A single course of corticosteroids given to women before an anticipated preterm birth reduces the incidence of respiratory distress syndrome, intraventricular hemorrhage, and neonatal death. For women who do not deliver within 1 week, repeated courses of corticosteroids have become common obstetric practice, despite little evidence of efficacy. Emerging data suggest little benefit and potential harm from that practice, so corticosteroids to improve perinatal outcomes should be restricted to a single course unless future randomized trial data prove additional courses to be beneficial.