Introduction: Exomphalos is a common fetal anomaly, whose outcome depends on its contents and the presence of a chromosomal, genetic or other structural abnormality. The aim of this study is to compare the outcome of antenatally diagnosed exomphalos in isolated and complex (those with associated abnormalities) cases, and the effects of herniation of liver in these groups.

Material and methods: This is a retrospective review of antenatally diagnosed exomphalos.

Results: There were 64 cases of exomphalos, 40.6% were isolated and 59.4% were complex. In the isolated group, there was spontaneous resolution in 61.5% cases in those containing bowel only and 0% in those with liver herniation. There were no chromosomal abnormalities in the isolated group. All isolated cases had successful defect closures. Of the complex exomphalos' 68.2% had chromosomal abnormities. Those containing bowel only and a normal karyotype follow a similar post-operative course as their isolated equivalents. The complex cases with liver herniation and a normal karyotype have an increased risk of neonatal death (NND).

Conclusion: Our study shows that in isolated exomphalos, with or without liver herniation, the risk of chromosomal abnormality is low (0%). Those with liver herniation had a higher surgical morbidity but not mortality. In complex exomphalos it is necessary to exclude an abnormal karyotype, and then counselling needs to be individualised depending on the coexisting anomalies.

Keywords: First trimester, Exomphalos, Isolated, Complex, Bowel, Liver herniation, Spontaneous resolution, Karyotype, Outcome

Abbreviations: CRL: Crown Rump Length; NT: Nuchal Translucency; OEIS: Omphalocoele-Extrophy-Imperforate Anus-Spinal Defects; LB: Live Birth; TOP: Terminations of Pregnancy; IUD: Intrauterine Death

INTRODUCTION

Exomphalos is a common fetal anomaly with a reported prevalence at 11-14 weeks of 1 in 1000 [1,2] which decreases to between 1 in 5000 and 1 in 7000 at birth [3-5]. Recent evidence suggests that the incidence of exomphalos at birth has risen to 1 in 3500 [6] due to increasing maternal age.

The outcome of antenatally diagnosed exomphalos is dependent on its contents and the presence of a chromosomal, genetic or other structural abnormality. The rate of spontaneous resolution by 12 weeks of gestation in exomphalos containing bowel only has been stated to be between 31% [7] and 92%, however, there is no reported spontaneous resolution of exomphalos containing liver [2,7].

The presence of the fetal liver in the exomphalos has been reported to have varying effects on perinatal and neonatal outcomes. St-Vil et al. [8] showed that the absence of liver herniation increased the risk of aneuploidy, whereas Iacovella et al. [7] suggested that exomphalos contents do not correlate with the risk of aneuploidy. There is also a discrepancy in reported postoperative neonatal outcomes, with one study suggesting a lower survival rate in cases with liver herniation (42% vs. 82%) [13], whilst others have suggested the exomphalos contents do not affect survival rate [9].

The aim of this study is to compare the outcome of antenatally diagnosed exomphalos in isolated and complex (those with associated abnormalities) cases and the effects of herniation of liver in these groups.

PATIENTS AND METHODS

A retrospective review of all consecutive fetuses with antenatally diagnosed exomphalos managed in the Fetal Medicine Unit, University College London Hospitals and subsequently treated at the Pediatric Surgical Unit, Great Ormond Street Hospital, London, between 2009 and 2014 was performed. Initial diagnosis was made by a sonographer, and then reviewed by a fetal medicine specialist to confirm the diagnosis, assess liver herniation (Figures 1 and 2) and exclude other structural anomalies. A repeat scan one week later was offered to those with a crown rump length (CRL) of less than 55 mm and having an exomphalos containing bowel only to assess for spontaneous resolution. Invasive testing and a detailed cardiac scan by a specialist were offered in all cases. Patients who wished to continue their pregnancies had regular follow up at 2 to 4 week intervals and were counseled by pediatric surgeons about neonatal surgical management and outcome. All scan data and surgical reviews were documented on a fetal medicine database.

