Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Edinburgh, Scotland

Transcription

1 J Neurosurg 77:55-61, 1992 The effect of changes in cerebral perfusion pressure upon middle cerebral artery blood flow velocity and jugular bulb venous oxygen saturation after severe brain injury KWAN-HON CHAN, F.R.C.S., J. DOUGLAS MILLER, M.D., PH.D., F.R.C.S., F.R.C.P., N. MARK DEARDEN, B.Sc., F.F.A.R.C.S., PETER J. D. ANDREWS, F.F.A.R.C.S., AND SUSAN MIDrLEV, F.F.A.R.C.S. Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Edinburgh, Scotland v" Middle cerebral artery blood flow velocity and jugular bulb venous oxygen saturation (SJO2) were measured by transcranial Doppler (TCD) ultrasonography and continuous venous oximetry, respectively, in 41 severely brain-injured patients. The purpose of the study was to examine the relationships between TCD flow velocity, SJO2, and alterations in blood pressure (BP), intracranial pressure (ICP), and cerebral perfusion pressure (CPP). In these patients, CPP was reduced either by rising ICP or by falling BP. Both forms of reduction of CPP resulted in a greater fall in diastolic flow velocity than other flow parameters. As CPP decreased below a critical value of 70 mm Hg, a progressive increase in TCD pulsatility index (PI) was observed (r = , p < ), accompanied by a fall in SJO2 (r = 0.78, p < ). At pressures above 70 mm Hg, there was no correlation of either PI or SJO2 with CPP. The relationship between PI and CPP held true in patients with both focal and diffuse pathologies and was the same whether changes in CPP resulted from alterations in ICP or BP. The PI and SJO2 correlated better with CPP than with ICP or BP. Transcranial Doppler ultrasonography can identify states of reduced CPP. Decreases in SJO2 with falling CPP suggested progressive failure of cerebral blood flow to meet metabolic demands. Monitoring of TCD and SJO2 may be used to define the optimum CPP level for management of severely brain-injured patients. KEY WORDS " autoregulation 9 cerebral blood flow 9 cerebral perfusion pressure intracranial pressure 9 head injury 9 jugular bulb venous oxygen saturation 9 ultrasound I N the past decade, transcranial Doppler (TCD) ultrasonography has allowed repeated and continuous bedside assessment of blood flow velocity in major basal intracranial vessels.~ Continuous recording of arterial and jugular bulb venous blood oxygen saturation (SJO2) permits calculation of the cerebral arteriovenous oxygen content difference (AVDO:). This reflects the ratio of global cerebral oxygen supply to demandfl Whether these investigative tools can be used as a practice aid to the management of patients with severe brain injury remains controversial. The relationships between TCD blood flow velocity, SJO: changes, and alterations in intracranial pressure (ICP) have not been fully defined. The present study aims to define in patients with severe brain injury the relationships of alterations in ICP, mean arterial blood pressure (MABP), and cerebral perfusion pressure (CPP) to changes in TCD blood flow velocity and SJO2. Clinical Material and Methods Between August, 1989, and June, 1990, 41 patients with severe closed brain injury (defined on admission as a postresuscitation Glasgow Coma Scale score < 8, with no eye opening) and ICP monitoring were entered into this study. There were 34 males and seven females with a mean age of 29 years (range 6 to 59 years). All patients underwent computerized tomography (CT) on admission. The intracranial pathology was classified according to CT findings as focal (hematoma and/or unilateral contusion or swelling) or diffuse injury. All patients were managed by a standard regimen that included artificial ventilation under paralysis and sedation with continuous infusions of pancuronium and phenoperidine, respectively. Blood pressure, arterial oxygen saturation (SaO2) by pulse oximetry, endtidal CO2 concentration, body temperature, and bipa- J. Neurosurg. / Volume 77~July,

