Transcranial doppler ultrasound in neurocritical care
Overview
In 1982, Aaslid et al1 authored a paper with the title “Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries,” describing the successful insonation and blood flow velocity (FV) measurement of the basal cerebral arteries with a range-gated Doppler transducer. These authors located an “ultrasonic window” above the zygomatic arch and 1 to 5 cm anterior to the ear, through which a 2-MHz ultrasonic pulse could be emitted and recorded. The velocity and direction of blood flow were recorded as a spectral display as recorded by the ultrasound transducer. Measurement of FV as well as direction of flow during unilateral common carotid artery (CCA) compression enabled Aaslid et al1 to describe collateral flow as well as “steal” dynamics in real time by using transcranial Doppler (TCD). They noted that compression of the CCA resulted in decreased velocity in the ipsilateral middle cerebral artery (MCA), reversal of flow or “steal” in the ipsilateral terminal internal carotid artery, as well as increase in velocity in both the ipsilateral anterior and posterior cerebral arteries, suggesting a contribution to collateral flow through the circle of Willis.1
Basic principles
TCD technology is based upon the Doppler effect principle, in which the ultrasound transducer emits a frequency, fo, and this frequency is reflected back to the probe as fe. The difference between the emitted and received frequencies, or the Doppler shift, fd, can be calculated as fd = fe − fo (see Chapter 1). Pulsed wave Doppler refers to an ultrasound transducer that emits and receives the reflected ultrasound pulse. By using a pulsed wave Doppler, TCD can be performed at variable depths to follow the course of cerebral blood vessels. The frequency refers to the number of cycles a sound wave goes through per second. A higher frequency is used to insonate more peripheral vessels, and a lower frequency is able to insonate deep cerebral vessels. The sample volume size refers to the width of the area being insonated and is measured in millimeters (e.g., a sample volume size of 2 mm will give a more precisely localized signal compared with a sample volume size of 6 mm). The intensity or power of the ultrasound wave refers to the energy emitted through the tissue being insonated. This energy is absorbed by the tissue and converted mainly to heat. The U.S. Food and Drug Administration (FDA) regulates the amount of energy able to be transmitted by ultrasound equipment to ensure patient safety (ALARA [As Low As Reasonably Achievable] principle). Ultrasound over the eye or under the chin must be used at a lower power because these locations are not covered by bone, thus exposed to greater intensity. Attenuation refers to the decrease in intensity as the ultrasound wave passes through tissue and is higher for muscle and bone and lower for fluid-filled vessels. Based on attenuation, the reflected wave will be weaker for deeper vessels.2
Acoustic windows
Acoustic windows are naturally occurring areas of cerebral bone thin enough to allow transmission of ultrasound waves. There are three commonly used acoustic windows: transtemporal, transorbital, and transforaminal. Up to 10% of people may not have adequate acoustic windows. The transtemporal window allows insonation of the anterior, middle, and posterior cerebral arteries; the transorbital window is used to insonate the ophthalmic artery as well as the cavernous portion of the internal carotid artery; and the transforaminal window allows insonation of the vertebral and basilar arteries (Figure 2-1).
Transcranial doppler interpretation
Waveform morphology
The waveform recorded by TCD reflects both systole and diastole, with systole represented by the upstroke and peak of the wave, and diastole represented by the decelerating downslope of the wave (Figure 2-2). The morphology of the waveform demonstrates valuable information regarding cerebral blood flow, with a normal systolic upstroke being a quick upstroke climaxing into a peak. An upstroke that is slow and dull could be representing a proximal obstruction or focal stenosis when seen in a single vessel or may be an indication of a global low-flow state resulting from cardiac dysfunction when a widespread finding. Hassler et al3 described TCD waveform changes in the setting of intracranial hypertension. As diastolic pressure rises to approach the intracranial pressure (ICP), end-diastolic flow decreases and results in three stages of waveform changes: initially a decrease, followed by cessation, and lastly a reversal of flow (Figure 2-3). This reversal is seen in severe intracranial hypertension near cerebral circulatory arrest (see Chapter 4), when the diastolic pressure rises higher than the ICP and has been coined “diastolic flow obliteration.”3
The pulsatility index