The Fetal Medicine Unit database was searched to identify all cases of exomphalos. Maternal characteristics collected included age, ethnicity, mode of conception and parity. Fetal data collected included gestational age at diagnosis, CRL, nuchal translucency (NT) measurement, other structural anomalies, fetal karyotype, subsequent ultrasound and post-mortem findings (where available). Obstetric, neonatal and paediatric surgical notes were examined for gestational age at delivery, mode of delivery, birth weight, gender and surgical outcome. All parents of neonates requiring surgery were contacted by telephone and data on associated anomalies, length of hospital stay and number of procedures required for closure was collected. There was a minimum follow-up period of 1 year and the maximum 6 years.

Statistical analysis included comparison of perinatal outcomes between the isolated and complex cases. In addition, comparison was made between the presence and absence of liver herniation within the isolated and complex groups. Mann-Whitney U, non-parametric test was used for continuous ordinal data and Z Test was used for categorical data to check two population proportions.

RESULTS

There were a total of 82 cases referred with exomphalos between January 2009 and January 2014. Of these, 18 were excluded from the study population due to an antenatal ultrasound diagnosis of body stalk anomaly (n=12), bladder extrophy (n=3), omphalocoele-extrophy-imperforate anus-spinal defects (OEIS) (n=2) and conjoined twins (n=1). Of the 64 remaining cases, 26 (40.6%) were isolated and 38 (59.4%) were complex with other associated abnormalities (Figures 3 and 4).

ISOLATED EXOMPHALOS

The 26 cases of isolated exomphalos were subdivided into 13 (50%) with liver herniation and 13 (50%) without liver herniation. The characteristics of the patients can be seen in Table 1.

In the isolated exomphalos group without liver herniation, the mean maternal age (28.6 vs. 32.8 years; p=0.03) and the mean gestational age at diagnosis (12.0 vs. 14.6 weeks; p=0.04) were statistically significantly lower than in those with liver herniation. The mean NT was not statistically significantly different between those without and with liver herniation (1.8 mm vs. 1.5 mm; p=0.28), however, the mean NT was lower in the isolated cases than the complex ones (1.6 mm vs. 2.3 mm, p=0.003).

Spontaneous resolution occurred in 8 (61.5%) isolated cases without liver herniation. Of the remaining 5 cases, all were karyotyped, with 4 having a normal karyotype and one having an abnormal result (a maternally inherited Para centric inversion of chromosome 9). All 13 (100%) cases resulted in a live birth (LB). Of the 5 non-spontaneously unresolved cases one baby found to have a cleft palate postnatally. Three required surgical correction, all of which were done as a single uncomplicated procedure. The remaining two did not necessitate treatment; one had a non-surgical tuck procedure and the other was found to be only an umbilical hernia.

In the group of 13 cases of isolated exomphalos with liver herniation, there were no cases of spontaneous resolution. 11 (84.6%) opted for karyotyping (all proved normal) and 2 (15.4%) declined (1 terminated and 1 had a live birth with a normal karyotype postnatally). There were 8 (61.5%) had live births, 4 (30.8%) terminations of pregnancy (TOP) and 1 (7.7%) intrauterine death (IUD). All the live births were diagnosed postnatally with further anomalies (Table 2) and all required surgical correction.

The postnatally diagnosed anomalies included a single case of congenital diaphragmatic hernia diagnosed at 18 months’ post-delivery investigated due to repeated respiratory infections post-surgery. There were two small ventricular septal defects, neither of which required any intervention, and three patent ductus arteriosus managed conservatively. There were 2 cases complicated by pulmonary hypoplasia and 2 cases of intestinal malrotation. The single case of IUD in this group was diagnosed with right isomerism and intestinal malrotation at post-mortem.

The surgical outcome in the isolated group without liver herniation showed a statistically significantly shorter median hospital stay (12 days vs. 49 days; p=0.01) and fewer median operations to complete abdominal closure (1 vs. 7; p=<0.01) as compared to the isolated cases with liver herniation.

