Hylke Salverda

73 4 Comparison of two AOCs in oxygen targeting in preterm infants during admission Modes of respiratory support used during both epochs were invasive mechanical ventilation (volume targeted, high frequency oscillation (HFO)), continuous positive airway pressure or non-invasive positive pressure ventilation, high flow nasal cannula (HFNC), and low flow. The SLE6000 supports all these modes, whereas during the AVEA epoch HFNC was administered via the Optiflow system (Fisher & Paykel Healthcare, Auckland, New Zealand) and HFO via a Sensormedics 3100 A ventilator (Vyaire, Yorba Linda, California, USA). Both the AVEA and the SLE6000 ventilator have Masimo SET technology on board to measure SpO2, but with different fixed settings for averaging time during AOC (SLE6000: 2-4 seconds, AVEA: 8 seconds). Automated oxygen control was not available during HFNC or HFO support during the AVEA period, and these periods were excluded in the primary analysis for both epochs. The SpO2 target range during CLiO2 control was 90%-95%, which needed to be changed to 91%-95% on introduction of the SLE6000 as this is a pre-set target range. For the purpose of the primary analysis the target range was considered to be 91%-95%. Local protocol is to disable automated oxygen control when infants received no supplemental oxygen while saturating >98% for more than 30 minutes or a few days prior to transfer to a different hospital. Our local protocol is to set SpO2 alarm limits to 88%-98%. Data collection and analysis Patient characteristics and vital parameters up to a postmenstrual age of 30 weeks were sourced from our patient data management system (PDMS Metavision; IMDsoft, Tel Aviv, Israel). The instantaneous SpO2 and FiO2 were stored once per minute, we recently demonstrated that there were no significant differences when using one-per-minute vs. one-per-second data for descriptive statistics such as time within target range.16 Small for gestational age was defined as a birth weight under p10 in the Hoftiezer curves.17 There is an incongruency in the incoming FiO2 values between the SLE6000 and AVEA. In case of the SLE6000, data forwarded to our patient data management system consists of measured FiO2 by the SLE6000’s oxygen cell. For this cell the accuracy is 3%.18 In some cases this leads to situations where the ventilator is providing no supplemental oxygen, or 21% of oxygen, while the recorded FiO2 is 23%. Contrary to the SLE6000, AVEA’s recorded values for FiO2 are based on the intended, or set, FiO2 rather than the measured FiO2. Therefore we chose to define room air, or no supplemental oxygen, as any FiO2 value of 23% or lower.