Apnoea-triggered increase in fraction of inspired oxygen in preterm infants: a randomised cross-over study

Objectives To investigate the impact of a pre-emptive apnoea triggered oxygen response on oxygen saturation (SpO(2)) targeting following central apnoea in preterm infants. Design Interventional crossover study of a 12-hour period of automated oxygen control with an apnoea response (AR) module, nested within a crossover study of a 24-hour period of automated oxygen control compared with aggregated data from two flanking 12-hour periods of manual control. Setting Neonatal intensive care unit Patients Preterm infants receiving non-invasive respiratory support and supplemental oxygen; median (IQR) birth gestation 27 (26-28) weeks, postnatal age 17 (12-23) days. Intervention Automated oxygen titration with an automated control algorithm modified to include an AR module. Alterations to inspired oxygen concentration (FiO(2)) were actuated by a motorised blender. Desired SpO(2) range was 90-94%. Apnoea detection was by capsule pneumography. Main outcome measures Duration, magnitude and area under the curve (AUC) of SpO(2) deviations following apnoea; frequency and duration of apnoeic events. Comparisons between periods of manual, automated and automated control with AR module. Results In 60 studies in 35 infants, inclusion of the AR module significantly reduced AUC for SpO(2) deviations below baseline compared with both automated and manual control (manual: 87.1%+/- 107.6% s, automated: 84.6%+/- 102.8% s, AR module: 79.4%+/- 102.7% s). However, there was a coincident increase in SpO(2) overshoot (AUC (SpO(2)>SpO(2(onset))); manual: 44.3 +/- 99.9% s, automated: 54.7%+/- 103.4% s, AR module: 65.7%+/- 126.2% s). Conclusion Automated control with a pre-emptive apnoea-triggered FiO(2) boost resulted in a modest reduction in post-apnoea hypoxaemia, but was followed by a greater SpO(2) overshoot.