CLINICAL REPORT
Transcranial Color Duplex Ultrasound: A Reliable Tool for
Cerebral Hemodynamic Assessment in Brain Injuries
Guillaume Dupont, MD, Laetitia Burnol, MD, Richard Jospe, MD, Terrana Raphael, MD,
Christian Auboyer, MD, Serge Molliex, MD, PhD, Laurent Gergele´, MD,
and Je´rôme Morel, MD, PhD
Background: Transcranial color duplex ultrasound (TCCD) is
becoming an important tool for cerebral monitoring of brain-
injured patients. To date, TCCD reproducibility has been
studied in healthy volunteers or patients with subarachnoid
hemorrhage and its efficiency in many brain injuries has not
been proved. Our aim was to evaluate TCCD interobserver
agreement in different brain injuries.
Patients and Methods: We performed a prospective monocentric
trial conducted from January 2014 to September 2014 in in-
tensive care unit (ICU) of Saint-Etienne university teaching
hospital, France. Brain-damaged patients admitted in ICU were
included, excluding those with decompressive craniectomy. Two
randomized operators among the ICU medical staff consec-
utively performed measurements of cerebral blood flow veloc-
ities with TCCD.
Results: One hundred measurements were obtained from 42
patients. Hemodynamic and end-tidal CO2 pressure were similar
between both measurement set. The results obtained with the
Bland-Altman method showed bias at 0.52 (95% confidence
interval [CI], 4.19 to 3.16), 0.53 (95% CI, 1.86 to 2.92), and
0.002 (95% CI, 0.06 to 0.06) for mean velocity, diastolic ve-
locity, and pulsatility index, respectively. The limits of agree-
ment were ( 32.4; 31.4), ( 20.4; 21.4), ( 0.5; 0.5) for mean
velocity, diastolic velocity, and pulsatility index, respectively.
The Passing and Bablok regression have shown a quasilinear
relationship between measurements.
Conclusions: We reported the reliability of TCCD interobserver
agreement in brain-damaged patients.
Key Words: transcranial Doppler, interobserver agreement,
brain injuries, cerebral blood flow velocities, reproducibility,
pulsatility index, mean velocity
(J Neurosurg Anesthesiol 2016;28:159–163)
Received for publication April 20, 2015; accepted September 21, 2015.
From the Department of Anesthesiology and Critical Care, University
Hospital of Saint Etienne, Saint-Étienne Cedex, France.
The authors have no funding or conflicts of interest to disclose.
Reprints: Guillaume Dupont, MD, Department of Anesthesiology and
Critical Care, University Hospital of Saint Etienne, Avenue Albert
Raimond 42270, Saint-Priest en Jarez, France (e-mail: dupont_gui
llaume@live.fr).
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
J Neurosurg Anesthesiol  Volume 28, Number 2, April 2016
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
T ranscranial color duplex ultrasound (TCCD) is a
noninvasive and easy access technique for cerebral
blood flow monitoring and indirectly cerebral perfusion
pressure. TCCD is a combination of a Doppler and ul-
trasound techniques. Comparing with conventional
transcranial Doppler (TCD), this process allows the vis-
ualization of the circle of Willis and permits registration
of angle-corrected flow velocities in basal cerebral ar-
teries.1 Studies evaluating TCD versus TCCD concluded
that TCCD was more efficient particularly to detect cer-
ebral vasospasm during subarachnoid hemorrhage
(SAH).2–4 TCD/TCCD is currently recommended as a
part of the monitoring of SAH patients.5 It also allows
the detection of patients at risk for secondary neuro-
logical deterioration after mild to moderate traumatic
brain injury.6 Preservation of cerebral perfusion is es-
sential for the management of patients with severe brain
injury. In these patients, TCD/TCCD allows an early
evaluation of brain hemodynamic, while intracranial
pressure (ICP) monitoring is being performed.7 TCD has
been used for 30 years but only few studies have evaluated
interobserver reproducibility and all but 1 with TCCD
were carried out in healthy volunteers.8–11 Moreover, it
appeared that there have been more measurement errors
in patients with the highest velocities.11 Brain-injured
patients represent a heterogenous population requiring a
close follow-up monitoring. There are many reasons for
performing TCCD and the parameter of interest (veloc-
ities or pulsatility index [PI]) is different depending on the
cerebral pathology. However, whatever the reason for
doing TCCD, the measurement should be reliable and
reproducible. The primary aim of this study was to
evaluate TCCD interobserver agreement in patients ad-
mitted in intensive care unit (ICU) for different brain
injuries.
PATIENTS AND METHODS
Study Population
The study group consisted of 42 patients admitted
in ICU with brain injury from January 2014 to September
2014. Each patient needing a cerebral monitoring was
included. The protocol was approved by the local ethics
committee (IORG0007394 on December 4, 2014) and
informed consent was deemed not to be required.
www.jnsa.com | 159
Dupont et al J Neurosurg Anesthesiol  Volume 28, Number 2, April 2016

