Download Atrial septostomy benefits severe pulmonary - AJP

Am J Physiol Heart Circ Physiol 302: H2654–H2662, 2012.
First published April 27, 2012; doi:10.1152/ajpheart.00072.2012.
Atrial septostomy benefits severe pulmonary hypertension patients by increase
of left ventricular preload reserve
Yvette Koeken,1 Nico H. L. Kuijpers,1 Joost Lumens,2 Theo Arts,1 and Tammo Delhaas1
1
Department of Biomedical Engineering, Maastricht University, Maastricht, The Netherlands; and 2Unité de Rythmologie et
Stimulation Cardiaque, Hôpital Cardiologique du Haut-Lévêque, Bordeaux, France
Submitted 24 January 2012; accepted in final form 14 April 2012
computer modeling; exercise; blood pressure; circulation; oxygen
delivery
SHORTNESS OF BREATH, FATIGUE, dizziness, and syncope are symptoms related to severe pulmonary hypertension (PH). These
symptoms probably originate from decreased capacity to increase right ventricular (RV) and, hence, left ventricular (LV)
output during exercise (10). On the basis of clinical observations that PH patients with a patent foramen ovale, allowing
blood flow from the right to the left atrium (LA), have better
life expectancy (3, 18), severe PH patients without this opening
often undergo atrial septostomy (AS) to artificially create an
atrial septal defect (ASD) (Fig. 1). After this procedure, symptoms of severe PH diminish and survival is prolonged (10). The
beneficial effects of AS for PH patients are ascribed to
1) volume unloading of the RV, 2) facilitated LV filling,
3) increased systemic output in rest and the possibility to
further increase systemic output during exercise to maintain
Address for reprint requests and other correspondence: Y. Koeken, Dept. of
Biomedical Engineering, Maastricht Univ., P.O. Box 616, 6200 MD, Maastricht, The Netherlands (e-mail: [email protected]).
H2654
cerebral blood pressure, and 4) increased oxygen delivery to
the peripheral tissues (3, 8 –10, 19, 23).
It should be noted, however, that the ASD causes mixing of
oxygenated pulmonary venous blood with low saturated systemic venous blood in the LA. As a result, arterial oxygen
saturation is reduced (8, 14, 17, 19, 22). In clinical trials it was
found that the relative decrease in arterial oxygen saturation
was less than the relative increase in systemic output. Hence, it
was concluded that systemic oxygen delivery, defined as systemic flow multiplied by the arterial oxygen saturation, was
increased (17, 19). It is believed that the improvement of the
patient’s condition after AS is explained by this increase in
systemic oxygen delivery (6, 10, 17, 19, 23). However, because
venous oxygen saturation is not increased in PH patients after
AS (4), improvement of the patient may not be the result of
increased oxygen delivery. It has also been suggested that the
improvement of the patient’s condition and the relief of syncope may be ascribed to conservation of systemic blood pressure during exercise. The additional filling of the LV through
the ASD facilitates the mandatory increase in systemic flow to
maintain cerebral blood pressure during exercise (4, 12).
To the best of our knowledge, it is still unclear whether the
benefits of AS in PH patients are explained by the increase of
oxygen availability in the tissues, the possibility to maintain
blood pressure during exercise, or a combination of these
factors. In the present study, we evaluated effects of AS using
a multiscale computational model of the cardiovascular system
(1, 13). Variations of the degree of PH, amount of pulmonary
flow, and ASD size were simulated to investigate whether their
effects on hemodynamics and oxygen distribution may explain
the beneficial effects of AS in severe PH. Additionally, hemodynamics were evaluated in severe PH patients during exercise
before and after AS.
METHODS
We applied the multiscale CircAdapt model of the cardiovascular
system to investigate the effects of AS on hemodynamics and oxygen
distribution in severe PH. The CircAdapt model enables simulation of
beat-to-beat dynamics of the four cardiac cavities and the systemic
and pulmonary circulations (1, 13). It comprises myocardial walls,
large blood vessels, peripheral resistances, and cardiac valves. Mechanical ventricular interaction is accounted for by relating global
ventricular pump mechanics to local myofiber mechanics in the three
ventricular walls, i.e., the LV free wall, the interventricular septum,
and the RV free wall (13). In each wall, the myofiber stress-strain
relation is determined by a three-element Hill model that describes
active and passive cardiac myofiber mechanics (13), including the
Frank-Starling relation. An important characteristic of the CircAdapt
model is the reduced number of required input parameters, which is
achieved by adaptation of cardiac and vascular wall size and mass to
mechanical load (2). More specifically, parameters describing cardiac
0363-6135/12 Copyright © 2012 the American Physiological Society
http://www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
Koeken Y, Kuijpers NH, Lumens J, Arts T, Delhaas T. Atrial
septostomy benefits severe pulmonary hypertension patients by increase of left ventricular preload reserve. Am J Physiol Heart Circ
Physiol 302: H2654 –H2662, 2012. First published April 27, 2012;
doi:10.1152/ajpheart.00072.2012.—At present, it is unknown why
patients suffering from severe pulmonary hypertension (PH) benefit
from atrial septostomy (AS). Suggested mechanisms include enhanced filling of the left ventricle, reduction of right ventricular
preload, increased oxygen availability in the peripheral tissue, or a
combination. A multiscale computational model of the cardiovascular
system was used to assess the effects of AS in PH. Our model
simulates beat-to-beat dynamics of the four cardiac chambers with
valves and the systemic and pulmonary circulations, including an
atrial septal defect (ASD). Oxygen saturation was computed for each
model compartment. The acute effect of AS on systemic flow and
oxygen delivery in PH was assessed by a series of simulations with
combinations of different ASD diameters, pulmonary flows, and
degrees of PH. In addition, blood pressures at rest and during exercise
were compared between circulations with PH before and after AS. If
PH did not result in a right atrial pressure exceeding the left one, AS
caused a left-to-right shunt flow that resulted in decreased oxygenation and a further increase of right ventricular pump load. Only in the
case of severe PH a right-to-left shunt flow occurred during exercise,
which improved left ventricular preload reserve and maintained blood
pressure but did not improve oxygenation. AS only improves symptoms of right heart failure in patients with severe PH if net right-to-left
shunt flow occurs during exercise. This flow enhances left ventricular
filling, allows blood pressure maintenance, but does not increase
oxygen availability in the peripheral tissue.
