Ultrasound in the Neurointensive Care Unit

Rajajee, Venkatakrishna

doi:10.1007/978-1-4939-1723-5_15

Venkatakrishna Rajajee M.D.⁴

Part of the book series: Respiratory Medicine ((RM))

3568 Accesses

Abstract

Bedside ultrasound plays an important role in the neurointensive care unit. Patients with aneurysmal subarachnoid hemorrhage (aSAH) are at risk for cerebral vasospasm and delayed cerebral ischemia (DCI). The use of transcranial Doppler to detect and monitor vasospasm is well established. Transcranial color-coded sonography (TCCS) permits B-mode imaging of cerebral parenchyma and duplex imaging of intracranial vessels. The use of TCCS permits angle correction for more accurate determination of mean cerebral blood flow velocities (mCBFV). Measured mCBFV greater than 200 cm/s is highly suggestive of severe vasospasm with the consequent risk for cerebral infarction. Other recognized applications for TCCS include confirmation of brain death and assessment of recanalization following thrombolysis for ischemic stroke. Distension of the optic nerve sheath (ONS) identified on ocular ultrasound may be useful for the detection of raised intracranial pressure (ICP), with an ONS diameter ≥0.50 cm predicting ICP > 20 mmHg in the ventilated patient. The use of B-mode imaging of cerebral parenchyma to detect intracerebral hematomas, measure midline shift, and monitor lateral ventricle size in hydrocephalus is under investigation. The use of bedside echocardiography in aSAH patients may be useful for the identification of a Takotsubo pattern of cardiomyopathy as well as for titration of hemodynamic augmentation therapy for DCI. The use of thoracic ultrasound in aSAH may be useful for the detection and monitoring of neurogenic as well as cardiogenic pulmonary edema through the presence of B-lines. Other applications of ultrasound in the neuroICU include assessment of shock, the BLUE protocol for acute respiratory failure, eFAST in trauma, detection of deep venous thrombosis, and performance of bedside procedures.

Access provided by Autonomous University of Puebla. Download chapter PDF

Brain Death

Role of Transcranial Doppler Ultrasonography in Neuroanesthesia

Transcranial Doppler in Subarachnoid Hemorrhage

Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Bedside ultrasound has long played an important role in the neurointensive care unit, although mostly in the form of transcranial Doppler (TCD) for the detection of cerebral vasospasm following subarachnoid hemorrhage (SAH) [1]. More recently, however, the range of potential clinical applications for bedside ultrasound in the neurointensive care unit has greatly expanded. Optic nerve ultrasound (ONUS) has been extensively investigated as a noninvasive tool for the detection of raised intracranial pressure (ICP) [2–6]. Similar to patients in any other unit, bedside ultrasound is also invaluable for the assessment of shock, acute hypoxic respiratory failure, suspected deep venous thrombosis, and for the performance of several common bedside procedures. Patients with aneurysmal subarachnoid hemorrhage (aSAH) may particularly benefit from the use of bedside ultrasound: in addition to the risk of cerebral vasospasm, they are at considerable risk for severe left ventricular dysfunction and neurogenic pulmonary edema. While cranial ultrasound in the neonate and carotid ultrasound in ischemic cerebrovascular disease are among the most common applications of ultrasound relevant to neurological disease, they are beyond the scope of this chapter, which will focus on common applications of multipurpose point-of-care (POC) ultrasound in the adult neurointensive care unit.

Transcranial Color-Coded Sonography

Technical and Patient Considerations

Transcranial Doppler vs Transcranial Color-Coded Sonography

Transcranial Doppler sonography has traditionally been performed using dedicated portable machines and transducers that display a spectral Doppler waveform only. Acquisition of the Doppler signal from intracranial vessels with these machines requires “blind” manipulation of the probe and adjustment of depth of the sample volume. “Blind” TCD can therefore be technically challenging, particularly for inexperienced operators. While traditional TCD is an established and well-validated tool, this technique requires dedicated equipment separate from the multipurpose point-of-care machines typically available in the intensive care unit (ICU). The angle of insonation of vessels cannot be known, and an insonation angle of greater than 60° may easily occur, leading to marked inaccuracy in measured velocities.

