Previous prognostic studies of infants with intrauterine growth retardation (IUGR) have not adequately considered the heterogeneity of IUGR in terms of cause, severity, and body proportionality and have been prone to misclassification of IUGR because of errors in estimation of gestational age. Based on a cohort of 8719 infants with early-ultrasound-validated gestational ages and indexes of body proportionality standardized for birth weight, the consequences of severity and cause-specific IUGR and proportionality for fetal and neonatal morbidity and mortality were assessed. With progressive severity of IUGR, there were significant (all P less than .001) linear trends for increasing risks of stillbirth, fetal distress (abnormal electronic fetal heart tracings)O during parturition, neonatal hypoglycemia (minimum plasma glucose less than 40 mg/dL), hypocalcemia (minimum Ca less than 7 mg/dL), polycythemia (maximum capillary hemoglobin greater than or equal to 21 g/dL), severe depression at birth (manual ventilation greater than 3 minutes), 1-minute and 5-minute Apgar scores less than or equal to 6, 1-minute Apgar score less than or equal to 3, and in-hospital death. These trends persisted for the more common outcomes even after restriction to term (37 to 42 weeks) births. There was no convincing evidence that outcome among infants with a given degree of growth retardation varied as a function of cause of that growth retardation. Among infants with IUGR, increased length-for-weight had significant crude associations with hypoglycemia and polycythemia, but these associations disappeared after adjustment for severity of growth retardation and gestational age.(ABSTRACT TRUNCATED AT 250 WORDS)