H2655
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
d关O2兴LA
dt
⫽
1
共max共0, QPV兲 · 共关O2兴PV ⫺ 关O2兴LA兲
VLA
(2)
⫹ max共0, QSHUNT兲 · 共关O2兴RA ⫺ 关O2兴LA兲
⫹ max共0, QMV兲 · 共关O2兴LV ⫺ 关O2兴LA兲兲
S O2 ⫽
Fig. 1. Schematic overview of the circulation with a right-to-left shunt. LA,
left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle; QS,
systemic blood flow; QP, pulmonary blood flow; QRL, right-to-left shunt
blood flow through the atrial septal defect (ASD); V̇O2, total oxygen
consumption by the body. In contrast with carbon dioxide removal and
oxygen uptake and delivery that are determined by the effective pulmonary
flow (location A), the delivery of nutrients (location B) and the removal of
metabolic waste products (location C) are determined by systemic flow.
Part of the venous return is redirected through the ASD directly to the left
atrium, thereby bypassing the pulmonary circulation (dashed line).
[Adapted with permission from Diller et al. (4).]
and vessel mechanics are adapted to obtain physiological stresses and
strains. More detailed theoretical descriptions of the CircAdapt and
TriSeg model have been published previously (1, 2, 13). Specific input
parameters used in this study are shown in Table 1.
Oxygen distribution model. The hemodynamic pressure difference
over the ASD determines the magnitude and direction of blood flow
between the atria. Thus ASD flow in healthy persons will be from left
to right. However, in severe PH, ASD flow reverses due to the high
right atrial pressure. As a result, poorly saturated blood from the right
atrium (RA) and highly saturated blood from the pulmonary veins are
mixed in the LA.
The dynamic behavior of the CircAdapt model allows blood to
move in both directions through all cavities and vessels during the
complete cardiac cycle. Mixing of flows with different oxygen concentrations in a cavity is described by:
d关O2兴CAV
dt
⫽
1
VCAV
n
max共0, Qi兲共关O2兴i ⫺ 关O2兴CAV兲
兺i⫽1
关 O 2兴
K · 关Hb兴
· 100%
Oxygen uptake in the lungs was assumed to render an oxygen
saturation of 98% in the pulmonary veins (5, 7). The oxygen consumption in the systemic tissue was set equal to the oxygen consumption of an adult male (70 kg) in rest (300 ml/min; Refs. 4, 21).
Simulation setup. In the normal circulation, stroke volume was set
to 85 ml and heart rate to 70 beats/min, resulting in a cardiac output
of 5.95 l/min. The pulmonary resistance (Rpulm) was 18 MPa s/m3 [2.3
Wood units (WU)]. To mimic chronic compensated PH, the vascular
and cardiac wall masses and cardiac cavity volumes were adapted to
a pulmonary resistance of 60 MPa s/m3 (7.6 WU). Further increase in
pulmonary resistance in absence of structural adaptation will lead to
decompensated PH, clinically characterized by decreased pulmonary
flow. To simulate decompensated PH, pulmonary flow (Qp) was
reduced to 2.8 l/min by limiting RV stroke volume to 40 ml, while the
pressure difference between the pulmonary artery and LA was set at
14 kPa (105 mmHg). This resulted in further increase of the pulmonary resistance to 340 MPa s/m3 (46 WU; Ref. 11). To mimic AS, a
connection with an adjustable diameter was created between the two
atria.
To assess the effect of AS on systemic flow (Qs) and oxygen
delivery in PH, variations in ASD diameters (0 –16 mm), pulmonary
flows (40, 50, 60 ml/s), and degrees of PH (Rpulm; 80 –340 MPa s/m3;
11– 46 WU) were simulated. Systemic flow was defined as the
summation of flow through the ASD and predefined pulmonary flow.
Arterial oxygen saturation was calculated from the oxygen concentrations in the systemic arteries. The systemic venous saturation was
calculated from the oxygen concentration in the systemic veins. The
set of differential equations describing pressure, volume, and oxygen
Table 1. Input parameters
Value
(1)
with Qi positive for inflow and negative for outflow, [O2]CAV the
oxygen concentration in the cavity, [O2]i oxygen concentration of the
inflow, and n the number of flows entering the cavity with volume
VCAV. As defined by Eq. 1, the oxygen concentration in a cavity
depends only on the inlet blood flow sizes and oxygen concentrations.
When Eq. 1 is applied to the LA in a patient with right-to-left shunt
flow it results in:
(3)
Parameter
Mean systemic arterial blood
pressure, mmHg
Mean pulmonary arteriovenous
pressure drop, mmHg
Mean systemic blood flow, ml/s
Cardiac cycle time, s
Rest
Compensated PH
92
92
11
85
0.85
37.5
85
0.85
PH, pulmonary hypertension.
