OBJECTIVE: The aim of the study was to determine whether infants weighing ≤1000 g after birth who are born to women who undergo indicated preterm delivery have different neonatal outcomes than do those born as a result of either spontaneous preterm labor or preterm premature rupture of membranes. STUDY DESIGN: In a 1-year observational study (1992-1993] the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network collected outcome data for 799 infants whose birth weights were ≤1000 g. Only singleton infants with gestational age >20 weeks who were not produced as the result of an induced abortion were included. Our analysis was further limited to infants without major congenital anomalies who survived >2 days, were deemed potentially viable by the obstetrician, and would have undergone a cesarean delivery for fetal indications (N = 411). The primary reason for delivery was categorized as indicated delivery, spontaneous preterm labor, or spontaneous preterm premature rupture of membranes. Selected neonatal outcomes were evaluated among infants born to women in each of these groups. Logistic regression analyses were used to control for the effects of other potentially confounding variables. RESULTS: A total of 156 of the 411 infants were born to women who underwent an indicated preterm delivery, whereas 160 were born after spontaneous preterm labor and 95 were delivered after preterm premature rupture of membranes. Univariate analyses revealed significantly lower incidences of grade II or IV intraventricular hemorrhage, grade III or IV retinopathy of prematurity, and seizure activity among infants born in an indicated preterm delivery than among those born after spontaneous preterm labor or preterm premature rupture of membranes. However, infants of women who underwent indicated preterm delivery had a more advanced mean gestational age at birth than did those born after spontaneous preterm labor or preterm premature rupture of membranes (28 ± 2 weeks, 26 ± 2 weeks, and 26 ± 1 weeks, respectively, P < .001). Multiple logistic regression analysis was therefore used to control for the disparity in gestational age. Multivariate analyses did not confirm the apparent improvement in neonatal outcome in the indicated delivery group. CONCLUSION: In this population of infants weighing ≤1000 g, selected neonatal outcomes did not differ according to birth by indicated preterm delivery, spontaneous preterm labor, or preterm premature rupture of membranes.