Hylke Salverda

161 9 General discussion and future perspectives term clinical outcomes were not significantly different between the two groups, and neurodevelopmental outcome at two years is not yet available. The reduction in retinopathy of prematurity is plausible. In this thesis we reported tighter target range adherence (i.e. less fluctuation of oxygenation) and less frequent and shorter episodes of both hypoxaemia and hyperoxaemia while using OxyGenie.39 Hypoxaemia, hyperoxaemia, and fluctuation of oxygenation have all been associated with an increased rate of ROP.34, 40, 41 Early after preterm birth, a varying oxygenation of the retina might lead to decreased retinal vascular growth and blood vessel loss, leaving the retina more susceptible to damage due to hypoxia. In a later phase, this increases the risk of uncontrolled neovascularisation and retinal detachment.35 Less frequent and shorter episodes of hyperoxaemia during OxyGenie control may also contribute to the reduction in ROP. We did not have data on cardiotonic medication, but the other known risk factors (postnatal steroids, sepsis, NEC and mechanical ventilation > 3 days) for ROP were not different between cohorts. Limitations Although there are currently no alternatives available, the use of a pulse oximeter is a limitation when performing studies to measure the effect of tighter control of oxygenation. A proxy for oxygenation status, oxygen saturation measured with pulse oximetry (SpO2), is used in chapter 3, chapter 4, and chapter 5 for continuous non-invasive monitoring of oxygenation, but is limited in accuracy.42-44 The FiO 2SpO2 relationship shows substantial intra-subject variability in the change of the infants’ SpO2 following an adjustment in FiO2. 10 Many factors will influence the SpO 2 response, including for example, the changes in the oxygen-dissociation relationship during transition from foetal to adult haemoglobin. This shift will be quite pronounced in preterm infants who receive transfusions of adult blood. In chapter 3, chapter 4, and chapter 8 we compared two ventilators rather than purely the effect of the AOC algorithms on outcome. It is possible that ventilator mechanics also played a role in the effectiveness of oxygen control, as well as other aspects of ventilator function including the circuit flow characteristics.45 However, this was a pragmatic choice as license agreements precluded us from implementing two algorithms in one ventilator. Severalmeasureswere taken tominimise the risk of bias associatedwith retrospective chart studies as reported in chapter 4, chapter 6, chapter 7, and chapter 8. For all respiratory support data we used automatically stored data in our patient data