Correlation of noninvasive near-infrared spectroscopy with outcomes in pediatric traumatic brain injury.

OBJECTIVE

Pediatric traumatic brain injury (TBI) is a major cause of morbidity and mortality. Preventing cerebral hypoxia and hypoperfusion is crucial. Near-infrared spectroscopy (NIRS), which measures regional cerebral oxygen saturation (rSO2), may provide a noninvasive method of identifying impaired cerebral oxygenation. Whether NIRS correlates with physiological variables or clinical changes following pediatric TBI is unknown, and its relationship with invasive intracranial pressure (ICP), cerebral perfusion pressure (CPP), and brain tissue oxygenation (PbtO2) is poorly defined.

METHODS

This retrospective exploratory study used electronic health records from 2016 to 2022 at a quaternary children's hospital. Patients ≤ 18 years of age with moderate or severe TBI (admission Glasgow Coma Scale [GCS] score ≤ 12) who underwent bilateral frontal NIRS monitoring and intubation for ≥ 48 hours were included. The repeated-measures correlation test, Mann-Whitney U-test, and Kruskal-Wallis test assessed associations and differences between NIRS rSO2 and 1) physiological values (respiratory, hemodynamic, systemic temperature), 2) ICP, CPP, and PbtO2 at clinically relevant thresholds, 3) EEG-confirmed seizures, 4) intracranial hemorrhage, and 5) outcomes (admission GCS score, 6-month Glasgow Outcome Scale-Extended [GOS-E] score, and length of stay [LOS] in the pediatric intensive care unit [PICU] and hospital). Significance was defined as p < 0.05 and correlation coefficient |r| > 0.10.

RESULTS

Seventy-two patients were included. NIRS rSO2 positively correlated with systolic blood pressure (right, r = 0.12, p < 0.001) and PbtO2 ≥ 20 mm Hg (left, r = 0.30; right, r = 0.25; p < 0.001), and NIRS rSO2 correlated negatively with systemic temperature (left, r = -0.14, p < 0.001) but not with ICP ≥ 20 mm Hg, mean arterial pressure, or suboptimal/ischemic PbtO2 levels. While not statistically significant, NIRS rSO2 trended higher following seizure and lower with hemorrhage. Bilateral NIRS rSO2 positively correlated with initial GCS score (left, r = 0.393, p = 0.001; right, r = 0.30, p = 0.018) and GOS-E score at 6 months (left, r = 0.315, p = 0.0161; right, r = 0.429, p = 0.0011) but not with PICU or hospital LOS.

CONCLUSIONS

NIRS rSO2 demonstrated significant correlations with select physiological parameters, including invasive optimal PbtO2 levels and acute outcomes (GCS and GOS-E scores), but not with respiratory measures, invasive neuromonitoring during ICP crisis or brain tissue ischemia, seizures, or intracranial hemorrhage. These findings suggest that noninvasive NIRS in its current state cannot reliably predict many physiological or clinical changes important for moderate to severe pediatric TBI or substitute invasive neuromonitoring.