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Decompensated PH
92
105
40
0.85
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
where [O2]LA is left atrial oxygen concentration and VLA the left atrial
blood volume, QPV is blood flow entering the LA from the pulmonary
veins with oxygen concentration [O2]PV, QSHUNT is blood flow
entering the LA through the ASD with oxygen concentration [O2]RA,
and QMV is blood leaving the LA through the mitral valve with
oxygen concentration [O2]LA. In case of mitral regurgitation, QMV is
negative and contains LV oxygen concentration [O2]LV. The oxygen
distribution model is applied to the four cardiac cavities and the
systemic and pulmonary arteries and veins.
Oxygen saturation (SO2) in the blood is determined by the oxygen
concentration ([O2]) and the capacity of blood to carry oxygen.
Transport capacity of blood for oxygen depends on hemoglobin
concentration ([Hb]), set to 9.4 mol Hb/m3 (15 g/l), and its capacity to
carry oxygen (␬, 4 mol O2/mol Hb; Refs. 4, 20). The oxygen
saturation in a cavity is calculated from the oxygen concentration as
follows:
H2656
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
RESULTS
Pressure-volume relations (PV-loops) for both the LV and
RV are shown in Fig. 2 under normal circumstances (A),
compensated PH (B), decompensated PH (C), decompensated
PH with decreased pulmonary flow (D), and decompensated
PH with decreased pulmonary flow with an ASD of 14 mm in
diameter (E).
Figure 2A shows the LV and RV pressure-volume relations
for the normal healthy circulation, with high pressures (125
mmHg) generated by the LV and lower pressures (25 mmHg)
by the RV. Cardiac output and heart rate were set to normal
values at a systemic flow of 5.95 l/min with 70 beats/min,
respectively. In Fig. 2B, cardiac mechanics of compensated PH
with a pulmonary resistance of 60 MPa s/m3 (7.6 WU) is
shown. Compensated PH was associated with increase of RV
systolic pressure (⫹105%) compared with the normal circulation (Fig. 2A) and with ⬃120% increase of RV wall volume. A
shift of the RV pressure-volume relation to higher RV cavity
volumes was observed. End-diastolic RV cavity volume increased from 120 up to 140 ml and end-systolic volume
increased from 48 to 66 ml. The area enclosed by the RV
pressure-volume relation, representing external stroke work
generated by the RV, was increased compared with normal.
Decompensated PH (Fig. 2C) showed further increase of endsystolic and end-diastolic RV cavity volumes to 163 and 248
ml, respectively, and an increase of RV peak pressure from 26
to 134 mmHg. Severely decompensated PH is shown in Fig.
2D, where pulmonary flow was limited to 2.8 l/min and Rpulm
was 340 MPa s/m3 (46 WU). The reduction of pulmonary flow,
hence RV stroke volume, resulted in a decrease of workload
for the RV with 65% compared with the RV workload of the
RV in decompensated PH (Fig. 2C). Ventricular pressurevolume relations after the creation of an ASD with a diameter
of 14 mm in decompensated PH are shown in Fig. 2E. This
simulation demonstrated that the ASD enables right-to-left
shunting, causing increased mitral valve flow. As a result, LV
end-diastolic volume was increased from 84 ml before AS to
91 ml after AS. Since end-systolic volume remained unchanged, LV stroke volume increased from 40 to 47 ml,
thereby increasing systemic outflow from 2.8 to 3.3 l/min.
Stroke work as generated by the LV was increased with 19%
and RV peak pressure remained relatively unaffected by AS.
Subsequently, systemic flow (Fig. 3, A–C) and arterial oxygen saturation (Fig. 3, D–F) were studied for combinations of
different pulmonary resistances (Rpulm), pulmonary flows (Qp),
and ASD diameters. Figure 3, A and D, shows the systemic
flow and arterial oxygen saturation for a circulation with an
ASD of 8 mm in diameter. A constant Qp in combination with
an increase in Rpulm resulted in an increase of systemic flow
(Fig. 3A) and a decrease of arterial saturation (Fig. 3D). The
grey shaded area denotes simulations with right-to-left shunt
flow. The lower border of this area indicates the critical
resistance (Rcrit) at which Qs and Qp are equal, i.e., net shunt
flow is zero. If Rpulm is smaller than Rcrit, the shunt flow across
the ASD will be left-to-right, and, consequently, Qs will
decrease. Rcrit was dependent on Qp; for Qp of 60, 50, and 40
ml/s, Rcrit was 180 MPa s/m3 (23 WU), 230 MPa s/m3 (29
WU), and 290 MPa s/m3 (37 WU), respectively. It should also
be noted that Rcrit increased with increasing ASD size (Fig. 3,
A–C). Figure 3D shows that for Qp of 40 ml/s, decrease of
arterial oxygen saturation occurred at higher Rpulm compared
with a Qp of 60 ml/s. Figure 3, D–F, shows arterial oxygen
saturation in relation to Rpulm and pulmonary flow (Qp) for
three different ASD diameters. With increasing ASD diameter,
arterial oxygen saturation decreased in decompensated PH.
Shunt direction. Figure 4 shows the systemic flow as well as
the arterial and venous oxygen saturation as a function of ASD
diameter for three different values of Qp in a circulation with a
left-to-right shunt (Rpulm⬍ Rcrit; Fig. 4, A–C) and a right-to-left
shunt (Rpulm ⬎ Rcrit; Fig. 4, D–F). Qp was kept constant at 60,
50, and 40 ml/s, respectively, while Qs was allowed to increase
or decrease as a result of shunt flow. If the systemic flow
decreased due to left-to-right shunt flow (Fig. 4A), arterial
Fig. 2. Pressure volume relations (PV-loop)
A: normal (Qp ⫽ 85 ml/s, Qs ⫽ 85 ml/s, and
Rpulm ⫽ 18 MPa s/m3). B: compensated pulmonary hypertension (PH; Qp⫽ 85 ml/s, Qs ⫽ 85
ml/s, and Rpulm ⫽60 MPa s/m3). C: decompensated PH (Qp ⫽ 85 ml/s, Qs ⫽ 85 ml/s, and
Rpulm ⫽ 165 MPa s/m3). D: severe decompensated PH with decreased cardiac output (Qp ⫽
40 ml/s, Qs ⫽ 40 ml/s, and Rpulm ⫽ 340 MPa
s/m3). E: same as D, with an ASD of 14 mm in
diameter (Qp ⫽ 40 ml/s, Qs ⫽ 47 ml/s, and
Rpulm ⫽ 340 MPa s/m3).
