Chronic lung disease (CLD) remains a common problem in neonatal intensive care units. Corticosteroids are being used increasingly to prevent or treat CLD. Uncertainties remain including the timing, duration and route of administration, and ratio of benefits to costs. This paper reviews the outcome of 39 randomized clinical trials of postnatal corticosteroid treatment. Twenty-five trials studied systemic steroids at three different postnatal ages; early (<96 h), moderately early (7-14 days), and delayed (>3 weeks). Fourteen-trials studied inhaled steroids (early: <7 days; late: >14 days) and inhaled versus systemic steroids. Systemic steroids have short-term beneficial effects improving gas exchange and lung mechanics to facilitate earlier extubation when used at any postnatal age in infants who are ventilator dependent. Early and moderately early steroids also reduce the risk of CLD at both 28 days and 36 weeks. For moderately early steroid use, there is also a reduction in neonatal mortality with 1 extra survivor for approximately every 16 babies treated (95% confidence interval 9-55). Proven adverse effects are either gastro-intestinal (bleeding) or metabolic (hyperglycaemia and hypertension). Unproven but potential adverse effects include decreased brain and lung growth. Inhaled steroids have been studied less thoroughly. Some studies report improvement in gas exchange and lung mechanics, but long-term benefits are not apparent to date in the published material. Trials comparing inhaled and systemic steroids suggest a more rapid response with the latter, but no significant differences in the rate of CLD at either 28 days or 36 weeks were found. Further research is necessary to define the roles of systemic and inhaled steroids in the prevention and treatment of CLD and to allow comparison of benefits to costs.