Article Text
PDF
Abstract
Design All fetuses diagnosed with ‘absent stomach’ at anomaly screening over an 8-year period were identified using the University College Hospital fetal medicine database. These were cross-referenced with records from the paediatric surgical unit at Great Ormond Street Hospital and pathology department at University College Hospital to ascertain postnatal or postmortem diagnosis and outcome in each case.
Results Of the 84 cases identified, eight were found to have normal stomachs on subsequent antenatal scans, while 76 had persistent non-visualisation of the stomach. Underlying diagnoses included 24 gastro-intestinal tract and/or respiratory anomalies, 22 aneuploidies, six neuromuscular syndromes, three central nervous system anomalies, seven renal anomalies and five genetic syndromes. Seven cases had no identifiable postnatal abnormalities, 26 pregnancies were terminated and nine fetuses died in utero. Of the 33 live births, eight died in the neonatal period and three died in infancy. Only 28 survived into childhood. Two patients were lost to follow up.
Conclusions Persistent non-visualisation of the fetal stomach in the antenatal period was associated with a wide range of underlying diagnoses. In many cases, prognosis was poor. Only 37% of pregnancies resulted in liveborn infants surviving more than 6 months. The incidence of an abnormal karyotype was 29%. Diagnosis and outcome was normal in only 9.2% of cases. We propose an algorithm for the management of persistent non-visualisation of the fetal stomach on antenatal ultrasound.
Statistics from Altmetric.com
Introduction
Identification of the fetal stomach is an established component of fetal anomaly ultrasonography (figure 1). It is also routinely used as a landmark for measuring abdominal circumference in the assessment of fetal growth.1
Normal fetal stomach on antenatal ultrasound.
The ‘absent stomach’ (figure 2) can reveal a wide range of underlying pathologies. This poses significant problems in diagnosis, counselling and management of these pregnancies. Only three publications have addressed this issue to date,2,–,4 the most recent of which was published over 15 years ago. All cases presenting to our unit over an 8-year period were retrospectively reviewed to define and quantify diagnoses and outcomes, and propose a management strategy for clinicians faced with non-visualisation of the fetal stomach on antenatal scans.
Absent fetal stomach on antenatal ultrasound.
Methods
The Fetal Medicine Unit (FMU) of the Elizabeth Garrett Anderson Wing of University College Hospital is a tertiary referral centre in London, UK receiving patients from a large catchment area covering London, and south east and central England. A prospectively collected computerised database of all cases seen in the FMU was instituted in March 2000. All cases with suspicion of ‘absent fetal stomach’ or ‘collapsed fetal stomach’ on antenatal scans performed after 18 weeks gestation in the FMU between March 2000 and June 2008 were included in this study.
What is already known on this topic
▶ Predicting diagnosis and outcome in fetuses with an ‘absent stomach’ is challenging.
▶ Only three publications have addressed this issue, the most recent being over 15 years ago.
▶ One study reported abnormal outcome in 100% of cases with absent stomach, another identified underlying pathology in 85%, while the third reported an abnormal outcome in only 48%
What this study adds
▶ In this the largest series of ‘absent stomach’ to date (n=76), only 37% of pregnancies resulted in liveborn infants surviving more than 6 months.
▶ The incidence of an aneuploidy was 30%.
▶ Diagnosis and outcome were normal in only 9.2% of cases.
Patients attending the clinic were seen jointly by the fetal medicine specialist and the paediatric surgeon from Great Ormond Street Hospital (GOSH). The ultrasound devices used to perform the examinations were the Acuson Sequoia (Siemens, Munich, Germany) and ProSound Alpha 10 (Aloka, Tokyo, Japan) systems. All relevant sonographic features were recorded, in particular the presence of any other structural anomalies and the quantitative assessment of liquor volume, together with any differential diagnoses made at the time of the antenatal assessment. Non-visualisation of the fetal stomach was confirmed by two examinations on the same day at least 1 h apart, followed by another within 10 days. Invasive karyotype assessments were performed as appropriate and the results of these included in this study. Gestational and maternal ages were calculated as mean and range in weeks and years, respectively. The antenatal sonographic features were compared to postnatal clinical features in all pregnancies which continued to term. Follow-up occurred either at GOSH or at the referring hospital.
