Escolar Documentos
Profissional Documentos
Cultura Documentos
different pumping-activation functions (VAD motion profiles) on the cardiovascular response, three different VAD
motion profiles are investigated. The numerical results
suggest that Hermitian type motion profiles (smooth
curves skewed toward early systole) have the advantage of
requiring minimum power to the VAD, and producing the
minimum after-load to the left ventricle, minimum ventricular wall stress, and minimum ventricular work to the
diseased ventricle; while sawtooth type motions need
slightly more power input, and induce slightly increased
aortic pressure in diastole, thus improving coronary
perfusion. Key Words: Numerical simulationCardiovascular dynamicsHeart failureVentricular dysfunctionPulsatileVentricular assist deviceHeart valve
dynamics.
929
930
the systemic circulation and the coronary circulation, and a pulsatile catheter pump type VAD
model, to study the influence of the VAD on the
native cardiovascular system. Other researchers
(2,10) simplified the cardiovascular system into a
Windkessel model or a Westkessel model combined
with the left heart, while some researchers developed complete cardiovascular system models
including both the systemic and pulmonary circulation (3,7,1113). From the point of view of accuracy,
the complete cardiovascular system model as well
as the complete VAD model are necessary in the
numerical simulation.
Among the above numerical models, some
studied the interaction between the native cardiovascular system and the nonpulsatile, impeller-type
VADs (3,13), some studied the impeller-type VADs
but changed the rotating speed of the impeller in
the heart cycle to produce pulsatile type VAD
action (2,3), and some investigated the assisting
action of the intra-aortic balloon pump or of displacement type pulsatile VADs (7,912). ndar (8)
suggested that physiological pulsatile pumps
provide the best hemodynamic VAD support
mechanism. To date the dynamic interaction of pulsatile and nonpulsatile rotary-type VADs with the
cardiovascular system has been systematically investigated with both numerical and experimental
studies, while for the physiological pulsatile type
VADs, only the traditional intra-aortic balloon
pump has been extensively studied with numerical
investigations (10,12) as well as experimental study
(14). Numerical study of other displacement type
VADs involve complex dynamic characteristics
(7,9,11) and thus received less attention, but due to
their special hemodynamic advantage they still need
further investigation.
The present article presents numerical models of
the cardiovascular system and of physiological type
pulsatile VADs, and investigates the interaction of
reciprocating valve pulsatile VADs with the native
cardiovascular system. Atrial contraction and heart
valve dynamics, which affect and contribute to the
pulsatility of the physiological flow, are specially considered in the model. Various types of prototype
models for reciprocating valve VADs can be found in
the literature (e.g., [15,16]). In these designs any type
of prosthetic heart valve (monoleaflet, bileaflet,
PTFE membrane, bioprosthetic, etc.) is mounted on
an annular magnet, and it is moved in a reciprocating
motion along the length of the VAD. The prosthetic
valve is closed and acts as a piston pump in one
direction of motion along the length of the VAD,
while it opens to let flow through the valve in the
931
(a)
(b)
(c)
FIG. 1. Illustration of the simulation model. (a) Schematic of the native circulation system; (b) schematic of the reciprocating valve pump
type VAD; (c) illustration of typical forces acting on a heart valve leaflet.
932
Heart chambers
In modeling the heart, basic chamber characteristics are described with the pressure-volume relation
with the widely used Suga et al.s variable elastance
model (17), in which the chamber pressure is
described as a linear function of the chamber volume
and the chamber elastance. The chamber volume is
decided by the flow-rate difference between the inlet
and the outlet of the chamber, while the chamber
elastance varies in a heart cycle representing the
action of the heart muscle.
Take the left ventricle for example. The instantaneous volume change in the left ventricle is equal to
the flow-rate difference between mitral and aortic
valves:
dVlv
= Qmi Qao
dt
(1)
elv ( t ) = Elv,d +
Elv,s Elv,d
elv ( t )
2
(2)
The activation function elv(t) describes the contraction and the relaxation changes in the ventricular
muscle. A commonly used activation function is
adopted:
1 cos t
Ts1
t Ts1
elv ( t ) = 1 + cos
Ts2 Ts1
: 0 t < Ts1
(3)
: Ts2 t < T
(4)
t Tpwb
ela ( t ) = 1 cos
2
Tpww
: 0 t < Tpwb
: Tpwb t < Tpwb + Tpww
: Tpwb + Tpww t < T
(5)
(6)
(7)
(8)
The systemic vein is modeled as a compliance combined with a resistance. In the vein the pressure is:
(9)
Qsvn =
Psvn Pra
Rsvn
(10)
933
: Plv Psas
: Plv Psas
(11)
ARao =
2
(1 cos )2
[ rao(1 cos )]
=
2
(1 cos max )2
[ rao(1 cos max )]
(12)
934
Iao
d2
d
(13)
= kp,ao ( Plv Psas ) Aao cos kf,ao
dt 2
dt
The number of coefficients is minimized by substituting Kp,ao = kp,aoAao/Iao and Kf,ao = kf,ao/Iao:
d2
d
= K p,ao ( Plv Psas ) cos K f,ao
2
dt
dt
A dx + CQ AR
vad
vad
vad dt
: Psas Psat
Psas Psat
Qvad =
Avad dx + CQvad ARvad
dt
: Psas < Psat
Psat Psas
(14)
(15)
Qvad = Qsas
(16)
X vad t
: 0 t Ttwe
Ttwe
x (t ) =
X vad X vad ( t Ttwe ) : Ttwe < t < T
T Ttwe
(17)
935
(b)
FIG. 2. Pumping action profiles of the left ventricular assist device (LVAD). (a) Motion profiles of the LVAD; (b) velocity profiles of the
LVAD.
