Transcranial Doppler Ultrasonography in Anesthesia and Neurosurgery

Introduced in 1982 by Aaslid and colleagues, transcranial Doppler (TCD) ultrasonography, sometimes termed a “stethoscope for brain,” has become one of the most useful methods of noninvasive examination of cerebral circulation. Provided that the limitations of this technology are recognized, information about cerebral hemodynamics can be obtained that can be used in the perioperative and intensive care of neurologically injured patients (those with head injury, subarachnoid hemorrhage, poor grade stroke, etc.) and in the prevention of neurologic insult in patients at risk for cerebral ischemia (those undergoing carotid endarterectomy, cardiopulmonary bypass, liver transplant, etc.). This chapter discusses the principles and limitations that govern the use of TCD and describes current and potential future applications of this reliable, noninvasive, continuous, although indirect measure of cerebral blood flow (CBF).

Principles of transcranial Doppler ultrasonography

TCD ultrasonography calculates the velocity of red blood cells (FV) flowing through the large vessels at the base of the brain by means of the Doppler principle. This principle, first described by Christian Doppler in 1843, relates to the shift in the frequency of any wave, including an ultrasound wave, when either the transmitter or the receiver is moving relative to the wave-propagating medium. The change in the frequency of emitted pulse of ultrasound reflected by red blood cells is proportional to FV. By convention, the shift in Doppler frequency is expressed in centimeters per second to allow comparison of readings from instruments that operate at different emission frequencies. The frequency best suited for TCD applications is on the order of 2 MHz.

A constant vessel diameter and an unchanged angle of insonation are the two main assumptions that govern the use of TCD as an indirect measure of CBF. The velocity detected by the probe as a fraction of the real velocity depends on the cosine of the angle insonation compared to a vector of FV (measured velocity = real velocity × cosine of angle of incidence). Therefore at 0 angle, the detected and true red cell velocities are equal (cosine of 0 = 1), whereas at 90 degrees, no detection of velocity is possible. Fortunately, the anatomic limitations of transtemporal insonation of the middle cerebral artery (MCA) are such that signal capture is possible only at narrow angles (< 30 degrees). Thus the detected velocity is a very close approximation of the true velocity (87–100%). Furthermore, as long as the angle of insonation is kept constant by fixing the probe in position (and vessel diameter remains constant during examination – see later comments), changes in the detected velocity closely reflect changes in the true velocity.

The other main factor that affects the interpretation of TCD measurements is the cross-sectional area of the insonated vessel. The volume passing through a particular segment of a vessel depends on the velocity of red cells and the diameter of the vessel. Therefore for velocity to be a true reflection of flow, the diameter of the vessel must not change significantly during the measurement period. Factors that may affect the diameter of vessels are arterial carbon dioxide tension (PaCO ₂ ), blood pressure, anesthetic agents, and vasoactive drugs. The basal cerebral arteries, being conductance vessels, do not dilate or constrict as the vascular resistance changes (smaller resistive vessel). It has been shown angiographically and through direct observation during brain surgery that change in PaCO ₂ , one of the most important determinants of cerebrovascular resistance (CVR), has no or very limited effect on the diameter of the basal arteries. Moreover, CO ₂ reactivity studies using TCD have demonstrated values similar to those obtained with conventional CBF measurements. Similarly, changes in blood pressure have negligible influence on the diameter of the proximal segments of the basal arteries. The effect of vasoactive drugs on cerebral conductance vessels is variable. Although sodium nitroprusside and phenylephrine do not significantly affect the proximal segments of the MCA, significant vasodilation occurs when nitroglycerine is administered to healthy volunteers.

The effect of anesthetic agents on the diameter of the basal vessels remains controversial. The intravenous agents are devoid of direct cerebrovascular effects, and it is accepted that these agents do not affect the diameter of the conductance vessels. The situation is less clear-cut with the inhalational agents, with most but not all the evidence suggesting that they have negligible effects on the diameter of the conductance vessels. It is generally accepted that during steady-state anesthetic conditions, changes in FV can be interpreted to mean corresponding changes in cortical CBF.

