Transcranial Doppler Ultrasonography

NEUROSURGICAL INTENSIVE CARE 1042-3680/94 $0.00 + .20

TRANSCRANIAL DOPPLER ULTRASONOGRAPHY

David W. Newell, MD

BASIC PRINCIPLES OF of the examination techniques have been pub TRANSCRANIAL DOPPLER lished elsewhere and describe the various ULTRASONOGRAPHY methods for obtaining signals from different intracranial arteriesY Various techniques are Transcranial Doppler ultrasonography available for identification of the individual (TCD) is a method introduced by Aaslid et intracranial arteries. The most common is the 1 5 al • to record blood flow velocity in the basal hand-held technique, which utilizes the ex cerebral arteries in humans. Early investiga aminer's knowledge of the vascular anatomy tors using ultrasound to measure blood flow and vascular interrelationships as well as the concluded that recording from the intracranial characteristic signal generated by the various arteries would be impossible due to the barri arteries examined. 1• 15 Additional techniques ers imposed by the human skull. It was, how include vessel mapping as well as transcranial ever, possible to evaluate the cerebral circula color-coded real-time ultrasonography (color 5 10 tion by recording from the carotid arteries in flow) techniques. • the neck, and this application has been refined Transcranial Doppler study makes use of extensively and widely used in clinical medi two important principles to measure patho cine. The groundwork for the development logical changes in the cerebral circulation. of transcranial applications was made possible First, under conditions of constant flow, the by improvements in Doppler equipment velocity through a vessel will increase in pro along with the use of a relatively low fre portion to the decrease in cross-sectional area quency (2 MHz) and the introduction of spec produced by vessel narrowing. Second, in the tral analysis and pulsed range-gated Doppler. absence of any significant changes in vessel Initial recordings were made of the middle diameter, changes in flow will be directly pro cerebral artery through the transtemporal portional to changes in the average cross window. It was subsequently recognized that sectional velocity. Changes in the maximal a more complete examination of the basal ce flow velocity (V max or spectral outline) will be rebral arteries could be obtained by recording proportional to changes in average velocity in through the orbit via the transorbital route the absence of any turbulence or altered flow and also through the foramen magnum via patterns in the vessel. 6 It is also essential to the transforaminal route. Utilizing these three maintain a fixed probe position in relation to windows, extensive examination of the basal the artery during any calculation of relative cerebral arteries can be accomplished. Details flow changes.

From the Department of Neurological Surgery, University of Washington, Seattle, Washington

NEUROSURGERY CLINICS OF NORTH AMERICA

VOLUME 5 • NUMBER 4 • OCTOBER 1994 619 620 NEWELL

2 54 55 The velocity signal in the basal cranial arter neurosurgeons. • • The subsequent ex ies, particularly in the middle cerebral artery, tended applications of TCD have made it of can therefore be monitored and used to indi interest also to anesthesiologists, intensivists, cate relative flow changes under a variety of and vascular surgeons. Vascular technology circumstances. 57 This information can be used training programs are now including TCD in the intensive care unit setting to assess cere methodology in their curricula. Many vascu lar laboratories in the United States now offer bral reactivity in response to C02 changes, blood pressure changes, and changes in intra TCD as a clinical service. We established a cranial pressure (ICP), as well as during vari transcranial Doppler laboratory in the Depart ous spontaneous waves in the ICP and in ment of Neurological Surgery at the Univer sity of response to certain medications. TCD moni Washington in 1986 and have found it essential to have highly skilled personnel with toring can therefore be used to accomplish a background in vascular ultrasound who are many of the objectives that previously re trained to perform the examinations. TCD is quired more cumbersome techniques for mea a difficult vascular examination and requires surement of cerebral blood flow (CBF). Mea training and practice. Maintenance of a high surement of CBF following head injury has quality laboratory requires a high volume of been a subject of intense investigation. After examinations as well as constant feedback in the collection of much data, Langfitt and the form of correlation with clinical condition 3 Obrist2 defined some of the potential clinical and radiographic studies including angio applications for CBF studies as follows: graphy. 1. To ensure sufficient CBF to meet meta Some of the requirements for obtaining con bolic demands of the brain. tinuous monitoring tracings were incorpo 2. To aid in predicting clinical outcome. rated into the first transcranial Doppler design 3. To assess vasoreactivity in the cerebral by Aaslid. 5 These features included a monitor circulation by testing autoregulation and ing headband to ensure probe stability as well as the ability to sample spectral information C02 reactivity. 4. To evaluate the effect of various thera and the spectral outline as analog informa pies for management of ICP. tion. A variety of devices can be used to record Transcranial Doppler monitoring cannot give quantitative blood flow data and, there fore, cannot be used for the first two applica tions mentioned. The ability to calculate rela tive blood flow changes in response to C02 as well as perfusion pressure changes and the ability to calculate relative blood flow changes to various therapeutic regimens enables TCD to be used for the third and fourth applications mentioned. 42 Advantages over the radioactive xenon CBF method include the fact that TCD provides continuous relative blood flow infor mation, can be performed relatively easily us ing portable equipment in the intensive care Spectrum Outline unit, and does not involve the use of radioac 100 tive isotopes.

