Ischemia is one of the major factors causing secondary brain damage after severe head injury. We have investigated the value of continuous partial pressure of brain tissue oxygen (PbrO2) monitoring as a parameter for cerebral oxygenation in 22 patients with severe head injury (Glasgow Coma Scale score, < or = 8). Jugular bulb oxygenation, intracranial pressure, and cerebral perfusion pressure were simultaneously recorded. O2 and CO2 reactivity tests were performed daily to evaluate oxygen autoregulatory mechanisms. PbrO2 monitoring was started an average of 7.0 hours after trauma with a mean duration of 74.3 hours. No complications were seen, and the calibration of the catheters after measurement showed a zero drift of 1.2 +/- 0.8 mm Hg and a sensitivity drift of 9.7 +/- 5.3%. In 86% of patients, PbrO2 was < 20 mm Hg in the acute phase. Mean PbrO2 significantly increased during the first 24 hours after injury. Two distinct patterns of change of PbrO2 over time were noted. The first pattern was characterized by normal stable levels after 24 hours, and the second was characterized by transiently elevated levels of PbrO2 during the second and third days. PbrO2 values < or = 5 mm Hg within 24 hours after trauma negatively correlated with outcome. O2 reactivity was significantly lower in patients with good outcomes. CO2 reactivity showed no constant pattern of change over time and was not correlated with outcome. Increased hyperventilation was shown to decrease PbrO2 in some patients. Accurate detection of the moment of cerebral death was possible on the basis of the PbrO2 measurements. The correlation between PbrO2 and other parameters, such as intracranial pressure and cerebral perfusion pressure, was weak. We conclude that PbrO2 monitoring is a safe and clinically applicable method in patients with severe head injury. The early occurrence of ischemia after head injury can be monitored on a continuous basis. Deficiency of oxygen autoregulatory mechanisms can be demonstrated, and their occurrence is inversely related to outcome. For practical clinical use, the method seemed to be superior to jugular oximetry.