X vad
2
x (t ) =
X vad
2
t
1 cos
Ttwe
: 0 t Ttwe
( t Ttwe )
1 + cos
T Ttwe
X 1 2 t Tr + S ( t T )
r
r
r
Tr
2
t
0 t Tr
T
r
t Tr
X r 1 2 T T +
r
twe
x (t ) =
2
S ( t T ) t Ttwe + X
vad
r
r
Tr Ttwe
2
t Ttwe t Tr
Tr < t Ttwe
1
Ttwe Tr Ttwe Tr
X vad
( t Ttwe )
1 + cos
Ttwe < t < T
T Ttwe
2
(19)
System parameters
In this work, most of hemodynamic variables used in
the numerical model of the cardiovascular system are
assigned values by referring to published materials.
Most researchers used variable elastance model to
describe the heart dynamics. By referring to parameter selection in Ursino (21) and Lu et al. (22) in the
current research, chamber elastance values are
assigned as shown in Table 1. Parameter settings as
shown in Table 2 for systemic and pulmonary loops
are based on the works of Ursino (21) and Lu et al.
(22). In modeling heart valve dynamics, coefficients
for valve motion equations are selected through
Artif Organs, Vol. 30, No. 12, 2006
936
Value
Parameter
Value
Unit
CQao
CQmi
Elv,s
Elv,d
Plv,0
Vlv,0
Ela,max
Ela,min
Pla,0
Vla,0
350
400
2.5
0.1
1.
5.
0.25
0.15
1.
4.
CQpo
CQti
Erv,s
Erv,d
Prv,0
Vrv,0
Era,max
Era,min
Pra,0
Vra,0
350
400
1.15
0.1
1.
10
0.25
0.15
1.
4.
mL/(s mm Hg0.5)
mL/(s mm Hg0.5)
mm Hg/mL
mm Hg/mL
mm Hg
mL
mm Hg/mL
mm Hg/mL
mm Hg
mL
Value
Parameter
Value
Unit
0.08
0.003
0.000062
1.6
0.05
0.0017
0.5
0.52
0.075
20.5
800
Cpas
Rpas
Lpas
Cpat
Rpat
Lpat
Rpar
Rpcp
Rpvn
Cpvn
Vrv0
0.18
0.002
0.000052
3.8
0.01
0.0017
0.05
0.25
0.006
20.5
500
mL/mm Hg
mm Hg s/mL
mm Hg s2/mL
mL/mm Hg
mm Hg s/mL
mm Hg s2/mL
mm Hg s/mL
mm Hg s/mL
mm Hg s/mL
mL/mm Hg
mL
Value
Parameter
Value
Unit
Kp,ao
Kf,ao
Kp,mi
Kf,mi
Kp,vad
Kf,vad
5500
50
6000
40
3500
25
Kp,po
Kf,po
Kp,ti
Kf,ti
5000
60
5800
45
rad/(s2 mm Hg)
L/s
rad/(s2 mm Hg)
L/s
rad/(s2 mm Hg)
L/s
Value
0.0008
0.0001
0.5
1.
0.3
0.92
0.09
0.28
0.3
0.45
0.22
0.42
0.2
0.11
0.0605
Unit
2
m
s
mm Hg/mL
s
s
s
s
s
s
s
s
s
s
m
m
937
938
(d)
(b)
(e)
(c)
(f)
FIG. 3. System response with cardiovascular model in healthy conditions. (a) Pressure, left heart; (b) flow rate, left heart; (c) volume, left
heart; (d) pressure, right heart; (e) flow rate, right heart; (f) volume, right heart.