Transcranial color-coded sonography (TCCS) is performed using multipurpose machines with standard phased array probes to obtain grayscale images of cerebral parenchyma and duplex images of intracranial vessels, permitting rapid acquisition and identification of multiple intracranial vessels. The pulse Doppler spectral waveform is obtained guided by the duplex images, with appropriate consideration for the angle of insonation. More recently, TCCS capability in hand-carried POC ultrasound machines more familiar to intensive care professionals has become more widely available, obviating the need for dedicated TCD equipment. Image-guided acquisition of Doppler waveforms with TCCS can be much easier than “blind” acquisition of Doppler signal using traditional TCD, particularly for trainees and inexperienced sonographers. For these reasons, this chapter will focus on the use of TCCS instead of “blind” TCD. Potential disadvantages of TCCS are the relatively larger footprint of the probe and the relative insensitivity of duplex ultrasound for low flow signals.

Equipment

A phased array (5–1 MHz) transducer (Fig. 15.1) is used, and the transcranial preset (Fig. 15.2) selected. Transcranial imaging cannot be performed with the cardiac and abdominal presets available on most multipurpose POC machines.

Patient Factors

The patient is positioned supine with the head resting comfortably back on the bed during transtemporal insonation. During transforaminal insonation the patient is ideally positioned sitting up with the head flexed forward; however, the patient can also be positioned on his/her side with the chin in contact with the chest. While no specific adverse effects have been reported with the diagnostic use of standard TCD/TCCS probes, the general principle of ALARA (as low as reasonably achievable) must be kept in mind, with the duration of insonation restricted to that which is clinically essential. This is particularly true because transcranial insonation typically requires maximal acoustic power output. About 5–20 % of patients have an inadequate transtemporal window for performance of TCCS [7, 8]. Risk factors for absence of transtemporal window include older age, female sex, and measured skull thickness [7, 8].

Image Acquisition and Interpretation

Image Acquisition

The sector-shaped image on the screen obtained from the phased array probe is oriented with the index mark on the upper left aspect of the image. The examiner ideally stands behind the patient’s head, although any position that allows the elbow to rest on the patient’s bed or be held close to the examiner’s side is acceptable; arm fatigue can be expected to develop rapidly. Insonation is initially performed through the transtemporal acoustic window, with the transducer placed immediately anterior to the upper border of the tragus and superior to the posterior zygomatic arch (Fig. 15.3). The index mark on the probe is directed anteriorly and rotated at a 10–20° angle superior to the axial plane (Fig. 15.4) with the transducer itself tilted at a 10–20° angle cephalad to the axial plane (Fig. 15.5). If the desired image is not obtained, the probe can be moved within the approximate area delineated by the circle in Fig. 15.3 until the desired initial grayscale image is obtained. If no recognizable anatomical features are seen, it is possible that the patient does not have an acoustic window on that side.

