2011, Tonguc et al, 2011, and Valli et al, 2004) offered relevant data, which, when pooled, suggested a decreased, yet not significantly so, probability of preterm labour in women who had been treated with hysteroscopic resection of the septum compared with women who had not been treated (RR 0.66, 95% CI 0.29 to 1.49; heterogeneity I2= 0%; six datasets) ( Figure 4c ). The subgroup analysis per type of population analysed suggested the presence of not significantly different effect sizes across subgroups ( Table 2 ).

Discussion

Presentation of the main findings

This systematic review and meta-analysis highlights the importance of the presence of CUA in the reproductive performance of women. On the basis of the results of this systematic review, the presence of CUA is associated with a decreased, albeit not significantly so, probability of pregnancy achievement both in natural and in assisted reproductive techniques cycles. This finding reached statistical significance only when the two aforementioned groups of studies were combined.

It seems that the presence of CUA has important negative consequences on the ability of a woman to maintain her pregnancy. Especially in women with a septated or bicornuate uterus, a higher probability of spontaneous abortion is present even in the first trimester, whereas in women with arcuate, didelphys and unicornuate uterus the effect size is small and not statistically significant. During the second trimester of pregnancy, when the space requirements of the developing fetus are increased, it seems that the detrimental effect of the presence of CUA on maintaining the pregnancy is also more pronounced in all types of CUA. Similarly, CUAs that cause a greater deformity of the uterine cavity (e.g. septated and bicornuate uterus) are associated with a higher probability of spontaneous abortion during the second trimester.

Moreover, a clear association of the presence of CUA and adverse obstetric outcome is also suggested by this meta-analysis. Women with CUA seem to be in at increased risk of preterm or premature delivery and giving birth to a baby with low or very low birth weight. At the same time, the risk of malpresentation at delivery, which might seriously complicate labour, is also increased uniformly across all types of CUA. Most importantly, though, perinatal mortality seems to be almost 2.5-fold higher in women with CUA compared with women without CUA. This means that further to the increased probability of subfertility and spontaneous abortion, these women have also a higher risk of not having a take-home baby, despite giving birth.

A plausible biological mechanism linking the presence of CUA with an impaired reproductive potential (e.g. implantation failure, spontaneous abortion and impaired obstetrical outcome) of the women has not been yet established. Several hypotheses have been suggested to explain these findings. It has been supported that once the epithelial barrier has been overcome it may be that the uterine vasculature and stroma carry out a subsequent barrier or ‘interrogative’ functions towards the implanting conceptus ( Crocker et al., 2005 ). Thus, infertility and pregnancy losses in patients with uterine anomalies may be associated with abnormalities in the later vascular stages of implantation. Different vascular beds differ in receptivity to the invading trophoblast; uterine septum, uterine defective walls, or both, represent locations with alterations of endometrial vascularization indicating an impaired vascular bed (Dabirashrafi et al, 1995 and Fedele et al, 1996). Furthermore, differences in the structure of the uterine musculature could alter normal uterine contractility in some categories of CUA (e.g. septate and bicornuate uteri), whereas the decrease of endometrial cavity volume could be an important factor in others (e.g. unicornuate, didelphys uteri) ( Mollo et al., 2009 ). Other, more sophisticated hypotheses, have also been suggested: the normal expression ofHOXgenes is important for the proper development of the female genital tract and also for the development of the endometrium, thus playing a major role in fertility ( Taylor, 2000 ). Hence, a disrupted expression ofHOXgenes might be the link between CUA and impaired fertility in these women. It is important to note that the accumulation of further relevant epidemiological data might elucidate the exact role of each type of uterine anomaly on reproductive potential of the woman, and enhance our understanding of the pathophysiology and the exact contribution of each proposed mechanism.