Approval for the study was obtained from the GOSH Clinical Ethics Committee. Pregnancy outcomes were established from paediatric records (identified from GOSH Patient Information Management Solutions (PiMS; Productive Data Management, Rowland Heights, California, USA)) and the perinatal pathology database. We also contacted general practitioners by telephone and fax where we were otherwise unable to identify the outcome.
Results
Eighty-four cases were referred during the study period. Eight of these were subsequently found to have a normal stomach appearance on further antenatal ultrasound scanning; information from the general practitioners indicated that all these children had normal outcomes. Of the remaining 76 cases, only two were lost to follow up (2.6%). Mean maternal age was 31 years (range 15–43).
Pregnancy outcomes
Thirty-nine (51.3%) of the 76 cases resulted in live births at a mean gestational age of 37 weeks (range 27–41) (table 1, figure 3). There were nine (11.8%) intrauterine deaths from 24 to 40 weeks (mean 32.4 weeks), and the remaining 26 (34.2%) pregnancies were terminated from 16 to 36 weeks (mean 22.7 weeks). Of the 39 live births, eight (10.5%) died in the neonatal period and another three (3.9%) died within 6 months. Therefore at 6 months after delivery there were only 28 (36.8%) surviving children (of whom seven (9.2%) were found to be normal).
Pregnancy outcomes in all cases of persistent non-visualisation of the fetal stomach (n=76).
Outcome in all cases of persistent non-visualisation of the fetal stomach by gestational age (n=76)
Underlying diagnoses
Of the 39 live born babies, diagnoses were obtained in all cases (table 2). Seven (9.2%) were found to have no identifiable postnatal abnormality. Investigations included gastrointestinal contrast imaging, ultrasound of the renal tracts and echocardiography. These babies are now between 2 and 8 years of age and are all well.
Definitive diagnoses in all cases of persistent non-visualisation of the fetal stomach (n=76)
Invasive assessment was routinely offered and karyotyping by amniocentesis or chorionic villus sampling was performed in 33 cases (43.4%). Twenty-one (27.6%) cases were found to have abnormal karyotypes; the remaining 12 were euploid. The most common aneuploidies were Edwards syndrome (13.2%) and Down syndrome (6.6%). There was one case each of Patau syndrome (trisomy 13) and rarer aneuploidies: trisomy 22, mosaic trisomy 15, tetrasomy 12p mosaic (Pallister-Killian syndrome), Wolf-Hirschhorn syndrome, partial monosomy 10 and partial trisomy 16. In the 10 cases of Edwards syndrome, all displayed multiple abnormalities, most frequently cardiac anomalies (90%).
In 24 (31.6%) cases, a respiratory or upper gastrointestinal anomaly was present: four isolated oesophageal atresia (5.3%), two oesophageal atresia with tracheo-oesophageal fistula and a cardiac anomaly (2.6%), two oesophageal atresia with tracheo-oesophageal fistula as part of VACTERL association (2.6%), three Fraser syndrome (3.9%), two congenital diaphragmatic hernia (2.6%), two hydrops fetalis (2.6%), two exomphalos (2.6%), and one each (1.3%) of cleft palate, Pierre Robin syndrome, neck tumour, gastroschisis, abdominal tumour, congenital microgastria (figure 4) and jejunal atresia. In the cases with congenital diaphragmatic hernia, exomphalos and gastroschisis, the stomach was subsequently found within the herniated viscera at the time of operation.
Congenital microgastria on barium swallow.
Seven fetuses (9.2%) were found to have renal anomalies. One had bilateral renal agenesis (1.3%), two had polycystic kidneys (2.6%), two had bilateral severe renal dysplasia (2.6%) and two had lower urinary tract obstruction with hydronephrosis (2.6%).
Six fetuses (7.9%) had an underlying neuromuscular syndrome, the fetal akinesia sequence (also known as Pena-Shokeir syndrome). None survived. Three were intrauterine deaths, two women opted for termination of pregnancy, and the sixth case resulted in a neonatal death.
Three cases (3.9%) were found to have severe central nervous system anomalies: two had anencephaly (one termination and one in utero death) and one had Dandy-Walker sequence (neonatal death).
