Surg Neurol 1992;38:433-6
433
Transcranial Doppler Waveform Differences in Hyperemic and Nonhyperemic Patients After Severe Head Injury K w a n - H o n C h a n , F . R . C . S . , N . M a r k D e a r d e n , B.Sc., F . F . A . R . C . S . , J. D o u g l a s M i l l e r , M . D . , P h . D . , F . R . C . S , , F . R . C . P . , S u s a n M i d g l e y , F . F . A . R . C . S . , a n d I a n R. P i p e r , B . S c . , P h . D . Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Edinburgh, Scotland, United Kingdom
Chan KH, Dearden NM, MillerJD, MidgleyS, Piper IR. Transcranial doppler waveformdifferencesin hyperemicand nonhyperemicpatients after severe head injury. Surg Neurol 1992;38:433-6. Although increased cerebral blood flow velocity is readily measured by transcranial doppler ultrasonography (TCD), the causes of the velocity elevation may differ. After severe head injury, increased blood flow velocity can develop both in patients with global hyperemia (suggestive of vasodilation) and in those without hyperemia (suggestive of vasospasm). The present study attempts to determine whether TCD can differentiate these two mechanisms of velocity increase. Fourteen severely brain-injured patients who developed increased middle cerebral artery blood flow velocity (timeaveraged mean velocity > 100 cm/s) were studied. Eight cases were nonhyperemic and six were hyperemic as defined by arterial-jugular venous oxygen content differences of more than 4 mL/dL and less than 4 mL/dL, respectively. The TCD waveform of all eight nonhyperemic cases showed a diastolic notch, which was absent in all six hyperemic patients (p = 0.00066). TCD waveform profile appears to provide a noninvasive means of differentiating at the bedside the two causes of increased flow velocity. If associated with raised intracranial pressure, these require different treatment. KEY WORDS: Cerebral blood flow velocity; Head injury; Transcranial doppler ultrasound
The recent introduction of the bedside transcranial doppler ultrasound (TCD) method of measurement of cerebral blood flow velocity (BFV) enables an important intracranial circulatory indicator to be monitored nonin-
Address reprint requests to: ProfessorJ.D. Miller, Department of Clinical Neurosciences, Universityof Edinburgh, Western General Hospital, Edinburgh,EH4 2XU, United Kingdom. ReceivedJanuary 27, 1992; accepted May 4, 1992.
© 1992by ElsevierSciencePublishingCo., Inc.
vasively. Since BFV is proportional to cerebral blood flow and inversely related to the caliber of the blood vessel being insonated, an increase in BFV may signify either increased cerebral blood flow caused by vasodilation or narrowing of the insonated vessel caused by vasospasm. After severe head injury, hyperemia as a result of cerebral vasodilation is an important cause of raised intracranial pressure [10]. Vasospasm of the intracranial arteries may cause brain ischemia [ 1]. Differentiation of these two conditions is important because intracranial pressure can be increased in both situations and because treatment of intracranial hypertension due to hyperemia involves agents that involve cerebral vasoconstriction, which is contraindicated when vasospasm is present. To correctly differentiate hyperemia from vasospasm requires invasive and multiple monitoring techniques, such as angiography and measurement of cerebral blood flow or jugular venous oxygen saturation (SJO2). In an earlier study on head-injured patients with increased intracranial BFV, we found that hyperemia was associated with bilateral increase in BFV in the insonated vessels and arterial-jugular venous oxygen content difference (AVDO2) below 4 mL/dL. In contrast, vasospasm was suggested by increased BFV usually limited to a single vessel in nonhyperemic cases with normal or high values of AVDO 2 [5]. The aim of this study was to determine whether TCD alone can differentiate hyperemic from nonhyperemic increases in intracranial BFV after severe head injury. Clinical Material and Methods Fourteen patients with severe brain injury, defined as postresuscitation admission Glasgow Coma Score of less than or equal to 8 points, who developed increased flow velocity (time-averaged mean velocity > 100 cm/s) on TCD were included in the study. There were 12 males and two females, with a mean age of 21 years (range 0090-3019/92/$5.00
434
Surg Neurol 1992;38:433-6
Chan et al
Table 1. Doppler Velocity Measurements and CerebralArterial-Jugular Venous Oxygen Content Differences in Nonhyperemic and Hyperemic Patients with Increased Flow Velocity
Nonhyperemic (n = 8) Hyperemic (n = 6)
