Using near-infrared spectroscopy to measure cerebral metabolic rate of oxygen under multiple levels of arterial oxygenation in piglets.

Improving neurological care of neonates has been impeded by the absence of suitable techniques for measuring cerebral hemodynamics and energy metabolism at the bedside. Currently, near-infrared spectroscopy (NIRS) appears to be the technology best suited to fill this gap, and techniques have been proposed to measure both cerebral blood flow (CBF) and cerebral metabolic rate of oxygen (CMRO2). We have developed a fast and reliable bolus-tracking method of determining CMRO2 that combines measurements of CBF and cerebral venous oxygenation [venous oxygen saturation (CSvO2)]. However, this method has never been validated at different levels of arterial oxygenation [arterial oxygen saturation (SaO2)], which can be highly variable in the clinical setting. In this study, NIRS measurements of CBF, CSvO2, and CMRO2 were obtained over a range of SaO2 in newborn piglets (n=12); CSvO2 values measured directly from sagittal sinus blood samples were collected for validation. Two alternative NIRS methods that measure CSvO2 by manipulating venous oxygenation (i.e., head tilt and partial venous occlusion methods) were also employed for comparison. Statistically significant correlations were found between each NIRS technique and sagittal sinus blood oxygenation (P<0.05). Correlation slopes were 1.03 (r=0.91), 0.73 (r=0.73), and 0.73 (r=0.81) for the bolus-tracking, head tilt, and partial venous occlusion methods, respectively. The bolus-tracking technique displayed the best correlation under hyperoxic (SaO2=99.9±0.03%) and normoxic (SaO2=86.9±6.6%) conditions and was comparable to the other techniques under hypoxic conditions (SaO2=40.7±9.9%). The reduced precision of the bolus-tracking method under hypoxia was attributed to errors in CSvO2 measurement that were magnified at low SaO2 levels. In conclusion, the bolus-tracking technique of measuring CSvO2, and therefore CMRO2, is accurate and robust for an SaO2>50% but provides reduced accuracy under more severe hypoxic levels.