Perspectives in Medicine (2012) 1, 132136

Bartels E, Bartels S, Poppert H (Editors):

New Trends in Neurosonology and Cerebral Hemodynamics an Update.
Perspectives in Medicine (2012) 1, 132136

journal homepage: www.elsevier.com/locate/permed

How local hemodynamics at the carotid bifurcation

inuence the development of carotid plaques
Dieter Liepsch a,, Andrea Balasso b, Hermann Berger b,
Hans-Henning Eckstein b

a
Insitut f. Biotechnik, e.V. Lothstr. 34, 80335, University of Applied Sciences, Munich, Germany
b
Interdisciplinary Research Laboratory of the Klinikum rechts der Isar der TU Mnchen,Germany

KEYWORDS Summary A short introduction is given of how uid dynamics forces and velocity distribu-
Carotid artery tion inuence the development of plaque in the carotid bifurcation. The ow parameters are
models; discussed. Flow visualization techniques and also laser-Doppler-anemometer measurements
Flow visualization; demonstrate the importance of the ow. This will be shown in true-to-scale, physiological
Velocity accurate models of the carotid arteries. These models have the same compliance as the real
measurement; blood vessel. Some applications are shown e.g. patches, stents and lters. The most impor-
Stents; tant factors are the ow rate ratio and geometry, unsteady pulsatile ow, wall elasticity and
Patch plastic non-Newtonian ow behavior of blood.
2012 Elsevier GmbH. Open access under CC BY-NC-ND license.

Introduction and bifurcations, the secondary ow cannot be neglected.

Cell/cell and cell/vessel wall interactions have been the
subject of investigation and discussion for more than 40
years. It has been shown that low and high shear regions
caused by ow separation regions and oscillatory ow are
primarily responsible for chemical reactions which con-
tribute to the formation of arterial plaques.
Most previous shear stress studies have only measured
the axial velocity component at a few local points. They
calculated the shear stresses with the velocity gradients
using a constant viscosity. Accurate three-dimensional or,
at least, two-dimensional velocity measurements are neces-
sary to calculate the shear stresses. This is because, at bends
Corresponding author at: Fakultt 05, Lothstr. 34, 80335
Germany. Tel.: +49 89 1265 1522; fax: +49 81573160.
E-mail addresses: dliepsch@t-online.de, liepsch@hm.edu
(D. Liepsch).
Numerical studies very often neglect the real, local viscos-
ity of blood and the compliance of the vessel wall which
shows a hysteresis. It is also very important that the non-
Newtonian ow behavior of blood be considered, especially
in ow separation regions.
We have studied the ow behavior in more than 200 arte-
rial models with a different geometry and different ow rate
ratios. The principles of hemodynamics, such as the forces
on uid elements, are important. These forces are
the volume forces (such as gravity force or centrifugal
forces);
the pressure forces (normal pressure multiplied with the
area of the uid element);
friction forces (shear stresses multiplied with the surface
of the cell uid element).
Normally, in larger blood vessels, the blood cells are con-
centrated in the center whereas the plasma ows near the

2211-968X 2012 Elsevier GmbH. Open access under CC BY-NC-ND license.