A factor that may affect the reliability of TCD measurements as true representations of CBF variation is the presence of intracranial pathology. Intracranial lesions and cerebral vasospasm have all been identified as factors that affect the accuracy of FV measurements, as a representative variable for changes in CBF.

Measurements using transcranial Doppler ultrasonography

The Examination

Three main pathways for accessing the intracranial arteries are (1) the transtemporal route through the thin bone above the zygomatic arch to the anterior, middle, and posterior cerebral arteries, (2) the transorbital approach to the carotid siphon, and (3) the suboccipital route to the basilar and vertebral arteries.

Although a complete diagnostic examination usually incorporates all three approaches, because the probe can be easily secured in position once a signal is obtained, intraoperative monitoring usually utilizes the transtemporal route. In expert hands, it is possible to transtemporally insonate the proximal segment (M1) of the MCA in more than 90% of people. The MCA carries about 60–70% of the ipsilateral carotid artery blood flow and can be regarded as representative of hemispheric CBF. However, because the successful transmission of ultrasound through the skull depends on the thickness of the skull, which varies with gender, race, and age, the failure rate can be as high as 10–30%. The theoretical risk of eye damage limits the use of the transorbital route, and the lack of suitable means to secure the probe in position makes the suboccipital route impractical.

Through the temporal window, the MCA, anterior cerebral artery (ACA), and posterior cerebral artery (PCA) can be readily examined. In each patient, the same insonation window should be used throughout the entire study period. This can be accomplished by putting a small marker at the patient’s temporal region. The TCD examination begins with the identification of the bifurcation of the intracranial portion of the internal carotid artery (ICA) into the MCA and ACA according to the method described by Aaslid. This bifurcation can usually be identified at a depth of 60–65 mm. The typical Doppler signal from the carotid bifurcation, which consists of images above and below the zero line of reference, represents the flow directions toward and away from the ultrasound probe of the MCA and ACA, respectively. The depth of insonation is then reduced to follow the upward deflection image of the MCA FV as the vessel runs toward the skull. The MCA can usually be traced up to a depth of 30 mm, which is beyond the bifurcation of the MCA into the peripheral branches. The proximal portion of the main trunk of the MCA (the M1 segment) can be located at a depth of around 45–55 mm. The depth that gives the highest velocity is usually chosen for measurement. In children, this depth is usually 10 mm less than that in adults, but the same principles apply. This method of obtaining the MCA signal eliminates the possibility of mistaking the PCA for the MCA, because for anatomic reasons, the PCA signal cannot be obtained at a depth less than 55 mm.

After the MCA signal is obtained, the depth of insonation is increased so that the image of the carotid bifurcation can be seen again. The depth is increased further with the probe directed slightly anteriorly so that the ACA image can be found. The first part of the ACA (the A1 segment) is recognized from a direction of flow away from the probe. With the identification of the ACA, the depth of insonation is decreased until the carotid bifurcation signal is obtained. The probe is then angled slightly posteriorly until the signal of the PCA is seen. The PCA can be distinguished from the MCA signal because it has a lower FV and because the PCA signal cannot be obtained when the depth of insonation is decreased to less than 55 mm. A more detailed description of the TCD examination can be found in a standard textbook. Fig. 7.1 illustrates various signals obtained from the most commonly insonated vessels.

Fig. 7.1, Schematic representations of blood flow velocity (FV) traces obtained from the internal carotid artery bifurcation at 5.5 to 6.5 cm (A), the M1 segment of the middle cerebral artery (MCA) at 3 to 6 cm (B), the posterior cerebral artery (P1) at 6 to 8 cm (C), and the anterior cerebral artery (A1) at 6 to 8 cm (D). Flow above the horizontal is toward the probe, and flow below the horizontal is away from the probe.

Velocity Measurements

Although the most physiologic correlate with actual CBF is the weighted mean velocity (FV _mean ), which takes into consideration the different velocities of the formed elements in the blood vessel insonated, the maximal FV (FV _max as depicted by the spectral outline) is generally used because of the higher signal-to-noise ratio. A good correlation also exists between the FV _max and FV _mean in the basal cerebral arteries as flow is usually laminar. The mean FV with time-averaged FV usually refers to the mean velocity of FV _max . The time-averaged FV _max is determined from area under the spectral curve.