The value of the hysteroscopic removal of a uterine septum has also been examined in this review by analysing only comparative studies. It seems that the hysteroscopic removal of a uterine septum leads to a decreased probability of spontaneous abortion compared with women that are not treated. On the other hand, hysteroscopic removal of the septum does not lead to an increased probability of pregnancy achievement and delivering at term. The effect sizes observed in this analysis (that imply a beneficial effect), and the fact that both achievement of pregnancy and preterm delivery have been shown to be associated with the presence of a uterine septum, however, mandate the accumulation of further evidence to properly assess the value of hysteroscopic septotomy for these indications.

Strengths and weaknesses of the present review

Attempts to critically appraise the literature on this issue have been made in the past. Grimbizis et al. (2001) evaluated the studies that had been published up to that point, and concluded that the presence of CUA is associated with impaired pregnancy outcome. In 2011, a systematic review and meta-analysis examined nine comparative studies (n= 3850 patients in total) and concluded that CUA are associated with poor reproductive outcome ( Chan et al., 2011a ). Our meta-analysis confirms these findings by evaluating the clinical implications of CUA through the assessment of a wider array of clinically relevant outcomes, such as preterm delivery, placental abruption, occurrence of IUGR, occurrence of low or very low birth weight, and perinatal mortality rates. In addition, the current systematic review evaluates the effect of CUA in additional specific subpopulations (i.e. infertile patients and those experiencing recurrent spontaneous abortions). As a result, this meta-analysis includes 25 comparative studies (3766 women with CUA and 160,070 patients in total), and thus provides an updated and more comprehensive evaluation of the available evidence. Moreover, the current meta-analysis assesses the potential benefit of hysteroscopic treatment of uterine septa by analysing only parallel comparative studies.

In this review, an attempt has been made to provide answers to clinically relevant questions on the implications of CUA by using an evidence-based approach. Combining the results of individual studies with the use of meta-analytic methodology increases the statistical power, while at the same time allows for a more robust estimation of the actual effect sizes. The inclusion and exclusion criteria were a-priori defined in such a way so that the best available evidence is analysed. Non-comparative studies and studies comparing before and after states in the same subjects were not analysed in this review, as they are known to lead to exaggerated effects ( Eccles et al., 2003 ).

At the same time, through exploratory subgroup analyses, hypotheses can be made about the presence of associations in specific types of CUA or types of population in larger sample sizes. Although, in most cases, effect sizes between the various types of CUA or types of population analysed in this review were not significantly different ( Table 2 ), it should be stressed that the limited number of studies might have not allowed observed differences in effect sizes to be statistically detected and, thus, further evidence is required before solid conclusions can be drawn about the associations of specific subtypes of CUA with specific clinical outcomes.

Besides the limited number of studies, however, this systematic review also highlights the significant heterogeneity present in the available studies. This might be attributed to the fact that the eligible studies analyse patients treated over a period of almost 50 years (1962–2011). In these studies, the population characteristics were variable, as well as the mode of diagnosis of CUA, which was suboptimal in some of them. Furthermore, variability was present in other characteristics of the eligible studies, such as the criteria used to classify an anomaly, the definition of the outcome measures and the period of follow-up. This heterogeneity should be taken into account when interpreting the results of this meta-analysis. For this reason, available data were analysed with the use of a random-effects model, which uses a conservative approach by acknowledging that the samples evaluated in the individual studies might not all originate from the same population. At the same time, the subgroup analyses conducted in this meta-analysis might give the reader some insight about the potential moderating effect of some of these characteristics on the observed effect sizes.

