Reliability of pulse oximetry in hypoxic newborn pigs.

OBJECTIVE: The reliability of pulse oximetry in the lower SaO2-ranges has not been fully investigated. We wanted to investigate pulse oximeter performance in a pig model of perinatal asphyxia.

METHODS: Asphyxia was induced in 22 newborn pigs. Pulse oximetry-values for SpO2, as well as SaO2 from blood gas analyses were recorded multiple times before, during and after asphyxiation. The relationship between SpO2 and SaO2 in normoxia and hypoxia in this model was quantified. Calculations were made for 5% increments of SaO2 increasing successively from ≥95% to ≥20%.

RESULTS: The pigs became severely hypoxic, acidotic and hypotensive. The bias in the total data was +13.7 (p < 0.001), i.e. SpO2 was on average higher than SaO2 for the entire interval of SaO2. However, bias was not significantly different from zero in most SpO2-ranges. Accuracy and precision were considerably higher than stated by the manufacturer in all SpO2-ranges, with both accuracy and precision increasing when SpO2-values down to 60% and lower were included.

CONCLUSION: Pulse oximetry was less reliable at SaO2-values below 60%. Tissue acidosis and reduced peripheral perfusion may have contributed to this. It is unknown whether the difference between SpO2 and SaO2 is clinically relevant.