The volume of blood flowing through a vessel depends on the velocity of the moving cells and the diameter of vessel concerned. For a given blood flow the narrower the vessel, the higher the velocity. Although CBF in millilitres per minute per 100 g of brain tissue is relatively constant under conditions of constant brain metabolism and arterial content of carbon dioxide and oxygen, FV in the MCA ranges from 35 to 90 cm/sec in the awake resting state. This range is due to inter-individual variations in vessel diameter and angles of insonation and probably accounts for the poor correlation between absolute FV and CBF in any given population. However, relative changes in FV accurately reflect variations in CBF.

Mean FV varies with age. MCA red cell velocity (FV _MCA ) rises from 24 cm/sec at birth to a peak of 100 cm/sec at age 4 to 6 years. Thereafter the FV _MCA decreases steadily to about 40 cm/sec during the seventh decade of life. ^, Some of the reduction is the result of the genuine decrease in hemispheric CBF, which has been reported by several investigators. Overall, the velocity trend seen in the MCA with age is similar to that in hemispheric CBF.

Mean FV is higher during hemodilution. A reduction in hematocrit has been shown to increase CBF in a linear fashion and probably accounts for the greater velocities reported in low hematocrit states. However, low hematocrit may present diagnostic difficulties in patients with potential arterial stenotic lesions because the increase in velocity observed may be incorrectly interpreted as vessel stenosis. An example is subarachnoid-related vasospasm.

Women have higher hemispheric CBF than men, which is reflected in 3–5% higher mean FV _MCA values. Although a convincing explanation for this difference in velocity has not yet been found, a lower hematocrit and slightly higher arterial CO ₂ tension found in premenopausal women may partly explain this increase in velocity.

Waveform Pulsatility

Pulsatility describes the shape of the maximal shift (the envelope) of the Doppler spectrum from peak systolic pressure to end-diastolic pressure with each cardiac cycle. The FV waveform depends on the arterial blood pressure (ABP) waveform and the viscoelastic properties of the cerebrovascular bed, provided that blood rheology remains constant. The absence of vessel stenosis or vasospasm, with constant pulsatility of ABP, during constant cerebral perfusion pressure (CPP), and constant arterial blood CO ₂ concentration, suggests that changes in pulsatility may reflect the changes in product of distal CVR, compliance of blood vessels and heart rate. Two derived indices have been used to quantify pulsatility. The pulsatility index (PI), or Gosling index, is calculated as follows:

PI = \frac{(F V_{sys} - F V_{dias})}{F V_{mean}}

$PI = \frac{(F V_{sys} - F V_{dias})}{F V_{mean}}$

where FV _dias is diastolic FV and FV _sys is systemic FV. The resistance index (RI), or Pourcelot index, is calculated with the following equation:

RI = \frac{(F V_{sys} - F V_{dias})}{F V_{sys}}

$RI = \frac{(F V_{sys} - F V_{dias})}{F V_{sys}}$

In a highly pulsatile spectrum, FV _sys is peaked and much greater than end-FV _dias , whereas FV _dias greater than 50% of FV _sys gives a “damped” waveform. Normal PI ranges from 0.5 to 1 with no significant side-to-side or cerebral interarterial differences.

In general, PI and RI correspond to each other, reflecting changes in central or cerebral hemodynamics. However, neither index provides meaningful information about the cause of the change; for example, an increase in PI can be caused by cerebral vasoconstriction (intrinsic, as in hyperventilation) or vasodilatation at low CPP, when vessels dilate because of the autoregulatory response. Furthermore, PI is very sensitive to changes in heart rate and its values are best compared when measured during periods of similar heart rates. The advantage of PI is that it is dimensionless and, therefore, is not affected by the angle of insonation because the equation used to calculate PI has the cosine of the angle of incidence in both the numerator and the denominator. A PI value above 1.5 with normal or increased mean arterial pressure (MAP) in a normocapnic patient with normal pulsatility of blood pressure, may indicate elevated ICP. Also, asymmetry in PI values greater than 0.5 between the left and right hemispheres may give rise to concern about clinically relevant asymmetry of cerebral hemodynamics (unilateral carotid artery stenosis, acute subdural hematoma, brain contusion with a midline shift, etc.). PI is only superficially simple. It depends on numerous interrelated factors: ABP pulsatility, heart rate, PaCO ₂ , CPP, hematocrit, body temperature, CVR, compliance of the proximal cerebral vessels, ICP, and compliance of the cerebrospinal space.