(d)
(b)
(e)
(c)
(f)
939
FIG. 4. System response of cardiovascular model with left ventricular failure. (a) Pressure, left heart; (b) flow rate, left heart; (c) volume,
left heart; (d) pressure, right heart; (e) flow rate, right heart; (f) volume, right heart.
940
(d)
(b)
(e)
(c)
(f)
FIG. 5. System response with left ventricular failure and sawtooth type left ventricular assist device motion. (a) Pressure, left heart; (b)
flow rate, left heart; (c) volume, left heart; (d) pressure, right heart; (e) flow rate, right heart; (f) volume, right heart.
(a)
(d)
(b)
(e)
(c)
(f)
941
FIG. 6. System response with left ventricular failure and sine type left ventricular assist device motion. (a) Pressure, left heart; (b) flow
rate, left heart; (c) volume, left heart; (d) pressure, right heart; (e) flow rate, right heart; (f) volume, right heart.
942
(d)
(b)
(e)
(c)
(f)
FIG. 7. System response with left ventricular failure and Hermitian type ventricular assist device motion. (a) Pressure, left heart; (b) flow
rate, left heart; (c) volume, left heart; (d) pressure, right heart; (e) flow rate, right heart; (f) volume, right heart.
chambers in the model are affected by the combined effect of the equations in the numerical
model, Eqs. 119, and particularly by Eqs. 14.) The
shape of the aortic flow-rate response in the LVAD
assist condition, as shown in parts (b) of Figs. 57,
becomes smoother and is different from that of
Fig. 3b in the healthy condition. Nevertheless, the
volume changes as discussed above suggest that the
cardiac output is not impaired by the change in
Artif Organs, Vol. 30, No. 12, 2006
(d)
(b)
(e)
(c)
(f)
943
FIG. 8. Comparison of cardiovascular system response under various conditions. (a) Left atrial pressure; (b) left ventricular pressure; (c)
pressure in ventricular assist device inlet; (d) pressure in systemic artery; (e) pressure in pulmonary vein; (f) pressure-volume loops.
944
(b)
FIG. 9. Estimation of ventricular assist device (LVAD) characteristic variables. (a) Estimation of LVAD ideal power requirement; (b)
estimation of left ventricular wall stress.
(20)
Plv Dlv2
4w (w + Dlv )
(21)
945
946
947
lv = left ventricle
lvf = left ventricular failure
max = maximum value
min = minimum value
mi = mitral valve
p = effect of pressure force
par = pulmonary arterioles
pas = pulmonary artery sinus
pat = pulmonary artery
pcp = pulmonary capillary
po = pulmonary valve
pvn = pulmonary vein
pwb = beginning of P wave
pww = duration of P wave
r = a specified intermediate instant in early systole
ra = right atrium
rv = right ventricle
s = systolic phase
s1 = peak of systolic phase
s2 = end of systolic phase
sar = systemic arterioles
sas = systemic aortic sinus
sat = systemic artery
scp = systemic capillary
svn = systemic vein
ti = tricuspid valve
twb = beginning of T wave
twe = end of T wave
tww = duration of T wave
vad = ventricular assist device
REFERENCES
1. Reul HM, Akdis M. Blood pumps for circulation support.
Perfusion 2000;15:295311.
2. Vandenberghe S, Segers P, Meyns B, Verdonck P. Unloading
effect of a rotary blood pump assessed by mathematical
modeling. Artif Organs 2003;27:1094101.
3. Pekkan K, Frakes D, de Zelicourt D, Lucas CW, Parks WJ,
Yoganathan AP. Coupling pediatric ventricle assist devices to
the fontan circulation: simulation with a lumped-parameter
model. ASAIO J 2005;51:61828.
4. Saito S, Westaby S, Piggot D, et al. End-organ function during
chronic nonpulsatile circulation. Ann Thorac Surg 2002;74:
10805.
5. Chow G, Roberts IG, Edwards AD, et al. The relation between
pump flow rate and pulsatility on cerebral hemodynamics
during pediatric cardiopulmonary bypass. J Thorac Cardiovasc
Surg 1997;114:56877.
6. Sezai A, Shiono M, Orime Y, et al. Major organ function under
mechanical support: comparative studies of pulsatile and nonpulsatile circulation. Artif Organs 1999;23:2805.
7. Drzewiecki GM, Pilla J, Welkowitz W. Design and control
of the atrio-aortic left ventricular assist device based on o2
consumption. IEEE Trans Biomed Eng 1990;37:12837.
8. ndar A. Myths and truths of pulsatile and nonpulsatile perfusion during acute and chronic cardisc support. Artif Organs
2004;28:43943.
9. Verkerke GJ, Geertsema AA, Mihaylov D, Blanksma PK,
Rakhorst G. Numerical simulation of the influence of a left
Artif Organs, Vol. 30, No. 12, 2006
948
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.