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
concentration was solved using the MATLAB ode113 solver (The
Mathworks, Natick, MA). Oxygen saturations were averaged over
time within one cardiac cycle. Hemodynamic steady state was
achieved within 100 cardiac cycles.
Simulating exercise. Exercise was simulated in a circulation with
decompensated PH by incrementally reducing systemic peripheral
resistance, down to 50% of its normal resting value, while keeping
pulmonary resistance constant. Regulation of blood pressure by venoconstriction and renal retention are both included in the model. The
total blood volume in a subject can be divided into two volume
partitions, which are stressed and unstressed blood volume. Stressed
blood volume is defined as the volume of blood that must be removed
from the vasculature to decrease the transmural pressure of the vessels
from the existing value to zero (16). Unstressed volume, i.e., the
volume of blood in the vessels at zero transmural pressure, is disregarded in the model. Venoconstriction was modeled by lumping
recruitment of unstressed volume into increase of stressed blood
volume. Renal retention increases total circulating blood volume and,
hence, also stressed blood volume. Heart rate was allowed to increase
linearly up to 140 beats/min. Adaptation of the cardiac walls to the
increased hemodynamic load was disregarded. Simulations with and
without ASD of 16 mm were performed.
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
H2657
oxygen saturation (Fig. 4B) also slightly decreased. Venous
oxygen saturation (Fig. 4C) decreased with decreasing Qp and
with increasing ASD size. In the presence of a right-to-left
shunt, systemic output increased with shunt flow. The shunt
flow increased proportionally with increase of ASD diameter
and Qp (Fig. 4D). Arterial oxygen saturation decreased proportionally with increase of ASD diameter (Fig. 4E) due to
mixing of oxygen poor and rich blood in the LA. Venous
oxygen saturation showed to be proportionally related to Qp
and to be independent of ASD size (Fig. 4F).
Exercise. In Fig. 5, simulations in which systemic resistance
incrementally decreased to 50% of Rsys in rest, are shown for
circulations with and without ASD. The two most extreme
situations are presented. Figure 5, B and E, shows the results of
simulations with maintenance of mean systemic arterial pressure only by means of increased stressed volume and, hence,
Fig. 4. Venous and arterial oxygen saturation for 3 different pulmonary flows (Qp; dotted: 40 ml/s; dashed: 50 ml/s; and solid: 60 ml/s) with a constant oxygen
consumption for different ASD diameter with a left-to-right shunt flow (Rpulm ⬍ Rcrit; top) and a right-to-left shunt flow (Rpulm ⬎ Rcrit;, bottom). Pulmonary flow
is kept constant resulting in decreased systemic flow in a left-to-right shunt flow and increased systemic flow with right-to-left shunt flow. Systemic flow
decreases for larger ASDs in left-to-right shunting and increases with right-to-left shunting for increasing ASD diameter (A and D). Arterial oxygen saturation
shows decrease in both cases as dynamic mixing of blood is allowed (B and E). Venous oxygen saturation appears to be related to ASD size for Rpulm ⬍ Rcrit,
unrelated to ASD size for Rpulm ⬎ Rcrit, and related to Qp regardless of shunt flow direction (C and F).
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
Fig. 3. Systemic flow (A–C) and arterial oxygen saturation (D–F) in relation to pulmonary resistance (Rpulm) and pulmonary flow (Qp) for ASD diameter 8 mm
(A and D), 12 mm (B and E), and 16 mm (C and F). Grey shadow area indicates occurrence of right-to-left shunt flow.
H2658
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
Fig. 5. Aortic valve flow (QAO) and pressures in the aorta (pAO) and left ventricle (pLV) for patient without (A–C) and with an ASD (D–F) while systemic
resistance was decreased to 95% (A and D) or 50% (B, C, E, and F) of its value at rest. Heart rate (HR)was kept constant at 70 beats/min (A, B, D, and E) and
was increased to 140 (C and F). G: stressed blood volume increase in relation to the decrease in systemic resistance for simulations without and with ASD. Solid
lines indicate simulations with a constant heart rate of 70 beats/min and dashed lines with an increasing heart rate (Table 2). Increase of left ventricular
contractility resulted in a decrease of stressed volume.