AVD02(mL/dL)
Peak systolic velocity (cm/s)
End-diastolic velocity (cm/s)
Time-averaged
mean velocity (cm/s)
Pulsatility
6.1 (5.8-6.5)*
179 (165-192)
75 (64-86)
113 (104-122)
0.9 (0.8-1.0)
2.9 (2.4-3.3)
165 (152-179)
73 (63-82)
106 (100-112)
0.9 (0.8-1.0)
index
limits). Abbreviation:AVDO2,arterial-jugularvenousoxygencontentdifference. ° Mean (95% confidence
9 - 3 4 years). All patients were managed according to a standard regime that included artificial ventilation with muscle paralysis and sedation. They had continuous monitoring of arterial blood pressure, intracranial pressure, arterial oxygen saturation, end-tidal carbon dioxide concentration, body temperature, and SJO2 (Oximetrix 3, Abbott Laboratories, Chicago), as reported previously [2]. The AVDO2 was calculated by multiplying the difference between arterial oxygen saturation and SJO2 by the daily hemoglobin concentration plus 1.39, and dividing by 100. Global cerebral hyperemia was defined as AVDO2 of less than 4 mL/dL Transcranial doppler (Medasonics, Mountain View California) measurement of both middle cerebral artery (MCA) BFV was first performed within 24 hours of admission according to the method described by Aaslid et al [1]. Thereafter, it was repeated at least daily during the patients' entire hospital stay. Continuous TCD monitoring was performed during periods of increased intracranial pressure or decreased cerebral perfusion pressure. Velocity variables measured included the peak systolic, end-diastolic, and time-averaged mean velocities, from which the pulsatility index (peak systolic minus end-diastolic velocities divided hy time-averaged mean velocity) could be defined. TCD measurements were averaged over 15 cardiac cycles during periods of cerebral and hemodynamic stability, and arterial oxygen saturation and end-tidal carbon dioxide concentration were maintained above 95% and between 4 and 4.5 kPa, respectively. The coefficient of variation of repeated measurement was 5 %. The normal time-averaged velocity measured in our department is 60.1 cm/s (95% confidence limits 56.9-63.3 cm/s). A time-averaged mean velocity greater than 100 cm/s is considered abnormally high [5]. Statistical analyses were by Fisher exact and Wilcoxon tests. Results Of the 14 patients studied, eight were judged from AVDO2 to be nonhyperemic and six were found to have global hyperemia (Table 1). The mean values of the
velocity variables in the nonhyperemic and hyperemic patients are shown in Table 1. There were no significant differences between peak systolic, end-diastolic, and time-averaged velocities and pulsatility index between the two groups by the Wilcoxon test (p > 0.05). Although the velocity measurements were comparable in the two groups, the TCD velocity waveforms were distinctly different in the cases of nonhyperemic and hyperemic increases in BFV. A notch at diastole was observed in all eight nonhyperemic patients and was absent in all six hyperemic cases (Figure 1). The presence of a diastolic notch was significantly related to nonhyperemic increase in BFV (p = 0.00066). In three of the nonhyperemic cases, cerebral infarction developed in the territory of the MCA with abnormally high flow velocity. None of the hyperemic cases developed this complication. The relationships between increased TCD velocity and intracranial pressure, cerebral perfusion pressure, and time after injury have been reported previously [5]. Discussion Cerebral hyperemia can be defined independently by cerebral blood flow and AVDO2 measurements [10]. Various methods of cerebral blood flow measurement are prone to error and are difficult to perform by the bedside [3]. Retrograde withdrawal of blood by aspiration from the jugular bulb for measurement of SJO 2 may be contaminated by extracerebral sources, making AVDO2 calculation inaccurate [4]. Recent studies have shown that continuous recording of SJO2 by a fiberoptic catheter placed in the jugular venous bulb provides an accurate measurement of global SJO2, thereby eliminating the sources of error previously encountered [2,6]. The present study used this new method for classification of patients as hyperemic or nonhyperemic. Bilateral increases in MCA velocity are also associated with hyperemic increases in BFV, but not exclusively [5]. The present study suggests that TCD velocity waveform provides a better way of separating the two different types of elevated BFV. Increased BFV is also de-
Doppler Waveform in Head Injury
Surg Neurol 1992;38:433-6
435
Nonhyperaemic
cm/s
100
patients
1
2
3
4
5
6
7
8
Hyperaemic
cm/s
3..