REQUIREMENT FOR 0 EXAMINATIONS AND MONITORING A The development of TCD has been a natural Figure 1A. Recording from the middle cerebral artery extension of vascular ultrasound. The initial that yields a spectral tracing. The spectral tracing can application of TCD to detect cerebral vaso then be assigned a spectral outline that can be used spasm following subarachnoid hemorrhage for monitoring relative changes in blood flow. made it of much interest to neurologists and Illustration continued on opposite page TRANSCRANIAL DOPPLER ULTRASONOGRAPHY 621

continuously either the spectral signal or the as the audio portion using a video recorder. analog signal from the spectral outline. 30• 42 To record relative blood flow changes, the Different recording devices are needed for spectral outline which can be optionally as various applications. For example, when per signed to the spectral signal can be recorded forming monitoring of emboli, it is most use ful to record the entire spectral signal as well using a variety of devices including magnetic tape and strip charts or by converting it to a

150

100

-ABP

0 5 10 8 Seconds

150

150 200 250 300 c Seconds

Figure 18. Simultaneous recordings of arterial blood pressure, bilateral middle cerebral artery velocity, intracranial pressure, and end tidal C02 are illustrated. C. Recording of similar signals as in 8 as a longer trend. ABP = arterial blood pressure; MCA = middle cerebral artery; LMCA = left middle cerebral artery; RMCA = right middle cerebral artery; ICP = intracranial pressure; ETC0 = end tidal C0 • 2 2 622 NEWELL digital signal and using computerized storage setting and that may have an impact on TCD devices. The last option has been refined and velocity values. is now being incorporated into TCD equip ment that offers software for continuous mon itoring. Computerized storage of digitized sig VASOREACTIVITY nals offers trend analyses as well as software to analyze various short intervals that are The ability of the cerebral circulation to un commonly used for reactivity testing. The de dergo vasoconstriction and vasodilatation in velopment of multi-channel Doppler ultraso response to various stimuli has been termed nography has allowed continuous recording vasoreactivity. Many of the currently used of multiple channels, most commonly, both treatments for lowering ICP, such as hyper middle cerebral arteries. This feature allows ventilation, mannitol therapy, and barbiturate simultaneous testing of both hemispheres therapy, rely on intact vasoreactivity for their during reactivity testing (Fig. 1). action.11. 12,32-35,45 Conditions leading to hospi talization in the neurosurgical intensive care unit such as subarachnoid hemorrhage, intra PHYSIOLOGIC FACTORS cranial hemorrhage, or head injury can impair AFFECTING BLOOD vasoreactivity and therefore make it difficult FLOW VELOCITY to control ICP. By monitoring relative ch. nges in CBF velocity, TCD can be used to evaluate Normative data on blood flow velocities in C02 reactivity, autoregulation, and responses 16 42 the basal intracranial arteries of humans have to various medications. • 18• 30• 37, -44 been collected by various investigators and Transcranial Doppler ultrasonography can summarized by Adams et aU Some of the be used to calculate C02 reactivity in the inten variables affecting velocities have included sive care unit patient by monitoring the blood age, gender, hematocrit, and metabolic fac flow velocity in the middle cerebral arteries tors such as C02 and oxygen concentration. while inducing a brief period of hypocapnia Physiologic factors include physiologic blood (usually 2 to 3 minutes by manual hyperventi flow fluctuations as well as ICP, blood pres lation with a ventilation bag). During changes sure, and the state of cerebral autoregulation. in C02 concentration, major changes in cereb Table 1 summarizes some of the factors that rovascular resistance are produced at the mi may have a role in the intensive care unit crocirculatory level, and there is little or no change in the diameter of the main trunk 17 21 of the middle cerebral artery. • This has been documented by several studies in hu Table 1. CONDITIONS IN NEUROINTENSIVE CARE 17 21 INFLUENCING BLOOD FLOW VELOCITY mans, • and C02 reactivity measured using TCD correlates well with C02 reactivity calcu Condition Effect on Flow Velocity lated with other blood flow methods.9. 57 The t Age ~ relative change in velocity through the middle ~ Hematocrit t cerebral artery in response to C0 changes ~ Cerebral metabolism ~ 2 ~ PC02 ~ is calculated on a percentage basis, and this ~ 02 t appears to correlate with relative changes in I f Intracranial pressure No change between CPP 9 57 i 50-150 with intact CBF. · The reactivity is expressed as percent ( autoregulation change in CBF per mm C02 • Normal C02 reac ~ Below CPP of 50 3% c With abolished tivity in humans is approximately change ~ 35 iJ autoregulation per mm C02• There are three potential uses Mannitol administration f Greater increase with t; of calculating C02 reactivity in intensive care impaired patients. h autoregulation n Barbiturate administration ~ First, if hyperventilation is to be used for Spontaneous fluctuations, Velocity fluctuates ICP control, knowledge of the C0 reactivity ti B waves synchronously with ICP B 2 b waves may be helpful. Hyperventilation for ICP con A waves ~ At peak of wave trol has been the subject of recent controversy. ti Hyperventilation in the acute setting with t< f = increase; ~ = decrease; PC0 = partial pressure of 2 head injury as well as other conditions leading tl carbon dioxide; 0 2 = oxygen; CPP = cerebral perfusion pres sure; ICP = intracranial pressure. to high ICP has been standard therapy in most S] TRANSCRANIAL DOPPLER ULTRASONOGRAPHY 623

