Effect of hyperventilation on regional cerebral blood flow in head-injured children.

OBJECTIVES To study cerebral blood flow and cerebral oxygen consumption in severe head-injured children and also to assess the effect of hyperventilation on regional cerebral blood flow. DESIGN Prospective cohort study. SETTING Pediatric intensive care unit at a tertiary-level university children's hospital. PATIENTS Twenty-three children with isolated severe brain injury, whose admission Glasgow Coma Scores were <8. INTERVENTIONS PaCO2 was adjusted by altering minute ventilation. Cerebral metabolic measurements were made at three levels of PaCO2 (>35, 25 to 35, and <25 torr [>4.7, 3.3 to 4.7, and <3.3 kPa]) after allowing 15 mins for equilibrium. MEASUREMENTS AND MAIN RESULTS Thirty-eight studies (each study consisting of three sets of measurements at different levels of PaCO2) were performed on 23 patients. At each level of PaCO2, the following measurements were made: xenon-enhanced computed tomography scans; cerebral blood flow; intracranial pressure; jugular venous bulb oxygen saturation; mean arterial pressure; and arterial oxygen saturation. Derived variables included: cerebral oxygen consumption; cerebral perfusion pressure; and oxygen extraction ratio. Cerebral blood flow decreased below normal after head injury (mean 49.6 +/- 14.6 mL/min/100 g). Cerebral oxygen consumption decreased out of proportion to the decrease in cerebral blood flow; cerebral oxygen consumption was only a third of the normal range (mean 1.02 +/- 0.59 mL/min/100 g). Neither cerebral blood flow nor cerebral oxygen consumption showed any relationship to time after injury, Glasgow Coma Score at the time of presentation, or intracranial pressure. The frequency of one or more regions of ischemia (defined as cerebral blood flow of <18 mL/min/100 g) was 28.9% during normocapnia. This value increased to 73.1% for PaCO2 at <25 torr. CONCLUSIONS Severe head injury in children produced a modest decrease in cerebral blood flow but a much larger decrease in cerebral oxygen consumption. Absolute hyperemia was uncommon at any time, but measured cerebral blood flow rates were still above the metabolic requirements of most children. The clear relationship between the frequency of cerebral ischemia and hypocarbia, combined with the rarity of hyperemia, suggests that hyperventilation should be used with caution and monitored carefully in children with severe head injuries.