Cerebral vascular response to hyperoxia in head injured patients

Introduction As brain tissue pO₂ (PbtO₂) monitoring is becoming more wide-spread, it has been observed that PbtO₂ values are influenced by changes in arterial pO₂ (PaO₂) to a greater extent that can be explained by changes in oxygen delivery, especially in the damaged brain. It has been proposed that this may reflect an impaired ability of the cerebral vessels to vasoconstrict in response to an increasing PaO₂. The purpose of this study was to compare the changes in PbtO₂ to changes in cerebral blood flow velocity during a step change in PaO₂. Methods Twenty-eight patients in coma from severe head injury were studied serially during the first 10 days after injury. In addition to the usual physiological monitors, a Licox PbtO₂ monitor was placed in the brain, usually in white matter underlying an evacuated hematoma, or in tissue surrounding a contusion. Flow velocity in the middle cerebral artery (mcaFV), PaO₂, jugular venous pO₂ (PjvO₂), and PbtO₂ were measured at the patient's baseline FiO₂ (usually 0.4), then at an FiO₂ of 1.0, then again after returning the FiO₂ to the patient's baseline. Results PaO₂ increased from 146±3 to 330 ±8 mmHg, PbtO₂ increased from 32 ±3 to 57 ±4 mmHg, and PjvO₂ increased from 45±1 to 67±6 mmHg as the FiO₂ was increased to 1.0. In response to these changes in oxygenation, mcaFV fell from 75 ±2 to 73 ±2 cm/sec on the injured side and from 79 ±2 to 75 ±2 cm/sec on the uninjured side (both p<.001). ICP also significantly decreased from 19±1 to 17±1 mmHg (p<.001). There was no relationship between the change in mcaFV and the change in PbtO₂. Conclusions Hyperoxia causes a small, but detectable reduction in mcaFV, even on the most severely injured side of the brain. This vascular response to hyperoxia does not seem to explain the variability in the response of PbtO2 to hyperoxia that occurs after injury, and perhaps other factors such as diffusion characteristics of the injured brain tissue are responsible.