Most premature babies require oxygen therapy. There is uncertainty about what oxygen levels are the best. The oxygen levels in the blood are measured using a monitor called a saturation monitor and the oxygen the baby breathes is adjusted to keep the level in a target range. Although there is evidence that lower oxygen levels maybe harmful, it is not known how high they need to be for maximum benefit. Very high levels are also harmful. Saturation monitors are not very good for checking for high oxygen levels. For this a different kind of monitor, called a transcutaneous monitor, is better.

Keeping oxygen levels stable is usually done by nurses adjusting the oxygen levels by hand (manual control). There is also equipment available that can do this automatically (servo control). It is not known which is best.

Studies of automated control have shown that infants spend more time within their intended target oxygen saturation range. These have not included measurements of transcutaneous oxygen.

The investigators aim to show the transcutaneous oxygen levels as well as the oxygen saturation levels when babies have their oxygen adjusted manually or automatically.

Presently oxygen is titrated against saturation (SpO2) by manual adjustment. Automated or servo-control systems have been developed that result in tighter control of SpO2 and more time spent in the intended target range. These systems are already in clinical use. Automated systems produce quite large fluctuations in fraction of inspired oxygen (FiO2) in order to keep SpO2 in range. It is possible that this could result in short periods of high or low oxygen tension (PO2) that are undetectable using saturation monitoring. Studies to date have examined the effects of manual and automated (servo) oxygen targeting on SpO2 but not on transcutaneous oxygen tension (TcPO2).

There is a need to determine the achieved SpO2 and TcPO2 distributions associated with the use of manual and automated control as a first step in planning trials comparing these approaches over many weeks. When this is measured over a small number of hours it is not anticipated that this would have an influence on clinical outcome.

This study is a prospective, single centre, randomised crossover trial of automated (servo) control versus manual oxygen titration. Each infant will act as their own control. Infants born at less than 29 weeks gestation, greater than 48 hour of age and receiving supplementary oxygen will be eligible for inclusion.

The study will be undertaken in the Neonatal Unit at the Simpson Centre for Reproductive Health at the Royal Infirmary of Edinburgh.

Total study time is 12 hours for each infant. Infants will be randomised to commence on either automated (servo) control or manual mode. SpO2 (range 90-95%) will be continuously monitored as per normal standard of care. A second pulse oximetry probe will be place for servo control input.

Additional monitoring will be carried out as shown below:

1. TcPO2 monitoring
2. FiO2 monitoring
3. Heart rate monitoring (used to validate SpO2 readings)
4. Arterial gas sampling (only if conducted by the direct care team as part of the routine care of the infant; no extra blood samples will be taken as part of the study)

In manual mode, all oxygen adjustments will be made by clinical/nursing staff. In automated mode, oxygen will be adjusted by the respiratory support device.

In automated mode (servo control), oxygen adjustments will be made by one of two devices (depending on the clinical requirements of the baby) - the IntellO2 device (IntellO2, Vapotherm) or Leoni plus CLAC (Closed-Loop Automated oxygen Control) ventilator (Leoni plus, Löwenstein Medical).

SpO2 and TcPO2 readings will be downloaded directly from the multiparameter patient monitor. SpO2 will be measured using a Phillips MX500 multiparameter monitor. TcPO2 will be measured using a SenTec Digital Monitoring System with OxiVent sensor. TcPO2 is calculated by dynamic fluorescence quenching which measures oxygen molecules present in the vicinity of a fluorescent dye incorporated within the sensor surface. The sensor is operated at a constant temperature of 43 degrees Celsius. Control of sensor temperature and application duration are designed to meet all applicable standards and this monitoring device is used routinely in many neonatal units.

Inclusion Criteria:

1. Infants born at less than 29 weeks gestation
2. Infants greater than 48 hours of age
3. Infants who are receiving supplementary oxygen
4. Person with parental responsibility able to give consent

Exclusion Criteria:

1. Congenital anomalies that would prevent targeting SpO2 to 90-95% (e.g. cardiac defects)
2. Clinical condition of an infant would impair accurateTcPO2 measurement (e.g. impaired perfusion or requirement of inotropic or vasopressor support)

The study is randomised but not blinded. Infants will be randomised to commence on either automated (servo) control or manual mode, and then cross-over to the alternative range after 6 hours of monitoring (total study time of 12 hours). SpO2 (range 90-95%) will be continuously monitored as per normal standard of care.