http://dx.doi.org/10.1016/j.permed.2012.04.005
How local hemodynamics at the carotid bifurcation inuence the development of carotid plaques
wall. The blood cells are deformed in capillaries where phys-
ical/chemical reactions take place. However blood cells are
also occasionally transported into these recirculation zones
in larger blood vessels, at bends and bifurcations. The cells
remain in the recirculation zones over several pulse cycles
and are subjected to both high and low shear stresses.
Many papers use the term turbulent ow, however a
true turbulent ow is found only in the ascending aorta and
this is not fully developed because of the entrance length.
Everywhere else you will have a nominal, laminar or transi-
tional ow.
The denition for laminar and turbulent ow is:
Laminar ow The uid elements move parallel to
each other in distinct paths. In all
layers the velocity (uid elements)
moves tangentially to the main ow.
Nominal laminar Small velocity uctuations are added
to laminar ow. This ow is
characterized by small velocity
disturbances.
Transitional ow is laminar ow with spatial and
temporal velocity disturbances
(uctuations), which decreases
relatively quickly distal to the local
ow disturbance. It is a ow between
laminar and turbulent, where ow
disturbances disappear over time.
Turbulent ow Three-dimensional, spatial and
temporal velocity uctuations are
superimposed on the main ow
direction. The ow becomes irregular
and chaotic.
A fully developed laminar prole creates a parabolic
velocity prole (1) and a fully turbulent ow creates a very
at velocity prole (2). The ow behavior can be calculated
with a dimensionless parameter called Reynolds number
Figure 1 Velocity prole of a laminar and turbulent ow.
133
(Re-number). The Re-number can be calculated with the
average velocity over the cross section of the vessel, the
diameter and the kinematics viscosity. Re = (ud/) = (Fig. 1)
For pulsatile ow the Reynolds number should be calcu-
lated with a ow rate over one pulse cycle
4V d 4V
u = V/A Re = =
d 2  d
Normally, you will never nd Reynolds numbers higher
than 2300 in blood vessels using the above denition. The
entrance length is too short and the pulse wave cannot
develop into a turbulent ow.
The non-Newtonian ow behavior of blood can be
neglected in straight pipes because the prole is only 34%
different compared to a fully developed paraboloid in a
straight pipe (Fig. 1 right, white arrow).
Methods and models
The inuence of the bifurcation angle and the stenosis
degree were studied.
We used 1:1 true-to scale, elastic silicon rubber models
with a compliance similar to that of the arterial wall. This
special technique was described in Biorheology 23, 1986.
The surface in the model reproduces the biological ves-
sel surface. The carotid artery models were installed in a
physiologically accurate circulatory system.
The uid was a polyacrylamid mixture and a water solu-
tion which shows a ow behavior similar to that of human
blood. Only the thixotropic ow behavior could not be sim-
ulated (that means the coagulation of blood). The uid is
transparent and has the same refraction index as the model
wall. This is important for the laser measurements. The
laser light will not be absorbed and the laser beam is not
deected. Measurements were done with 3D-LDA ber optic
system (DANTEC) in a physiological healthy carotid artery
134
model with a bifurcation angel of 37 between the internal
and external carotid artery.
In addition to this model, we studied models with a bifur-
cation angle of 29 and 41 and also with a 90% stenosis
in the internal carotid artery, and with a 80% stenosis in
the internal and external carotid artery (Fig. S1 online
supplementary le).
The ow rate ratio was mostly 70:30 in the internal to-
external carotid artery, but also other ow rate ratios were
tested. In earlier studies we used 90 , 60 and 45 bifurca-
tions to study the inuence of the different ow parameters
separately
pulsatile, unsteady ow;
non-Newtonian behavior of blood;
elasticity of the vessel wall;
change of blood caused by shear stresses.
We found that the endothelial cell layer was elongated in
the ow direction; however in the ow separation area the
endothelial cells have a rounded form and are not packed
closely together, so small leaks can be found. That means,
in this area, material transport from inside into the wall or
from outside into the blood can easily occur. At the stagna-
tion point, the endothelial cells are packed closely together
and are also around the apex of the inner wall of the ow
divider (Fig. 2).
The ow was visualized using dyes for steady ow, and
with a photoelasticity apparatus and a birefringent solution
to visualize the unsteady pulsatile ow.
Fig. S2 (left) (online supplementary le) demonstrates
the inuence of the ow rate ratio. The ow separation
zone starts at a ow of 30% into the branch and increases
with higher ow rate in to the branch. On the right, a short
demonstration is shown under pulsatile conditions for a ow
rate ratio of 0.3.
It is well known that vessel blockage is caused by the
growth of plaque. First, a small atherosclerotic plaque can
Figure 2 Sketch of the stream paths in a 90 bifurcation such as the left descending coronary artery.
D. Liepsch et al.
be found at a bifurcation which creates damage to the intima
(ulceration). Fibrin platelet aggregation can be created
leading to additional thrombus formation. Finally particles
are released from the plaque or parts of the thrombus which
can lead to a total blockage of the vessel (thrombus, throm-
bus emboli). This can be clearly observed with our ow
visualization techniques.
Results and applications
Basic ow studies
Fig. 3 shows ow, with a dye, hitting the apex of the carotid
bifurcation model. The dye separates into two parts, ow-
ing into the internal and external carotid artery. Because
the velocity at the inner side of the internal carotid artery
is high, an area with a lower pressure is created on the
opposite side; therefore the blue dye spreads out. The
ow, however, is still very smooth and therefore, fur-
ther downstream, the blue dye momentarily spreads over
the whole diameter before returning to a smooth laminar
ow.
The velocity measurements were done using a laser-