The single most important anatomical feature to identify at this point is the midbrain (Fig. 15.6), at a depth of about 5–8 cm, with the “Mickey Mouse ears” of the crus cerebri seen turned on their side. This is comparable, but not identical, to the image obtained with computed tomography (CT) imaging at this level (Fig. 15.7). Once this plane of imaging is obtained, other prominent anatomical features can be rapidly identified, including the lesser sphenoid wing of the temporal bone and the basal cisterns, particularly the interpeduncular and quadrigeminal cisterns (Figs. 15.6 and 15.7). The normal anatomical distribution of the major intracranial arteries and the circle of Willis is shown in Fig. 15.8. Once the midbrain is identified on the grayscale image, the color box is positioned over the expected anatomical location of the vessel(s) of interest, with “red” typically depicting flow toward and “blue” flow away from the probe. The most important intracranial artery to identify is the middle cerebral artery (MCA), seen as a curved vessel extending out toward the probe with flow directed toward the probe (depicted in red in Fig. 15.9) from a point anterior and lateral to the midbrain that is the point of bifurcation of the internal carotid artery, at about 4–6 cm depth. The A1 segment of the anterior cerebral artery (ACA) is seen extending from the internal carotid artery (ICA) bifurcation in the opposite direction, anteriorly, and toward the midline, with flow away from the probe (depicted in blue in Fig. 15.9). The ipsilateral (blue) and contralateral (red) A1 segments appear to meet in the midline (where the anterior communicating artery is typically not visualized) with the two A2 segments (which can often not be individually distinguished) extending anteriorly from this point, in the midline, with flow away from the probe (blue in Fig. 15.9). Minimal downward (caudal) angulation of the probe from this plane will reveal the distal internal carotid artery (ICA) following its entry into the skull, in continuity with proximal MCA, with flow directed toward the probe. Placing the color box around the midbrain itself, in the original plane of insonation, will reveal the posterior cerebral artery (PCA) curving around the crus cerebri and the midbrain from its point of origin (the bifurcation of the basilar artery) in the interpeduncular cistern. The ipsilateral PCA P1 segment demonstrates flow toward the probe (red in Fig. 15.9), which changes to flow away from the probe (blue) in the P2 segment following the completion of the bend in the PCA around the crus cerebri.

Increasing the depth of insonation to 13 cm or greater and extending the color box will sometimes reveal the contralateral intracranial vessels, depending on the quality of the acoustic window. This is most feasible when performing insonation through a large decompressive craniectomy defect. When visible, the contralateral ACA A1 segment and PCA P2 segments are seen with flow toward the probe (red), while the contralateral MCA and PCA P1 segment are seen with flow away from the probe (blue), as seen in Fig. 15.9. An attempt at interrogation of the contralateral vessels may be the only recourse when a patient has a viable transtemporal acoustic window on only one side.

To visualize the vertebrobasilar circulation, the patient’s head is flexed forward and the transducer firmly positioned in the suboccipital region with use of copious conductive gel to permit acoustic transmission through the hair at the nape of the neck. The probe is angled so that the long axis of the probe passes through the nasion, permitting insonation through the foramen magnum (Fig. 15.10). The most important structure seen on the grayscale image will be the foramen magnum, a large circular hypoechoic structure in the midline between about 2–7 cm of depth (Fig. 15.11). The color box is placed so that it covers the foramen magnum and the expected location of the basilar artery deeper to the foramen, within the cranial vault. The two vertebral arteries are visualized on either side of the foramen magnum, with flow away from the probe (blue in Fig. 15.11), extending intracranially to a point of confluence at about 6–8 cm of depth, beyond which the basilar artery is seen extending distally into the cranial vault, also with flow away from the probe (blue in Fig. 15.11). This confluence typically results in the “Y” shape seen in Figs. 15.11 and 15.12. The distal bifurcation of the basilar artery is typically not visualized with TCCS.

Once the vessels of interest are identified, pulse Doppler is selected with a sample volume size of about 3–10 mm. The sample volume is then placed on the arterial segment of interest (Fig. 15.13). Since the anatomical orientation of the artery can be clearly visualized, an effort must be made to obtain an insonation angle as close as possible to zero degrees for flow toward the probe and 180° for flow away from the probe (Figs. 15.13 and 15.14). When this is not possible, most POC machines with TCCS capability permit angle correction through manual delineation of the direction of flow within the vessel. The machine then corrects the calculated velocity using a formula that includes the angle of insonation. Using angle correction is generally not recommended when the angle of insonation is greater than 60° or when a sufficiently long arterial segment is not available to accurately depict the direction of flow. Once the sample volume is appropriately positioned (and angle correction performed if so desired), the spectral Doppler reading is obtained (Fig. 15.15). The Doppler frequency shift recorded within the sample volume is converted to flow velocity, depicted on the y-axis, against time, on the x-axis. One complete cardiac cycle on the spectral Doppler waveform is then traced out and the machine calculates the parameters of interest, including the peak systolic velocity (PSV), end-diastolic velocity (EDV), the mean cerebral blood flow velocity (mCBFV – otherwise known as the time averaged peak or TAP), and the pulsatility index {PI = (PSV-EDV)/mean flow velocity or MFV}.