The exclusion criteria were patients below 18 years

TABLE 1. Patients’ Characteristics
old and those with decompressive craniectomy.
Overall Brain-damaged Patients
Study Design Parameters (n = 42)
Five operators took part in the study, all are Age (y) 63 (46-68)
members of the ICU medical staff and were trained to Sex—male 32 (76.2)
Hypertension 19 (45.2)
perform TCCD (the technique was learned during their Diabetes 1 (2.4)
medical studies and practiced daily). As described below, SAPS II on admission 39 (25-48)
practices were standardized before the beginning of the 30-day mortality 9 (21.4)
study. For each measurement, 2 observers were Main diagnosis
randomized among the physicians present in the unit at Aneurysmal subarachnoid 13 (31)
hemorrhage
the moment of the measurement by drawing pieces of Intracerebral hematoma 9 (22)
paper from a bag. They successively performed TCCD on Traumatic brain injury 6 (14)
the same patient, and were blinded to each other’s results. Ischemic stroke 5 (12)
Measurements were undertaken bilaterally. The mean Other 9 (21.4)
Initial surgery
arterial blood pressure, ICP (if assessed), and end-tidal External ventricular drain 11 (26)
CO2 pressure (PETCO2) were noted. Insonation of the Aneurysm embolization or 9 (22)
middle cerebral artery, M1 segment, was achieved surgical clip
through the transtemporal acoustic window using Other 5 (12)
color-coded duplex ultrasound (S5-1.1 MHz probe; Phi- None 16 (38)
lips Healthcare CX50, DA Best, the Netherlands). Data are presented as median (interquartile range, 25% to 75%) or number of
The butterfly-shaped mesencephalic brainstem was cases (percentage).
sought as an anatomic B-mode landmark before color-
coded identification of the middle cerebral artery. An
optimal signal was sought using angle correction when SAH. ICP was assessed in 21 patients at the moment of
necessary (left to the discretion of the operator) (pref- TCCD measurements. The mean arterial blood pressure
erably in a depth of 4.5 to 5.5 cm), the screen was then (MAP), ICP, and PETCO2 were not significantly different
frozen, and the mean flow velocities (systolic, diastolic, between both timepoints. The mean MAP was
and mean) and PI during 10 cardiac cycles were auto- 93 ± 15.2 mm Hg (n = 50) for observer 1 and
matically measured. Only measurements without artifacts 93 ± 14.2 mm Hg for observer 2 (P = 0.22). The mean
were taken into account. Regarding the SAH patients, the ICP was 10 ± 5.6 cm H2O (n = 21) for observer 1 and
measurements were performed during the first 15 days 13 ± 6.9 cm H2O for observer 2 (P = 0.37). The mean
from the ictus. PETCO2 was 34 ± 6.8 mm Hg (n = 20) for observer 1
and 34 ± 3.3 for observer 2 (P = 0.37). No therapeutic
Statistical Analysis was administrated between both timepoints. The average
Statistical analyses were performed with XLSTAT time between both the operators’ measurements was
software (Addinsoft, Wales, UK) for Mac version 8 ± 0.01 minutes.
2014.4.05 and Medcalc (Medcalc Software, Ostend, The Bland-Altman plot for the mean velocity, diastolic
Belgium) version 15.6. A Bland-Altman plot with multi- velocity, and PI are shown in Figure 1. The bias of mean
ple measurements per subject was performed. Interob- velocity, diastolic velocity, and PI was 0.52 (95% CI, 4.19
server agreement was also assessed with the Passing and to 3.16), 0.53 (95% CI, 1.86 to 2.92), and 0.002 (95% CI,
Bablok regression. This test represents a linear regression 0.06 to 0.06), respectively. The limits of agreement were
procedure with no special assumption regarding the dis- ( 32.4; 31.4) for mean velocity, ( 20.4; 21.4) for diastolic
tribution of the samples and the measurement errors. The velocity, and ( 0.5; 0.5) for PI (Figs. 1A–C).
result does not depend on the assignment of the methods Using the Passing and Bablok regression, the slope
to X and Y. The slope and intercept are calculated with coefficient of right diastolic velocity, right mean velocity,
their 95% confidence interval (CI). These CIs are used to and right PI was 0.93 (95% CI, 0.80-1.05), 0.90 (95% CI,
determine whether there is only a chance difference be- 0.76-1.02), and 0.97 (95% CI, 0.81-1.17), respectively
tween the slope and 1 and between the intercept and 0. (Figs. 2A–C). The intercept was 1.14 (95% CI, 4.5 to
The analysis was performed for each side separately. The 6.7), 6.8 (95% CI, 2.1 to 15.6), and 1 (95% CI, 0.8-1.2)
Student test was performed to compare blood pressure, for right diastolic velocity, right mean velocity, and right
ICP, and PETCO2 at both timepoints. Data are presented PI, respectively (Figs. 2A–C). The slope coefficient of left
as median (interquartile range 25% to 75%) or number of diastolic velocity, left mean velocity, and left PI was 0.99
cases (percentage). (95% CI, 0.75-1.22), 1.04 (95% CI, 0.86-1.30), and 0.85
(95% CI, 0.72-1.02), respectively (Figs. 2D–F). The in-
RESULTS tercept was 0.55 (95% CI, 8.98 to 7.75), 2.92 (95%
One hundred measurements were obtained from 42 CI, 20.1 to 8.16), and 0.16 (95% CI, 0.02 to 0.30) for
patients whose characteristics are presented in Table 1. left diastolic velocity, left mean velocity, and left PI,
Thirty-one percent of the patients were admitted for respectively (Figs. 2–F). We observed an acceptable