100
= patients
1
3
Figure 1. Diagram of the doppler velocity waveforms of the nonhyperemic patients (showing the diastolic notch) and those of the hyperemic cases, showing a gradual fall in velocity during diastole without the presence of a notch.
scribed in arteriovenous malformations (AVM), in which there is vasodilation and increased cerebral blood flow, and in vasospasm after subarachnoid hemorrhage, when vessel narrowing is present [ 1,8]. The TCD pulsatility index is one way of assessing waveform shape. The pulsatility index has been found to be significantly higher in cases with vasospasm than in those with AVM [8]. However, there is a large degree of overlap in the noted pulsatility index values. As a result, pulsatility index cannot be used to differentiate the two conditions in individual patients. Our findings in head injury are in partial agreement with these results. This may be explained by the fact that only the peak systolic, end-diastolic, and time-averaged mean velocities are used in calculating the pulsatility index, without taking into consideration the whole shape of the doppler waveform. In a recent study, Fourier analysis was used to analyze the TCD waveforms of AVM and vasospasm after subarachnoid hemorrhage [9]. Patients with AVM had a significantly lower Fourier coefficient when compared with those with vasospasm. The degree of vascular distensibility of the cerebral vasculature, inferred from the shape of the Fourier analysis curve, was different in the two groups, suggesting that vasodilation and constriction, respectively, accounted for these differences. A diastolic notch in the doppler velocity waveform has previously been described in recordings from the human uterine artery. It was attributed to an increase in distal vascular resistance as a result of spasm of the downstream vessels [7]. Although the distal vasculature could not be insonated in our patients, spasm of the insonated portions of the vessel may produce a signifi-
4
time
6
cant increase in vascular resistance, with resultant formation of a diastolic notch. Although we did not have angiographic proof of the presence ofvasospasm in nonhyperemic patients in this study, the finding that ischemic infarctions were encountered in some of these cases and never in hyperaemic patients provided indirect support for this assumption [5]. In conclusion, TCD waveform provides a simple and noninvasive method of differentiating hyperemic from nonhyperemic increases in BFV at the bedside. Both conditions may be associated with raised intracranial pressure. This will have significance for the acute management of patients with severe head injury because the two underlying conditions, which require different treatment strategies, can be distinguished. This study was supported by the Medical Research Council special project grant (SPG 880 9197). K.H. Chan was supported by the Croucher Foundation Research Fellowship and the University of Hong Kong.
References 1. Aaslid R, Huber P, Nornes H. Evaluation of cerebrovascular spasm with transcranial doppler ultrasound. J Neurosurg 1984;60:37-41. 2. Andrews PJD, Dearden NM. Validation of the oximetrix III for continuous monitoring of jugular bulb oxygen saturation: comparison with IL282 in vitro co-oximeter. Br J Anaesth 1990;64:393-4. 3. Betz E: Cerebral blood flow: Its measurement and regulation. Physiol Rev 1972;52:595-630. 4. Bruce DA, Langfitt TW, Miller JD, Schutz H, Vapalahti MP, Stanek A, Goldberg HI: Regional cerebral blood flow, intracranial pressure and brain metabolism in comatose patients. J Neurosurg 1973;38:131-44. 5. Chan KH, Dearden NM, Miller JD: The significance of posttraumatic increase in cerebral blood flow velocity: a transcranial doppler ultrasound study. Neurosurgery 1992;30:697-700.
436
Surg Neurol 1992;38:433-6
6. Chan KH, Miller JD, Deacden NM, Andrews PJD, Midgley S: the effect of changes in cerebral perfusion pressure on middle cerebral artery blood flow velocity and jugular bulb venous oxygen saturation. J Neurosurg 1992;77:55-61. 7. Fleischer A, Schulman H, Farmakides G, Bracero L, Grunfeld L, Rochelson B, Koenigsberg M: Uterine artery doppler velocimetry in pregnant women with hypertension. Am J Obstet Gynecol 1986;4:806-14.
Chan et al
8. Giller CA, Hodges K, Bat jer HH: Transcranial doppler pulsatility in vasodilation and stenosis. J Neurosurg 1990;72:901-6. 9. Njemanze PC, Beck OJ, Gomez CR, Horenstein S: Fourier analysis of the cerebrovascular system. Stroke 1991;22:721-6. 10. Obrist WD, LangfittTW, Jaggi JL, Cruz J, GennarelliT: Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 1984;61:241-53.