Describing cerebral haemodynamics: testing and monitoring

All CBF-regulatory relationships—between CBF and CPP (autoregulation), CBF and PaCO ₂ and PO ₂ (chemoregulation), probably CBF and brain metabolism (flow-metabolism coupling) and autonomic activity (neuroregulation)—have non-linear characteristics. Most frequently tested is cerebral pressure autoregulation, with the static CBF–CPP relationship described as Lassen’s curve.

Almost all cerebrovascular reactivity testing methods rely on comparing changes in CBF velocity to changes (provoked or spontaneous) in controlling variables: CPP, PaCO ₂ , PaO ₂ or other (breath holding, Valsalva maneuver, etc., make quantification more difficult).

However, cerebral vessels’ reactivity responses have their own time inertia (from 6 to 12 seconds). Therefore, changes in CBF provoked by changes in CPP that are slower than 6 seconds will contain information about cerebral autoregulation, and changes provoked by CPP variations faster than that will describe the mechanoelastic properties of the cerebrovascular bed, like cerebrovascular resistance and compliance. Domain demarcation frequency is around 0.05 Hz, which means that slower waves of CPP and CBF carry information about cerebral autoregulation, while faster waves carry information about the mechanoelastic properties of the cerebrovascular system.

Cerebrovascular Reactivity to CO ₂

Cerebrovascular reactivity to CO ₂ describes the relationship between arterial CO ₂ tension and CBF. Within the limits of mild hypocapnia to mild hypercapnia, a slow change in Paco ₂ produces almost proportional change in CBF and FV. In deep hypocapnia and hypercapnia, this linear relationship saturates. The reactivity can be tested by observation of the change in CBF in response to a change in Paco ₂ . If we accept that the diameter of the basal arteries is unaffected or is affected to a negligible degree by changes in arterial CO ₂ tension, then TCD is particularly suitable for such investigations because multiple paired measurements are taken and regression lines can be constructed more accurately than with a limited number of conventional blood flow measurements. The percentage change in FV with a change in Paco ₂ shows a low dependence on the baseline value and is, therefore, a valid indicator of CO ₂ reactivity and a more appropriate variable for use in comparing clinical conditions.

In normal individuals CBF (or FV) changes by approximately 2.5–3% for every mmHg change in Paco ₂ . TCD can, therefore, be used in many clinical situations to assess cerebrovascular reserve, such as for patients with carotid artery stenosis and after head injury. The effect of anesthetics and vasoactive drugs on cerebral vasoreactivity to CO ₂ can also be easily examined using TCD. ^, An induced change in Paco ₂ usually provokes a change in ABP. In such cases CO ₂ reactivity values should be adjusted accordingly ( Fig. 7.2 ). Increasing and decreasing CO ₂ tension test both the vasodilatory and vasoconstrictive capabilities of the cerebral circulation.

Fig. 7.2, CO 2 reactivity test performed in patient with right side common carotid artery stenosis (90%). Right side reactivity was 27%/kPa, and left side reactivity 14%/kPa. After corrections for change in arterial pressure, reactivity at left was 19%/kPa and right 9%/kPa, indicating that right side reactivity was severely depleted. ABP, arterial blood pressure, FVl, blood flow velocity in left middle cerebral artery (MCA); FVr, blood flow velocity in right MCA; EtCO 2 , end-tidal CO 2 pressure.

Cerebral Pressure Autoregulation

Cerebral autoregulation, a sensitive mechanism that can be impaired by pathologic processes and inhalational anesthesia, minimizes deviations in CBF when CPP changes between 50 and 170 mmHg. Cerebral autoregulation has been traditionally assessed by repeated static measurements of CBF during a period of hypotension or hypertension. Before the TCD era, in addition to the bulky equipment or radioactive material necessary for these measurements, the process was labor intensive and assumed that cerebral autoregulation was a uniform and slow-acting process. Furthermore, the drugs used to induce hypertension or hypotension may have influenced cerebrovascular tone.