160 | www.jnsa.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.

J Neurosurg Anesthesiol  Volume 28, Number 2, April 2016 Transcranial Color Duplex Ultrasound

FIGURE 1. Interobserver measurements. A–C, Interobserver analysis with the Bland-Altman plot. A, Mean velocity interobserver
agreement, limit of agreement is ( 32.4; 31.4). B, Diastolic velocity interobserver agreement, limit of agreement is ( 20.4; 21.4).
C, Pulsatility index (PI) interobserver agreement, limit of agreement is ( 0.5; 0.5).

interobserver agreement whatever the pathology (SAH
and other brain injuries) (data not shown).
DISCUSSION
Few information is available on interobserver re-
producibility of TCCD measurements in clinical practice.
Indeed, most studies have assessed interobserver agree-
ment in healthy volunteers with a limited numbers of
patients and without high velocity or abnormal PI as seen
in brain death or SAH.8–10,12 We reported here for the
first time a reasonable interobserver agreement for
TCCD’s derived parameters in neurocritical care patients
admitted after different brain injuries.
Our data are consistent with those of Staalsø et al11
obtained in SAH patients. They found an overall good
interobserver agreement except in patients with vaso-
spasm as compared with those without. Those patients
have usually high mean velocities (>120 cm/s). We did
not observe such dichotomy between measurements but
only few points above 100 cm/s were observed in our
study, which constitutes a limitation. In their study, the 2
measurements were taken within 15 minutes. They have
reported a moment-to-moment variability also described
by Venkatesh et al.13 They explained this variability by
medical interventions like calcium antagonists admin-
istration or triple H therapy, which are more aggressively
administrated in patients with vasospasm. In our work,
the average period between 2 operators’ measurements
was 8 minutes; no difference was observed considering the
parameters affecting the cerebral blood flow (PETCO2,
MAP, ICP). Moreover, no therapeutic was administrated
between 2 measurements. This could explain the differ-
ences observed with older studies.
Our results show that TCCD is a reliable but not
necessarily accurate tool. Indeed, the bias obtained with
the Bland-Altman plot is close to zero but with wide
limits of agreements (Fig. 1). With healthy volunteers,

Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. www.jnsa.com | 161

Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.

Dupont et al J Neurosurg Anesthesiol  Volume 28, Number 2, April 2016

FIGURE 2. A–F, Interobserver analysis with Passing and Bablok method: the solid line was obtained from the 2 observers’
measurements. The gray lines are the 95% confidence interval. We observed a near linear relationship for mean velocities,
diastolic velocities, and pulsatility index (PI).

McMahon et al10 found a limit of agreement of
± 22.1 cm/s for mean velocity, we found ( 32.4;
31.4) cm/s. The level of error measurement clinically
acceptable is difficult to define. More data could have
improved these outcomes.
Although the usefulness of TCCD has been eval-
uated in brain-injured patients, its use is only recom-
mended to monitor the cerebral vasospasm occurring
with SAH.5 Recently, Wakerley et al14 showed that
PIZ1.26 reliably predicts cerebrospinal fluid pressure

162 | www.jnsa.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.

