24-hour oxygen saturation recordings in preterm infants: editing artefact

To create editing guidelines for artefact removal in preterm infant pulse oximetry recordings.

38 preterm infants ready for discharge home from the neonatal intensive care unit underwent 24-hour pulse oximetry recording using the Masimo^® Rad-8 device. An expert panel determined editing rules based on clinical protocols. For each recording, three reports were generated, 'raw' no editing, 'auto' using the software editing feature and 'manual' reviewed and edited according to the rules. Primary outcome measures were desaturation indices including desaturation index 3% and 4%. Secondary measures included heart rate, mean oxygen saturation and time below 90%.

While all oximetry outcomes differed significantly between editing modes, the majority were not considered likely to influence clinical management. Use of the auto editing compared to no editing did alter by >5%: Time spent <90% oxygen saturation and Desaturation index 4% >10 seconds. The use of manual editing removed extremely low pulse values that were considered unphysiological in this group of otherwise healthy infants.

We recommend that oximetry recordings to determine cardiorespiratory stability in newborn infants ready for discharge from the neonatal unit have software editing features applied. This will remove artefact without consuming time in a busy unit.