2 K. H. Chan, et al. rietal cortical electrical activity by cerebral function monitor were recorded continuously. The lower-border voltage of the cerebral function monitor was measured. The ICP was obtained by subdural transducers in 35 patients or by subdural fluid-filled catheters* in six patients. The CPP was derived from the difference between the MABP and the mean ICP. In addition, 22 patients had unilateral continuous monitoring of SJO2 following percutaneous retrograde insertion of a 40-cm optical catheter via the internal jugular vein on the side of predominant venous drainage.t This was determined by the ICP response to neck compression] The position of the tip of the catheter in the jugular bulb was confirmed by x-ray study. In vivo calibration was performed every 12 hours by measuring SJO2 in vitro. The AVDO2 was calculated by multiplying the difference between SaO2 and SJO2 by the daily hemoglobin concentration and 1.39, and dividing by 100. Global cerebral hyperemia was defined as an AVDO2 of less than 4 ml/dl and ischemia as an AVDO2 of greater than 9 ml/dl. Due to the reciprocal relationship between AVDO2 and SJO2, a low AVDO2 is associated with a high SJO2 and vice versa. Treatment was initiated when ICP exceeded 25 mm Hg or CPP fell below 60 mm Hg. Data obtained after ICP therapy were excluded from this study. Transcranial Doppler insonation~ of both middle cerebral arteries (MCA's) was performed within 24 hours of admission according to the method described by Aaslid, et al.~ The depth of insonation giving the highest mean flow velocity was chosen for recording, and subsequent daily measurements used the same window and depth. Velocity parameters measured included the peak systolic (S), end-diastolic (D), and timed mean (M) velocities from which the pulsatility index (PI) could be calculated: S - D/M. Intermittent TCD measurements were averaged over at least 15 cardiac cycles during periods of cerebral and hemodynamic stability. The MCA peak systolic, end-diastolic, and mean flow velocities and the PI (mean values + standard deviations) obtained in 20 normal subjects were, respectively: 98 _ 16, 44 _+ 7, and cm/sec, and 0.9 _ The coefficient of variation of repeated velocity measurements was 5%. An MCA mean flow velocity of greater than 100 cm/sec was considered abnormally high. During the TCD examinations, SaO2 was maintained above 95%, end-tidal CO2 concentrations varied from 2.7 to 4.7 kpa (mean 3.5 kpa), body temperatures were in the range of 37* to 39"C, heart rate varied from 70 to 105 beats/min (mean 83 beats/min), and hemoglobin * Subdural transducers manufactured by Camino Laboratories, San Diego, California; catheters manufactured by Cordis Laboratories, Brentford, England. Oximetrics 3 O2-monitoring system manufactured by Abbott Laboratories, Chicago, Illinois. :~ Transcranial Doppler ultrasound system manufactured by Medasonics, Mountain View, California. concentrations varied from 9.9 to 14 gm/dl (mean 12.1 gm/dl). In 30 patients (22 with SJO2 monitoring), continuous recordings of MCA flow velocity were obtained during periods of changing CPP by mounting the ultrasound probe in a head band. Recordings were on the same side as the lesion for focal pathology and on the same side as the ICP monitor in cases of diffuse injury. Blood flow velocity and SJO2 measurements were made at the points of maximum and subsequent minimum CPP before treatment was instigated. Statistical analysis was by correlation regression. Results Intracranial pathology was classified as focal in 25 patients and diffuse in 16 patients. In 25 patients, increases in ICP to above 25 mm Hg were observed on 92 occasions, decreases in MABP to below 70 mm Hg were noted in 26 instances, and falls in CPP to less than 60 mm Hg were recorded 78 times. Relationship Between ICP, MABP, CPP, and TCD Changes In general, as ICP increased or MABP decreased, accompanied by a decrease in CPP, there was a reduction in flow velocity, with diastolic velocity falling more than systolic velocity. This resulted in a decrease in mean flow velocity, an increase in the systolic-diastolic velocity difference, and an increase in PI. As ICP exceeded the diastolic blood pressure, systolic velocity remained positive but diastolic velocity became negative (reverberant