It should also be emphasized that, although the prevalence of CUA does not seem to be small (Chan et al, 2011b, Grimbizis et al, 2001, and Saravelos et al, 2008a), and their potential implications might be quite significant, even the best available evidence analysed in this review present methodological problems. Most of the available studies are retrospective in design, which is known to be prone to various sources of bias. Specific measures (e.g. matching procedures or multivariate analyses) against bias originating from known confounders (e.g. age, body mass index, smoking and socioeconomic status) were taken in a limited number of studies. The quality assessment conducted for the purpose of this meta-analysis suggested that most of the studies were of average quality, with only three out of 25 studies being graded as of high quality. Whether the methodological problems of the individual studies have biased the results cannot be assessed, although it might be considered reassuring that, in most cases, the exclusion of the low-quality studies did not materially alter the conclusions drawn. It should be made clear, however, that the shortcomings of the individual studies might affect the observed associations, and this should also be considered when interpreting the results of this systematic review and meta-analysis.

Regardless of the limitations of the individual studies, however, this systematic review critically appraises the available literature and summarizes the best available evidence on a topic that has been highly controversial for decades. Hence, the reader can be informed about the direction of the effect and the most likely answer to important clinical questions, and also, about the level of the existing evidence and the limitations of it. Although the absence of high-quality studies must be considered when interpreting the results of this systematic review, it should not serve as a rationale for ignoring the existing evidence, as by doing so treatment will continue to be mostly empirical.

Implications for clinical practice and future research

On the basis of the results of this systematic review, it seems that the presence of CUA has significant implications for the reproductive potential of women. Clinicians should be aware of these, and be alert for the presence of CUA, always supplementing, routine gynaecological examination of women with a two-dimensional ultrasound. When a CUA is suspected through the gynaecological examination or two-dimensional ultrasound, a confirmatory investigation might be warranted, depending on the patient's wish and her medical history. Non-invasive procedures (i.e. sonohysterography and three-dimensional ultrasound) with high accuracy ( Saravelos et al., 2008b ) are now available and can help the clinician in the diagnosis of CUA. If such a diagnosis is made, then the patient should be properly counselled and informed about potential risks and management options.

Most importantly though, it is clear that further research in this field is urgently needed. Apparently, a fundamental part of future research should be the criteria used to diagnose and classify these anomalies, which should be homogeneous. These criteria should reflect accumulated knowledge on the pathogenesis of CUA and they should be universally accepted. The recently published consensus from the European Society of Human Reproduction and Embryology and the European Society of Gynaecological Endoscopy on the classification of female genital tract congenital anomalies might provide clinicians and researchers with such criteria (Grimbizis et al, 2013a and Grimbizis et al, 2013b). Moreover, the diagnosis of CUA should be made with highly accurate methods ( Saravelos et al., 2008b ) so that the internal validity of future studies and the homogeneity of patient population among them is increased.

The evaluation of the association between the presence of CUA and the probability of pregnancy requires a carefully selected population pursuing pregnancy and their proper observation over a sufficiently long follow-up period. Similarly, the investigation of the effect of CUA on spontaneous abortion rates and the obstetric outcome should be conducted through a well-designed, prospective observational study that will take into account potential confounders. Last but not least, high-quality evidence on the value of hysteroscopic removal of a uterine septum can be produced only through a sufficiently powered randomized controlled trial.

Conclusions

This systematic review and meta-analysis highlights the potential clinical importance of congenital uterine anomalies for the reproductive performance of women. It also emphasizes the significant gap in the literature of high-quality evidence, as most of the available studies are retrospective and do not adequately control for potential confounders. Further well-designed prospective studies are warranted to assess these associations with more confidence. Furthermore, it is clear that the value of the hysteroscopic removal of a uterine septum in improving reproductive outcome can be properly assessed only through a large randomized controlled trial.

Currently, the best available evidence suggests that the presence of certain types of CUA might be associated with a detrimental effect on the probability of pregnancy achievement, spontaneous abortion and on the obstetric outcome of each pregnancy. Furthermore, hysteroscopic removal of a septum seems to reduce the probability of a spontaneous abortion. These findings should be taken into account when deciding on the optimal management of patients with CUA.

Appendix. Supplementary material

The following is the supplementary data to this article:

Download file

Supplementary Table S1 Qualitative assessment of eligible studies.

