Value of transcranial Doppler ultrasonography for patients with acute stroke.

Razumovsky A.Y..

The Johns Hopkins Medical Institutions, Baltimore, MD, USA

Currently, acute cerebral infarction is presumed when a patient presents with an acute neurological deficit and when other diagnostic possibilities are excluded by computed tomography (CT) or magnetic resonance imaging (MRI) and laboratory tests. However, in the first 24-48 hours following stroke the CT often fails to reveal any detectable abnormality, since it is negative in 25-50% of cases [1] Other imaging tests, such as positron emission tomography or single photon emission computer tomography, are frequently neither available nor practical. Angiography during this period carries a 1.2% complication rate and may be of limited availability [2]. MRI has been shown to be more sensitive in the imaging of acute stroke, up to 82% of MRI studies being abnormal on admission compared with 58% of CT scans during the first 24 hours [1]. Magnetic resonance angiography (MRA) is a non invasive method of visualizing the cerebrovascular anatomy and preliminary studies of patients with cerebrovascular disease have been performed [3]. Transcranial Doppler ultrasonography (TCD) is another non invasive, non ionizing and inexpensive method of assessing cerebral blood flow velocities (CBFV). The TCD technique has been used in the diagnosis of intra and extracranial disease and for the sequential monitoring of patency of cerebral arteries in stroke [4-6]. MRA and TCD, or recently emerged transcranial color-coded Doppler ultrasonography permit non invasive detection of intracranial vascular stenosis or occlusion with considerable accuracy, but each of these techniques has specific and different limitations. The limitations of all the methods can produce mistakes or misinterpretations in the acute period of cerebral ischemia because of possible early spontaneous revascularization, development of collateral pathways, evolution of infarction, time of scanning, etc.

TCD is a technique to define CBFV within the vessels at the skull base. TCD allows quantitative noninvasive evaluation of the cerebral hemodynamic consequences of cerebrovascular disease (CVD). Because blood flow through an artery is the product of mean velocity and cross sectional area, TCD ultrasound does not directly measure cerebral blood flow (CBF). However, with the assumption that vessel diameter does not change, the CBFV is directly proportional to flow, and so a CBFV change accurately detects any change in CBF and could be informative about blood flow in the stroke region. TCD evaluation of the circle of Willis arterial CBFV's is ideally suited for the detection of the acute occlusion of the carotid or vertebral arteries, because a proximal obstruction will be detected as an immediate attenuation of CBFV in more distal segments. It is well established that specificity and sensitivity of TCD vary from one segment to another. Middle cerebral artery (MCA) occlusion was detected with specificity exceeding 98%, while internal carotid artery (ICA) proximal and distal stenosis were detected with specificity of 88% and 97%, respectively [7]. For a given patient, the side-to-side difference of CBFV's can serve as the most sensitive and reliable criterion for the presence, or absence of blood flow abnormalities [7].

According to the literature, decreased CBFV in the territory of the basal cerebral arteries involved in the circle of Willis, correlated well with clinical findings. This fact confirms the potential importance of TCD during acute stroke which may have practical importance for emergency medical or surgical revascularization. Several TCD limitations should be noted - it is good for large vessel occlusion, but smaller, more distally located vessel occlusions will be missed. Also, the TCD technique is quite subjective and dependent upon the capabilities of the person performing the test. Our recent study suggests that in the first 24 hours after acute stroke TCD sensitivity was 100%, while specificity of TCD in detecting low CBFV due to the obstruction of the intracranial vessels or extracranial disease is about 33% [4]. This is lower than the figures reported by Camerlingo et al. [8], who found a specificity for TCD of about 92%. However, the former authors considered only patients with a middle cerebral artery (MCA) stroke during first six hours after onset and they used as a standard arteriography. In selected (from our population) patients with only MCA affected territory we found a specificity of 100% and a sensitivity of 93% [4], that correlate well with Ley-Pozo et al. [7], Camerlingo et al. [8], and Alexandrov et al. [5] data. Among the possible reasons that can contribute to a low specificity, is a collateralization (anterior-to-anterior, posterior-to-anterior or vice versa).

In triage of acute stroke patients, the first step is to perform a patient history and physical, complete with a neurologic examination. These should then be followed by CT of the brain and TCD screening of the vasculature. For example, today in Germany, the current approach is that a CT is performed first, but the TCD can be done initially to save time in case obtaining a CT is delayed [9]. Currently it seems that TCD may be as reliable as arteriography for the majority of stroke-associated arterial lesions. This of course, is only true if the TCD laboratory has sufficient experience. TCD helps to prevent arteriorgraphy and may even be complimentary to it. If the results of TCD are normal, there is at least a 94% chance that angiographic studies will be negative [6]. A similar conclusion would likely apply to MRA, owing to the insensitivity of MRA compared to the angiography. It is possible to significantly reduce the risk and cost for individual patients if performance of a TCD was utilized as an initial evaluation for patients with acute stroke or TIA. TCD will permit the use of fewer angiographic studies by allowing a more rapid, less expensive and non-invasive evaluation of cerebrovascular anatomy. A decisive advantage of TCD studies is that they may be repeated, and are therefore ideal to guide any aggressive treatment. TCD is expected to improve acute stroke care by permitting embolus detection and potentially identify the active source of embolism in a very cost-effective manner.