In four cases (5.3%), a specific genetic diagnosis with a normal karyotype was made. These were two lysosomal storage disease (2.6%), one Ivemark syndrome (1.3%) and one fragile X syndrome (1.3%).
Two fetuses (2.6%) had complex cardiac anomalies with normal karyotypes.
Liquor volume
In 38 cases (50%), there was an abnormality of liquor volume (table 3): 24 (31.6%) had polyhydramnios and 14 (18.4%) had oligohydramnios.
Diagnosis in polyhydramnios (n=24) and oligohydramnios (n=14) subgroups
Discussion
Confirmation of the presence of a ‘stomach bubble’, a sonolucent spherical structure in the left upper abdomen below the heart, forms an integral part of a structured fetal sonogram. Its appearance is sought at the initial 12-week booking scan, the 20-week anomaly scan and any subsequent examination. It forms a key landmark for the standard plane of measurement of the fetal abdominal dimensions, used in all standard algorithms to estimate fetal mass.
The sonolucency is caused by the presence of amniotic fluid within the lumen of the stomach. Its appearance implies that functional fetal renal and urinary systems have produced amniotic fluid, that central nervous and neuromuscular pathways have enabled fetal swallowing, and that a patent conduit exists between the mouth and the stomach (albeit via the tracheo-oesophageal fistula in most cases of oesophageal atresia).
However, the sporadic nature of fetal swallowing may result in transient non-visualisation of the fetal stomach in the absence of underlying pathology. A large, rounded, full stomach will be seen immediately after the ingestion of amniotic fluid, whereas, if no fluid has been swallowed for some time, the stomach may not be seen or appear ‘collapsed’. It is therefore important to reassess the stomach should it not be seen initially, as it will usually become visible over time. Pretorius et al2 repeated ultrasonography at 30–60 min and 1 week. Both Millener et al3 and McKenna et al4 used a 45 min interval scan. Our patients were asked to wait at least an hour before a rescan. If the stomach was still not seen, then another appointment was scheduled within 10 working days.
All studies distinguished between the presence and absence of the fetal stomach. However, McKenna included a group of ‘small’ fetal stomachs. Although nomograms of fetal stomach dimensions have been published,5 6 there is good evidence that it is not possible to arrive at reliable or reproducible biometry because of the dynamic nature of the stomach, so we did not include fetuses with ‘small stomach’ in this analysis.7
Assessment of liquor volume correlated to some extent with underlying diagnosis. Each member of the an/oligohydramnios subgroup had postnatal abnormalities, most commonly renal, although one only had mild intrauterine growth retardation. Polyhydramnios prompted suspicion of a structural gastro-intestinal defect, as the association between polyhydramnios, absent stomach bubble and isolated oesophageal atresia has been well described.8 All cases of oesophageal atresia in our series, whether isolated or associated with tracheo-oesophageal fistula or VACTERL syndrome, were indeed associated with increased liquor volume, but these eight cases represented only a third of the polyhydramnios subgroup.
The proportion of aneuploid cases in our series (30%) was comparable to that seen by McKenna et al.4 In their series of 27 cases of persistent non-visualisation, they identified eight aneuploid fetuses (38%). Millener et al3 (n=31) described one infant with Down syndrome but did not report karyotyping uptake or results. Similarly, Pretorius et al2 (n=19) only mention one case of triploidy. Measurement of liquor volume did not assist in identification of these aneuploid fetuses in our series, as they constituted 20% of the an/oligohydramnios group, 25% of the polyhydramnios group and 30% of the group with normal liquor volume.
We advocate referral to a fetal medicine centre for detailed assessment of the fetal anatomy and liquor volume, karyotyping and multidisciplinary input, as appropriate, by paediatric surgeons, neonatologists and paediatric radiologists (figure 5). Where no other abnormality is identified, plans should be made for postnatal paediatric assessment to exclude underlying pathology before feeding is started.
Proposed management pathway.
Footnotes
-
Competing interests None.
-
Ethics approval This study was conducted with the approval of the GOSH Clinical Ethics Committee.
-
Provenance and peer review Not commissioned; externally peer reviewed.