emergency rooms and trauma centers. More changes, 37, 42 induced blood pressure prolonged hyperventilation for ICP control changes, 3, 4, 44 or transient reduction in local may be unsound on a theoretical basis, and, perfusion pressure by carotid compression. 16 in addition, there is evidence to suggest that All of these methods make use of the principle it is not of benefit in head-injured patients. 36 that relative velocity changes through the Muizelaar et al34 have shown that the vasocon middle cerebral artery reflect relative flow striction induced by acute hypocapnia is usu changes through this vessel in response to ally reversed within 24 hours, and, therefore, changing CPP.4, 26,44 Aaslid et al3 introduced the reduction in ICP by this mechanism is not a method to lower the blood pressure tran sustained. A recent prospective study showed siently by rapidly deflating bilateral thigh that treatment with hyperventilation for 5 cuffs in observing the blood flow velocity re days in head-injured patients was not found sponse to rapid step change in arterial blood to be beneficial. 36 Shorter periods of hyper pressure (Fig. 2). Despite concerns to the con ventilation in the intensive care unit setting trary,22 the validity of this method has been may be appropriate for transport of patients confirmed experimentally in several groups of or for lowering ICP on an acute basis. If hyper human subjects. 4, 44 The evaluation of cerebral ventilation is to be considered for these pur autoregulation following head injury may poses, it may be beneficial to know the C02 have several clinical applications. reactivity in advance. In severe head injury, 1. Pressure autoregulation may have impli C02 reactivity can be abolished or severely cations in the ability to control ICP ade impaired, ll, 12, 13 and it would not be useful quately using various maneuvers. Mu to employ hyperventilation in this subset of izelaar et al33 have shown that mannitol patients. is less effective in reducing ICP when Second, C02 reactivity may correlate with cerebral autoregulation is impaired. prognosis in head-injured patients as well as 2. Increasing arterial blood pressure to im other patients with neurosurgical condi 13 prove CPP results in reduced ICP if the tions. , 14, 51, 52 Previous studies using CBF autoregulation is intact and may increase methods have shown that C0 reactivity can 2 ICP if autoregulation is defective.35 become impaired or absent in patients with 3. It may be useful to identify patients with severe brain damage, and C0 reactivity has 2 absent autoregulation to prevent ele been correlated with outcome. 52 Recent stud vated blood pressures, which may ag ies using TCD for the same purpose have con firmed this relationship.1s, 43, 51 gravate edema and cause secondary hemorrhages. This also may help the cli Third, C02 reactivity may correlate with the nician be more vigilant in preventing hy response to barbiturate therapy and may be potension, which may be induced by useful in predicting the effectiveness of this 45 secondary surgical procedures or other therapy. preventable causes. 53 Cerebral autoregulation refers to the ability 4. The effectiveness of strategies to im of the cerebral circulation to maintain a con prove autoregulation, such as hyperven stant blood flow during changes in cerebral tilation, can be assessed. 59 perfusion pressure (CPP). 24, 48 This mecha nism appears to be different from C02 reactiv ity and is believed to operate by a separate EVALUATION OF VASOSPASM control mechanism. Cerebral autoregulation can be evaluated in head-injured patients us One of the original applications of TCD was ing several different CBF methods. Most of in the detection of cerebral vasospasm. 2 Sub the data on impaired autoregulation following sequently, many studies have confirmed its head injury has been obtained using the xe validity and usefulness in detecting vaso non CBF method, which employs radioac spasm following subarachnoid hemor 13 33 47 tive xenon. ' , More recently, TCD has rhage.19, 27, S4, 55 In addition, vasospasm follow been utilized to evaluate cerebral autoregula ing head trauma has been recognized to be a tion. 3, 16, 37, 44 Several methods to measure au clinical entity. 25, 31 Vasospasm can cause dete toregulation have been described, including rioration and infarction following closed head the observation of velocity changes in re injury. 19, 42 Vasospasm following subarach sponse to spontaneous blood pressure noid hemorrhage most commonly affects the 624 NEWELL

