Prenatal Ultrasound Diagnosis of Gastrointestinal Malformations By Simon Phelps, Ross Fisher, Ann Partington, and Evelyn Dykes

London, England

l Prenatal ultrasonography is commonly used to detect fetal structural abnormalities. The aim of this study was to evaluate the reliability of fetal sonography in the detection of congenital gastrointestinal malformations (GIM). From a prospective database of all major structural fetal anomalies detected by ultrasound over 3 years in a region with a population of 3.5 million and an annual birth rate of 52,000, 294 reports of GIM were identified and reviewed. There were 220 confirmed cases of GIM, of which only 35 (16%) had been correctly identified prenatally. Of 84 prenatal diagnoses of GIM, only 35 (42%) were confirmed postnatally. Prenatal ultrasound was most reliable in the detection of duodenal obstruction (55% confirmed cases identified prenatally) and least reliable in malformations of the hindgut. Of 44 cases of nonspecific sonographic bowel abnormality (eg, echogenic bowel) with adequate follow-up, only 12 (27%) had a con- firmed GIM after delivery. This study demonstrates that the prenatal ultrasound scan is still unreliable in the detection or exclusion of fetal GIM, and consequently clinicians involved in prenatal sonography or counselling should exercise cau- tion in making such diagnoses. Copyright o 1997 by W.B. Saunders Company

INDEX WORDS: Prenatal diagnosis, ultrasonography, gastro- intestinal malformation.

T THE PRENATAL ULTRASOUND scan is com- monly used to detect fetal structural abnormalities.

The resulting diagnoses may have a profound effect on parental counselling, fetal outcome, and perinatal manage- ment. The aim of this study was to evaluate the reliability of prenatal ultrasound scan in the detection of congenital gastrointestinal malformations.

MATERIALS AND METHODS

A prospective database (Congenital Malformation Registry) of all major fetal structural abnormalities identified through prenatal ultra- sound or clinically after delivery was established in the South East Thames region of England in 1992. This region has a population of 3.5 million and an annual birth rate of 52,000. The Registry is notified of all

From the South East Thames Congenital Malformation Registry, Children’s Hospital Lewishum, London, England.

Presented at the 43rd Annual International Congress of the British Association of Paediatric Surgeons, St Heliel; Jersey, Channel Islands, July 16-19, 1996.

The South East Thames Congenital Malformation Registp is funded by the South Thames Regional Health Authority.

The opinions expressed in this paper are those of the authors and do not necessarily represent the views of the Registry Steering Group or South Thames Regional Health Authority.

Address reprint requests to E. Dykes, FRCS, Senior Lecturer in Paediatric Surgery, Children’s Hospital Lewisham, London SE13 6LH, England.

Copyright o 1997 by WB. Saunders Company 0022-3468/97/3203-0014$03.00/O

438

malformations identified by prenatal sonographers, obstetricians, labour wards, pathologists, geneticists, paediatricians, and paediatric surgeons from the 17 maternity and children’s departments in the region. Once notified of an abnormality, the registry coordinator actively collects data on prenatal treatment and postnatal outcome up to 1 year of age.

For the purpose of this study, all diagnoses of gastrointestinal malformation (GIM) reported over a 3-year period (1992 to 1995) were reviewed and analysed to establish the proportion of prenatal diagnoses that were confirmed postnatally and the proportion of confirmed malformations that had been correctly identified prenatally.

RESULTS

Of the 3,975 diagnoses of fetal malformation that were made in 3,146 pregnancies during the 3-year period under review, 294 (7.4%) were related to the digestive system including 25 hepatobiliary anomalies and 269 GIM. There were 220 confirmed cases of GIM, only 35 (16%) of which had been correctly identified prenatally (Table 1). Of 84 prenatal diagnoses of fetal GIM, only 3.5 (42%) were subsequently confirmed postnatally (Table 1).

The prenatal ultrasound scan was most reliable in the detection of duodenal obstruction (79% prenatal diag- noses confirmed, 55% proven cases diagnosed prena- tally) and least reliable in abnormalities of the hindgut (83% prenatal diagnoses false-positive and less than 1% of confirmed malformations identified prenatally). Of the 44 fetuses reported prenatally to have abnormally echo- genie bowel, 24 were normal at delivery, and 12 (27%) had postnatal problems; 8 were found to have small intestinal obstructions (only 1 of which was meconium ileus), and 4 had anorectal malformations. In 8 cases no follow-up or postmortem data were available.

Four of eight (50%) prenatal diagnoses of hepatobili- ary abnormalities were confirmed postnatally, but only four of 21 cases identified postnatally had been diagnosed prenatally (19%).

DISCUSSION

The widespread use of prenatal ultrasonography for the detection of fetal abnormality and the increasing number of anomalies detected in this way has led to a public (and occasionally professional) perception that the technique is both sensitive and specific. Those involved in prenatal scanning and postnatal management are aware of its limitations, but the relatively small numbers of cases seen by individual institutions makes it difficult to assess the true reliability of prenatal ultrasound scan for different anomalies.