COMPLEX EXOMPHALOS

The 38 cases of exomphalos with associated malformations were subdivided into 22 without (57.9%) and 16 (42.1%) with liver herniation (Figure 4). The characteristics of the patients can be seen in Table 1.

There was no statistical significant difference between those without and with liver herniation in the mean maternal age at diagnosis (34.5 vs. 34.0 years; p=0.810), the mean gestational age of diagnosis (14.6 vs. 14.8 weeks; p=0.960) and the mean NT (3.3 mm vs. 2.6 mm; p=0.276). There were 2 cases with CRL’s below 45 mm at diagnosis, however, both had multiple abnormalities so a chromosome abnormality with growth restriction were strongly suspected and these were included in the analysis. There were no cases of spontaneous resolution.

All 22 cases without liver herniation opted for karyotype, with 15 (68.2%) having an abnormal and 7 (31.8%) a normal karyotype. The abnormal karyotypes were as follows: 11 (50%) cases of trisomy 18, 2 (9.1%) of trisomy 13, 1 (4.5%) of triploidy and 1 (4.5%) Turner’s syndrome. The outcomes of those with abnormal karyotype were 12 TOP’s and 3 IUDs. Of those with a normal karyotype, 3 were terminated for other abnormalities (acrania, spina bifida and megacystis) and 4 (18.2%) were live births. Each of the 4 live births were found to have additional postnatally diagnosed abnormalities, these included Beckwith Wiedemann Syndrome, cleft palate, hypospadias, undescended testis, duplication of kidneys, renal dysplasia and single umbilical artery. These findings are presented in Table 3.

In the 16 complex cases with liver herniation 15 opted for karyotype, with 3 (18.8%) having an abnormal and 12 (75%) a normal karyotype. There were statically significantly more abnormal karyotypes in the complex without liver herniation group (68.2% vs. 18.8%, p=0.003). One case declined karyotype and opted for TOP due to spina bifida. All 3 abnormal karyotypes were trisomy 18 and terminated. Of the 12 with a normal karyotype, 5 (31.3%) terminated and 7 (43.8%) had live births, 3 (18.8%) of which died within the neonatal period. There was no difference in the number of live births in the two complex exomphalos groups (p=0.085). All the live births were found to have associated abnormalities postnatally. Amongst the surviving neonates these included a right aortic arch, and overriding aorta, duplex kidneys, inguinal hernia, undescended testis, mal rotation of the bowel and talipes. These findings are presented in Table 4.

The surgical outcome in the associated anomalies group without liver herniation showed a shorter median hospital stay (21 days vs. 42 days) and fewer median number operations to complete abdominal closure (1.5 vs. 4) as compared to the isolated cases with liver herniation. The sample size was too small to comment on any significance.

DISCUSSION

Our study presents a comprehensive counselling framework for antenatally diagnosed exomphalos depending on first trimester ultrasound findings. We have used the ultrasound marker of presence or absence of liver herniation in isolated and complex exomphalos to predict antenatal, postnatal and surgical outcomes.

A similar study divided exomphalos into isolated and complex, however, it is limited as it did not contain any fetuses with liver herniation in the isolated group and did not look at the long-term outcome [10].

Some studies have attempted to predict the risk of aneuploidies in exomphalos. Khalil et al. [11] compared their patients in three groups: isolated exomphalos with a normal NT, isolated exomphalos with an increased NT and complex exomphalos. However, this series used NT as a marker for outcome and did not evaluate surgical outcome [11]. Iacovella et al. [7] used the exomphalos contents and NT to predict the risk of aneuploidy but did not have a long-term follow up of the patients [7]. Finally, Tassin et al. [12] examined the surgical outcome of exomphalos based on its size in the first trimester; the information regarding the content was only available retrospectively [12].