ABP

50 L MCA Velocity

0 100 ABP /\['J'\/' 'v-..JJJ Absent Autoregulation

0 .0 "'~R MCA J " X , c n,.• 4 ' - :;o V e I OCI"t y

0 0 50 100 c Seconds

Figure 2A. CT scan of a patient with an acute subdural hematoma. Note early low density changes in the right hemisphere. 8, Postoperative CT scan 1 day later showing decompression of the subdural hematoma with some residual low density changes in the right hemisphere. C, Demonstration of cerebral autoregulation on the left and right side assessed by using rapid induced blood pressure drops. Upper tracing illustrates intact autoregulation on the left side. Lower tracing illustrates completely abolished autoregulation on the right side despite an adequate cerebral perfusion pressure. ASP = arterial blood pressure; LMCA = left side middle cerebral artery; RMCA = right side middle cerebral artery. basal cerebral vessels. 40 TCD is most useful in narrowing of cerebral vessels and elevation of detecting vasospasm of the middle cerebral TCD velocities have revealed close correla 4 27 39 56 56 artery and distal internal carotid artery, but it tions. • • • Sloan et al revealed an 84% is also useful in evaluating other intracranial sensitivity and 89% specificity when compar arteries. Correlations between angiographic ing angiography with TCD in the middle cere- TRANSCRANIAL DOPPLER ULTRASONOGRAPHY 625

7 bral artery. Lindegaa~d et aF have reported found this method useful in detecting vaso a close correlation between elevated velocities spasm in head-injured patients.31, 58 and angiographic vasospasm. Their study re vealed that TCD had a sensitivity of 94% and a specificity of 90% for detecting angiographic INTRACRANIAL PRESSURE spasm of the middle cerebral artery. Advan tages of intermittent monitoring of vasospasm Transcranial Doppler monitoring can be using TCD include the ability to follow the used to reflect various hemodynamic changes development and resolution of the spasm by in the cerebral circulation that occur with repeated examinations. Knowledge of the changes in ICP. Lundberg28 published an ex time course of the vasospasm can be helpful tensive review of his initial observations using in determining the risk of delayed ischemic 19 ICP monitoring in humans. Characteristic neurologic deficit. Vasospasm following fluctuations occurred in the ICP, and several subarachnoid hemorrhage may affect the dis waves were regularly observed in neurosurgi tal cerebral vasculature selectively in certain cal patients. Lundberg A waves or plateau isolated cases. Isolated distal vasospasm fol waves are characterized by a sudden increase lowing subarachnoid hemorrhage causing in ICP that may last from several minutes up neurologic deficit is uncommon, however, 40 to 20 minutes followed by an abrupt decline. and frequently is associated with distally lo Subsequent investigators have documented cated blood on the original CT scan. that during Lundberg A waves or plateau When using TCD to evaluate head-injured waves, there is a marked increase in cere patients for vasospasm, it is recommended bral blood volume as well as a decrease in that velocity recordings also be obtained from 29 49 CBF. , It is believed that these waves result the distal internal carotid artery in the neck from an instability in the cerebral vascular as described by Aaslid et al4 and Lindegaard 27 control mechanism that leads to a positive et al. By constructing a ratio of the recordings feedback cycle under conditions of reduced in the middle cerebral artery versus the inter intracranial compliance. 46, 5° Cerebral vaso nal carotid artery, velocity correlations can be dilatation recorded during Lundberg A made for hyperemia, which is common in waves reveals a marked decrease in velocity head-injured patients. Several studies have and increase in pulsatility index (systolic

100

A Wave

0 10 20 30 40 Minutes

Figure 3. Illustration of multimodality recording of an intracranial pressure A wave showing marked increase in intracranial pressure and reduction in middle cerebral velocity during the A wave. ABP = arterial blood pressure; MCV = middle cerebral velocity; ICP = intracranial pressure. 626 NEWELL

100 B Waves

ABP

50 MCU

ICP

C02

0 50 100 150 200 250 Seconds

Figure 4. Illustration of multimodality recording during B waves indicating synchronous fluctuations of the middle cerebral velocity with intracranial pressure B waves. Note that there are no significant changes in end tidal C02 in this ventilated patient and no significant changes in arterial blood pressure. ABP = arterial blood pressure; MCV = middle cerebral velocity; ICP = intracranial pressure. velocity - diastolic velocity/mean velocity) frequency between 0.5 and 2 per minute, during the peak of the wave (Fig. 3). The most which he termed B waves. 28 B waves have been likely explanation is that the middle cerebral observed in conditions associated with re velocity is reflecting a decrease in CBF duced intracranial compliance and are be through the middle cerebral trunk. Following lieved to be the result of vasomotor waves of 8 41 the cessation of the wave, a hyperemic phase the small regulating vessels. • We observed can often be seen, characterized by an increase synchronous velocity fluctuations in the mid in velocity above the values recorded immedi dle cerebral artery that were in phase with ately preceding the wave. Lundberg also de the ICP B waves in a series of head-injured scribed smaller amplitude periodic fluctua patients41 (Fig. 4). Velocity waves of similar tions in the ICP that usually occurred at a frequency also were present in a cohort of