flow), giving rise to a greater decrease in mean flow velocity and a sharp increase in PI) ~ When rises in ICP were not associated with reductions in CPP, no appreciable changes in velocity or PI were observed (Fig. 1). A linear correlation was noted between ICP (r = , p < 0.001), MABP (r = 0.271, p < 0.001), CPP (r = 0.477, p < 0.001), and mean flow velocity. Similar correlations existed between ICP (r = 0.439, p < 0.001), MABP (r = , p < 0.001), CPP (r = , p < ), and PI. As a result, CPP correlated better with PI (Fig. 2 left). This relationship between CPP and PI held true for both focal and diffuse lesions. Similar changes in PI were observed whether CPP was compromised as a result of decreased MABP or increased ICP (Fig. 2 right). An alternative analysis of the PI/CPP curve revealed a triphasic relationship with two apparent breakpoints. Linear regression analysis was applied to different segments of the curve until the most significant correlation coefficients were obtained. Breakpoints were noted at CPP values of 20 and 70 mm Hg. At CPP values above 70 mm Hg, there was no correlation between CPP and PI (r = , p = not significant). As CPP decreased from 70 to 20 mm Hg, the strongest inverse correlation between CPP and PI was obtained (r = , p < ); as CPP decreased below 20 mm Hg, a further sharp increase in PI occurred (r = -0.73, p = 0,01). 56 J. Neurosurg. / Volume 77~July, 1992

3 Cerebral flow velocity and perfusion pressure FIG. 1. Graphs showing changes in mean Doppler ultrasound blood flow velocity (vel, cm/sec) and pulsatility index (PI) during alterations in mean arterial pressure (MAP, mm Hg), intracranial pressure (ICP, mm Hg), and cerebral perfusion pressure (CPP, mm Hg). Two patients with focal head injury had high PI measurements when the recorded CPP values exceeded 70 mm Hg (represented by filled triangles in Fig. 2 left). Despite this, both patients had other evidence to suggest raised ICP, consisting of ipsilateral pupillary dilatation on the side of pathology and midline shift, absence of third ventricle, and effacement of perimesencephalic cisterns on CT scans. The ICP monitoring devices were therefore suspected of giving inaccurate pressure recordings. In another six patients, the PI levels remained low despite a progressive fall in CPP (shown by filled circles in Fig. 2 left). They had an abnormally high mean flow velocity (> 100 cm/sec) when CPP exceeded 60 rnra Hg and AVDO2 values were in the hyperemic range. Relationships Between ICP, MABP, CPP, and S J02 Linear correlations were noted between ICP (r = , p < 0.001), CPP (r = 0.521, p < 0.001), and SJO2, but not MABP (r = 0.124, p = not significant). Fro. 2. Left: Graph plotting cerebral perfusion pressure (CPP) versus pulsatility index (PI) in patients with diffuse (crosses) or focal (open triangles) lesions. Two patients considered to have inaccurate intracranial pressure (ICP) recordings are shown by closed triangles, and levels of patients with global hyperemia and increased velocity are shown by closed circles. Right: Graph plotting CPP versus PI in patients before and after CPP was compromised as a result of increased ICP (open circles) or fall in mean arterial blood pressure (closed circles). J. Neurosurg. / Volume 77~July,

4 K. H. Chan, et al breakpoint of 70 mm Hg (r = , p < ). Above the CPP breakpoint, there was no correlation between CPP and either PI (r = , p = not significant) or SJO2 (r = , p = not significant). FIG. 3. Flow chart showing the relationship between increased velocity to above 100 cm/sec, pulsatility index (PI), and jugular venous oxygen saturation (SJO2) in hyperemic and nonhyperemic patients and reduction in cerebral perfusion pressure (CPP) to below 60 mm Hg. The lower-border voltage of the cerebral function monitor showed no major changes when CPP was reduced. Interrelationships Between CPP, S J02, and TCD Findings Of the 22 patients with SJO2 monitoring, 10 developed global hyperemia (AVDO2 < 4 ml/dl) during periods of recording and 12 remained nonhyperemic (AVDO2 4 to 9 ml/dl). Eight of the 12 nonhyperemic patients and six of the 10 hyperemic patients had increased flow