This is a cross-over randomized study. Eligible participants are preterm infants born at less than 37 weeks gestation (23+0 to 36+6 weeks), who are receiving conventional mechanical ventilation through an endotracheal tube and have a need for supplemental oxygen at the time of enrolment. The planned sample size is 19 subjects completing the study with both arms (38 study periods). The objective of this crossover study is to evaluate the efficacy of the automatic oxygen control function with or without Volume Guarantee®(automatic control of ventilator pressure to deliver the set volume) mode of ventilation in keeping oxygen levels in the safe target range (90 to 95%) in ventilated preterm infants requiring oxygen therapy.

Supplemental oxygen remains by far the most commonly used 'drug' in neonatal intensive care units. The goal of oxygen therapy is to maintain normal oxygenation while minimizing hyperoxaemia and hypoxemia. Preterm infants are particularly vulnerable to oxygen toxicity and oxidative stress leading to retinopathy of prematurity (ROP), bronchopulmonary dysplasia (BPD), and periventricular leukomalacia (PVL)[1]. Similarly, exposure to hypoxemic episodes can lead to increased mortality [2, 3]. Traditionally oxygen saturation (SpO2) targeting is carried out by manual adjustment of fraction of inspired oxygen (FiO2) by the caregiver. However, in practice this is only partially achieved during routine care[4]. Hagadorn et al conducted a study in 14 centres and showed that preterm infants under 28 weeks' gestation receiving oxygen spent on average only 48% of the time with SpO2 within the prescribed target range, about 36% of the time above and 16% of the time with SpO2 below the target range There was a wide and significant variation in SpO2 target range compliance between the participating centres[5].

Preterm infants have frequent fluctuations in SpO2 due to their respiratory instability requiring frequent adjustments of FiO2 [6]. Consequently, these particularly vulnerable infants spend significant time with SpO2 outside intended range and are often exposed to extremes of hypoxemia and hyperoxaemia. To this end, it is now possible to have automated control of inspired oxygen using a device (CLiO2™) incorporated in Avea® ventilator. The safety, feasibility and efficacy of this device have already been established [7-12]. There has been further improvement in the algorithm of the pulse oximeter incorporated in Avea® ventilator to achieve a better normative distribution around the median SpO2 value[13]. Automated control of FiO2 significantly improves compliance of oxygen saturation targeting and significantly reduces exposure to hypoxemia as well as hyperoxaemia [7-12, 14].

Another mechanism of respiratory instability and hypoxemia is wide fluctuation in tidal volume in ventilated preterm infants. Volume-targeted modes of ventilation have been used for several years to aim to more accurately control the tidal volumes delivered to ventilated infants. Jain et al showed a reduction in duration of hypoxaemic events when using volume targeted ventilation as compared to pressure controlled ventilation[15]. Avea ventilators can deliver several types of volume-targeted ventilation, including Volume Guarantee® (VG) and volume controlled ventilation (VCV). VCV aims to deliver a set volume of gas irrespective of the lung compliance whereas VG® uses a servo-controlled feedback loop to automatically adjust inspiratory pressures to aim to control tidal volume delivery.

There is no data available currently on whether automatic control algorithm adjustment of inspired oxygen and tidal volume together leads to further improvements in maintaining SpO2 profile within prescribed target range and more importantly to reduce episodes of prolonged hypoxemia and hyperoxaemia in preterm ventilated infants. Hence the investigators propose this study.

The aim of this study is to examine whether automatic control of inspired oxygen and tidal volume together leads to further improvements in maintaining SpO2 profile within the prescribed target range in preterm, ventilated infants.

The objective is to evaluate the efficacy of the automatic oxygen control function with or without VG® (automatic control of ventilator pressure to deliver the set volume) mode of ventilation in keeping oxygen levels in the safe target range (90 to 95%) in ventilated preterm infants requiring oxygen therapy.

This study will be conducted in a tertiary neonatal intensive care unit at the James Cook University Hospital, Middlesbrough, United Kingdom after approval by the Local Research Ethics Committee and the Institutional Review Board.

This study will be completed over 2 consecutive 12-hour periods in randomly assigned sequence of automatic FiO2 control (CLiO2™) with VG® and automatic FiO2 control without VG®.

The Automated FiO2 system CLiO2™ is integral to the Avea® infant ventilator (CareFusion, Yorba Linda, CA) and allows automated FiO2 adjustment aiming to maintain SpO2 within assigned target range using Radical neonatal pulse oximeter (Masimo, Irvine, CA). When first started it adopts the FiO2 previously set by the clinician as the initial 'Baseline FiO2' level. Thereafter, the changes to the FiO2 and their frequency depend on whether SpO2 is below, above or within the target range, the trend in SpO2 and all changes are proportionate to the 'Baseline FiO2' level. FiO2 is reduced in a step wise fashion when SpO2 exceeds the target range. The step decrements in FiO2 result in a gradual reduction in the FiO2. When SpO2 falls below the target range the increments in FiO2 are generally larger and more frequent. These are proportionate to the difference between SpO2 and the target range and the declining or increasing trend in SpO2. The changes in FiO2 generally range between 0.01 and 0.05. Their frequency increases to achieve a faster rate of change of FiO2 and can be as frequent as one per second. The Baseline FiO2 level is gradually adjusted to changes in the infant's needs for FiO2 to keep SpO2 in range. The pulse oximeter default settings are normal sensitivity, 8 second averaging time, 20 second alarm delay, and a tight alarm limit of 90% and 95% SpO2. Masimo neonatal probe (LNOP Neo-L) will be applied to the right wrist whenever possible. In the event of SpO2 signal loss (saturation 'dropouts') the fail-safe mechanism adjusts the backup FiO2 at the median level in the preceding 15 seconds, or to the baseline if higher.

