Hylke Salverda

18 Chapter 1 Aim and outline of this thesis The general aim of this thesis was to evaluate the effects on outcome after automated oxygen control for preterm infants. This thesis aims to: describe currently available automated oxygen control algorithms and what to expect when they are used (Part II); compare effectiveness of automated oxygen control algorithms on oxygenation in the NICU (Part III); and investigate clinical and long-term outcome after using automated oxygen controllers (Part IV). This thesis comprises of observational studies and a randomised clinical trial. Part II consists of Chapter 1 in which an overview of approaches for algorithm design are described, after which the details on six commercially available oxygen control algorithms are set out. Per algorithm an outline follows on how the algorithm works, and what clinical effects were reported. In this narrative review we conclude that although all available controllers seem to improve time within target range and have a beneficial effect on the occurrence of hypoxia and hyperoxia, the most effective strategy is unknown, as available clinical studies were heterogenous. Part III reports on the effects of automated oxygen control on oxygenation. The NICU of the LUMC was the first to implement automated oxygen titration as standard of care. As of August 2015, all infants requiring respiratory support with supplemental oxygen received automated oxygen titration by the CLiO2 algorithm built into the AVEA ventilator (Vyaire, Yorba Linda, California, USA). In November 2018, all AVEA ventilators were replaced with SLE6000 ventilators (SLE Limited, South Croydon, UK), employing the OxyGenie algorithm for automated oxygen control. This led to a unique setting in which caregivers were competent to handle both ventilators allowing for a comparative, randomised, crossover trial. In Chapter 2 the results of this randomised crossover trial are presented. The effectiveness of an automated oxygen controller may vary depending on the postnatal age of the infant. The condition of the lungs and frequency of apnoea can change markedly during the course of admission. With the exception of one study, all studies report on achieved target range times while using automated oxygen control in an experimental setting and for a short period of time (maximum 24 hours). In Chapter 3 the achieved times within certain SpO2 ranges from birth to 32 weeks of postmenstrual age are compared when either using the CLiO2 or the OxyGenie controller as standard of care. Contemporary patient data management systems for intensive care may be limited to storing vital parameters once per minute, which is not sufficient to register all variation in vital parameters such as SpO2 and FiO2. In Chapter 4 we investigate whether one-per-minute data can be used to perform