Oxygenation saturation index in neonatal hypoxemic respiratory failure

Background This study aimed to assess the validity of the oxygenation saturation index (OSI) and the ratio of oxygen saturation to the fraction of inspired oxygen (FIO2) (S/F ratio) with percutaneous oxygen saturation (OSISpO2 and the Sp/F ratio) and to evaluate the correlation between these values and the oxygen index (OI). It also determined their cut‐off values for predicting OI in accordance with neonatal hypoxic respiratory failure severity. Methods We reviewed the data of 77 neonates (gestational age 31.7 ± 6.1 weeks; birthweight, 1768 ± 983 g) requiring invasive mechanical ventilation between 2013 and 2020, 1233 arterial blood gas samples in total. We calculated the OI, OSISpO2, OSI with arterial oxygen saturation (SaO2) (OSISaO2), Sp/F ratio, and the ratio of SaO2 to FIO2 (Sa/F ratio). Results The regression and Bland–Altman analysis showed good agreement between OSISpO2 or the Sp/F ratio and OSISaO2 or the Sa/F ratio. Although a significant positive correlation was found between OSISpO2 and OI, OSISpO2 was overestimated in SpO2 > 98% with a higher slope of the fitted regression line than that below 98% of SpO2. Furthermore, receiver‐operating characteristic curve analysis using only SpO2 ≤ 98% samples showed that the optimal cut‐off points of OSISpO2 and the Sp/F ratio for predicting OI were: OI 5, 3.0 and 332; OI 10, 5.3 and 231; OI 15, 7.7 and 108; OI 20, 11.0 and 149; and OI 25, 17.1 and 103, respectively. Conclusion The cut‐off OSISpO2 and Sp/F ratio values could allow continuous monitoring for oxygenation changes in neonates with the potential for wider clinical applications.