increased stroke volume. Figure 5, C and F, shows results of
simulations with maintenance of mean systemic arterial pressure by increase of heart rate up to 140 beats/min. In the
presence of an ASD, increase in stressed blood volume and,
hence, of systemic venous pressure, was present, but not as
abundant as in the simulations without an ASD (Fig. 5G and
Table 2). Due to the increased RV preload, pulmonary flow in
the simulations with ASD increased up to 130% if systemic
resistance was reduced to 50% of its value at rest. The combination of increased systemic venous pressure and decreased
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
H2659
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
Table 2. Simulated values for flows, pressures, and volumes in rest and during exercise
No ASD
Rest
100
50
50
0
70
1.00
109
80
2
92
98
193
115
2
2
300
98
49
298
98
50
Exercise
90.0
55
55
0
84
1.04
108
79
2
92
109
206
125
2
2
300
98
53
331
98
50
70
69
69
0
112
1.23
109
79
3
92
137
254
153
5
3
300
98
63
415
98
50
Rest
50
86
86
0
140
1.64
112
78
6
92
173
338
188
15
7
300
98
70
514
98
50
100.0
50
52
⫺2
70
1.00
109
77
2
92
98
191
114
2
2
300
93
47
313
93
45
Exercise
90.0
53
58
⫺5
84
1.01
109
78
2
92
104
196
119
2
2
300
93
51
347
92
44
70.0
59
74
⫺15
112
1.08
110
78
2
92
116
210
125
3
3
300
93
60
444
90
42
50.0
66
98
⫺32
140
1.25
114
76
4
92
134
239
148
5
5
300
90
65
589
82
34
ASD, atrial septal defect; Qp, pulmonary flow; Qs, systemic flow; QASD, shunt flow; HR, heart rate; V, normalized stressed blood volume; SBP, systemic blood
pressure; DBP, diastolic blood pressure; CVP, central venous pressure; MAP, mean arterial pressure; pPA, pulmonary artery pressure; EDVRV, right ventricular
end-diastolic volume right ventricle; pRV, systolic right ventricular pressure; pRA, right atrial pressure; pLA, mean left atrial pressure; V̇O2, oxygen consumption;
Art. sat., mean arterial oxygen saturation; Ven. sat., mean venous oxygen saturation.
systemic resistance led to an increase in right-to-left shunt flow
across the ASD. The ASD flow was responsible for 70% of the
almost twofold increase in systemic flow needed in simulations
with 50% decreased systemic resistance.
The exponential relation between stressed blood volume and
systemic resistance is steeper for the simulations without ASD
compared with the ones with ASD (Fig. 5G). In the absence of
an ASD, decrease of Rsys and, hence, in mean systemic arterial
pressure was mainly compensated for by an increase in stressed
blood volume, leading to increased systemic venous pressure.
On the one hand, this increased systemic venous pressure will
lower the pressure difference across the peripheral circulation.
Consequently, a smaller increase in systemic output is needed
to maintain systemic arterial pressure in case of a decreased
systemic resistance. On the other hand, the increased systemic
venous pressure will lead to an increased RV preload and,
hence, increased pulmonary flow. In a circulation without an
ASD, increase of heart rate resulted in a large increase of
stressed volume. However, the increase of stressed blood
volume was less than in a simulation with constant heart rate
(Fig. 5, C, F, and G, dotted line).
In a circulation with an ASD, stressed blood volume did not
differ between simulations with constant and increased heart
rate. The effect of increased LV contractility on the reduction
of stressed blood volume increase was limited. A twenty-five
percent increase in active stress generation resulted in 1%
reduction of the stressed volume increase in a circulation
without ASD, while the reduction was 4% in a circulation with
ASD. Although systemic arterial pressure could be maintained
in simulations without an ASD, it can be appreciated, both
from Table 2 and Fig. 5, that this process is accompanied by an
increase of stressed blood volume of ⬎30%, which is above
values mentioned in literature (15, 16).
DISCUSSION
The CircAdapt model of the human heart and circulation has
been extended with an oxygen distribution model. The resulting model has been used to evaluate the effect of AS on both
hemodynamics and oxygen transport in patients with severe
PH. In our simulations, the direction of shunt flow was dependent on both pulmonary flow and resistance. Shunt flow direction reversed when pulmonary vascular resistance (Rpulm) exceeded a critical value, Rcrit. This critical resistance value had
a distinct inverse proportional relation with pulmonary flow
(Qp) and was not related to ASD size. Although arterial oxygen
saturation was drastically decreased by right-to-left shunt flow,
venous oxygen saturation did not change, suggesting similar
oxygen delivery to the tissue. Simulations of exercise with
severe PH in which pulmonary resistance exceeds Rcrit showed
that these PH patients benefit from an ASD because right-toleft atrial shunt flow facilitates maintenance of systemic arterial
pressure.
Shunt direction. Most patients with severe PH suffer from
syncope, fatigue, and right heart failure (10). The presence of
right heart failure suggests that the RV is working at maximal
capacity in these patients. We therefore assumed that pulmonary flow and pulmonary vascular resistance do not change
after AS and retain their maximal value. In contrast, the
systemic flow could change in our simulations, enabling investigation of the effect of AS on systemic flow.
ASD flow direction is determined by the pressure difference
between the RA and LA, whereas amount of flow is determined by ASD size and the pressure difference across the
ASD. Normally, LA pressure exceeds RA pressure and an
ASD will result in a left-to-right shunt flow and an increased
pulmonary flow with respect to the systemic flow. In a PH
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
Exercise, %Rsys
Qp, ml/s
Qs, ml/s
QASD, ml/s
HR, beats/min
V, ⫺
SBP, mmHg
DBP, mmHg
CVP, mmHg
MAP, mmHg
pPA, mmHg
EDVRV, ml
pRV, mmHg
pRA, mmHg
pLA, mmHg
V̇O2 constant, ml/min
Art. sat., %
Ven. sat., %
V̇O2 variable, ml/min
Art. sat., %
Ven. sat., %
ASD 16 mm
H2660
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
Fig. 6. Oxygen transport in a circulation with a constant pulmonary flow of 2.5
l/min, constant oxygen consumption (V̇O2; 0.2 mmol/s), and right-to-left shunt
flow through the ASD. Systemic oxygen transport (SOT) increases with
increasing systemic flow (line B, corresponding with location B in Fig. 1), but
also more oxygen is returned to the heart through the venae cavae (venous
return; line C, corresponding with location C in Fig. 1). The amount of oxygen
that can be used by the systemic tissue, the effective systemic oxygen transport
(SOTeff), is unchanged (line A, corresponding with location A in Fig. 1).
