OBJECTIVE: To determine if antenatal sonographic findings or selected obstetric factors predict neonatal outcome in cases of fetal gastroschisis. STUDY DESIGN: We retrospectively studied all (N = 59) cases of isolated fetal gastroschisis with prenatal diagnosis and follow-up at our institution from 1986-1996. Obstetric factors included gestational age (GA) at deliven, labor, meconium and oligohydramnios. Abnormal sonographic findings were defined as subjective or objective (15 mm max. lumen diameter) bowel dilatation and subjective bowel obstruction or bowel wall thickening. Indicators of neonatal morbidity included a staged repair (silo) or bowel resecti-jn and the interval from deliver)' to tolerance of full enterai feedings. RESULTS: The mean GA at delivery was 36 ±1.8 wks (range 28-39), and the mean birthweight was 2527 ±546 g (range 1135-3760); all but two patients underwent cesarean. Fourteen babies required a silo, and 5 required bowel resection. The median interval to tolerance of enterai feedings was 24 days (range 8-193). There was no significant relationship between GA at delivery and neonatal outcome. However, infants with oligohydramnios (p = .04), labor (p = .006), and meconium (p = .002) had longer intervals to full enterai feedings; labor was also associated with an increased need for bowel resection (p = .05). In the 47 (80%) women whose last scan was performed 4 wks prior to delivery, the only sonographic abnormality related to morbidity was subjective obstruction, which increased both the need for a silo (p = .05) and the interval to full enterai feedings (p = .03). In a linear regression model, controlling for GA at delivery, the presence of labor still predicted a longer time to enterai feeding (p = .007). CONCLUSIONS: In cases of fetal gastroschisis, oligohydramnios, labor and meconium were associated with increased neonatal morbidity. Of the various sonographic findings, only subjective bowel obstruction predicted neonatal morbidity.