URRENT treatments for comatose head-injured pa- tients often include three vasoreactivity-based ther- apies. Mild hyperventilation is used to help control intracranial hypertension and assumes effective CO 2 reac- tivity. 10,13,49 Maintenance of an adequate CPP is advocated to optimize CBF and assumes a relative degree of pre- served PA. 52 Use of metabolic suppressive therapy with high-dose pentobarbital or propofol for refractory intra- cranial hypertension is based on the assumption that there is a degree of retained metabolic suppression reactivity. 9,53 These components of vasoreactivity form the foundation for current head injury treatment strategies, namely hyper- ventilation therapy, CPP therapy, and pharmacological metabolic suppressive therapy. Despite widespread use of these therapies, it is generally acknowledged that vasore- activity is often seriously impaired after injury. Vascular dysfunction has been implicated in the development of posttraumatic cerebral ischemia, 7,8,37 cerebral hyper- emia, 25,30,32,46 intracranial hypertension, 6,25 and poor out- come. 29,44 Additionally, there remains considerable confu- sion about how to determine, for a given patient, the most appropriate degree of hyperventilation, 15,42,51 the optimal J Neurosurg 95:222–232, 2001 222 Carbon dioxide reactivity, pressure autoregulation, and metabolic suppression reactivity after head injury: a transcranial Doppler study JAE HONG LEE, M.D., M.P.H., DANIEL F. KELLY , M.D., MATTHIAS OERTEL, M.D., DAVID L. MCARTHUR, PH.D., THOMAS C. GLENN, PH.D., P AUL VESPA, M.D., W. JOHN BOSCARDIN, PH.D., AND NEIL A. MARTIN, M.D. Brain Injury Research Center; Cerebral Blood Flow Laboratory; Departments of Epidemiology and Biostatistics; Division of Neurosurgery, Center for Health Sciences; and Harbor–UCLA Medical Center and Research and Education Institute, University of California at Los Angeles, California Object. Contemporary management of head-injured patients is based on assumptions about CO 2 reactivity, pressure autoregulation (PA), and vascular reactivity to pharmacological metabolic suppression. In this study, serial assessments of vasoreactivity of the middle cerebral artery (MCA) were performed using bilateral transcranial Doppler (TCD) ultra- sonography. Methods. Twenty-eight patients (mean age 33  13 years, median Glasgow Coma Scale score of 7) underwent a total of 61 testing sessions during postinjury Days 0 to 13. The CO 2 reactivity (58 studies in 28 patients), PA (51 studies in 23 patients), and metabolic suppression reactivity (35 studies in 16 patients) were quantified for each cerebral hemi- sphere by measuring changes in MCA velocity in response to transient hyperventilation, arterial blood pressure ele- vation, or propofol-induced burst suppression, respectively. One or both hemispheres registered below normal vaso- reactivity scores in 40%, 69%, and 97% of study sessions for CO 2 reactivity, PA, and metabolic suppression reactivity (p  0.0001), respectively. Intracranial hypertension, classified as intracranial pressure (ICP) greater than 20 mm Hg at the time of testing, was associated with global impairment of CO 2 reactivity, PA, and metabolic suppression reactiv- ity (p  0.05). A low baseline cerebral perfusion pressure (CPP) was also predictive of impaired CO 2 reactivity and PA (p  0.01). Early postinjury hypotension or hypoxia was also associated with impaired CO 2 reactivity (p  0.05), and hemorrhagic brain lesions in or overlying the MCA territory were predictive of impaired metabolic suppression reac- tivity (p  0.01). The 6-month Glasgow Outcome Scale score correlated with the overall degree of impaired vasoreac- tivity (p  0.05). Conclusions. During the first 2 weeks after moderate or severe head injury, CO 2 reactivity remains relatively intact, PA is variably impaired, and metabolic suppression reactivity remains severely impaired. Elevated ICP appears to affect all three components of vasoreactivity that were tested, whereas other clinical factors such as CPP, hypotensive and hypoxic insults, and hemorrhagic brain lesions have distinctly different impacts on the state of vasoreactivity. Incorpo- ration of TCD ultrasonography–derived vasoreactivity data may facilitate more injury- and time-specific therapies for head-injured patients. KEY WORDS • cerebral vasoreactivity • carbon dioxide reactivity • hyperventilation • intracranial hypertension • autoregulation • propofol • transcranial Doppler ultrasonography • traumatic brain injury C J. Neurosurg. / Volume 95 / August, 2001 Abbreviations used in this paper: CBF = cerebral blood flow; CPP = cerebral perfusion pressure; CT = computerized tomogra- phy; eCVR = estimated cerebrovascular resistance; eCVR1 = base- line eCVR; eCVR2 = final eCVR; ETCO 2 = end-tidal carbon diox- ide; GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale; ICA = internal carotid artery; ICH = intracranial hemorrhage; ICP = intracranial pressure; ICU = intensive care unit; MAP = mean arterial pressure; MAP1 = baseline MAP; MAP2 = final MAP; MCA = middle cerebral artery; PA = pressure autoregulation; SAH = subarachnoid hemorrhage; SE = standard error; SjvO 2 = jugular venous O 2 saturation; TCD = transcranial Doppler; UCLA = University of California at Los Angeles; V MCA = flow velocity of the MCA averaged over four cardiac cycles; %∆ = percentage change.