100

("v·.·v·,;·· ...... r·!v""'-~\'\_,,1\.l I Cp tW'-'"'-""~H'i"IW"""'"""""''""-~' ~' 0-

0 50 100 150 200 250 Seconds

Figure 5. Illustration of the effect of increasing intracranial pressure (decreased cerebral perfusion pressure) on the middle cerebral artery velocity in a patient with impaired autoregulation. Note the concordance of the arterial blood pressure tracing and middle cerebral velocity tracing previous to tlie test. There is a marked decrease in middle cerebral velocity induced by the increase in intracranial pressure. ABP = arterial blood pressure; MCV = middle cerebral velocity; ICP = intracranial pressure. TRANSCRANIAL DOPPLER ULTRASONOGRAPHY 627

normal volunteers, indicating that they may 'be a common phenomenon. It appears that BFV/~J the B waves in the ICP associated with velocity fluctuations usually become amplified in the ICP tracing in states of reduced intracranial compliance. 41

CPP~ ALTERATIONS IN VELOCITY WITH INCREASED INTRACRANIAL PRESSURE

Characteristic changes in the velocity in the middle cerebral artery have been recorded Figure 6. Illustration of the effect of decreasing cerebral 20 perfusion pressure on the middle cerebral artery veloc with increased ICP. The changes seen will ity waveform. Note that with progressive decreases in be dependent, in part, on the level of the CPP cerebral perfusion pressure, a marked decrease in dia and also the state of the cerebral autoregula stolic velocity occurs, as does a slight increase in sys tion. In patients with intact cerebral autoregu tolic velocity. With progressive decreases in cerebral perfusion pressure, the diastolic velocity becomes re lation, increases in ICP will cause an increase versed and progresses to an oscillating pattern. With in the pulsatility index and no change in the further compromise of the circulation, only small sys mean flow velocity if the CPP remains in the tolic peaks are eventually seen which indicate a very autoregulatory range (50 to 150 mm Hg). Fur short burst of antegrade movement of blood in ther increases in ICP which reduce the CPP the vessel in systole. BFV = blood flow velocity; CPP = cerebral perfusion pressure. below the range of autoregulation will cause a further increase in pulsatility index as well as a decrease in mean velocity. In patients with absent autoregulation, any decrease in

2.5 80

2.0 60 X Q) ""0 OJ c: J: "E -~ 1.5 .§. ~ 40 en ~ :J :; en Q_ en ~ Q_ 1.0 2.()

0.5 -+------.------.------..------L.., 0 Before A During Intracranial Pressure Wave

Figure 7A. This figure indicates the increase in pulsatility index seen with marked increases in intracranial pressure as seen during an intracranial pressure A wave. Note the marked increase in pulsatility index seen with increases in intracranial pressure.

Illustration continued on following page 628 NEWELL

''l1.5 ~ 20 Cl X J: Q) "0 E -c .s ~ 1.0 15 ::;Q) (/) o; (/) !!2 ::> ~ a.. a.. 10

0.5 5

0.0 0 Before After B Hyperventilation

2.0- .... 25

20 1.5

Cl X J: Q) 15 "0 E E .s -~ ~ ::> (/) o; (/) (/) :; 10 ~ a.. - a..

0.5 5

0.0 -'------r----...... ------r---...... 0 Before After c Barbiturate Administration

Figures 78 and 7C. Effect of hyperventilation in barbiturate administration on pulsatility index and intracranial pressure. B, After hyperventilation, there is a marked increase in pulsatility index; however, there is a decrease in intracranial pressure in contrast with the response seen in Figure 7A. C, Increase in pulsatility index seen after barbiturate administration and simultaneous decrease in intracranial pressure. B and C illustrate the effect of increased vascular tone induced by intrinsic vascular factors which increase the pulsatility index and decrease intracranial pressure. These figures demonstrate that pulsatility index alone is a nonspecific indicator of increased intracranial pressure and cannot be used to reliably predict ICP. Triangles = intracranial pressure; ovals = pulsatility index. TRANSCRANIAL DOPPLER ULTRASONOGRAPHY 629