Our overall aneuploidy rate was 29.6%. Most studies report aneuploidy rates in exomphalos of between 55-62% [7,11]. However, Khalil et al. [11] included all cases with increased NT's, whilst Iacovella et al. [7] included all those with associated anomalies. The inclusion of these cases would increase the presence of a chromosomally abnormal fetus. The use of these heterogeneous groups does not allow direct comparison of isolated and complex anomalies.

Isolated exomphalos

We have shown that antenatally diagnosed isolated exomphalos containing bowel only has a spontaneous resolution rate of 61.5%. Previous literature on the rate of spontaneous resolution of exomphalos without liver herniation varies from 31% to 92% [2,7]. The lower quoted rate is a combination of isolated and complex exomphalos, whilst the higher rate includes those diagnosed at an earlier gestation when physiological exomphalos is more prevalent. Our results are consistent with Khalil et al. [11] who had a similar spontaneous resolution rate of 64% in isolated exomphalos with a normal NT [11]. Our policy of rescanning all cases of isolated exomphalos without liver herniation and a CRL of less than 55 mm is effective at both reassuring the patient and reducing invasive testing. Neither complex nor isolated exomphalos containing liver show spontaneous resolution, which is consistent with previous studies [2,7,11].

In isolated cases, it was 0% in both the bowel only and liver herniation groups, with only one case not tested as the patient proceeded straight to TOP. There are consistent results to those reported by Blazer et al. [10]. In our series, there was a single case of a maternally inherited paracentric inversion of chromosome 9, which was reviewed by a geneticist who concluded that this chromosomal abnormality was not the cause of the exomphalos.

In the isolated group with liver herniation, 11 accepted karyotyping, which was normal in all cases. Of the two that declined one was shown to have a normal karyotype postnatally and the other was terminated without karyotyping. Previous studies have reported rates of aneuploidy of between 0-53% in those with liver herniation. Blazer et al. [10] showed the rate of aneuploidy in 16 fetuses with isolated exomphalos containing liver herniation to be 0% [10]. This finding was also shown by Khalil et al. [11] where, in 23 cases of exomphalos containing liver herniation the rate of aneuploidy was 0%, even though these were a mixture of complex and isolated cases [11]. However, Kagan et al. [2] study was a heterogeneous group of both isolated and complex exomphalos with liver herniation, and hence explains the higher rate, 53%, of chromosomal abnormalities [6].

In terms of surgical repair, our results show that only 60% of bowel only exomphalos required repair, and that these were single-stage procedures with a short postoperative stay. These results were comparable to those of Blazer et al. [10] who presented 16 cases of isolated exomphalos containing only bowel. Of the 7 neonates requiring operations, all had an uncomplicated single-stage surgical repair. St-Vil et al. [8] also showed that bowel only containing exomphalos were more likely to have a single primary closure and a shorter hospital stay [8].

Each of the eight antenatally diagnosed isolated cases with liver herniation was found to have associated anomalies after delivery (Table 2). Though these had an impact on the postnatal course (increased hospital stay and operations), they were not shown to alter the mortality rate. Other studies have revealed postnatally diagnosed anomalies in neonates thought to have isolated exomphalos alone. A study by Durfee et al. [13] showed that up to 33% of antenatally diagnosed isolated exomphalos can have additional abnormalities diagnosed postnatally; these included genitourinary, cardiac or central nervous system abnormalities [13]. Kominiarek et al. [14] also noted additional postnatally diagnosed anomalies including Beckwith-Weidemann Syndrome, cleft palate, atrial-septal defects, inguinal hernias and hydronephrosis [14].

The isolated exomphalos containing liver group had a more complex surgical course, with complications such as delayed wound healing, wound infection and complications of tight closure (including CDH, lung collapse, splenic rupture and inguinal hernia). Our findings of increased surgical morbidity are consistent with the findings of St-Vil et al. [8] who showed that those exomphalos containing liver were more likely to have multiple closures and a longer hospital stay [13]. Tassin et al. [12] also documented a longer hospital stay, an increased need for respiratory assistance and a greater need for parenteral feeding in those with an exomphalos containing liver [12] whilst, Nicholas et al. [5] stated that these cases were also more likely to have gastrointestinal atresia and sepsis.