Doppler-anemometer. The ow, pressure and velocity
curve over one pulse cycle is shown in Fig. S3 (online
supplementary le).
Fig. S4 (online supplementary le) shows the axial veloc-
ity distribution over one pulse cycle at different phases
2.5 mm distal to the apex. The velocity at the inner wall is
very high (up to 1 m/s) and, at the outer wall, very low dur-
ing the peak systolic phase (60 ), as already demonstrated
with dyes.
Fig. S5 (online supplementary le) shows the velocity
measurements over the cross-section in color, and Fig. S6
(online supplementary le) shows the secondary ow which
is very high during peak systolic phase and decreases dur-
ing the diastolic phase. The velocities in a 90% stenosed
model are 45 times higher than normal, with velocities
How local hemodynamics at the carotid bifurcation inuence the development of carotid plaques
Figure 3 Visualization using a dye in a native and a 90% model with stenoses.
up to 45 m/s and with high velocity uctuations further
downstream, just behind the stenosis.
Studies in stents
The uid dynamic inuence of several stents were tested
in transparent models. The inuence of stents is demon-
strated using dyes. Fig. S7 (online supplementary le) shows
the angiogram of a stenosed artery (left side and with the
inserted stent on the right side. Fig. 4 shows the inuence of
the stent. The dye spreads slightly into the external carotid
artery compared to the healthy model. This is caused by the
threads of the stent.
The wire geometry, direction of wires, mesh of wires,
the in- and outow, and the stretching and surface rough-
ness was tested. We tested several stents including covered
and uncovered stents. The experiments were carried out
with the stents in various positions. Fig. 5 demonstrates the
velocity distribution 5 mm distal to the apex in the internal
Figure 4 Dye in a healthy model and in a model with a stent.
135
carotid artery model for two different stents compared to a
model without a stent.
Other therapeutic procedures in the carotid artery
Surgical procedures in the carotid artery such as endarterec-
tomy for treatment of such conditions as stenosis,
aneurysms, thrombosis and cerebral ischemia are risky and
may lead to improvement or not. The following study shows
the ow and velocity distribution of endarterectomy which
is the standard procedure to treat patients with high degree
stenosis in the carotid artery (an alternative is to use patch
plastics from articial and biological materials).
Fig. 6 shows a ow, visualized with a dye, at the point
marked in the cross-section of the model, in a healthy com-
mon carotid artery, a model with a wide patch and a narrow
patch. The differences can be clearly seen. The model with
the narrow patch shows the same ow behavior as the
healthy model; whereas the wide patch creates ow distur-
bances. Fig. S8 (online supplementary le) shows the pulse
cycle. At the beginning of the diastolic phase (90 ) a back-
ward ow can be seen in the model with the wide patch.
Again, the model with a narrow patch shows ow behav-
ior similar to that in the healthy model. The secondary ow
demonstrates this also (Fig. S9 online supplementary le).
Filters are inserted into the carotid artery before the
stent is opened to avoid a particle ow in to the brain. There
is controversy in discussions about this procedure because
many physicians report that the complication rate is almost
the same as without a lter.
Fig. S10 (online supplementary le) demonstrates that
particles captured in the lter can escape if the closing
of the lter occurs during the diastolic phase. It has been
demonstrated that it is crucial that the lter is closed during
the systolic phase to prevent this escape.
136
Figure 5 Velocity distribution in the internal carotid artery with two different types of stents.
Figure 6 Dye in a healthy model, a model with a wide patch
and narrow patch.
Discussion and outlook
From these experiments the following can be clearly seen:
All three velocity components have to be measured. The ow
rate ratio between the internal and external carotid artery
is the most important and signicantly inuences the ow
separation region. The experiments show that particles in
ow separation regions sometimes rotate over several pulse
cycles before they are washed away. They can, however,
suddenly adhere to the wall and remain there. The pulse
D. Liepsch et al.
wave is not strong enough to wash these small deposits away.
The procedure of plaque formation starts. More particles are
attracted to and adhere to this area and the ow rate ratio
is altered because of the higher resistance caused by the
deposits. This effect continues and the stenosis enlarges.
The geometry only plays a signicant role in these regions
with larger bifurcation angles, >40 , where a backward ow
is created.
In 3D measurements, the calculated shear stresses are
up to 20% higher than those found when measuring only the
axial velocity component. With an increasing ow rate, the
separation region is slightly reduced but the shear stresses
increase. 1016 Pa are the highest shear stresses in a healthy
carotid artery and are found just at the apex. Shear stresses
higher than 180250 Pa have been measured in models with
90% stenosis for 100200 ms.
Downstream of such stenoses, vortices are created where
particles can remain over several pulse cycles. They can also
adhere to the wall, creating a growing stenosis. Oscillation
causes shear stresses between 140 Pa in such recirculation
zones. Biochemical reactions are released.
It is very important that stents have to be placed pre-
cisely. End threads or wires should never reach into the
vessel lumen. Filters have to be closed during the sys-
tolic phase, so that no particles escape during the diastolic
phase, before they can be pulled out. Experimental studies
including MRI, ultrasound measurements and new ultrasound
imaging which can measure all three velocity components
will be increasingly important in the future to aide in training
and renement of diagnostic and therapeutic procedures.
Appendix A. Supplementary data
Supplementary data associated with this article (Figs.
S1S10) can be found, in the online version, at
http://dx.doi.org/10.1016/j.permed.2012.04.005.