According to modern concepts, the TCD test can speed initial diagnostic work-up and thus indirectly decrease the risk of fibrinolysis. The impact of early TCD upon stroke subtype diagnosis is unknown, and may affect therapeutic strategies. In the recent study [10] the diagnostic usefulness of TCD on stroke subtype diagnosis according to the criteria of the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) study was investigated in patients (n=50) with acute cerebral ischemia. Initial TOAST stroke subtype diagnosis (ITSSD) was based upon clinical presentation, and initial brain imaging studies. Modified TOAST subtype diagnosis (MTSSD) was determined subsequently after additional review of the TCD examination. Final TOAST subtype diagnosis (FTSSD) was determined at hospital discharge, incorporating all diagnostic studies. Using FTSSD as the “gold standard” ITSSD and MTSSD were compared in order to determine additional benefit from the information obtained by TCD. ITSSD classified patients correctly in 23/50 (46%) of cases. After TCD, 30/50 (60%) of patients were classified correctly, for an absolute benefit of 14%, and a relative benefit of 30% (p = 0.018). Most benefit from TCD was observed in the TOAST stroke subtype category large artery atherosclerosis in particular in patients with intracranial vascular disease. In this category, ITSSD had a sensitivity of 27% which increased to 64% after TCD (p = 0.002). TCD within 24 hours of symptom onset greatly improves the accuracy of early stroke subtype diagnosis in patients with acute cerebral ischemia due to large artery atherosclerosis. This may have clinical implications for early therapeutic interventions. Large artery atherosclerosis is a common mechanism of stroke, that is likely to respond to a number of therapeutic interventions in the acute setting. Early identification of stroke subtype allows the assessment of specific therapeutic measures that aim at acting upon the underlying stroke mechanism. For example, there are some data to support that patients with large artery atherosclerosis may benefit more from anticoagulation than antiplatelet therapy. Thus, early identification of patients with cerebral ischemia due to large artery atherosclerosis may proof to be of substantial clinical importance. TCD assesment within 24 hours of symptom onset significantly improves the accuracy of stroke subtype diagnosis in patients with acute cerebral ischemia due to large artery atherosclerosis. This may have important clinical implications for early interventional opportunities.

Combination of MRI, MRA and TCD could greatly increase the ability to specify the pathophysiology of acute stroke. Ideally, MRI/MRA/TCD comparisons should be made in studies performed within 2 - 4 hours of each other to limit discrepancies caused by the natural history of thrombosis. While there have been no large studies of acute stroke with MRA and TCD, it is likely that both these techniques will become part of the initial study of acute stroke in addition to MRI, which evaluates parenchymal changes. Diffusion-weighted (DW) MRI has been recently introduced for the acute evaluation of focal brain ischemia [11]. Preliminary studies in acute human stroke have demonstrated that ischemic regions are identifiable before T2 increases occur [10]. However, there are theoretical and practical concerns for DW MRI, among them quantitation of the absolute diffusion constants, and rapid changes in physiological variables (breathing, brain pulsation).

The lack of correlation between initial clinical presentation and imaging findings further justifies obtaining appropriate studies to define information about cerebrovascular hemodynamics and vessels. Rapid, repeatable measurements of cerebral hemodynamics may offer new insight into the process of acute stroke and provide guidance for and monitoring of therapeutic interventions. Noninvasive identification of major cerebrovascular lesions in the acute stroke with MRA and quantitative measurements of the CBFV with TCD could be a powerful tool for use in selecting patients for different therapies to facilitate recanalization.



  1. Bryan RN, Levy LM, Whitlow WD, Killian JM, Preziosi TJ, Rosario JA. Diagnosis of acute cerebral infarction: comparison of CT and MR imaging. AJNR 1991;12:611-20.
  2. Executive Committee for the asymptomatic carotid atherosclerosis study: Endarterectomy for asymptomatic carotid artery stenosis. JAMA 1995;273:1421-28.
  3. Johnson BA, Heiserman JE, Drayer BP, Keller PJ. Intracranial MR angiography: its role in the integrated approach to brain infarction. AJNR 1994;15:901-8.
  4. Razumovsky AY, Gillard JH, Bryan RN, Hanley DF, Oppenheimer SM. TCD, MRA and MRI in acute cerebral ischemia. Acta Neurol Scand 1999;99: 65-76.
  5. Alexandrov AV, Demchuk AM, Wein TH, Grotta JC. Yield of Transcranial Doppler in acute cerebral ischemia. Stroke 1999;30:1604-1609.
  6. Demchuk AM, Christou I, Wein TH, Felberg RA, Malkoff M, Alexandrov AV. Accuracy and criteria for localizing arterial occlusion with transcranial Doppler. J Neuroimag 2000;10:1-12.
  7. Ley-Pozo J, Ringelstein EB. Noninvasive detection of Occlusive Disease of the Carotid Siphon and Middle Cerebral Artery. Ann Neurol 1990;28:640-47.
  8. Camerlingo M, Casto L, Censori B, Ferraro B, Gazzaniga GC, Mamoli A. Transcranial Doppler in acute ischemic stroke of the middle cerebral artery territories. Acta Neurol Scand 1993;88:108-11.
  9. Ringelstein EB, von Kummer R, Baron J-C. Imaging and the early evaluation of stroke. In: Stroke. Emergency management and critical care. (Eds. Steiner T, Hacke W, Hanley DF), Springer, 1998; 43-60.
  10. Wijman C.A.C., McBee N.A., Keyl P.M., Varelas P.N., Williams M.A., Hanley D.F., Wityk R.J., Razumovsky A.Y. Diagnostic impact of early transcranial Doppler ultrasonography on TOAST classification subtype in acute cerebral ischemia. Cerebrovascular Dis, 2001 (accepted)
  11. Warrach S, Chien D, Li W, Ronthal MM, Edelman RR. Fast magnetic resonance diffusion weighted imaging of acute human stroke. Neurology 1992;42:1717-23.