BACKGROUND Nasal continuous positive airway pressure (nCPAP) stabilizes the residual volume and may decrease the risk of 'atelectotrauma', potentially promoting lung development in neonates. OBJECTIVE To assess whether replacing nCPAP by low-flow O2 by nasal cannula affects lung function expressed as the arterial/alveolar oxygen tension ratio (a/A pO2 ratio) on postnatal day 28. METHODS Preterm infants (birth weight <1,500 g and gestational age, GA >26 + 0 weeks) stable on nCPAP between postnatal days 4 and 7 were randomized to nCPAP or low-flow O2 by nasal cannula (<0.2 liters/min). Study criteria defined how to wean/restart respiratory support or change from low-flow O2 to nCPAP and vice versa. Transcutaneous monitoring was used for the assessment of the a/A pO2 ratio on day 28 using a head box for all infants for accurate measurement and to eliminate possible effects from nCPAP or low-flow O2 on oxygen requirement. RESULTS We enrolled 52 infants (nCPAP group n = 30 and low-flow O2 group n = 22). The a/A pO2 ratio at 28 days was 0.43 ± 0.17 (nCPAP group) versus 0.48 ± 0.18 (p = 0.36). The duration of nCPAP was 16.4 (low-flow group) versus 41.1 days (nCPAP group), p < 0.001. There was no difference between groups in the fraction needing any respiratory support at 36 weeks' corrected age, length of stay, weight at discharge, and relative weight gain. CONCLUSIONS Replacing nCPAP by low-flow O2 in preterm infants with GA >26 weeks at the end of the first week of life did not seem to affect the a/A pO2 ratio or weight gain negatively. Thus, prolonged nCPAP seems not to have a positive effect on lung function at 28 days of life and replacement by low-flow O2 could reduce the cost of equipment and increase the ease of nursing.