ObjectivesBrain-injured patients are susceptible to secondary brain damage related to decreased cerebral perfusion pressure associated with edema formation and increased intracranial pressure (ICP). Whenever conventional therapy fails to reduce elevated ICP, barbiturate coma represents an additional intervention that may control ICP. In patients suffering from severe traumatic brain injury, cerebrospinal fluid levels of glutamate, hypoxanthine, and lactate were measured during barbiturate coma and correlated to electroencephalographic recordings and ICP.DesignProspective, descriptive study.SettingTen-bed surgical intensive care unit in a university hospital.PatientsTwenty-one patients with severe traumatic brain injury (Glasgow Coma Scale score <or=to9); 11 required barbiturate coma because of refractory intracranial hypertension, and 10 were manageable with continuous administration of fentanyl and midazolam.InterventionsThiopental was administered continuously for increased ICP within the first 24 hrs after trauma and adjusted to the burst-suppression pattern (four to six bursts per minute) on continuous electroencephalographic monitoring.Measurements and Main ResultsGlutamate and hypoxanthine were analyzed using high-performance liquid chromatography, whereas lactate was measured enzymatically. Patients requiring thiopental presented with significantly higher ICP, glutamate, and hypoxanthine levels than patients receiving fentanyl and midazolam (p < .05). Within the first 24 hrs, thiopental significantly reduced cerebrospinal fluid glutamate and hypoxanthine levels in all patients, i.e., the burst-suppression pattern was successfully induced (p < .001). Interestingly, in five patients cerebrospinal fluid glutamate increased to initial values again despite unchanged neuronal activity. In these patients, ICP, hypoxanthine, and lactate remained significantly elevated compared with the six patients with steadily decreasing cerebrospinal fluid glutamate, hypoxanthine, lactate, and ICP values (p < .02).ConclusionsBarbiturate coma does not unequivocally preserve energetic stability despite successful suppression of neuronal activity. Despite the use of barbiturate coma in patients with refractory intracranial hypertension, persistent release or impaired uptake of glutamate may be associated with continuous anaerobic metabolism, as shown by increases in cerebrospinal fluid hypoxanthine and lactate levels. (Crit Care Med 1999; 27: 1351-1357)