will be the sum of Qs and the left-to-right shunt flow. In the
presence of a right-to-left shunt, desaturated blood will bypass
the pulmonary vascular bed and Qs will be the sum of Qp and
the right-to-left shunt flow. Therefore, using Qs as a representative of the amount of blood passing through the lungs that can
be oxygenated, is incorrect. We should look for the quantity of
systemic venous blood that is directed to the pulmonary circulation and oxygenated, the so-called effective pulmonary flow
(Qep). Consequently, the effective systemic oxygen transport
capacity (SOTeff) should be defined as the product of Qep and
pulmonary venous saturation. SOTeff can be applied to both
normal circulations and those with shunts. Reduction of RV
preload, which will occur during right-to-left shunt flow, will
not result in an increase in Qep. Although SOT will increase in
this situation, SOTeff will definitely not increase (Fig. 6).
Relieve of symptoms and the increase in exercise capacity in
PH patients after AS can therefore not be explained by changes
in oxygen distribution.
Nutrients and metabolic waste products transport. In contrast with oxygen uptake and delivery that are determined by
the effective pulmonary flow, the delivery of nutrients (Fig. 1,
location B) and the removal of metabolic waste products (Fig.
1, location C) are determined by systemic flow. Therefore, an
increase in systemic flow, induced by a right-to-left shunt flow
through the ASD, may ameliorate the condition in the patient
by improved delivery of nutrients and increased removal of
accumulated waste products. However, removal of carbon
dioxide still remains dependent on effective pulmonary flow
(Fig. 1, location A).
Exercise. Systemic arterial blood pressure is maintained
during exercise, because the decrease in systemic vascular
resistance will be compensated for by an increase in systemic
blood flow. In severe PH, however, increase of systemic output
is restricted due to the incapability of the RV to further increase
pressure and, hence, pulmonary flow through the highly resistant pulmonary vasculature. In these patients, resting heart rate
and finally stressed blood volume are increased to enlarge
venous return to the heart to maintain systemic arterial blood
pressure. Increase in stressed blood volume by venoconstriction and by renal retention are both included in the model. In
the absence of an ASD, a larger increase in stressed blood
volume is needed to compensate for the decrease in systemic
resistance and subsequent arterial pressure drop. The increase
in stressed blood volume is, however, not infinite. Considering
a healthy person with a total blood volume of 80 ml/kg, ⬃30%
(25 ml/kg) of this total volume is stressed volume (16). Venoconstriction can mobilize 7.5 ml/kg blood from the unstressed
blood volume and increase stressed volume to maintain blood
pressure (15). Therefore, an increase of stressed volume up to
30% is supposed to be reasonable. In a circulation without an
ASD, maximal recruitment of blood volume will occur at
lower exercise levels than in a circulation with an ASD (Fig.
5). Once PH patients without ASD start to exercise, they reach
either the limit of their capacity to increase their stressed blood
volume acutely, or the RV will be pushed towards its limits
with respect to cavity volume or pressure generation (Table 2).
Consequently, blood pressure will drop and the patient will
suffer from dizziness or fainting. After AS, the PH patient will
be able to increase systemic output without the need of maximal blood volume recruitment. Systolic blood pressure will
increase during exercise because LV stroke volume is in-
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
patient, an ASD will only result in right-to-left shunt flow,
when the PH is severe enough (6), so that RA pressure exceeds
LA pressure (Figs. 3 and 4). Due to right-to-left shunt flow,
systemic blood flow will increase, albeit that systemic arterial
blood oxygen content is reduced due to mixing of oxygen rich
and poor blood in the atria (Fig. 4E).
Although zero net shunt flow is hemodynamically unimportant because Qs and Qp remain unchanged, it can reduce
systemic arterial oxygen saturation (Fig. 3, D–F). Dynamic
movement of blood from left-to-right and vice versa allows
oxygen rich and oxygen poor blood to mix in the atria. This
mixing will increase oxygen saturation in the RA, but it will
also reduce LA saturation and, hence, arterial oxygen saturation (Fig. 3D). Conceivably, arterial oxygen saturation cannot
be used to quantify right-to-left shunt flow, as is common
clinical practice. To estimate net shunt flow and direction in the
clinical setting, we therefore recommend to use additional
measurements, for example of LV end-diastolic volume or
pressure, as done by Sandoval et al. (19).
Oxygen availability. Systemic oxygen transport (SOT) is
defined as systemic flow multiplied by the systemic arterial
oxygen saturation (Fig. 6). SOT defined as such will increase
in the presence of a right-to-left shunt after AS in severe PH
patients, because the decrease in arterial oxygen saturation is
overcompensated for by the increase in systemic flow (6, 10,
14, 17, 19, 23). However, the systemic venous blood that
passes through the shunt will travel round the circulation and
arrive back at the RA with the same oxygen content and,
therefore, will not add to the capacity to transport oxygen into
the tissue (12).