433
Transcranial Doppler Waveform Differences in Hyperemic and Nonhyperemic Patients After Severe Head Injury K w a n - H o n C h a n , F . R . C . S . , N . M a r k D e a r d e n , B.Sc., F . F . A . R . C . S . , J. D o u g l a s M i l l e r , M . D . , P h . D . , F . R . C . S , , F . R . C . P . , S u s a n M i d g l e y , F . F . A . R . C . S . , a n d I a n R. P i p e r , B . S c . , P h . D . Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Edinburgh, Scotland, United Kingdom
Chan KH, Dearden NM, MillerJD, MidgleyS, Piper IR. Transcranial doppler waveformdifferencesin hyperemicand nonhyperemicpatients after severe head injury. Surg Neurol 1992;38:433-6. Although increased cerebral blood flow velocity is readily measured by transcranial doppler ultrasonography (TCD), the causes of the velocity elevation may differ. After severe head injury, increased blood flow velocity can develop both in patients with global hyperemia (suggestive of vasodilation) and in those without hyperemia (suggestive of vasospasm). The present study attempts to determine whether TCD can differentiate these two mechanisms of velocity increase. Fourteen severely brain-injured patients who developed increased middle cerebral artery blood flow velocity (timeaveraged mean velocity > 100 cm/s) were studied. Eight cases were nonhyperemic and six were hyperemic as defined by arterial-jugular venous oxygen content differences of more than 4 mL/dL and less than 4 mL/dL, respectively. The TCD waveform of all eight nonhyperemic cases showed a diastolic notch, which was absent in all six hyperemic patients (p = 0.00066). TCD waveform profile appears to provide a noninvasive means of differentiating at the bedside the two causes of increased flow velocity. If associated with raised intracranial pressure, these require different treatment. KEY WORDS: Cerebral blood flow velocity; Head injury; Transcranial doppler ultrasound
The recent introduction of the bedside transcranial doppler ultrasound (TCD) method of measurement of cerebral blood flow velocity (BFV) enables an important intracranial circulatory indicator to be monitored nonin-
Address reprint requests to: ProfessorJ.D. Miller, Department of Clinical Neurosciences, Universityof Edinburgh, Western General Hospital, Edinburgh,EH4 2XU, United Kingdom. ReceivedJanuary 27, 1992; accepted May 4, 1992.
© 1992by ElsevierSciencePublishingCo., Inc.
vasively. Since BFV is proportional to cerebral blood flow and inversely related to the caliber of the blood vessel being insonated, an increase in BFV may signify either increased cerebral blood flow caused by vasodilation or narrowing of the insonated vessel caused by vasospasm. After severe head injury, hyperemia as a result of cerebral vasodilation is an important cause of raised intracranial pressure [10]. Vasospasm of the intracranial arteries may cause brain ischemia [ 1]. Differentiation of these two conditions is important because intracranial pressure can be increased in both situations and because treatment of intracranial hypertension due to hyperemia involves agents that involve cerebral vasoconstriction, which is contraindicated when vasospasm is present. To correctly differentiate hyperemia from vasospasm requires invasive and multiple monitoring techniques, such as angiography and measurement of cerebral blood flow or jugular venous oxygen saturation (SJO2). In an earlier study on head-injured patients with increased intracranial BFV, we found that hyperemia was associated with bilateral increase in BFV in the insonated vessels and arterial-jugular venous oxygen content difference (AVDO2) below 4 mL/dL. In contrast, vasospasm was suggested by increased BFV usually limited to a single vessel in nonhyperemic cases with normal or high values of AVDO 2 [5]. The aim of this study was to determine whether TCD alone can differentiate hyperemic from nonhyperemic increases in intracranial BFV after severe head injury. Clinical Material and Methods Fourteen patients with severe brain injury, defined as postresuscitation admission Glasgow Coma Score of less than or equal to 8 points, who developed increased flow velocity (time-averaged mean velocity > 100 cm/s) on TCD were included in the study. There were 12 males and two females, with a mean age of 21 years (range 0090-3019/92/$5.00
434
Surg Neurol 1992;38:433-6
Chan et al
Table 1. Doppler Velocity Measurements and CerebralArterial-Jugular Venous Oxygen Content Differences in Nonhyperemic and Hyperemic Patients with Increased Flow Velocity
Nonhyperemic (n = 8) Hyperemic (n = 6)
