Cerebral oxygenation and desaturations in preterm infants – a longitudinal data analysis

Abstract

Hypoxemic episodes commonly occur in very preterm infants and may be associated with several adverse effects. Cerebral tissue oxygen saturation (StO2) as measured by near infrared spectroscopy (NIRS) may be a useful measure to assess brain oxygenation. However, knowledge on variability of StO2 is limited in preterm infants at this time, so StO2 dependency on arterial oxygenation (SpO2) and heart rate (HR) was assessed in preterm infants using statistical methods of time series analysis. StO2, SpO2, and HR were recorded from 15 preterm infants every 2 seconds for six hours. Statistical methods of time series and longitudinal data analysis were applied to the data. The mean StO2 level was found as 72% (95% confidence interval (CI) 55.5% -85.5%) based on a moving average process with a 5 minute order. Accordingly, longitudinal SpO2 measurements showed a mean level of 91% (95% CI 69% -98%). Generally, compensation strategies to cope with both StO2 and SpO2 desaturations were observed in the studied patients. SpO2 had a significant effect on cerebral oxygenation (p < 0.001), but HR did not, which led to inconclusive results considering different time intervals. In infants with intermittent hypoxemia and bradycardia, we found a mean StO2 level of 72% and a strong correlation with SpO2. We observed large differences between individuals in the ability to maintain StO2 at a stable level.