Cerebral autoregulation is a complex process composed of several physiologic mechanisms operating possibly at different rates. TCD studies have estimated the time constant of fast autoregulatory responses; FV _MCA as an index of CBF was fully restored to the baseline value as early as 6–12 seconds after a step decrease in blood pressure. TCD allows noninvasive measurement of the autoregulatory response and can provide insight into both rapid and delayed components of cerebral autoregulatory mechanisms. Other continuous techniques, such as laser Doppler flowmetry and thermal methods, are invasive. Near-infrared spectroscopy seems to be the only competitor of TCD in this implementation, but its clinical use needs more validation.

Although many methods for the assessment of cerebral autoregulation have been described, only the methods most commonly employed are discussed here. Examples of testing dynamic and static autoregulation are shown in Fig. 7.3 .

Fig. 7.3, A, The recovery in blood flow velocity (FV) after thigh cuff deflation ( arrows point to start of decrease in arterial blood pressure [ABP] after deflation) is “steeper” in patients with intact dynamic autoregulation than the increase in ABP. B, In contrast, FV remains depressed after cuff deflation in patients with impaired autoregulation. C, Testing of static autoregulation during propofol anesthesia. Top trace represents mean blood pressure (MBP) in mmHg; bottom trace represents middle cerebral artery blood flow velocity. There is virtually no change in FV despite the increase in MBP as a result of dopamine infusion.

Leg-Cuff Test

Dynamic autoregulation is tested by the measurement of the recovery in FV after a rapid transient decrease in mean blood pressure that has been induced by the deflation of large thigh cuffs. These large blood pressure cuffs that have been modified with larger tubes are placed around one or both thighs, and inflated to 50 mmHg above systolic pressure for 3 minutes. Deflation of cuffs usually produces an approximately 20-mmHg drop in ABP. Through the use of an algorithm previously validated, ^, the FV response to the drop in blood pressure is fitted to a series of curves to determine the rate of dynamic cerebral autoregulation (dRoR) or autoregulation index (ARI). These curves are generated by a computer model of cerebral autoregulation that predicts the autoregulatory response on the basis of the continuous blood pressure record and compares its predictions with the measured response. The threshold for good and disturbed autoregulation is at an ARI value of around 5. The dRoR describes the rate of restoration of FV (percentage per second) with respect to the drop in MBP. The normal dRoR is 20%/sec (i.e., the process is complete within approximately 5 seconds).

The time for autoregulation to normalize FV occurs well within the period of hypotension achieved with cuff deflation before the mean ABP returns to baseline (10–20 seconds). Collection of autoregulation data in the first 10 seconds avoids the influence of the introduction of CO ₂ -rich blood from the legs after thigh cuff deflation. Figs. 7.3A and B show examples of leg cuff tests with intact (A) and defective (B) autoregulation.

Static Autoregulation

Static autoregulation can be tested by the induction of an approximately 20-mmHg increase in MBP through a 0.01% phenylephrine infusion with simultaneous recording of the FV. The FV and MBP recorded are then used for the subsequent calculation of the estimated CVR (CVRe; CVRe = ABP/FV). The static rate of autoregulation (SRoR) is the ratio of percentage change in estimated CVRe to percentage change in CPP, or if ICP is not measured, percentage change in MBP. Theoretically, no change in the FV would occur if the percentage change in CVRe was equal to the percentage change in MBP. Thus an SRoR of 100% implies perfect autoregulation, and an SRoR of 0% implies a complete disruption of autoregulation. Measurement of SRoR with MBP instead of CPP may cause an error called “false autoregulation.” It happens in a nonautoregulating brain when a change in MBP produces a 1:1 increase in ICP, leaving CPP constant. This obviously does not produce any change in CBF, giving a false SRoR value equal to 1. Fig. 7.3C shows an example of static testing when autoregulation is fully functional.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here