velocity to above 100 cm/sec when CPP was above 60 mm Hg (Fig. 3). Velocity decreased when CPP fell below 60 mm Hg. In the nonhyperemic cases, as CPP decreased due to either a rise in ICP or a fall in MABP, AVDO2 increased (SJO2 decreased) and PI increased (Fig. 4). As CPP fell in the hyperemic patients due to elevated ICP, two patterns of changes were observed. In cases without increased flow velocity, AVDO2 increased (SJO2 decreased) from below 4 ml/ dl to within normal or ischemic ranges and PI rose, a trend similar to that of the nonhyperemic patients; in patients with elevated flow velocity, AVDO2 decreased (SJO2 increased) further and PI remained constant (Figs. 2 left and 4). The SJO2 levels in the latter group of patients decreased with hyperventilation, which also reduced ICP and improved CPP. Since SJO2 and PI changes were different in the six hyperemic patients with increased velocity, they were analyzed separately. Correlations between CPP and PI (r = , p < ) and SJO2 (r = 0.685, p < ) were better. An alternative analysis of CPP against PI and SJO2 in the remaining 16 patients with concomitant continuous TCD ultrasound and SJO2 monitoring showed a progressive decrease in SJO2 as CPP fell from a breakpoint value of 71 mm Hg (r = 0.78, p < ), identified by sequential linear regression analysis, accompanied by a rise in PI from a CPP Discussion Methodological Considerations Transcranial Doppler Ultrasound Studies. Since intercompartmental differences in intermittent velocity recordings had been observed in our patients and in other series of patients with focal lesions causing intracranial hypertension, ~~ we employed continuous TCD recording from the same side as the focal injury. Continuous monitoring allows more reproducible recordings than with a hand-held probe because of the constancy of the position of the ultrasound probe in relation to the vessel being examined. '6 Use of the dimensionless PI has the advantage of eliminating errors in measurement due to insonation angle variations. 6 Simultaneous continuous monitoring of additional variables that may affect the TCD recordings is important, especially when serial measurements are compared. 6'18 These factors include MABP, ICP, CPP, CO2 level, SaO2, and hemoglobin concentration. Jugular Bulb Venous Oxygen Saturation Measurement. Anatomically, 80% to 90% of the blood draining from both cerebral hemispheres drains into one internal jugular vein. 3' Anatomical studies indicate that blood in each internal jugular vein is, however, representative of drainage from all parts of the brain. 29 Simultaneous blood sampling of both jugular venous bulbs in normal individuals shows that SJO2 is equal on both sides. 4 This pattern of drainage may vary in some but not all patients with focal brain lesions. 12,3~ Jugular venous blood taken from the side of the intracranial lesion is thought by some investigators 27 to be more representative of global SJO2; however, a collective review by Lassen 12 of a number of studies involving normal individuals and patients with various clinical conditions, including unilateral lesions, showed that no significant side-to-side differences in SJO2 have been found. Unilateral SJO2 measurement is considered representative of the global oxygen supply-and-demand status. The correlation observed in this study between regional TCD recordings and global SJO2 measurements lends further support to those findings. Previous studies have employed intermittent withdrawal of jugular bulb blood. Instances of contamination by extracerebral blood are not uncommon when samples are withdrawn during low cerebral blood flow (CBF) status. 3'27 Dye injection study has shown that blood in the jugular venous bulb contains very little blood derived from extracerebral sourcesfl 9 Continuous monitoring with the catheter tip position remaining unchanged in the jugular bulb and measurement independent of the speed of blood aspiration represents the most accurate way of monitoring SJO2, even when CPP and bqood flow are low. 2 Sampling error as a result of 58 J. Neurosurg. / Volume 77~July, 1992