In a normal circulation, the pulmonary and systemic circulations are connected in series and, hence, Qp and Qs are equal
and interchangeable. In the presence of a left-to-right shunt,
oxygenated blood will bypass the systemic vascular bed and Qp
H2661
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
Table 3. Simulations results compared with clinical data from literature
No ASD Present
Simulation
Kerstein et al. (10)
Simulation
Kurzyna et al. (11)
Kerstein et al. (10)
Means
Means
SD
Means
SD
Means
Means
SD
Means
SD
3.0
3.0
0
70
109
80
2
92
98
115
2
2
300
98
49
15
2.5
2.5
0
—
—
—
—
—
62.1
—
15.2
4.0
—
93
42
14.8
0.8
0.8
0
—
—
—
—
—
17.5
—
7.0
3.5
—
3
5
2.9
—
—
0
—
—
—
—
91.9
68.8
—
11.2
4.8
—
97.4
—
15.3
—
—
0
—
—
—
—
14.2
22.1
—
7.1
2.6
—
1.7
—
1.6
3.0
3.1
0.1
70
109
77
2
92
98
114
2
2
300
93
47
15
2.4
3.0
0.6
—
—
—
—
—
66.2
—
14.4
4.3
—
84.0
—
—
0.9
1.1
0.3
—
—
—
—
—
17.9
—
5.9
4.1
—
3.8
—
—
—
—
—
—
—
—
—
90.2
71.9
—
10.0
7.3
—
89.0
—
14.8
—
—
—
—
—
—
—
16.1
23.7
—
4.3
2.8
—
7.1
—
2.5
Hb, hemoglobin concentration.
creased. The increase in systemic flow is facilitated by the
blood flow through the ASD, which abolishes the preload
restriction of the LV (Fig. 5). PH patients with Rpulm close to
Rcrit in rest may benefit by AS during exercise, because shunt
direction may reverse due to decrease of systemic resistance.
As discussed above, blood passing right-to-left through the
ASD has low oxygen saturation and therefore will not contribute to additional oxygen delivery to the tissue. Over the long
term, however, the decreased arterial oxygen saturation might
induce hemoglobin production (10, 12). Although the resultant
increased hemoglobin levels will increase the capacity of blood
to carry oxygen, it will also increase blood viscosity, resulting
in higher workload for the ventricles and increased risk of
thrombosis.
Clinical relevance of model simulations. As compared with
a previous modeling study in this field (4), the present study
uses a multiscale model of cardiovascular system dynamics
that relates global pulmonary and systemic hemodynamics to
pump mechanics of the cardiac chambers and to local mechanics of the constituting myocardial tissue being responsible for
the pumping action of the heart. It allows for increase of
stressed volume by venoconstriction during exercise. The characteristics of the systemic and pulmonary circulation could be
adjusted to simulate various degrees of pulmonary hypertension and exercise. Moreover, the CircAdapt model gives insight in volumes and pressures which are mostly inaccessible
in patients. Table 3 shows the results of the simulations and
clinical data from literature (10, 11). Overall, simulated pressures and volumes are in agreement with clinical data. However, it should be mentioned that simulated mean RA pressure
is low and mean pulmonary artery pressure is high compared
with corresponding clinical data. Since atrial pressure is directly related to diastolic ventricular pressure, low RA pressure
implies low diastolic RV pressure. This suggests that the RV is
more compliant in the model than in severe PH patients. High
pulmonary artery pressure suggests a strong RV that is able to
produce high pressures. Despite these quantitative differences,
the qualitative effects of AS on cardiovascular function in our
simulations are in agreement with clinical findings. Simulating
a less compliant and weaker ventricle would not change these
qualitative effects but only amplify effects quantitatively.
Several clinical investigators evaluated atrial septostomy in
severe PH patients (9 –11, 19). Modeling studies are less
commonly used, although they can provide valuable mechanistic insights in cardiovascular function dynamics before and
after atrial septostomy both in rest and in exercise. Our study
shows that AS might only be beneficial to patients with severe
PH and that this benefit originates from an increase of LV
preload and not from an increase in oxygen delivery to the
tissue, as stated in previous studies (6, 10, 17, 19, 23). These
simulation results may lead to improved success rate of the
procedure by improving patient selection and, hence, withholding the procedure to likely nonresponders. Ultimately, an
inverse modeling approach, in which the model is personalized
by fitting simulated hemodynamics to a set of patient-specific
data, might facilitate patient selection for AS and prediction of
outcome by simulation of the intended intervention.
Conclusion. Our multiscale model of the cardiovascular
system enables the exploration of the effects of AS in severe
PH patients on oxygen distribution and cardiovascular mechanics. Atrial septostomy in PH patients results in a right-to-left
shunt flow if pulmonary vascular resistance exceeds a critical
value. This critical resistance value is inversely proportional
with pulmonary flow and is hardly influenced by ASD size.
Although right-to-left shunt flow does not lead to increased
oxygen availability, the condition of a PH patient may improve
because maintenance of systemic arterial pressure is facilitated
during exercise.
GRANTS
This work was supported by the Dutch Heart Foundation Grant 2007B203.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the author(s).
AUTHOR CONTRIBUTIONS
Author contributions: Y.K., T.A., and T.D. conception and design of
research; Y.K. performed experiments; Y.K., N.H.K., T.A., and T.D. analyzed
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
QP, l/min
QS, l/min
QASD, l/min
HR, beats/min
SBP, mmHg
DBP, mmHg
CVP, mmHg
MAP, mmHg
pPA, mmHg
pRV, mmHg
pRA, mmHg
pLA, mmHg
V̇O2, ml/min
Art. sat., %
Ven. sat., %
Hb, g/dl
Kurzyna et al. (11)
ASD Present
H2662
ATRIAL SEPTOSTOMY IN PULMONARY HYPERTENSION
data; Y.K., N.H.K., J.L., T.A., and T.D. interpreted results of experiments;
Y.K. and T.D. prepared figures; Y.K. and T.D. drafted manuscript; Y.K.,
N.H.K., J.L., T.A., and T.D. edited and revised manuscript; Y.K., N.H.K., J.L.,
T.A., and T.D. approved final version of manuscript.
