Chronic lung disease (CLD), defined as chronic oxygen dependency at 36 weeks postmenstrual age, is increasing and associated with chronic respiratory morbidity and high health care utilisation at follow up. Many strategies, tested in randomised trials, have failed to reduce CLD. In contrast, corticosteroids if given systemically within the first two weeks after birth reduce CLD and may also favourably influence survival. Unfortunately, systemically administered corticosteroids have many acute side-effects and adversely affect long term neurodevelopmental outcome. If given by inhalation, corticosteroids have fewer adverse effects, but are less efficacious. Further research is required to accurately identify infants at highest risk of developing CLD, the corticosteroid dosage associated with a positive risk: benefit ratio and preferably a safer and more effective alternative therapy.