Abstract We followed the clinical course of 21 infants with bronchopulmonary dysplasia enrolled in a prospective home O2 therapy programme during a 4-year-period. Mean gestational age was 28.5 weeks (range, 2536 weeks) and mean birth weight 1093g (range 6302750 g). Infants were regularly monitored to maintain pulse oximeter O2 saturation over 94%95%. The source of O2 was liquid oxygen and was delivered by nasal cannula. During the follow up oxygenation was assessed by SatO2 measurement, cardiac function by Doppler echocardiography and respiratory function by the occlusion technique. All patients had an ophthalmological follow up. The mean age of the infants at discharge was 3.7 months (range 1.78.6) and mean weight 2830g (range 21503780g). At discharge 8infants had right ventricular hypertrophy (RVH) and four of them had pulmonary hypertension. Mean duration of home O2 therapy was 97 days (range 15320days) and the mean age of discontinuation of O2 was 6.9months (range 314.7 months). The cardiological follow up was benign: the ECG signs of RVH disappeared by 12 months of age in six out of eight infants and the right ventricular pulmonary pressure, as measured by the Doppler method, normalised in the four patients in whom it was detected. No relationship was found between respiratory mechanics and the duration of O2 therapy. Weight gain was poor with mean growth at the 3rd percentile for females and just below the 3rd percentile for males. Twelve of the 21 infants required 25 rehospitalizations. No one presented deterioration of retinopathy of prematurity that was present in 16 infants at discharge; at 12 months retinopathy was resolved in 14 infants. A total of 2025 hospital days were saved, representing a significant financial saving. Conclusion Home O2 therapy permits the safe early discharge of O2-dependent BPD infants and it reduces significantly the length of time spent in hospital which represents a considerable financial saving.