There is considerable uncertainty regarding the oxygen saturation threshold below which additional inspired oxygen should be given to infants with acute or chronic lung disease. In the absence of data from controlled studies, recommendations can only be based on reference values for healthy infants and on observational studies regarding the pathophysiological effects of acute and chronic hypoxia. Reference values for pulse oximeter saturations (SpO2) in term and preterm infants show that during normal breathing 95% of infants maintain SpO2 at or above 93-97%, depending on age. Studies of infants with chronic lung disease (CLD) show that (1) when SpO2 was kept at > or =93% by administration of home oxygen, rates of sudden infant death were reduced; (2) weight gain was significantly better when SpO2 was maintained at > or =93-95%, (3) increasing SpO2 from 82 to 93% by delivering low-flow oxygen resulted in a 50% reduction in pulmonary artery pressure, (4) O2 administration to mildly hypoxemic infants (SPO2 89%) caused a 50% decrease in airway resistance, and (5) low-flow oxygen reduced the frequency of intermittent hypoxemic episodes, even in infants who had values of > or =90% at rest. Based on these data, it is recommended that oxygen therapy should be considered in infants whose baseline SpO2 is <93%, and that SpO2 should be maintained at > or =95% when infants are managed at home.