The incidence of ROP is birth weight dependent and restricting therapeutic oxygen levels has dramatically reduced the incidence of ROP in infants of birth weight >1000 g. However, the incidence of ROP has remained high in very low birth weight (VLBW) infants and this appears to be related to these babies being more ill. Several risk factors have been identified in this group, however oxygen variability, rather than high levels, has been correlated with severity of disease in recent clinical and animal studies. Difficulties in defining 'normal' oxygen in this group has meant the optimal range of oxygen therapy has not yet been defined. Clinical studies are now underway using even lower oxygen therapy ranges. The impact this may have on ROP, neurological and respiratory outcomes will require further study.