Absolute pulsatility is difficult to assess by Doppler ultrasound because the amplitude of the pulsatile blood flow velocity is dependent on the angle of insonation. Gosling and King4 proposed an angle-independent index, known as the Gosling index of pulsatility (PI), and defined it as the difference of peak systolic and lowest diastolic flow velocities referenced to time-averaged flow velocity ([FVsys − FVdia]/FV).4 A number of earlier studies linked the PI with distal cerebrovascular resistance (CVR), suggesting an increasing index as a reflection of an increasing resistance and vice versa.5,6 However, several settings have been reported where the link between PI and CVR was either weak or in the reverse direction than expected; giving rise to the suspicion that the PI is not a pure measure of downstream resistance.7,8 Czosnyka et al8 studied animal models in which CVR was manipulated in a controlled manner under different physiologic conditions, such as an increase in arterial carbon dioxide (CO2) tension, or a decrease in cerebral perfusion pressure (CPP) in autoregulating animals. Microvascular resistance was quantified as CPP divided by laser Doppler cortical flux. During the hypercapnic challenge, a significant, positive correlation was found between cortical resistance and Doppler flow pulsatility. On the contrary, in all groups in which cerebral perfusion pressure was reduced, a negative correlation between PI and cerebrovascular resistance was shown. These authors concluded that the PI cannot be interpreted simply as an index of CVR. The decrease in CVR when CPP decreases is followed by an increase in pulsatility, likely resulting from combined changes in vascular resistance and compliance of large cerebral arteries. The PI increases when cerebral autoregulatory reserve is compromised by a decrease in CPP; when CPP is stable, changes in PI may reflect changes in CVR.8
More recently, de Riva et al9 hypothesized that the PI is a complex function of various hemodynamic factors and explored the relationship PI-CVR by retrospectively comparing clinical data of two different physiologic situations where PI increases. The first one was intracranial hypertension, represented by ICP plateau waves (where a vasodilatory cascade is implicated); the second group involved patients submitted to a mild hypocapnic challenge, which is known to increase CVR. de Riva et al further sought to compare measured PI in both groups with a mathematical formula, expressing PI as a function of cerebrovascular impedance. Analysis of their model suggested that the PI is determined by the interplay of the value of CPP, pulsatility of arterial blood pressure (ABP), CVR, compliance of the cerebral arterial bed, and heart rate.9
Assessment of intracranial pressure
An accurate, precise and noninvasive alternative for measurement and monitoring of ICP/CPP remains a “holy grail” in neurointensive care. TCD FVs and the PI have been considered as potential surrogate candidates with variable success. Bellner et al10 reported on a cohort of 81 adult severe traumatic brain injury (TBI) and subarachnoid hemorrhage (SAH) patients, showing a strong correlation between PI and ICP (r = 0.938, P < .0001) and a sensitivity and specificity of 0.89 and 0.92, respectively, to detect ICP higher than 20 mm Hg. The excitement from this report has since been moderated in view of more recent studies demonstrating poor correlation between PI-based estimations of ICP and invasive ICP measurements; the large prediction confidence intervals and the poor sensitivity has lead authors to discourage the use of ICP-calculating formulas that are based solely on the PI.11,12 Zweifel et al12 analyzed prospectively collected TCD data from severe TBI patients, finding that if the PI is less than or equal to 1, there is a chance of about 15% that the CPP is less than 60 mm Hg; with the PI less than or equal to 0.8, there is a likelihood of about 10% that CPP is less than 60 mm Hg. On the other hand, if the PI is greater than or equal to 2.2, the probability of low CPP (less than 60 mm Hg) is about 50%, and for PI greater than or equal to 3, the probability increases to 80%.12
Notwithstanding the above caveats, TCD FVs and PI can still serve as noninvasive screening tests for compromised intracranial compliance and potentially inadequate cerebral perfusion for selected patients. This diagnostic capability can be further enhanced by the application of an advanced TCD technique named transcranial color-coded duplex (Figure 2 E-1), which is analyzed at the end of this chapter. In a clinical and physiologic setting, serial testing and correct interpretation of waveforms and values becomes essential. Recently, Bouzat et al13 showed that FV/PI measurements have prognostic value for secondary neurologic deterioration in patients with mild-to-moderate TBI. Using receiver-operating characteristic analysis, they found the best threshold limits to be 25 cm/sec (sensitivity, 92%; specificity, 76%; area under curve, 0.93) for diastolic cerebral blood flow velocity and 1.25 (sensitivity, 90%; specificity, 91%; area under curve, 0.95) for PI.13 Cerebral blood flow and blood flow velocity, oxygenation, and metabolism changes have been described after decompressive craniectomy (DC).14 TCD-derived variables can be useful in monitoring patients considered for or post-DC; studies have shown a correlation between decreasing ICP postdecompression and trends toward increasing FVs and decreasing PIs.14–17
