Diagnosis influences response of cerebral near infrared spectroscopy to intracranial hypertension in children

Abstract

Objective: To describe cerebral regional oxygen saturation measured by near infrared spectroscopy in the setting of normal and increased intracranial pressure in children to evaluate the association between cerebral regional oxygen saturation and intracranial pressure in comparison with other clinical variables.

Design: Prospective observational cohort study.

Setting: Two academic tertiary care centers' pediatric intensive care units.

Patients: Thirty patients with intracranial pressure and near infrared spectroscopy monitoring (median age, 11.5 yrs; interquartile range, 5.2-13 yrs) for a range of neurologic diagnoses, including brain tumor, trauma, intracerebral hemorrhage, and hydrocephalus.

Interventions: None.

Measurements and main results: Temporally correlated cerebral regional oxygen saturation with hematologic (hematocrit), biochemical (pH), and physiological (intracranial pressure, mean arterial pressure, cerebral perfusion pressure, temperature, heart rate, pulse oximetry and end-tidal carbon dioxide) variables. Cerebral regional oxygen saturation during episodes of increased intracranial pressure was lower than with normal intracranial pressure (mean +/- sd intracranial pressure >20 = 71% +/- 13% vs. intracranial pressure <20 = 75% +/- 10%), although the mean difference (-4%) is small compared with variability in the measurement. Neither isolated values nor change in cerebral regional oxygen saturation were significantly associated with intracranial pressure or cerebral perfusion pressure in the overall population. Isolated values and change in end-tidal CO2 were significantly correlated with cerebral regional oxygen saturation and change in cerebral regional oxygen saturation (all p < 0.01). In exploratory analyses, the diagnostic group significantly modified the effect of intracranial hypertension on regional oxygen saturation: regional oxygen saturation decreased during intracranial hypertension in patients with brain tumors (p = .05) and hydrocephalus (p < .001) but increased during intracranial hypertension in those with intracranial hemorrhage (p < .001).

Conclusions: These data suggest that cerebral regional oxygen saturation is lower with intracranial hypertension. However, the relationship between cerebral regional oxygen saturation and intracranial pressure is strongly influenced by diagnosis.

Figure 1

Representative plots of the continuous recordings of cerebral rSO₂ in a patient with hydrocephalus and ventriculoperitoneal shunt malfunction (A) and closed head trauma (B). Placement of external ventricular drain is noted in (A). ICP values are plotted in (B, solid black squares). Three events of ICP change > 10 mmHg are identified by rectangular boxes.

Figure 2

Histogram summarizing the range of values for cerebral rSO₂ during normal ICP (A). The population fits a normal distribution (fitted curve). Cerebral rSO₂ values during intracranial hypertension are summarized by histogram in (B). Cerebral rSO₂ during increased ICP is shifted to lower values compared with those observed during normal ICP (fitted curve from panel A). Bland-Altman plots (average vs. difference) of simultaneous right and left cerebral rSO₂ (C) compared with cerebral rSO₂ ipsilateral and contralateral to the ICP monitor (D).

Figure 3

Summary of the mean+/−standard error of the mean correlation coefficients between cerebral rSO₂ and ICP shifted by a series of time lags (−120 minutes to 120 minutes at 15 minute intervals) (A). Box plots (median, quartiles and outliers) of cerebral rSO₂ during periods of normal ICP (< 20 mmHg) and intracranial hypertension by diagnostic group. Measurement number is summarized above the box plots. Patient number is summarized below the X-axis category labels.