CPP will decrease the velocity in the middle levels. Figure 7 shows an example of the cerebral artery (Fig. 5). change in the pulsatility index that can be Characteristic changes in the velocity wave affected by an increase in ICP, hyperventila form have been described in patients who tion, and barbiturate administration. Future have progressed to brain death, usually from work will concentrate on the relationship of conditions which cause marked increases in the vascular tone and critical closing pressure ICP. 20 These initial changes are characterized of the cerebral vessels to ICP.6 An index de by marked increases in the pulsatility index, rived from these measurements may be much with a decrease in diastolic velocity of the ve more specific in predicting levels of ICP. locity waveform. The diastolic velocity then reaches zero, and, subsequently, the flow of velocity becomes reversed in diastole, eventu SUMMARY ally progressing to an oscillating waveform with a pendular flow and finally to a pattern Transcranial Doppler ultrasonography is an characterized by small sharp systolic peaks. extremely useful adjunct in neurosurgical in Finally, no velocity recording can be obtained tensive care. Continuous improvements in when the blood fails to fill the basilar intra TCD equipment as well as computer software cranial arteries (Fig. 6). These characteristic have improved examination success and also changes have been correlated with cerebral vessel identification. Recent expanding appli angiography as well as with radioisotope cations of TCD have also allowed the study of 20 38 disorders of control of the cerebral circulation. blood flow measurements. • Recognition of these characteristic patterns can be useful in TCD can be used to detect vessel narrowing identifying patients who have progressed to from a variety of causes, including vaso brain death. spasm, and also can be used to detect cerebral Changes in pulsatility alone, however, can emboli and to evaluate C02 reactivity, auto not be used to predict ICP. The pulsatility regulation, and the response to certain medi index is an attempt to evaluate the resistance cations, as well as to indicate progressive ob in the distal cerebral vessels and increases struction of the cerebral circulation as seen with progressive obstruction to blood flow. in conditions leading to brain death. In the The pulsatility index can be unreliable, how future, TCD may offer the ability to estimate ever, due to the fact that it can be very heavily the ICP using noninvasive means by evaluat influenced by cardiac factors and can change ing velocity in the middle cerebral artery and dramatically due to changes in contractility arterial blood pressure tracings. The noninva with no change in the cerebrovascular resis sive determination of cerebral autoregulation tance.6 A more important confounding factor may be useful in evaluating strategies to im is that the pulsatility index can be increased by prove cerebral autoregulation as well as aid increased vascular resistance due to increased in the optimal management of ICP control and vascular tone of the cerebral vessels produced preservation of optimal cerebral circulation. by vasoconstriction. It also can increase by increased vascular resistance due to extrinsic References compression of the vessels as seen in high ICP. For this reason, increases in pulsatility 1. Aaslid R, Markwalder T-M, Nornes H: Noninvasive index alone are nonspecific and are not neces transcranial Doppler ultrasound recording of flow sarily indicative of high ICP. One of the causes velocity in basal cerebral arteries. J Neurosurg 57:769-774, 1982 of increased vascular resistance in head 2. Aaslid R, Huber R, Nornes H: Evaluation of cerebro injured patients is hyperventilation. Another vascular spasm with transcranial Doppler ultra cause of increased vascular resistance is low sound. J Neurosurg 60:37-41, 1984 cerebral metabolism, which can occur in head 3. Aaslid R, Lindegaard K-F, Sorteberg W, eta!: Cere injured patients and also following the admin bral autoregulation dynamics in humans. Stroke 20:1005-1011, 1989 istration of barbiturates. Both hyperventila 4. Aaslid R, Newell OW, Stooss R, eta!: Simultaneous tion and barbiturate administration increase arterial and venous transcranial Doppler assessment the pulsatility index and decrease ICP. The of cerebral autoregulation dynamics. Stroke 22:1148- 1154, 1991 pulsatility index is therefore nonspecific and 5. Aaslid R: Developments and principles of transcra cannot be used to predict ICP in head-injured nial Doppler. In Newell OW, Aaslid R (eds): Transcra patients unless the ICP is at extremely high nial Doppler. New York, Raven Press, 1992, pp 1-8 630 NEWELL