References

  • Acien, 1993 Acien P. Reproductive performance of women with uterine malformations. Hum. Reprod. 1993;8:122-126
  • American Fertility Society, 1988 American Fertility Society. The American Fertility Society classifications of adnexal adhesions, distal tubal occlusion, tubal occlusion secondary to tubal ligation, tubal pregnancies, mullerian anomalies and intrauterine adhesions. Fertil. Steril. 1988;49:944-955
  • Ban-Frangez et al, 2009 Ban-Frangez H., Tomazevic T., Virant-Klun I., Verdenik I., Ribic-Pucelj M., Bokal E.V. The outcome of singleton pregnancies after IVF/ICSI in women before and after hysteroscopic resection of a uterine septum compared to normal controls. Eur. J. Obstet. Gynecol. Reprod. Biol. 2009;146:184-187 Crossref
  • Ben-Rafael et al, 1991 Ben-Rafael Z., Seidman D.S., Recabi K., Bider D., Mashiach S. Uterine anomalies: a retrospective, matched-control study. J. Reprod. Med. Obstet. Gynecol. 1991;36:723-727
  • Buttram, 1983 Buttram V.C. Jr. Mullerian anomalies and their management. Fertil. Steril. 1983;40:159-163
  • Buttram, Gibbons, 1979 Buttram V.C. Jr., Gibbons W.E. Mullerian anomalies: a proposed classification.(An analysis of 144 cases). Fertil. Steril. 1979;32:40-46
  • Chan et al, 2011a Chan Y.Y., Jayaprakasan K., Tan A., Thornton J.G., Coomarasamy A., Raine-Fenning N.J. Reproductive outcomes in women with congenital uterine anomalies: a systematic review. Ultrasound Obstet. Gynecol. 2011;38:371-382 Crossref
  • Chan et al, 2011b Chan Y.Y., Jayaprakasan K., Zamora J., Thornton J.G., Raine-Fenning N., Coomarasamy A. The prevalence of congenital uterine anomalies in unselected and high-risk populations: a systematic review. Hum. Reprod. Update. 2011;17:761-771 Crossref
  • Colacurci et al, 1996 Colacurci N., De Placido G., Mollo A., Carravetta C., De Franciscis P. Reproductive outcome after hysteroscopic metroplasty. Eur. J. Obstet. Gynecol. Reprod. Biol. 1996;66:147-150 Crossref
  • Cooney et al, 1997 Cooney M.J., Benson C.B., Doubilet P.M. Outcome of pregnancies in women with uterine duplication anomalies. J. Clin. Ultrasound. 1997;26:3-6
  • Crocker et al, 2005 Crocker I.P., Wareing M., Ferris G.R., Jones C.J., Cartwright J.E., Baker P.N., Aplin J.D. The effect of vascular origin, oxygen, and tumour necrosis factor alpha on trophoblast invasion of maternal arteries in vitro. J. Pathol. 2005;206:476-485 Crossref
  • Dabirashrafi et al, 1995 Dabirashrafi H., Bahadori M., Mohammad K., Alavi M., Moghadami-Tabrizi N., Zandinejad K., Ghafari V. Septate uterus: new idea on the histologic features of the septum in this abnormal uterus. Am. J. Obstet. Gynecol. 1995;172:105-107 Crossref
  • DerSimonian, Laird, 1986 DerSimonian R., Laird N. Meta-analysis in clinical trials. Control. Clin. Trials. 1986;7:177-188 Crossref
  • Eccles et al, 2003 Eccles M., Grimshaw J., Campbell M., Ramsay C. Research designs for studies evaluating the effectiveness of change and improvement strategies. Qual. Saf. Health Care. 2003;12:47-52 Crossref
  • Erez et al, 2007 Erez O., Dukler D., Novack L., Rozen A., Zolotnik L., Bashiri A., Koifman A., Mazor M. Trial of labor and vaginal birth after cesarean section in patients with uterine Mullerian anomalies: a population-based study. Am. J. Obstet. Gynecol. 2007;196:537 e531–e511
