Assessment of risk factors for cerebral oxygen desaturation during neonatal and infant general anesthesia: an observational, prospective study

Cerebral oxygen saturation (rSO₂c) decrease from baseline greater than 20 % during infant cardiac surgery was associated with postoperative neurologic changes and neurodevelopmental impairment at 1 year of age. So far, there is no sufficient evidence to support the routine monitoring of rSO₂c during general surgical procedures in children. We aimed to find out the frequency of cerebral desaturation 20 % or more from baseline and to identify possible predictors of change in cerebral oxygen saturation during neonatal and infant general surgery.

Forty-four infants up to 3 months of age were recruited. Before induction of anesthesia, two pediatric cerebral sensors were placed bilaterally to the forehead region and monitoring of regional cerebral saturation of oxygen was started and continued throughout the surgery. Simultaneously, mean arterial blood pressure (MAP), pulse oximetry (SpO₂), heart rate (HR), endtidal CO₂, expired fraction of sevoflurane and rectal temperature were recorded. The main outcome measure was rSO₂c value drop-off ≥20 % from baseline. Mann-Whitney U-test, chi-squared test, simple and multiple linear regression models were used for statistical analysis.

Forty-three infants were analyzed. Drop-off ≥20 % in rSO₂c from baseline occurred in 8 (18.6 %) patients. There were no differences in basal rSO₂c, SpO₂, HR, endtidal CO₂, expired fraction of sevoflurane and rectal temperature between patients with and without desaturation 20 % or more from baseline. But the two groups differed with regard to gestation, preoperative mechanical ventilation and the use of vasoactive medications and red blood cell transfusions during surgery. Simple linear regression model showed, that gestation, age, preoperative mechanical ventilation and mean arterial pressure corresponding to minimal rSO₂c value during anesthesia (MAP_minrSO2c) were associated with a change in rSO₂c values. Multiple regression model including all above mentioned variables, revealed that only MAP_minrSO2c was predictive for a change in rSO₂c values (β (95 % confidence interval) -0.28 (−0.52–(−0.04)) p = 0.02).

Cerebral oxygen desaturation ≥20 % from baseline occurred in almost one fifth of patients. Although different perioperative factors can predispose to cerebral oxygenation changes, arterial blood pressure seems to be the most important. Gestation as another possible risk factor needs further investigation.

The international registration number NCT02423369. Retrospectively registered on April 2015.