AVD02(mL/dL)
Peak systolic velocity (cm/s)
End-diastolic velocity (cm/s)
Time-averaged
mean velocity (cm/s)
Pulsatility
6.1 (5.8-6.5)*
179 (165-192)
75 (64-86)
113 (104-122)
0.9 (0.8-1.0)
2.9 (2.4-3.3)
165 (152-179)
73 (63-82)
106 (100-112)
0.9 (0.8-1.0)
index
limits). Abbreviation:AVDO2,arterial-jugularvenousoxygencontentdifference. ° Mean (95% confidence
9 - 3 4 years). All patients were managed according to a standard regime that included artificial ventilation with muscle paralysis and sedation. They had continuous monitoring of arterial blood pressure, intracranial pressure, arterial oxygen saturation, end-tidal carbon dioxide concentration, body temperature, and SJO2 (Oximetrix 3, Abbott Laboratories, Chicago), as reported previously [2]. The AVDO2 was calculated by multiplying the difference between arterial oxygen saturation and SJO2 by the daily hemoglobin concentration plus 1.39, and dividing by 100. Global cerebral hyperemia was defined as AVDO2 of less than 4 mL/dL Transcranial doppler (Medasonics, Mountain View California) measurement of both middle cerebral artery (MCA) BFV was first performed within 24 hours of admission according to the method described by Aaslid et al [1]. Thereafter, it was repeated at least daily during the patients' entire hospital stay. Continuous TCD monitoring was performed during periods of increased intracranial pressure or decreased cerebral perfusion pressure. Velocity variables measured included the peak systolic, end-diastolic, and time-averaged mean velocities, from which the pulsatility index (peak systolic minus end-diastolic velocities divided hy time-averaged mean velocity) could be defined. TCD measurements were averaged over 15 cardiac cycles during periods of cerebral and hemodynamic stability, and arterial oxygen saturation and end-tidal carbon dioxide concentration were maintained above 95% and between 4 and 4.5 kPa, respectively. The coefficient of variation of repeated measurement was 5 %. The normal time-averaged velocity measured in our department is 60.1 cm/s (95% confidence limits 56.9-63.3 cm/s). A time-averaged mean velocity greater than 100 cm/s is considered abnormally high [5]. Statistical analyses were by Fisher exact and Wilcoxon tests. Results Of the 14 patients studied, eight were judged from AVDO2 to be nonhyperemic and six were found to have global hyperemia (Table 1). The mean values of the
velocity variables in the nonhyperemic and hyperemic patients are shown in Table 1. There were no significant differences between peak systolic, end-diastolic, and time-averaged velocities and pulsatility index between the two groups by the Wilcoxon test (p > 0.05). Although the velocity measurements were comparable in the two groups, the TCD velocity waveforms were distinctly different in the cases of nonhyperemic and hyperemic increases in BFV. A notch at diastole was observed in all eight nonhyperemic patients and was absent in all six hyperemic cases (Figure 1). The presence of a diastolic notch was significantly related to nonhyperemic increase in BFV (p = 0.00066). In three of the nonhyperemic cases, cerebral infarction developed in the territory of the MCA with abnormally high flow velocity. None of the hyperemic cases developed this complication. The relationships between increased TCD velocity and intracranial pressure, cerebral perfusion pressure, and time after injury have been reported previously [5]. Discussion Cerebral hyperemia can be defined independently by cerebral blood flow and AVDO2 measurements [10]. Various methods of cerebral blood flow measurement are prone to error and are difficult to perform by the bedside [3]. Retrograde withdrawal of blood by aspiration from the jugular bulb for measurement of SJO 2 may be contaminated by extracerebral sources, making AVDO2 calculation inaccurate [4]. Recent studies have shown that continuous recording of SJO2 by a fiberoptic catheter placed in the jugular venous bulb provides an accurate measurement of global SJO2, thereby eliminating the sources of error previously encountered [2,6]. The present study used this new method for classification of patients as hyperemic or nonhyperemic. Bilateral increases in MCA velocity are also associated with hyperemic increases in BFV, but not exclusively [5]. The present study suggests that TCD velocity waveform provides a better way of separating the two different types of elevated BFV. Increased BFV is also de-
Doppler Waveform in Head Injury
Surg Neurol 1992;38:433-6
435
Nonhyperaemic
cm/s
100
patients
1
2
3
4
5
6
7
8
Hyperaemic
cm/s
3..