  • Fedele et al, 1993 Fedele L., Arcaini L., Parazzini F., Vercellini P., Di Nola G. Reproductive prognosis after hysteroscopic metroplasty in 102 women: life- table analysis. Fertil. Steril. 1993;59:768-772
  • Fedele et al, 1996 Fedele L., Bianchi S., Marchini M., Franchi D., Tozzi L., Dorta M. Ultrastructural aspects of endometrium in infertile women with septate uterus. Fertil. Steril. 1996;65:750-752
  • Grimbizis et al, 1998 Grimbizis G., Camus M., Clasen K., Tournaye H., De Munck L., Devroey P. Hysteroscopic septum resection in patients with recurrent abortions or infertility. Hum. Reprod. 1998;13:1188-1193 Crossref
  • Grimbizis et al, 2001 Grimbizis G.F., Camus M., Tarlatzis B.C., Bontis J.N., Devroey P. Clinical implications of uterine malformations and hysteroscopic treatment results. Hum. Reprod. Update. 2001;7:161-174 Crossref
  • Grimbizis et al, 2013a Grimbizis G.F., Gordts S., Di Spiezio Sardo A., Brucker S., De Angelis C., Gergolet M., Li T.C., Tanos V., Brolmann H., Gianaroli L., Campo R. The ESHRE/ESGE consensus on the classification of female genital tract congenital anomalies. Hum. Reprod. 2013;28:2032-2044 Crossref
  • Grimbizis et al, 2013b Grimbizis G.F., Gordts S., Di Spiezio Sardo A., Brucker S., De Angelis C., Gergolet M., Li T.C., Tanos V., Brolmann H., Gianaroli L., Campo R. The ESHRE-ESGE consensus on the classification of female genital tract congenital anomalies. Gynecol. Surg. 2013;10:199-212 Crossref
  • Heinonen, 1997 Heinonen P.K. Reproductive performance of women with uterine anomalies after abdominal or hysteroscopic metroplasty or no surgical treatment. J. Am. Assoc. Gynecol. Laparosc. 1997;4:311-317 Crossref
  • Hua et al, 2011 Hua M., Odibo A., Longman R., Roehl K.A., Macones G., Cahill A. Congenital uterine anomalies and adverse pregnancy outcomes. Am. J. Obstet. Gynecol. 2011;204:S334-S335 Crossref
  • Jarcho, 1946 Jarcho J. Malformations of the uterus: review of the subject, including embryology, comparative anatomy, diagnosis and report of cases. Am. J. Surg. 1946;71:106-166 Crossref
  • Jayaprakasan et al, 2011 Jayaprakasan K., Chan Y.Y., Sur S., Deb S., Clewes J.S., Raine-Fenning N.J. Prevalence of uterine anomalies and their impact on early pregnancy in women conceiving after assisted reproduction treatment. Ultrasound Obstet. Gynecol. 2011;37:727-732 Crossref
  • Lavergne et al, 1996 Lavergne N., Aristizabal J., Zarka V., Erny R., Hedon B. Uterine anomalies and in vitro fertilization: what are the results?. Eur. J. Obstet. Gynecol. Reprod. Biol. 1996;68:29-34 Crossref
  • Liang, Hu, 2010 Liang F., Hu W. Pregnancy complications and obstetric outcomes among women with congenital uterine malformations. Int. J. Gynecol. Obstet. 2010;109:159-160
  • Lin et al, 2009 Lin K., Zhu X., Xu H., Liang Z., Zhang X. Reproductive outcome following resectoscope metroplasty in women having a complete uterine septum with double cervix and vagina. Int. J. Gynecol. Obstet. 2009;105:25-28 Crossref
  • Mollo et al, 2009 Mollo A., De Franciscis P., Colacurci N., Cobellis L., Perino A., Venezia R., Alviggi C., De Placido G. Hysteroscopic resection of the septum improves the pregnancy rate of women with unexplained infertility: a prospective controlled trial. Fertil. Steril. 2009;91:2628-2631 Crossref