REFERENCES
AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00072.2012 • www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.33.2 on June 16, 2017
1. Arts T, Delhaas T, Bovendeerd P, Verbeek X, Prinzen FW. Adaptation
to mechanical load determines shape and properties of heart and circulation: the CircAdapt model. Am J Physiol Heart Circ Physiol 288: H1943–
H1954, 2005.
2. Arts T, Lumens J, Kroon W, Delhaas T. Control of whole heart
geometry by intramyocardial mechano-feedback: a model study. PLoS
Comput Biol 8: e1002369, 2012.
3. Austen WG, Morrow AG, Berry WB. Experimental studies of the
surgical treatment of primary pulmonary hypertension. J Thorac Cardiovasc Surg 48: 448 –455, 1964.
4. Diller GP, Lammers AE, Haworth SG, Dimopoulos K, Derrick G,
Bonhoeffer P, Gatzoulis MA, Francis DP. A modelling study of atrial
septostomy for pulmonary arterial hypertension, and its effect on the state
of tissue oxygenation and systemic blood flow. Cardiol Young 20: 25–32,
2010.
5. Diller GP, Uebing A, Willson K, Davies LC, Dimopoulos K, Thorne
SA. Analytical identification of ideal pulmonary-systemic flow balance in
patients with bidirectional cavopulmonary shunt and univentricular circulation: oxygen delivery or tissue oxygenation? Circulation 114: 1243–
1250, 2006.
6. Galie N, Hoeper MM, Humbert M, Torbicki A, Vachiery JL, Barbera
JA. Guidelines for the diagnosis and treatment of pulmonary hypertension: the Task Force for the Diagnosis and Treatment of Pulmonary
Hypertension of the European Society of Cardiology (ESC) and the
European Respiratory Society (ERS), endorsed by the International Society of Heart and Lung Transplantation (ISHLT). Eur Heart J 30: 2493–
2537, 2009.
7. Guyton AC, Hall JE. Textbook of Medical Physiology (12th ed.). Philadelphia, PA: W. B. Saunders, 2011.
8. Hausknecht MJ, Sims RE, Nihill MR, Cashion WR. Successful palliation of primary pulmonary hypertension by atrial septostomy. Am J
Cardiol 65: 1045–1046, 1990.
9. Keogh AM, Mayer E, Benza RL, Corris P, Dartevelle PG, Frost AE.
Interventional and surgical modalities of treatment in pulmonary hypertension. J Am Coll Cardiol 54: S67–S77, 2009.
10. Kerstein D, Levy PS, Hsu DT, Hordof AJ, Gersony WM, Barst RJ.
Blade balloon atrial septostomy in patients with severe primary pulmonary
hypertension. Circulation 91: 2028 –2035, 1995.
11. Kurzyna M, Dabrowski M, Bielecki D, Fijalkowska A, Pruszczyk P,
Opolski G. Atrial septostomy in treatment of end-stage right heart failure
in patients with pulmonary hypertension. Chest 131: 977–983, 2007.
12. Lammers AE, Haworth SG, Diller GP. Atrial septostomy in patients
with pulmonary hypertension: should it be recommended? Expert Rev
Respir Med 5: 363–376, 2011.
13. Lumens J, Delhaas T, Kirn B, Arts T. Three-wall segment (TriSeg)
model describing mechanics and hemodynamics of ventricular interaction.
Ann Biomed Eng 37: 2234 –2255, 2009.
14. Micheletti A, Hislop AA, Lammers A, Bonhoeffer P, Derrick G, Rees
P. Role of atrial septostomy in the treatment of children with pulmonary
arterial hypertension. Heart 92: 969 –972, 2006.
15. Rothe CF. Point-counterpoint: active venoconstriction is/is not important
in maintaining or raining end-diastolic volume and stroke volume during
exercise and orthostasis. J Appl Physiol 101: 1262–1264, 2006.
16. Rothe CF. Reflex control of veins and vascular capacitance. Physiol Rev
63: 1281–1342, 1983.
17. Rothman A, Sklansky MS, Lucas VW, Kashani IA, Shaughnessy RD,
Channick RN, Auger WR, Fedullo PF, Smith CM, Kriett JM, Jamieson SW. Atrial septostomy as a bridge to lung transplantation in
patients with severe pulmonary hypertension. Am J Cardiol 84: 682–686,
1999.
18. Rozkovec A, Montanes P, Oakley CM. Factors that influence the
outcome of primary pulmonary hypertension. Br Heart J 55: 449 –458,
1986.
19. Sandoval J, Gaspar J, Pulido T, Bautista E, Martinez-Guerra ML,
Zeballos M. Graded balloon dilation atrial septostomy in severe primary
pulmonary hypertension. A therapeutic alternative for patients nonresponsive to vasodilator treatment. J Am Coll Cardiol 32: 297–304, 1998.
20. Santamore WP, Barnea O, Riordan CJ, Ross MP, Austin EH. Theoretical optimization of pulmonary-to-systemic flow ratio after a bidirectional cavopulmonary anastomosis. Am J Physiol Heart Circ Physiol 274:
H694 –H700, 1998.
21. Sherwood L. Human Physiology (6th ed.). Belmont, CA: Thomson
Brooks/Cole, 2007.
22. Thanopoulos BD, Georgakopoulos D, Tsaousis GS, Simeunovic S.
Percutaneous balloon dilatation of the atrial septum: immediate and
midterm results. Heart 76: 502–506, 1996.
23. Zierer A, Melby SJ, Voeller RK, Moon MR. Interatrial shunt for chronic
pulmonary hypertension: differential impact of low-flow vs. high-flow
shunting. Am J Physiol Heart Circ Physiol 296: H639 –H644, 2009.