Volume Guarantee ventilation (VG)® is a volume-targeted mode of ventilation aimed at delivering the set tidal volume of gas by automatically adjusting the peak inspiratory pressure (PIP) on a breath-by-breath basis. Theoretically this should minimise variation in tidal volume delivery as lung compliance and the infant's condition changes. This function is achieved by an automated servo-controlled mechanism. The upper PIP limit can be set by the clinician as a safety mechanism. There is lack of data on VG® but it is thought to achieve comparable gas exchange at lower mean PIP levels[15].

Volume-Controlled ventilation (VCV) is a type of 'volume targeted' mode. It aims to target be supported the desired tidal volume by delivering a set volume (chosen by the clinician) irrespective of the underlying lung mechanics. The ventilator will generate whatever peak inspiratory pressure is necessary to deliver this volume. There is constant inspiratory flow pattern (a square flow waveform) and peak volume and inspiratory pressure delivery are achieved at the end of inspiration. During 'control period' of 12 hours without VG, infants will using VCV A/C (assist control) as is our current clinical standard.

Both VG® and VCV can be delivered to infants using the ventilators currently in use in our unit, the AVEA® ventilators (Carefusion, Yorba Linda, CA). All devices and equipment used for infants in either arm of the trial will be the same as those currently used on our unit. They will be used in line with Carefusion AVEA® ventilator systems operator's manual, L2786, revision M, July 2011.

Verbal and written information (in the form of the Participant Information Sheet) about the trial will be offered to parents at the earliest opportunity after the infant has been intubated. It will only be offered after all other information about the medical care of their infant and the progress has been discussed with them and only if they wish to receive trial information at that time. Consent will not be obtained before any infant is born.

After the infant has been intubated and ventilated, parents will be approached for participation consent only after relevant information about their infant's medical care and progress has been discussed. Parents will be offered written and verbal information about the trial. Parents giving consent for their baby to participate in the trial will be asked to sign a written consent form. Infants will only be randomised to either Volume Guarantee® or Volume-Controlled ventilation cross-over periods after obtaining written consent from parents.

If the infants' respiratory care is escalated to non-conventional mode of ventilation for clinical reasons, the time spent in the 12 hour cross over would be determined. If the infant has spent at least 50% of the time in a cross-over, this would be considered acceptable and data used for analysis. If the infant has spent less than 50%, then parents will be approached again to revalidate the consent for the study when the infant is ready for conventional ventilation.

When commencing Volume Guarantee® mode a setting of 5ml/kg will be the initial starting setting, with the option of decreasing to 4ml/kg or increasing to 6ml/kg as the clinical condition dictates. This is standard neonatal practice. Nursing allocation will remain at 1:1 (one nurse caring for 1 intensive care infant) as is our standard clinical practice during the study period. Target SpO2 range of 90 to 95% will be applicable to both study periods as is our current clinical practice. All elective / planned procedures will be performed before starting the study. All 'routine patient care and procedures' such as endotracheal tube suction, chest physiotherapy, oral care, change of position, kangaroo care, insertion of lines / cannula and catheters, blood sampling will be recorded during the study period. All babies will receive a loading dose of caffeine citrate (20mg/kg intravenously) followed by 5mg/kg once daily, as per standard practice.

Statistical Analysis:

Intention-to-treat analysis will apply for both 12-hour periods. Statistical analysis will consist of within-patient comparisons with paired t tests for normally distributed data or nonparametric Wilcoxon signed rank tests. Shapiro-Wilk test for normality will be used. Results will be presented as mean ± standard deviation (SD) or median and interquartile range. P values of less than 0.05 will be considered statistically significant. Descriptive statistics will be used for summarising the questionnaire responses. Stata® data analysis and statistical software version 11, Stata Corp LP Texas, USA will be used for all statistics.

Inclusion Criteria:

Preterm infants less than 37 weeks (23+0 to 36+6 weeks)

• Who are receiving conventional mechanical ventilation through an endotracheal tube.
• On supplemental oxygen at the time of enrolment (Defined as requiring more than 0.21 FiO2 to maintain saturation in the target range).

Exclusion Criteria:

• Infants more than or equal to 37weeks
• Preterm infants with congenital anomalies
• Infants on a non-conventional mode of ventilation Infants on inhaled Nitric Oxide