Despite a normal karyotype, a proportion of isolated exomphalos containing liver opted for a TOP. This may be a result of the counselling received and the anxiety that liver herniation is associated with significantly higher morbidity and mortality.

Our results emphasise that, firstly, in isolated exomphalos, regardless of the presence or absence of liver herniation, the rate of chromosomal abnormalities is low. Secondly that the presence of liver herniation increases the number of operations and length of hospital stay required for repair without an associated increase in mortality. Thirdly, that the presence of liver herniation should warrant a careful postnatal review for the possibility of other structural anomalies.

Complex exomphalos

There were no cases of spontaneous resolution in the complex exomphalos group, a finding consistent with previous studies [2,11]. One case of exomphalos containing bowel only and a single umbilical artery resolved by 30 weeks, but was subsequently found to have a cleft palate postnatally.

In the complex group, the overall rate of aneuploidy was 47.3%, which is less than the 61% quoted by Blazer et al. [10]. In the complex group without liver herniation, the rate of chromosomal abnormalities was 68.8%, which was significantly higher than that in the complex with liver herniation (18.8%).

It is difficult to compare our study to other publications, as they are heterogeneous groups of isolated and complex exomphalos. Kagan et al. [2] found the incidence of aneuploidy to be 55.6% in cases with an exomphalos containing bowel only and 52.9% in those with an exomphalos containing liver, whist Iacovella et al. [7] showed the incidences of 77.8% and 22.2% in the same groups respectively. These studies did not differentiate isolated and complex cases. As quoted above, Khalil et al. [11] stated the incidence of aneuploidy in isolated and complex exomphalos containing liver to be 0% [11], which is lower than our findings.

In the complex exomphalos group without liver herniation, all those with an abnormal karyotype were either terminated or had an IUD. All 3 IUD’s were of fetuses affected by trisomy 18. Of the 7 fetuses with a normal karyotype, 3 were terminated due to other associated structural anomalies (acrania, spina bifida and megacystis). Each of the 4 live births was found to have associated abnormalities post-delivery, these included Beckwith Wiedemann Syndrome, cleft palate, hypospadias, undescended testis, duplication of kidneys, renal dysplasia and single umbilical artery. These findings are shown in Table 3. Other studies have also commented on the additional abnormalities diagnosed postnatally. These abnormities have included cardiac anomalies, cystic hygroma, limb anomalies and renal anomalies [8,10].

In the 16 complex cases with liver herniation 15 opted for karyotype, with 3 having an abnormal and 12 a normal and karyotype. One case declined karyotype and proceeded directly to TOP for spina bifida. All 3 abnormal karyotypes were trisomy 18 and terminated. Of the 12 with a normal karyotype, 5 terminated and 7 had live births, 3 of which died within the neonatal period. All the live births were found to have associated abnormalities postnatally (Table 4). Amongst the surviving neonates these included a right aortic arch, and overriding aorta, duplex kidneys, inguinal hernia, undescended testis, malrotation of the bowel and talipes.

The surgical outcome in the complex group without liver herniation showed a shorter median hospital stay (21 days vs. 42 days) and fewer median number operations to complete abdominal closure (1.5 vs. 4) as compared to the isolated cases with liver herniation. The sample size was too small to comment on any significance.

Though the complex exomphalos without liver herniation had a statistically significantly higher abnormal karyotype rate than the complex with liver herniation, those born alive required fewer operations to correct the defect and had a shorter hospital stay. In addition, there were no NND’s in the complex exomphalos without liver herniation. Therefore, fetuses with a complex exomphalos without herniation of the liver, who have a normal karyotype, have a good surgical outcome.