When the blood flow velocity within an arterial segment is relatively high relative to the set color scale, aliasing may occur. Aliasing is visible as an abrupt change in color duplex examination within a segment of the artery (but not the entire artery) from the expected color appropriate to the direction of flow to the opposite color: red to blue or vice versa. The presence of aliasing can be clinically useful, since regions of aliasing are more likely to represent higher velocities and can be preferentially sampled when looking for vasospasm (Fig. 15.12). Most aliasing does NOT represent vasospasm; however, and if so desired, aliasing can be eliminated by increasing the range of the color scale.

Once the intracranial vessels have been interrogated, measuring the velocities in the distal cervical internal carotid arteries, immediately prior to their entry into the skull, should be performed when evaluating for vasospasm. These extracranial ICA (exICA) velocities are used to calculate the Lindegaard ratio (mCBFV_MCA/mCBFV_exICA). The transducer is placed in the submandibular region at the angle of the jaw and directed upward and slightly medially, toward the ipsilateral eye (Fig. 15.16). The exICA is a linear (sometimes tortuous) vessel with flow away from the probe (blue in Fig. 15.17) and is typically medial to the internal jugular vein, which is distinguished by the presence of flow toward the probe (red) and a venous pattern of pulsation. The most important distinguishing feature of the exICA, compared to the external carotid artery, is the presence of substantial diastolic flow in the exICA. As with intracranial vessels, obtaining an angle of insonation as close to 0° or 180° as possible is important; otherwise, perform angle correction.

Image Interpretation

The most important parameter obtained from TCCS is the mCBFV, since most angiographic correlation to vessel caliber has been performed with this parameter. The range of “normal” velocities used at the University of Michigan is shown in Table 15.1.

Table 15.1 Normal range of mean cerebral blood flow velocities

Full size table

1.
Cerebral vasospasm after aSAH

Reduction in vessel caliber caused by cerebral vasospasm following aSAH will result in an elevation in mCBFV, since velocity of blood flow is inversely proportional to the square of the vessel radius [9]. A mCBFV <120 cm/s in the MCA is reliable for the exclusion of hemodynamically significant vasospasm and mCBFV >200 cm/s is specific for severe vasospasm [1, 10]. Elevations in velocity between 120 and 200 cm/s may be caused by either hyperdynamic flow or vasospasm. Hyperdynamic flow frequently results from factors such as the catecholamine surge following aSAH, anxiety, pain, fever, anemia, and the use of inotropic agents. The Lindegaard ratio (mCBFV_MCA/mCBFV_exICA) can be used to differentiate hyperdynamic flow from vasospasm [1, 10]. A ratio >3 likely represents mild vasospasm and a ratio >6 severe vasospasm. Thresholds for the identification of vasospasm have been validated primarily in the MCA. However, normal MCA velocities are almost always higher than ipsilateral normal ACA, ICA, and PCA velocities. Therefore, an elevation in the velocity (or corresponding ratio to ipsilateral exICA velocity) in any of these other arteries above thresholds validated for vasospasm in the MCA are very likely to represent vasospasm in those vessels as well; this is especially true when the mCBFV in any of these other arteries exceeds 200 cm/s. Normal velocities in the posterior circulation are typically lower than in the anterior circulation. In one study, a basilar velocity >85 cm/s and a ratio of basilar to extracranial (measured in the neck, not transforaminal) vertebral mCBFV >3 had 92 % sensitivity and 97 % specificity for basilar narrowing of more than 50 % on angiography [11]. While mCBFV identifies large vessel vasospasm and not tissue ischemia, a PI ≤ 0.58 may identify patients with large vessel vasospasm at higher risk for subsequent clinical neurological decline [12].
2.
Determination of brain death