5 Cerebral flow velocity and perfusion pressure extracerebral contamination could account for some previously reported cases of hyperemia. 3 Hyperventilation has been used to differentiate true hyperemia from extracerebral contamination, assuming that CO: reactivity of the brain is still preserved. 1-' Apparent hyperemia may also be due to secondary causes, such as raised body temperature, seizures, and acid-base abnormalities. 5 Multimodality continuous monitoring allowed such secondary causes of hyperemia to be documented or excluded in the present study. Such factors may account for the greater association of hyperemia with elevated ICP reported previouslyy Relationship Between TCD Changes and Changes in MABP, 1CP, and CPP In this study, CPP correlated better with PI than with ICP or MABP. This is in contrast to previous reports that TCD changes are sensitive only to major changes in ICP. 6~-' ' The incidence of arterial hypotension may appear high; however, the data were obtained from a minute-by-minute printout of all monitored variables and carefully checked for accuracy? 5 In a comparison study between the efficiency of computer monitoring and data recorded by the intensive care nursing staff, episodes of arterial hypotension were frequently undetected by the nursing staff (unpublished observations). As CPP falls below the lower autoregulatory threshold, CBF progressively decreases, z~'24 A fall in volume flow is mainly a result of decrease in flow velocity, 26,28 which affects the diastolic velocity preferentially. 7,,4 The combination of widening systolic-diastolic velocity difference and the fall in mean velocity results in elevated pulsatility amplitude and an increase in PI. '~ Experimental and clinical studies have shown that blood flow velocity correlates with CBF only when the latter is low. 8"28 Our results are in full accord with these findings. Relationship Between SJO~_ and CPP By Fick's equation, AVDO2 represents the ratio of cerebral metabolism (CMR) to CBF. Since AVDO2 is derived from SJO2 and provided that the position of the oxygen dissociation curve and the hemoglobin concentration remain constant, the SJO2 is a measure of the ratio CBF:CMR. 5 A fall in SJOz indicates an increase in oxygen extraction by the brain to maintain its metabolic needs when CBF falls due to a reduction in CPP. When this mechanism is exhausted, CMR will fall? 4 The CMR was thought to remain relatively constant in this study because the physiological parameters and the lower-border voltage of the cerebral function monitor remained unchanged. The latter parameter has been shown experimentally to correlate with CMR. 2~ The physiological goal of autoregulation is to maintain a level of CBF adequate to meet tissue metabolic demands? 4 Cerebral resistance vessels are sensitive to both metabolic stimuli and changes in CPP? 4 The lower limit of autoregulation should be defined as the CPP threshold below which CBF adequate to meet tissue oxygen demands cannot be maintained. Increases in FIG. 4. Graph plotting cerebral perfusion pressure (CPP) versus jugular venous oxygen saturation (SJO2) in nonhyperemic (open circles) and hyperemic patients with blood flow velocity above 100 cm/sec (crosses) and below 100 cm/sec (closed circles). CPP above this threshold will not further augment global CBF and oxygen delivery. 2u4 Demonstration of a constant SJO2 or AVDO2 constitutes strong evidence of the constancy of CBF if CMR remains unchanged.'2 This must be distinguished from loss of autoregulation based on testing of CBF status with rapid changes in blood pressure, and represents only the instantaneous state of pressure autoregulation. Therefore, continuous SJO2 monitoring allows identification of a threshold CPP level above which CPP should be maintained during head-injury management. In this study, both TCD findings and SJO2 measurements identified a CPP threshold of 70 mm Hg. Relationship Between TCD and S J02 Measurements Below the CPP breakpoint of 70 mm Hg, the progressive increase in PI with decreasing CPP was an indication of exhaustion of the process of autoregulation. Our findings are in accord with the theoretical