J Neurosurg Anesthesiol  Volume 28, Number 2, April 2016
above 20 cm H2O (area under the curve, 0.84; 95% CI, 0.73-
0.94; P < 0.001). TCCD is also the only tool allowing to
evaluate cerebral perfusion before availability of invasive
cerebral monitoring and to potentially reduce the duration of
secondary brain injuries.15 Another challenging situation is
the detection of patients at risk for secondary neurological
deterioration after mild to moderate traumatic brain injury.
The use of TCD on hospital admission in complement with
brain computed tomography scan could accurately screen
patients at risk for secondary neurological deterioration.6,16
This study includes patients with a variety of diag-
nosis. This could be considered as a limitation, indeed
good agreement could be found in one clinical situation
but not in another. Nowadays, TCCD is used to monitor
cerebral hemodynamic after different brain injuries in
many neuro ICUs and could contribute to therapeutic
decision. In that context, reproducibility of measurement
is fundamental, no matter the cerebral disease. That is
why we decide to carry out this pragmatic study on
TCCD accuracy whatever the etiology of brain injury.
TCCD is becoming a key tool to monitor cerebral
perfusion. This study emphasizes its reasonable interob-
server agreement. This work was performed in a hetero-
genous group of brain-damaged patients and the
influence of a specific diagnosis on this reliability remains
unknown. Because of large limits of agreement between 2
measurements, small differences in TCCD recording
should be interpreted with caution and their under-
standing should be included in a multimodal monitoring.
The next step would be to integrate TCCD into a ther-
apeutic algorithm and to evaluate its impact on the out-
come of brain-damaged patients.
ACKNOWLEDGMENT
The authors wish to thank Emilie Presles for her
statistical assistance.
REFERENCES
1. Bogdahn U, Becker G, Winkler J, et al. Transcranial color-coded
real-time sonography in adults. Stroke. 1990;21:1680–1688.
2. Martin PJ, Evans DH, Naylor AR. Measurement of blood flow
velocity in the basal cerebral circulation: advantages of transcranial
color-coded sonography over conventional transcranial doppler.
J Clin Ultrasound. 1995;23:21–26.
Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Transcranial Color Duplex Ultrasound
3. Proust F, Callonec F, Clavier E, et al. Usefulness of transcranial
color-coded sonography in the diagnosis of cerebral vasospasm.
Stroke. 1999;30:1091–1098.
4. Schöning M, Buchholz R, Walter J. Comparative study of trans-
cranial color duplex sonography and transcranial Doppler sonog-
raphy in adults. J Neurosurg. 1993;78:776–784.
5. Connolly ES, Rabinstein AA, Carhuapoma JR, et al. Guidelines for
the management of aneurysmal subarachnoid hemorrhage a guide-
line for healthcare professionals from the American Heart Associ-
ation/American Stroke Association. Stroke. 2012;3:1–39.
6. Jaffres P, Brun J, Declety P, et al. Transcranial Doppler to detect on
admission patients at risk for neurological deterioration following
mild and moderate brain trauma. Intensive Care Med. 2005;31:
785–790.
7. Naqvi J, Yap KH, Ahmad G, et al. Transcranial Doppler
ultrasound: a review of the physical principles and major
applications in critical care. Int J Vasc Med. 2013;2013:1–13.
8. Baumgartner RW, Mathis J, Sturzenegger M, et al. A validation
study on the intraobserver reproducibility of transcranial color-
coded duplex sonography velocity measurements. Ultrasound Med
Biol. 1994;20:233–237.
9. Maeda H, Etani H, Handa N, et al. A validation study on the
reproducibility of transcranial Doppler velocimetry. Ultrasound Med
Biol. 1990;16:9–14.
10. McMahon CJ, Mcdermott P, Horsfall D, et al. The reproducibility
of transcranial Doppler middle cerebral artery velocity measure-
ments: implications for clinical practice. Br J Neurosurg. 2007;21:
21–27.
11. Staalsø JM, Edsen T, Romner B, et al. Transcranial Doppler
velocimetry in aneurysmal subarachnoid haemorrhage: intra- and
interobserver agreement and relation to angiographic vasospasm
and mortality. Br J Anaesth. 2013;110:577–585.
12. Shen Q, Stuart J, Venkatesh B, et al. Inter observer variability of the
transcranial Doppler ultrasound technique: impact of lack of
practice on the accuracy of measurement. J Clin Monit Comput.
1999;15:179–184.
13. Venkatesh B, Shen Q, Lipman J. Continuous measurement of
cerebral blood flow velocity using transcranial Doppler reveals
significant moment-to-moment variability of data in healthy
volunteers and in patients with subarachnoid hemorrhage. Crit
Care Med. 2002;30:563–569.
14. Wakerley BR, Kusuma Y, Yeo LLL, et al. Usefulness of trans-
cranial Doppler-derived cerebral hemodynamic parameters in the
noninvasive assessment of intracranial pressure: usefulness of
transcranial Doppler-derived cerebral hemodynamic parameters.
J Neuroimaging. 2014;00:1–6.
15. Ract C, Moigno S, Bruder N, et al. Transcranial Doppler
ultrasound goal-directed therapy for the early management of
severe traumatic brain injury. Intensive Care Med. 2007;33:645–651.
16. Bouzat P, Francony G, Declety P, et al. Transcranial Doppler to
screen on admission patients with mild to moderate traumatic brain
injury. Neurosurgery. 2011;68:1603–1609.
www.jnsa.com | 163