There is no consensus on home oxygen therapy in infancy. We hypothesised that this might lead to considerable variability in the practice of prescribing home oxygen to infants. To assess this variability, a structured questionnaire was sent to all departments of Paediatrics in Germany (n = 380). Response rate was 92% (n = 349). Indications were mostly (86%) based on long-term recordings of pulse oximeter saturation (S(P)O(2)). There was, however, a wide range (85-94%, mean 90%) of S(P)O(2)-values below which oxygen therapy was considered indicated. Hospitals more experienced with this therapy tended to prescribe oxygen at S(P)O(2) values closer to the physiological range. Similarly, S(P)O(2) values aimed for during therapy ranged from 86 to 100%, with hospitals prescribing home oxygen more frequently aiming for higher values. These differences would argue for concentrating home oxygen therapy to centers with the largest experience.