concept that cerebral autoregulation is not an all-ornone phenomenon, but a continuous process.17'2"24 The lower limit of autoregulation represents the CPP level (about 40 mm Hg in normal subjects) below which cerebral vasodilation and reduction in cerebrovascular resistance (CVR) can no longer compensate sufficiently for the decreasing CPP. Below the lower limit of autoregulation, maximum cerebral vasodilation and decrease in CVR is not observed until CPP falls to approximately 30 mm HgJ TM At this stage blood supply to the brain is markedly impaired and, in some cases, as CPP falls further it may come to a standstill. The column of blood in the major cerebral arteries may at this stage be oscillating with no net flow to the brain. This would account for the reverberant flow pattern often found below this level of CPP resulting in a sharp increase in PI. 1~ The CPP threshold of 70 mm Hg identified in this study suggests that a higher CPP than previously rec- J. Neurosurg. / Volume 77 ~duly,

6 K. H. Chan, et al. ognized is required for the successful management of severely head-injured patients. ~ This concept is in accord with the experimental observations that, with increasing levels of severity of brain tumors, CVR starts to rise at progressively high levels of CPP. ~3 Serial continuous TCD monitoring, like measurements of SJO2, may be used to define this lower autoregulatory limit. This relationship, however, was not observed in hyperemic patients with increased flow velocity. The CBF remained in excess of metabolic needs despite a reduction in CPP. Autoregulation may be impaired in these cases, since elevated CBF is associated with an increase in blood flow velocity. '5 Hyperemic patients with an abnormal rise in flow velocity may represent absolute hyperemia. This would be consistent with previous findings of progressive vasodilatation, reduced vascular resistance, increased cerebral blood volume, and loss of autoregulation during an absolute increase in global CBF complicating brain injury. 3'22"26 With a decrease in CVR, the diastolic velocity is preserved and the PI remains constant. TM A combination of TCD and SJO2 recording may permit the distinction between absolute and relative hyperemia. Prediction of CPP by TCD Ultrasonography The close correlation between PI and CPP, especially when the latter is low, makes it possible to use PI to estimate CPP. Since this can be done noninvasively, this capability is clinically important. Serial PI changes offer more valuable information than a single measurement. When there is no ICP monitor in place or there is clinical suspicion of an inaccurate ICP measurement, TCD ultrasonography may be useful in detecting low CPP. However, a normal PI in the absence of ICP and/ or SJO2 monitoring does not always indicate an adequate CPP, as illustrated by some cases of global hyperemia. Multimodality monitoring allows such situations to be identified and clarified. The optimum level of CPP for management can also be determined in individual patients after severe head injury. References l. Aaslid R, Markwalder TM, Nornes H: Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 57: , Andrews PJD, Dearden NM: Validation of the Oximetrics 3 for continuous monitoring of jugular bulb oxygen saturation after severe head injury: comparison with IL282 in vitro co-oximeter. Br J Anaesth 64:393P-394P, Bruce DA, Alavi A, Bilaniuk L, et al: Diffuse cerebral swelling following head injuries in children: the syndrome of "malignant brain edema." J Neurosurg 54: , Gibbs EL, Lennox WG, Gibbs FA: Bilateral internal jugular blood. Comparison of A-V differences, oxygendextrose ratios and respiratory quotients. Am J Psychiatry 102: , Gilbert J: Estimation of CBF by cerebral venous oxygen difference. J Neurosurg 71: , 1989 (Letter) 6. Giulioni M, Ursino M, Alvisi C: Correlations among intracranial pulsatility, intracranial hemodynamics, and transcranial Doppler wave form: literature review and hypothesis for future studies. Neurosurgery 22: , Greenfield JC, Tindall GT: Effect of acute increase in intracranial pressure on blood flow in the internal carotid artery of