  • Munoz Munoz et al, 2011 Munoz Munoz E., Ortiz Olivera G., Fernandez Lopez I., Martinez Martinez B., Aguilar Prieto J., Portela Perez S., Pellicer Martinez A. Do Mullerian anomalies affect embryo implantation?. Hum. Reprod. 2011;26:i211-i212
  • Pabuccu, Gomel, 2004 Pabuccu R., Gomel V. Reproductive outcome after hysteroscopic metroplasty in women with septate uterus and otherwise unexplained infertility. Fertil. Steril. 2004;81:1675-1678 Crossref
  • Pabuccu et al, 1995 Pabuccu R., Atay V., Urman B., Ergun A., Orhon E. Hysteroscopic treatment of septate uterus. Gynaecol. Endosc. 1995;4:213-215
  • Pang et al, 2011 Pang L.H., Li M.J., Li M., Xu H., Wei Z.L. Not every subseptate uterus requires surgical correction to reduce poor reproductive outcome. Int. J. Gynecol. Obstet. 2011;115:260-263 Crossref
  • Perino et al, 1987 Perino A., Mencaglia L., Hamou J., Cittadini E. Hysteroscopy for metroplasty of uterine septa: report of 24 cases. Fertil. Steril. 1987;48:321-323
  • Ravasia et al, 1999 Ravasia D.J., Brain P.H., Pollard J.K. Incidence of uterine rupture among women with mullerian duct anomalies who attempt vaginal birth after cesarean delivery. Am. J. Obstet. Gynecol. 1999;181:877-881 Crossref
  • Salim et al, 2003 Salim R., Regan L., Woelfer B., Backos M., Jurkovic D. A comparative study of the morphology of congenital uterine anomalies in women with and without a history of recurrent first trimester miscarriage. Hum. Reprod. 2003;18:162-166 Crossref
  • Saravelos et al, 2008a Saravelos S.H., Cocksedge K.A., Li T.C. Prevalence and diagnosis of congenital uterine anomalies in women with reproductive failure: a critical appraisal. Hum. Reprod. Update. 2008;14:415-429 Crossref
  • Saravelos et al, 2008b Saravelos S.H., Cocksedge K.A., Li T.C. Prevalence and diagnosis of congenital uterine anomalies in women with reproductive failure: a critical appraisal. Hum. Reprod. Update. 2008;14:415-429 Crossref
  • Saravelos et al, 2010 Saravelos S.H., Cocksedge K.A., Li T.C. The pattern of pregnancy loss in women with congenital uterine anomalies and recurrent miscarriage. Reprod. Biomed. Online. 2010;20:416-422 Crossref
  • Shuiqing et al, 2002 Shuiqing M., Xuming B., Jinghe L. Pregnancy and its outcome in women with malformed uterus. Chin. Med. Sci. J.. 2002;17:242-245
  • Sorensen, Trauelsen, 1987 Sorensen S.S., Trauelsen A.G.H. Obstetric implications of minor mullerian anomalies in oligomenorrheic women. Am. J. Obstet. Gynecol. 1987;156:1112-1118 Crossref
  • Sugiura-Ogasawara et al, 2010 Sugiura-Ogasawara M., Ozaki Y., Kitaori T., Kumagai K., Suzuki S. Midline uterine defect size is correlated with miscarriage of euploid embryos in recurrent cases. Fertil. Steril. 2010;93:1983-1988 Crossref
  • Taylor, 2000 Taylor H.S. The role of HOX genes in human implantation. Hum. Reprod. Update. 2000;6:75-79 Crossref
  • Tomazevic et al, 2007 Tomazevic T., Ban-Frangez H., Ribic-Pucelj M., Premru-Srsen T., Verdenik I. Small uterine septum is an important risk variable for preterm birth. Eur. J. Obstet. Gynecol. Reprod. Biol. 2007;135:154-157 Crossref
