PM1056: Prenatal Sonography in Suspected Proximal Gastrointestinal Obstructions: Diagnostic Accuracy and Neonatal Outcomes

Background: Although the “double bubble sign” and polyhydramnios are well-described prenatal sonographic findings suggestive of duodenal atresia, there is a paucity of data on their true predictive values for identifying proximal gastrointestinal obstructions (GIO). The purpose of this study was to determine the diagnostic accuracy of fetal ultrasound in suspected cases of proximal GIO and to describe neonatal outcomes.

Methods: After IRB approval, a single institution, retrospective chart review was conducted on all neonates with a prenatally suspected and/or postnatally confirmed proximal GIO (2011-2021). Ultrasound reports and images were queried for dilated stomach with adjacent proximal bowel dilatation (“double bubble”) and polyhydramnios as well as operative reports were queried for proximal GIO. Neonatal outcomes were assessed, and the diagnostic accuracy of fetal sonography was calculated.

Results: Among 53 neonates with suspected or confirmed proximal GIO, there were 27 males (51%) and 26 females (49%). The median birthweight and gestational age at birth were 2550 grams [interquartile range (IQR) 2028-3012] and 37 weeks (IQR 34-38), respectively. Fifteen (29%) had trisomy 21, and 18 (35%) had cardiac anomalies. Three (6%) had documented normal fetal ultrasounds but were diagnosed postnatally with a proximal GIO. Of 44 fetuses with “double bubble”, 36 (82%) were duodenal atresia, 1 (2%) was pyloric atresia, 5 (11%) were jejunal atresias, 1 (2%) was duodenal with jejunoileal, and only 1 (2%) had no obstruction. Double bubble had a sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 84%, 97%, 98%, and 84%, respectively (Table). The presence of both a double bubble and polyhydramnios had a sensitivity, specificity, PPV, and NPV of 68%, 100%, 100%, and 70%, respectively. Among the 46 confirmed duodenal atresias, 16 (35%) had combinations of other alimentary tract anomalies including 8 (17%) annular pancreas, 4 (9%) malrotations, 2 (4%) hepatic anomalies, and 2 (4%) with other intestinal atresias (Figure). The median postoperative length of stay of 27 days (IQR 19-42). Five (9%) patients required a combination of second operations: 3 (6%) were re-explorations (one of which was for ampulla injury concern with an identified choledochal cyst), 2 (4%) had gastrostomy tubes placed, and 1 (2%) had an anal pull-through for simultaneous anal atresia.

Conclusion: In this contemporary series of duodenal atresia, the prenatal double bubble sign shows high diagnostic accuracy, and the combination of double bubble and polyhydramnios is associated with 100% specificity and 100% positive predictive value. Additional important findings at operation were identified or confirmed in 35% of patients. These data are informative for pediatric surgeons and obstetrical sonologists in prenatal counseling and preoperative planning discussions with families given the high likelihood of requiring early postnatal surgical repair in affected fetuses.