man. J Ciin Invest 44: , Halsey JH, McDowell HA, Gelmon S, et al: Blood velocity in the middle cerebral artery and regional cerebral blood flow during carotid endarterectomy. Stroke 2@: 53-58, Harders AG: Neurosurgical Applications of Transcranial Doppler Sonography. Wien: Springer-Verlag, 1986, pp Hassler W, Steinmetz H, Gawlowski J: Transcranial Doppler ultrasonography in raised intracranial pressure and in intracranial circulatory arrest. J Neurosnrg 68: , Klingelhrfer J, Conrad B, Benecke R, et al: Evaluation of intracranial pressure from transcranial Doppler studies in cerebral disease. J Neurol 235: , Lassen NA: Cerebral blood flow and oxygen consumption in man. Physiol Rev 39: , Lewelt W, Jenkins LW, Miller JD: Autoregulation of cerebral blood flow after experimental fluid percussion injury of the brain. J Neurosarg 53: , Lindegaard KF, Grip A, Nornes H: Precerebral haemodynamics in brain tamponade. Part 1: Clinical studies on blood flow velocity. Nearochirurgia 23: , Lindegaard KF, Grolimund P, Aaslid R, et al: Evaluation of cerebral AVM's using transcranial Doppler ultrasound. J Neurosurg 65: , Lindegaard KF, Lundar T, Wiberg J, et al: Variations in middle cerebral artery blood flow investigated with noninvasive transcranial blood velocity measurements. Stroke 18: , MacKenzie ET, Farrar JK, Fitch W, et al: Effects of hemorrhagic hypotension on the cerebral circulation. I. Cerebral blood flow and pial arteriolar caliber. Stroke 10: , Markwalder TM, Grolimund P, Seiler RW, et al: Dependency of blood flow velocity in the middle cerebral artery on end-tidal carbon dioxide partial pressure -- a transcranial ultrasound Doppler study. J Cereb Blood Flow Metab 4: , McGraw CP: A cerebral perfusion pressure greater than 80 Hg is more beneficial, in Hoff JT, Betz AL (eds): Intraeranial Pressure VII. Berlin: Springer-Verlag, 1989, pp Michenfelder JD: The interdependency of cerebral function and metabolic effects following massive doses of thiopental in the dog. Anesthesiology 41: , Miller JD, Stanek AE, Langfitt TW: Concepts of cerebral perfusion pressure and vascular compression during intracranial hypertension. Prog Brain Res 34: , Muizelaar JP, Ward JD, Marmarou A, et al: Cerebral blood flow and metabolism in severely head-injured children. Part 2: Autoregulation. J Neurosurg 71:72-76, Obrist WD, Langfitt TW, Jaggi JL, el al: Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 61: , Paulson OB, Strandgaard S, Edvinsson L: Cerebral autoregulation. Cerebrovasc Brain Metab Rev 2: , Piper IR, Lawson A, Dearden NM, et al: Computerized data collection. Br J Intens Care 1:73-78, Risberg J, Ancri D, Ingvar DH: Regional cerebral blood 60 J. Neurosurg. / Volume 77 ~July, 1992

7 Cerebral flow velocity and perfusion pressure volume changes related to blood flow variations. Seand J Lab Clin Invest Suppl 102:XIC, Robertson CS, Narayan RK, Gokaslan ZL, et at: Cerebral arteriovenous oxygen difference as an estimate of cerebral blood flow in comatose patients. J Neurosurg 70: , Rowan JO, Harper AM, Miller JD, et al: Relationship between volume flow and velocity in the cerebral circulation. J Neurol Neurosurg Psychiatry 33: , Shenkin HA, Harmel MH, Kety SS: Dynamic anatomy of the cerebral circulation. Arch Neurol Psychiatry 60: , Shenkin HA, Spitz EB, Grant FC, et al: Physiologic studies of arteriovenous anomalies of the brain. J Neurosurg 5: , I. Williams PL, Warwick R, Dyson M, el al: Gray's Anat- omy, ed 37. New York: Churchill Livingstone, 1989, pp Manuscript received July 8, Accepted in final form December 6, This study was supported by Medical Research Council Special Project Grant Mr. Chan was supported by the Croucher Foundation Research Fellowship and the University of Hong Kong. Address for Mr. Chan: Division of Surgical Neurology, Department of Surgery, University of Hong Kong, Queen Mary Hospital, Hong Kong. This study will be submitted to the University of Hong Kong for consideration of the degree of Master of Surgery. Address reprint requests to: J. Douglas Miller, M.D., Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Edinburgh EH4 2XU, Scotland. J. Neurosurg. / Volume 77~July,