The American Academy of Neurology’s Quality Standards Subcommittee requires that the determination of brain death using TCD/TCCS be made on the basis of reverberating flow (Fig. 15.18) or short systolic spikes (Fig. 15.19) recorded in the anterior circulation (typically the MCA) bilaterally as well as in the posterior circulation (basilar) [1]. While the presence of such diffuse reverberating flow or short systolic spikes is considered incompatible with clinical recovery, the absence of flow signal cannot be used to determine brain death, since the inability to detect signal is commonly the result of inadequate windows or other technical factors.
Fig. 15.18
Reverberating flow in brain death. Reverberating systolic/diastolic flow seen in the MCA. When present in both MCAs and the basilar, this is incompatible with clinical recovery
Full size image

Fig. 15.19
Short systolic spikes in brain death. Short systolic spikes with no diastolic flow in the MCA. When present in both MCAs and the basilar, this is incompatible with clinical recovery
Full size image
3.
Recanalization following acute ischemic stroke

The spectral Doppler waveform obtained from the vessel targeted by reperfusion therapy, either intravenous tissue plasminogen activator (tPA) or endovascular therapy, may be used to assess the degree of recanalization. The consensus on grading intracranial flow obstruction (COGIF) score has been proposed as a tool for assessment of recanalization in clinical trials of acute ischemic stroke using TCCS and is shown in Table 15.2 [13].
Table 15.2 Consensus on grading intracranial flow obstruction (COGIF) scale for assessment of recanalization following thrombolysis for acute ischemic stroke
Full size table

Basic and Advanced Competencies in TCCS

There is, as yet, no consensus on what may constitute basic or core competencies in the performance and interpretation of TCCS in the neurointensive care unit. As the most common and well-established indication for TCCS in the neuroICU, the detection of vasospasm following aSAH would likely be considered a core competency, while the use of TCCS for brain death determination and assessment of recanalization are used less commonly than vasospasm monitoring and may therefore be considered relatively advanced competencies.

Evidence Review and Evidence-Based Use

Vasospasm of the large intracranial arteries evaluated by TCCS occurs in up to 70 % of patients with aneurysmal subarachnoid hemorrhage (aSAH) in a time window between 3 and 14 days following ictus (peak 8–10 days), although a majority of these patients remain asymptomatic [14–16]. About 20–30 % of patients will, however, develop delayed cerebral ischemia (DCI), presumably from severe vasospasm resulting in compromised distal flow to the brain, with 15–20 % of patients progressing to death or cerebral infarction [17–19]. Delayed infarction can result in severe morbidity and long-term functional disability [19].

The evidence that the use of TCD/TCCS is useful, although not definitive, for the prediction of large vessel (angiographic) vasospasm in the MCA following a SAH is robust [1, 10, 14, 15]. The predictive value appears to be substantially lower for vessels other than the MCA [10]. In one meta-analysis of 26 studies assessing the predictive value of TCD for vasospasm, sensitivity was 67 % (95 % CI 48–87 %), specificity 99 % (98–100 %), positive predictive value (PPV) 97 % (95–98 %), and negative predictive value (NPV) 78 % (65–91 %) for the MCA. For the ACA, however, sensitivity was 42 % (11–72 %), specificity 76 % (53–100 %), PPV 56 % (27–84 %), and NPV 69 % (43–95 %) [10]. The frequency at which TCD/TCCS should be performed following aSAH is unclear, although many institutions perform daily TCD/TCCS evaluation while the patient is within the risk period for vasospasm.

In 2004, the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology specified clinical indications for which the use of TCD/TCCS is well established on the basis of the existing evidence [1]. Detection and monitoring of large vessel vasospasm following aSAH was found to be a “setting in which TCD is able to provide information and in which its clinical utility is established,” based on a moderately high level of evidence. In contrast, confirmation of cerebral circulatory arrest in the patient with suspected brain death was determined to be a “setting in which TCD is able to provide information, but in which its clinical utility, compared with other diagnostic tools, remains to be determined,” while the monitoring of thrombolysis of acute MCA occlusions was determined to be a “setting in which TCD is able to provide information, but in which its clinical utility remains to be determined.”