6. Aaslid R: Cerebral hemodynamics. In Newell OW, 25. Lewis DH, Eskridge JM, Newell OW, et al: Single Aaslid R (eds): Transcranial Doppler. New York, Ra photon emission computed tomography, transcra ven Press, 1992, pp 49-55 nial Doppler ultrasound, and cerebral angioplasty 7. Adams RJ, Nichols FT, Hess DC: Normal values and for posttraumatic vasospasm. J Neuroimaging physiological variables. In Newell OW, Aaslid R 3:252-254, 1993 (eds): Transcranial Doppler. New York, Raven Press, 26. Lindegaard K-F, LundarT, Wiberg, J, eta!: Variations 1992, pp 41-48 in middle cerebral artery blood flow investigated with 8. Auer LM, Sayama I: Intracranial pressure oscillations noninvasive transcranial blood velocity measure (B-waves) caused by oscillations in cerebrovascular ments. Stroke 18:1025-1030, 1987 volume. Acta Neurochir 68:93-100, 1983 27. Lindegaard KF, Nornes H, Bakke SJ, eta!: Cerebral 9. Bishop CC, Powell S, Rutt D, et a!: Transcranial vasospasm diagnosis by means of angiography and Doppler measurement of middle cerebral artery blood velocity measurements. Acta Neurochir (Wien) blood flow velocity: A validation study. Stroke 100:12-24, 1989 28. Lundberg N: Continuous recording and control of 17:913-915, 1986 ventricular fluid pressure in neurosurgical practice. 10. Bogdahn U, Becker G, Winkler J, eta!: Transcranial Acta Psychiatr Scand (Suppl) 149:1-193, 1960 color-coded real-time sonography in adults. Stroke 29. Lundberg N, Cronqvist S, Kjallquist A: Clinical inves 21:1680-1688, 1990 tigation on interrelations between intracranial pres 11. Cold GE: Measurements of C02 reactivity and barbi sure and intracranial hemodynamics. Prog Brain Res turate reactivity in patients with severe head injury. 30:70-75, 1968 Acta Neurochir (Wien) 98:153-163, 1989 30. Lundar T, Lindegaard KF, Nornes H: Continuous 12. Cold GE, Jensen FT, Malmros R: The cerebrovascular recording of middle cerebral artery blood velocity in C02 reactivity during the acute phase of brain injury. clinical neurosurgery. Acta Neurochir (Wien) Acta Anaesth Scand 21:222-231, 1977 102:85-90, 1990 13. Envoldsen EM, Jensen FT: Autoregulation and C02 31. Martin N, Doberstein C, Zane C, eta!: Post-traumatic responses of cerebral blood flow in patients with vasospasm: Transcranial Doppler ultrasound, cere acute severe head injury. J Neurosurg 48:689-703, bral blood flow and angiographic findings. J Neuro 1978 surg 77:575-583, 1992 14. Envoldsen E: CBF in head injury. Acta Neurochir 32. Muizelaar JP, Wei EP, Kontos HA, et a!: Mannitol (Suppl) 36:133-136, 1986 causes compensatory cerebral vasoconstriction and 15. Fujioka KA, Douville CA: Anatomy and freehand vasodilation in response to blood viscosity changes. examination techniques. In Newell OW, Aaslid R J Neurosurg 59:822-828, 1983 (eds): Transcranial Doppler. New York, Raven Press, 33. Muizelaar JP, Lutz HA III, Becker DP, et a!: Effect 1992, pp 9-31 of mannitol on ICP and CBF and correlation with 16. Giller CA: A bedside test for cerebral autoregulation pressure autoregulation in severely head-injured pa using transcranial Doppler ultrasound. Acta Neu tients. J Neurosurg 61:700-706, 1984 rochir (Wien) 108:7-14, 1991 34. Muizelaar JP, van der Poe! HG, Li ZC, et a!: Pial 17. Giller CA, Bowman G, Dyer H, eta!: Cerebral artery arteriolar vessel diameter and C02 reactivity during diameters during changes in blood pressure and car prolonged hyperventilation in the rabbit. J Neurosurg bon dioxide during craniotomy. Neurosurgery 69:923-927, 1988 32:737-742, 1993 35. Muizelaar JP: Cerebral blood flow, cerebral blood vol 18. Grosset DG, Strebel S, Straiton J, eta!: Impaired car ume, and cerebral metabolism after severe head in bon dioxide reactivity predicts poor outcome in se jury. In Becker DP, Gudeman SK (eds): Textbook of vere head injury: A transcranial Doppler study. In Head Injury. Philadelphia, WB Saunders, 1989, pp Avezaat CJJ, van Eijndhoven JHM, Maas AIR, eta! 221-240 (eds): Intracranial Pressure VIII. Berlin, Heidelberg, 36. Muizelaar JP, Marmarou A, Ward JD, eta!: Adverse Springer-Verlag, 1993, pp 322-326 effects of prolonged hyperventilation in patients with 19. Grosset DG, Straiton J, McDonald I, et a!: Use of severe head injury: A randomized clinical trial. J Neu transcranial Doppler sonography to predict develop rosurg 75:731-739, 1991 ment of a delayed ischemic deficit after subarachnoid 37. Nelson RJ, Czosnyka M, Pickard JD, eta!: Experimen hemorrhage. J Neurosurg 78:183-187, 1993 tal aspects of cerebrospinal hemodynamics: The rela 20. Hassler W, Steinmetz H, Pirschel J: Transcranial tionship between blood flow velocity waveform and Doppler study of intracranial circulatory arrest. cerebral autoregulation. Neurosurgery 31:705-710, J Neurosurg 71:195-201, 1989 1992 21. Huber P, Haneda J: Effect of contrast material, hyper 38. Newell OW, Grady S, Sirotta P, eta!: Evaluation of capnia, hyperventilation, hypertonic glucose and pa brain death using transcranial Doppler. Neurosur paverine on the diameter of the cerebral arteries-an gery 24:509-513, 1989 giographic determination in man. Invest Radio! 39. Newell OW: Evaluation of vasospasm using transcra 2:17-32, 1967 nial Doppler. In Sano K, Takakura K, Kassell N, et 22. Kontos HA: Validity of cerebral arterial blood flow a! (eds): Cerebral Vasospasm. Proceedings of the IVth calculations from velocity measurements. Stroke International Conference on Cerebral Vasospasm, 20:1-3, 1989 Tokyo, 1990, pp 3-8 23. Langfitt TW, Obrist WD: Cerebral blood flow and 40. Newell OW, Grady MS, Eskridge JM, eta!: Distribu metabolism after intracranial trauma. Prog Neurol tion of angiographic vasospasm after subarachnoid Surg 10:14-48, 1981 hemorrhage: Implications for diagnosis by trans 24. Lassen NA: Cerebral blood flow and oxygen con cranial Doppler ultrasonography. Neurosurgery sumption in man. Physiol Rev 39:183-237, 1959 27:574-577, 1990 TRANSCRANIAL DOPPLER ULTRASONOGRAPHY 631