100
= patients
1
3
Figure 1. Diagram of the doppler velocity waveforms of the nonhyperemic patients (showing the diastolic notch) and those of the hyperemic cases, showing a gradual fall in velocity during diastole without the presence of a notch.
scribed in arteriovenous malformations (AVM), in which there is vasodilation and increased cerebral blood flow, and in vasospasm after subarachnoid hemorrhage, when vessel narrowing is present [ 1,8]. The TCD pulsatility index is one way of assessing waveform shape. The pulsatility index has been found to be significantly higher in cases with vasospasm than in those with AVM [8]. However, there is a large degree of overlap in the noted pulsatility index values. As a result, pulsatility index cannot be used to differentiate the two conditions in individual patients. Our findings in head injury are in partial agreement with these results. This may be explained by the fact that only the peak systolic, end-diastolic, and time-averaged mean velocities are used in calculating the pulsatility index, without taking into consideration the whole shape of the doppler waveform. In a recent study, Fourier analysis was used to analyze the TCD waveforms of AVM and vasospasm after subarachnoid hemorrhage [9]. Patients with AVM had a significantly lower Fourier coefficient when compared with those with vasospasm. The degree of vascular distensibility of the cerebral vasculature, inferred from the shape of the Fourier analysis curve, was different in the two groups, suggesting that vasodilation and constriction, respectively, accounted for these differences. A diastolic notch in the doppler velocity waveform has previously been described in recordings from the human uterine artery. It was attributed to an increase in distal vascular resistance as a result of spasm of the downstream vessels [7]. Although the distal vasculature could not be insonated in our patients, spasm of the insonated portions of the vessel may produce a signifi-
4
time
6
cant increase in vascular resistance, with resultant formation of a diastolic notch. Although we did not have angiographic proof of the presence ofvasospasm in nonhyperemic patients in this study, the finding that ischemic infarctions were encountered in some of these cases and never in hyperaemic patients provided indirect support for this assumption [5]. In conclusion, TCD waveform provides a simple and noninvasive method of differentiating hyperemic from nonhyperemic increases in BFV at the bedside. Both conditions may be associated with raised intracranial pressure. This will have significance for the acute management of patients with severe head injury because the two underlying conditions, which require different treatment strategies, can be distinguished. This study was supported by the Medical Research Council special project grant (SPG 880 9197). K.H. Chan was supported by the Croucher Foundation Research Fellowship and the University of Hong Kong.
References 1. Aaslid R, Huber P, Nornes H. Evaluation of cerebrovascular spasm with transcranial doppler ultrasound. J Neurosurg 1984;60:37-41. 2. Andrews PJD, Dearden NM. Validation of the oximetrix III for continuous monitoring of jugular bulb oxygen saturation: comparison with IL282 in vitro co-oximeter. Br J Anaesth 1990;64:393-4. 3. Betz E: Cerebral blood flow: Its measurement and regulation. Physiol Rev 1972;52:595-630. 4. Bruce DA, Langfitt TW, Miller JD, Schutz H, Vapalahti MP, Stanek A, Goldberg HI: Regional cerebral blood flow, intracranial pressure and brain metabolism in comatose patients. J Neurosurg 1973;38:131-44. 5. Chan KH, Dearden NM, Miller JD: The significance of posttraumatic increase in cerebral blood flow velocity: a transcranial doppler ultrasound study. Neurosurgery 1992;30:697-700.
436
Surg Neurol 1992;38:433-6
6. Chan KH, Miller JD, Deacden NM, Andrews PJD, Midgley S: the effect of changes in cerebral perfusion pressure on middle cerebral artery blood flow velocity and jugular bulb venous oxygen saturation. J Neurosurg 1992;77:55-61. 7. Fleischer A, Schulman H, Farmakides G, Bracero L, Grunfeld L, Rochelson B, Koenigsberg M: Uterine artery doppler velocimetry in pregnant women with hypertension. Am J Obstet Gynecol 1986;4:806-14.
Chan et al
8. Giller CA, Hodges K, Bat jer HH: Transcranial doppler pulsatility in vasodilation and stenosis. J Neurosurg 1990;72:901-6. 9. Njemanze PC, Beck OJ, Gomez CR, Horenstein S: Fourier analysis of the cerebrovascular system. Stroke 1991;22:721-6. 10. Obrist WD, LangfittTW, Jaggi JL, Cruz J, GennarelliT: Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 1984;61:241-53.