Pitfalls and Precautions

While considering the evidence base for the use of TCD/TCCS in prediction and monitoring of vasospasm following aSAH, several factors must be taken into account. While accurate at predicting large vessel vasospasm, TCD/TCCS is not consistently accurate at predicting the development of DCI and delayed infarction [20]. It is now recognized that DCI and angiographic vasospasm are distinct entities – patients with even severe angiographic vasospasm may continue to perfuse cerebral tissue adequately and DCI may occur in the absence of severe vasospasm [21]. The treatment of cerebral vasospasm with either hemodynamic augmentation or endovascular therapy, although widely performed, has not been evaluated in clinical trials with meaningful outcome measures. The presence of severe angiographic vasospasm, however, is strongly correlated with the subsequent occurrence of DCI [19], and the expectation that a severe reduction in luminal diameter will eventually compromise tissue perfusion is reasonable. The presence of TCCS evidence of severe vasospasm should, therefore, be a “red flag” that warrants close monitoring of the patient’s hemodynamic and neurological condition, with hemodynamic and endovascular intervention considered for any neurological decline.

As previously mentioned, an important limitation in the use of TCCS is the relatively frequent (5–20 %) absence of adequate acoustic windows [7, 8]. As with any other ultrasound examination, TCCS is an operator-dependent tool and is prone to error in the hands of less experienced operators. The ACA in particular is often challenging to detect on duplex examination. Angle-corrected velocities obtained from TCCS are significantly higher than those obtained from TCD and may lead to overestimation of vasospasm when using criteria established using “blind” TCD. Therefore, in the absence of large studies correlating angle corrected velocities to angiographic vasospasm, the University of Michigan protocol is to attempt to obtain as narrow an angle of insonation as possible using image guidance but not to perform angle-correction.

B-Mode Imaging of the Cerebral Parenchyma

While transcranial ultrasound is routinely used for the detection of intracranial hemorrhage in neonates, there has been interest more recently in the use of B-mode imaging of the cerebral parenchyma in adults, particularly when a decompressive craniectomy defect permits adequate acoustic transmission. Imaging of the cerebral parenchyma can be performed by first identifying the plane of the midbrain, as described above, then tilting the probe minimally upward/cephalad until the third ventricle and lateral ventricles are seen. This is typically possible only in the patients with the very best temporal acoustic windows, primarily patients with decompressive craniectomy and a large skull defect. While the lateral ventricles are hypoechoic from their fluid content, in the absence of hydrocephalus, it is the hyperechoic choroid plexus within the ventricles that is often visualized (Figs. 15.20 and 15.21). The third ventricle can be used to identify the midline, and midline shift in the setting of space-occupying lesions (such as malignant MCA infarction) is calculated as the difference in the distance from the probe to the midline from each side, divided by two (Figs. 15.20 and 15.21) with good correlation to measurement using CT (Fig. 15.22) [22]. Assessment using B-mode also has good correlation to CT imaging for determination of lateral ventricular size in hydrocephalus (Figs. 15.23 and 15.24) as well as for the detection and measurement of intraparenchymal hematomas (Figs. 15.25 and 15.26) [23, 24]. While B-mode imaging of the brain appears to hold significant promise, it is considered investigational at this time, and its clinical role not yet defined.

Optic Nerve Ultrasound

Management of raised intracranial pressure is a cornerstone of patient management in the neuroICU. The assessment of ICP in a variety of acute neurological conditions typically requires the placement of an invasive monitor, either an external ventricular drain or a parenchymal catheter. Since there is expense and risk associated with the placement of invasive catheters, there is significant interest in potential techniques for the noninvasive detection of raised ICP. Of the available noninvasive techniques, ONUS appears to be among the most promising. The optic nerve sheath is a continuation of the dura and contains the subarachnoid space. An elevation in the ICP is transmitted down the subarachnoid space and results in distension of the optic nerve sheath (ONS), which can be detected using ocular ultrasound. While the same mechanism eventually results in papilledema, this can take hours or days, and papilledema can persist for hours to days following return of ICP to normal levels.