This study of GIMs demonstrates the results of prena- tal ultrasound scan as practised in a large region with multiple antenatal centres and several tertiary paediatric

Journal ofpediatric Surgery, Vol32, No 3 (March), 1997: pp 438-440

PRENATAL DIAGNOSIS OF GUT ABNORMALITIES 439

Table 1. Accuracy of Prenatal Ultrasound Diagnosis in Congenital Gastrointestinal Malformations

No. Prenatally Total No. of No. of Correct Anomaly Confirmed Cases Diagnosed (%) Prenatal Diagnoses Prenatal Diagnoses (%)

Foregut

Oesophageal atresia Duodenal obstruction

Midgut Small bowel atresia

Malrotation Meckel’s diverticulum

Other small bowel

Hindgut Colonic atresia

Anorectal anomaly Hirschprung’s disease

Nonspecific

Hyperechogenic bowel Hepato-biliary

43 20

13 3 (23) 24 0 11 0

3 2 (67)

4 0

72 0 20 1 (5)

21

6 (14) 11 6 (55) 11 (55) 14 11 (79)

- 44 (36 followed up) 12 (27-33)

4 (19) 8 4 (50)

2 0 0 0

4 1 (25)

3 (50) 0

0

2 (67)

referral units. The results confirm that GIMs comprise a relatively small proportion of all fetal anomalies (less than 10%) and that sonography is currently unreliable in the detection or exclusion of most congenital malforma- tions of the gut, both in terms of its positive predictive value (prenatally diagnosed cases confirmed postnatally) and false- negative rate (confirmed cases missed prenatally).

There are clearly significant variations in sonographic reliability depending on the site of the anomaly and the gestational age at the time of the scan.’ Perhaps not surprisingly, duodenal and high intestinal obstructions are more readily diagnosed and excluded, even early in gestation.2,3 The ability of prenatal ultrasound scan to detect oesophageal atresia is heavily influenced by the presence or absence of a trachea-oesophageal fistula because if present, the fistula allows some amniotic fluid to enter the stomach, thereby disguising the most promi- nent sonographic sign of oesophageal obstruction.4 The majority of “missed” cases of oesophageal atresia in this series had a fistula; conversely, five of the six correct prenatal diagnoses were in pure oesophageal atresia.

Midgut abnormalities are uncommon and often diffi- cult to identify, except in severe cases in which volvulus or atresia results in prenatal perforation and gross abdomi- nal signs.2 Pathological processes in other organ systems including tortuous dilated ureters, multicystic dysplastic kid- ney, and intraabdominal cysts may mimic intestinal dilata- tion.3 Abnormalities such as Meckel’s diverticulum can- not realistically be expected to be identified prenatally.

The most commonly reported sonographic abnormali- ties of the midgut are nonspecific, including “echo- genie, ” “thickened,” or “mildly dilated” bowel. The possible association between hyperechogenic bowel and meconium ileus secondary to cystic fibrosis has been widely debated.5 Previous reports have shown fetal pathology including chromosome abnormalities in up to 40% of cases,6-10 but in our series less than one third of

infants had any abnormality after delivery, and only one of these had meconium ileus. It is important to distin- guish the gestational age at which the hyperechogenicity is observed and to distinguish transient from persisting sonographic signs,5 but overall, the low incidence of positive findings does not justify the invasive investiga- tion of asymptomatic infants after delivery.

Hindgut abnormalities are particularly difficult to diag- nose, because although the large bowel becomes visible sonographically as early as 22 weeks, it may not become enormously dilated even in the presence of complete obstruction.3 Nevertheless, it is particularly worrying to observe that several diagnoses of Hirschprung’s disease were made prenatally in this series. Aganglionosis is a histological diagnosis and at best could only have been visualised prenatally as probable large bowel obstruction. The dangers of ascribing a specific diagnosis to nonspe- cific appearances cannot be overemphasised; at best, subsequent parental counselling and perinatal treatment may be significantly distorted, and at worst, inappropriate prenatal interventions may ensue.

A basic knowledge of fetal anatomy and physiology along with an understanding of the limitations of sonogra- phy and an application of common sense should permit clinicians to predict the (usually low) probability of an accurate diagnosis of a GIM, but it is important to ensure that the level of diagnostic uncertainty is communicated honestly to parents and those responsible for postnatal management. It is likely that congenital malformations of the gastrointestinal system will remain amongst the most difficult to identify prenatally for some time to come.

ACKNOWLEDGMENT

The authors are grateful to the many clinicians, sonographers, and midwives in South East Thames who have contributed to this study by reporting patients to the Registry.

440 F‘HELPS ET AL

REFERENCES 1. Roberts AB, Hampton E, Wilson N: Ultrasound detection of fetal

structural abnomralities in Auckland 198889. NZ Med J 106:441-443, 1993

2. Robertson FM, Crombleholme TM, Paidas M, et al: Prenatal diagnosis and management of gastrointestinal anomalies. Semin Perina- to1 18:182-195, 1994

3. Hertzberg BS: Sonography of the fetal gastrointestinal tract: Anatomic variants, diagnostic pitfalls and abnormalities. AJR 162: 1175- 1182,1994

4. Pretorius DH, Drose JA, Dennis MA, et al: Tracheosophageal fistula in utero: Twenty-two cases. J Ultrasound Med 6509-513, 1987

5. Stringer MD, Thornton JG, Mason GC: Hyperechogenic fetal bowel. Arch Dis Child 74:Fl-F2, 1996

6. Dicke JM, Crane JP: Sonographically detected hyperechoic fetal

bowel: Significance and implications for pregnancy management. Obstet Gynecol80:778-782, 1992

7. Bromley B, Doubilet P, Frigoletto FD, et al: Is fetal hyperechoic bowel on second-trimester sonogram an indication for amniocentesis? Obstet Gynecol83:647-651, 1994

8. Nyberg DA, Dubinsky T, Resta RG, et al: Echogenic fetal bowel during the second trimester: Clinical importance. Radiology 188:527- 531,1993

9. Hill LM, Fries J, Hecker J, et al: Second-trimester echogenic small bowel: An increased risk for adverse perinatal outcome. Prenatal Diagn 142345-850, 1994

10. Scosia AL, Pretorius DH, Budorick NE, et al: Second-trimester echogenic bowel and chromosomal abnormalities. Am J Obstet Gyne- co1 167:889-894, 1992