CONCLUSION

Our study presents a comprehensive counselling framework for antenatally diagnosed exomphalos depending on first trimester ultrasound findings. We have shown that in isolated exomphalos, with or without liver herniation, the risk of chromosomal abnormality is low (0%). There is an increased risk of coexisting abnormalities found postnatally in those with liver herniation and a higher surgical morbidity but not mortality. In complex exomphalos it is necessary to exclude an abnormal karyotype. Those without liver herniation and a normal karyotype follow a similar post-operative course as their isolated equivalents. The complex cases with liver herniation have an increased risk of NND, therefore counselling needs to be individualised depending on the coexisting anomalies. 

1.       Baird PA, MacDonald EC (1981) An epidemiologic study of congenital malformations of the anterior abdominal wall in more than half a million consecutive live births. Am J Hum Genet 33: 470-478.

2.       Blazer S, Zimmer EZ, Gover A, Bronshtein M (2004) Fetal omphalocele detected early in pregnancy: Associated anomalies and outcomes. Radiology 232: 191-195.

3.       Durfee SM, Benson CB, Adams SR, Ecker J, House M, et al. (2013) Postnatal outcome of fetuses with the prenatal diagnosis of gastroschisis. J Ultrasound Med 32: 407-412.

4.       Heider AL, Strauss RA, Kuller JA (2004) Omphalocele: Clinical outcomes in cases with normal karyotypes. Am J Obstet Gynecol 190: 135-141.

5.       Iacovella C, Contro E, Ghi T, Pilu G, Papageorghiou A, et al. (2012) The effect of the contents of exomphalos and nuchal translucency vat 11-14 weeks on the likelihood of associated chromosomal abnormality. Prenat Diagn 32: 1066-1070.

6.       Kagan KO, Staboulidou I, Syngelaki A, Cruz J, Nicolaides KH (2010) The 11-13 week scan: Diagnosis and outcome of holoprosencephaly, exomphalos and megacystis. Ultrasound Obstet Gynecol 36: 10-14.

7.       Khalil A, Arnaoutoglou C, Pacilli M, Szabo A, David AL, et al. (2012) Outcome of fetal exomphalos diagnosed at 11-14 weeks of gestation. Ultrasound Obstet Gynecol 39: 401-406.

8.       Kominiarek MA, Zork N, Pierce SM, Zollinger T (2011) Perinatal outcome in the live-born infant with prenatally diagnosed omphalocele. Am J Perinatol 28: 627-634.

9.       Nicholas SS, Stamilio DM, Dicke JM, Gray DL, Macones GA, et al. (2009) Predicting adverse neonatal outcomes in fetuses with abdominal wall defects using prenatal risk factors. Am J Obstet Gynecol 201: 383.e1-383.e6.

10.    Paidas MJ, Crombleholme TM, Robertson FM (1994) Prenatal diagnosis and management of the fetus with an abdominal wall defect. Semin Perinatol 18: 196-214.

11.    Snijders RJM, Brizot ML, Faria M, Nicolaides KH (1995) Fetal exomphalos at 11 to 14 weeks of gestation. J Ultrasound Med 14: 569-574.

12.    Srivastava V, Mandhan P, Pringle K, Morreau P, Beasley S, et al. (2009) Rising incidence of gastroschisis and exomphalos in New Zealand. J Pediatr Surg 44: 551-555.

13.    St-Vil D, Shaw KS, Lallier M, Yazbeck S, Di Lorenzo M, et al. (1996) Chromosomal anomalies in new-borns with omphalocele. J Pediatr Surg 31: 831-834.

14.    Stoll C, Alembik Y, Dott B, Roth MP (2001) Risk factors in congenital abdominal wall defects (omphalocele and gastroschisi): A study in a series of 2,65,858 consecutive births. Ann Genet 44: 201-208.

15.    Tassin M, Descriaud C, Elie C, Houfflin Debarge V, Dumez Y, et al. (2013) Omphalocele in the first trimester: Prediction of perinatal outcome. Prenat Diagn 33: 497-501.