Cerebral oxygenation monitoring by near-infrared spectroscopy is not clinically useful in patients with severe closed-head injury: a comparison with jugular venous bulb oximetry

Abstract

Objective: To compare continuous jugular venous bulb oximetry and cerebral near-infrared spectroscopy in patients with severe closed head injury.

Design: A prospective observational study.

Setting: Intensive care unit of a major teaching hospital.

Patients: Adults (n = 10) with severe closed-head injury (Glasgow Coma Scale score of < or = 8).

Interventions: None.

Measurements and main results: Jugular venous bulb oximetry, cerebral near-infrared spectroscopy, and cerebral perfusion pressure were measured continuously. A total of 3,691 paired measurements of near-infrared spectroscopy and jugular venous bulb oximetry were analyzed. Poor correlation (r2 = .04) between paired measurements and wide limits of agreement (-13% to +21%) were demonstrated. The mean difference between measurements was +/- 4% and the standard deviation of the mean difference was +/- 8.69%. The data were subsequently grouped according to three clinically significant subgroups of jugular venous bulb oxygen saturation reflecting low ( < 55%), normal (55% to 75%) and high ( > 75%) saturation values. Poor correlation and wide limits of agreement between the two methods of measurement were observed in all groups. Values recorded by near-infrared spectroscopy did not significantly change between the groups, and 14 clinically significant episodes of jugular venous bulb desaturation were not detected by near-infrared spectroscopy.

Conclusions: Tissue oxygen saturation determined by near-infrared spectroscopy does not reflect significant changes in cerebral oxygenation detected by the global measurement of jugular venous bulb oximetry. This finding may be explained by inadequate signal detection and inaccuracies in the algorithm used to filter out extracranial components. Until these technical difficulties are addressed, near-infrared spectroscopy, as measured by the machine assessed in this study, cannot be routinely recommended for assessment of cerebral oxygenation in patients with acute head injury.