Introduction Pulse oximetry is an easily available, non- invasive continuous method for assessment and management of oxygenation. However, previous studies have reported discrepancies between pulse (SpO₂) and arterial hemoximetry saturations (SaO₂) with SpO₂ overestimating SaO₂ especially at low saturations. Whether new generation SpO₂ monitors are less prone to such discrepancies is not known.

Aim To prospectively assess correlation between SpO₂ and SaO₂ in term and preterm infants.

Methods Infants admitted to NICU over a period of 11 months, with an arterial line and SpO₂ monitoring were eligible for this study. Infants on iNO were excluded. During routine arterial blood gas analysis, measured SaO₂ was compared with stable SpO₂ value just prior to initiating blood gas draw. Data was analysed using Bland-Altman method to assess measurement agreement.

Results 221 paired measurements were obtained from 36 neonates. 77.8% were fit for analysis. Of these, 6.4% SpO₂ values were between 85–89%, 42.4% between 90–95%, 47.6% between 96–100% and the rest 3.6% were <85%. On an average, SaO₂ was lower than SpO₂ with a mean difference of -1.056. The 95% limits of agreement were –7.216 to 5.014 (SD =3.08). For the SpO₂ categories between 85–89%, 90–95%, 96–100%; 73%, 53% and 73% respectively correlated with the SaO₂ measurements. However, SaO₂ was lower than SpO₂ in 18%, 21% and 27% respectively.

Conclusion Though the study showed that SpO₂ and SaO₂ measurements did not correlate completely, the observed difference was not clinically significant. This could be because of small sample size or due to improved SpO2 technology.