41. Newell DW, Aaslid R, Stooss R, eta!: The relationship 50. Rosner MJ, Becker DP: Origin and evolution of pla of blood flow velocity fluctuations to intracranial teau waves: Experimental observations and a theoret pressure B waves. J Neurosurg 76:415-421, 1992 ical model. J Neurosurg 60:312-324, 1984 42. Newell DW, Seiler RW, Aaslid R: Head injury and 51. Sander D, Klingelhofer J: Correlation between C02 cerebral circulatory arrest. In Newell DW, Aaslid R reactivity, ICP and outcome in severe cerebral dis (eds): Transcranial Doppler. New York, Raven Press, ease. J Cardiovasc Tech 9:261-262, 1990 1992, pp 109-121 52. Schalen W, Messeter R, Nordstrom CH: Cerebral va 43. Newell DW, Aaslid R, Stooss R, et a!: Evaluation of soreactivity and the prediction of outcome in severe closed head injury patients using transcranial Dop traumatic brain lesions. Acta Anaesthesia! Scand pler monitoring. In Avezaat CJJ, van Eijndhoven 35:113-122, 1991 JHM, Maas AIR, eta! (eds): Intracranial Pressure VIII. 53. Simard JM, Bellefleur M: Systemic arterial hyperten Berlin, Heidelberg, Springer-Verlag, 1993, pp sion in head trauma. Am J Cardiol63:32C-35C, 1989 309-312 54. Seiler RW, Grolimund P, Aaslid R, et a!: Cerebral 44. Newell DW, Aaslid R, Lam A, et al: Comparison vasospasm evaluated by transcranial ultrasound cor of flow and velocity during dynamic autoregulation related with clinical grade and CT-visualized suba testing in humans. Stroke 25:793-797, 1994 rachnoid hemorrhage. J Neurosurg 64:594-600, 1986 45. Nordstrom C-H, Messeter K, Sundbarg G, eta!: Cere 55. Seiler RW, Newell DW: Subarachnoid hemorrhage bral blood flow, vasoreactivity, and oxygen consump and vasospasm. In Newell DW, Aaslid R (eds): Trans tion during barbiturate therapy in severe traumatic cranial Doppler. New York, Raven Press, 1992, pp brain lesions. J Neurosurg 68:424-431, 1988 101-108 46. Nornes H, Magnes B, Aaslid R: Observations of intra 56. Sloan MA, Haley EC Jr, Kassel NF, eta!: Sensitivity cranial pressure plateau waves. In Lundberg N, Pon ten U, Brock M (eds): Intracranial Pressure 2. Berlin, and specificity of transcranial Doppler ultrasonogra Heidelberg, New York, Springer-Verlag, 1975, pp phy in the diagnosis of vasospasm following suba 421-426 rachnoid hemorrhage. Neurosurgery 25:1514-1518, 47. Obrist WD, Langfitt TW, Jaggi JL, et a!: Cerebral 1989 blood flow and metabolism in comatose patients with 57. Sorteberg W: Cerebral blood velocity and cerebral acute head injury: Relationship to intracranial hyper blood flow. In Newell DW, Aaslid R (eds): Transcra tension. J Neurosurg 61:241-253, 1984 nial Doppler. New York, Raven Press, 1992, pp 57-66 48. Paulson OB, Strandgaard S, Edvinsson L: Cerebral 58. Weber M, Grolimund P, Seiler RW: Evaluation of autoregulation. Cerebrovasc Brain Metab Rev post-traumatic cerebral blood flow velocities by trans 2:161-192, 1990 cranial Doppler ultrasonography. Neurosurgery 49. Risberg J, Lundberg N, Ingvar DH: Regional cerebral 27:106-112, 1990 blood volume during acute transient rises of the intra 59. Weber JP, Newell DW, Watson R, eta!: Improvement cranial pressure (plateau waves). J Neurosurg in dynamic cerebral autoregulation with hyperventi 31:303-310, 1969 lation in head injured patients. Stroke 23:476A, 1992

Address reprint requests to David W. Newell, MD Department of Neurological Surgery 325 Ninth Avenue ZA-86 Seattle, WA 98104