  • Tomazevic et al, 2009 Tomazevic T., Ban-Frangez H., Virant-Klun I., Verdenik I. Spontaneous abortion in singleton pregnancies after IVF/ICSI in women before and after hysteroscopic resection of a uterine septum compared to normal controls. Gynecol. Surg. 2009;6:S97
  • Tomazevic et al, 2010 Tomazevic T., Ban-Frangez H., Virant-Klun I., Verdenik I., Pozlep B., Vrtacnik-Bokal E. Septate, subseptate and arcuate uterus decrease pregnancy and live birth rates in IVF/ICSI. Reprod. Biomed. Online. 2010;21:700-705 Crossref
  • Tonguc et al, 2011 Tonguc E.A., Var T., Batioglu S. Hysteroscopic metroplasty in patients with a uterine septum and otherwise unexplained infertility. Int. J. Gynecol. Obstet. 2011;113:128-130 Crossref
  • Valli et al, 2004 Valli E., Vaquero E., Lazzarin N., Caserta D., Marconi D., Zupi E. Hysteroscopic metroplasty improves gestational outcome in women with recurrent spontaneous abortion. J. Am. Assoc. Gynecol. Laparosc. 2004;11:240-244 Crossref
  • Wells et al, 2000 Wells, G., Shea, B., O'connell, D., Peterson, J., Welch, V., Losos, M., Tugwell, P., 2000. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses.
  • Woelfer et al, 2001 Woelfer B., Salim R., Banerjee S., Elson J., Regan L., Jurkovic D. Reproductive outcomes in women with congenital uterine anomalies detected by three-dimensional ultrasound screening. Obstet. Gynecol. 2001;98:1099-1103 Crossref
  • Zhang et al, 2010 Zhang Y., Zhao Y.Y., Qiao J. Obstetric outcome of women with uterine anomalies in China. Chin. Med. J.. 2010;123:418-422
  • Zlopasa et al, 2007 Zlopasa G., Skrablin S., Kalafatic D., Banovic V., Lesin J. Uterine anomalies and pregnancy outcome following resectoscope metroplasty. Int. J. Gynecol. Obstet. 2007;98:129-133 Crossref
  • Zupi et al, 1996 Zupi R., Solima E., Marconi D., Valli E., Romanini C. Uterine anomalies prevalence and reproductive outcome in women undergoing diagnostic hysteroscopy. Gynaecol. Endosc. 1996;5:147-150
rbmo1230-fig-5001

Christos A. Venetis obtained his MD degree from the Medical School of the Aristotle University of Thessaloniki, Greece. He subsequently, obtained an MSc in medical research methodology, a PhD, and completed his training in obstetrics and gynaecology at the same institution. He also recently completed his subspecialty training in reproductive medicine. Dr. Venetis is an author of more than 40 publications in international and national scientific journals and several chapters in books. He is a strong advocate of evidence-based medicine, and his research interests include ovarian stimulation, reproductive endocrinology and research methodology.

Footnotes

a First Department of Obstetrics and Gynecology, Medical School, Aristotle University of Thessaloniki, Thessaloniki, Greece

b School of Women's and Children's Health, UNSW Medicine, University of New South Wales, Sydney, Australia

c European Academy for Gynecological Surgery, Scientific project on Female Genital Tract Congenital Anomalies, Diestsevest 43/0001, 3000 Leuven, Belgium

d European Society for Gynecological Surgery, Special Interest Group on Reproductive Surgery, Project on Congenital Uterine Anomalies (CONUTA), Diestsevest 43/0001, 3000 Leuven, Belgium

* Corresponding author.