Background: Early-onset fetal growth restriction represents a particular dilemma in clinical management balancing the risk of iatrogenic prematurity with waiting for the fetus to gain more maturity, while being exposed to the risk of intrauterine death or the sequelae of acidosis. Objective: The Trial of Umbilical and Fetal Flow in Europe was a European, multicenter, randomized trial aimed to determine according to which criteria delivery should be triggered in early fetal growth restriction. We present the key findings of the primary and secondary analyses. Study Design: Women with fetal abdominal circumference <10th percentile and umbilical pulsatility index >95th percentile between 26-32 weeks were randomized to 1 of 3 monitoring and delivery protocols. These were: fetal heart rate variability based on computerized cardiotocography; and early or late ductus venosus Doppler changes. A safety net based on fetal heart rate abnormalities or umbilical Doppler changes mandated delivery irrespective of randomized group. The primary outcome was normal neurodevelopmental outcome at 2 years. Results: Among 511 women randomized, 362/503 (72%) had associated hypertensive conditions. In all, 463/503 (92%) of fetuses survived and cerebral palsy occurred in 6/443 (1%) with known outcome. Among all women there was no difference in outcome based on randomized group; however, of survivors, significantly more fetuses randomized to the late ductus venosus group had a normal outcome (133/144; 95%) than those randomized to computerized cardiotocography alone (111/131; 85%). In 118/310 (38%) of babies delivered <32 weeks, the indication was safety-net criteria: 55/106 (52%) in late ductus venosus, 37/99 (37%) in early ductus venosus, and 26/105 (25%) in computerized cardiotocography groups. Higher middle cerebral artery impedance adjusted for gestation was associated with neonatal survival without severe morbidity (odds ratio, 1.24; 95% confidence interval, 1.02–1.52) and infant survival without neurodevelopmental impairment at 2 years (odds ratio, 1.33; 95% confidence interval, 1.03–1.72) although birthweight and gestational age were more important determinants. Conclusion: Perinatal and 2-year outcome was better than expected in all randomized groups. Among survivors, 2-year neurodevelopmental outcome was best in those randomized to delivery based on late ductus venosus changes. Given a high rate of delivery based on the safety-net criteria, deciding delivery based on late ductus venosus changes and abnormal computerized fetal heart rate variability seems prudent. There is no rationale for delivery based on cerebral Doppler changes alone. Of note, most women with early-onset fetal growth restriction develop hypertension.