1
Chapter 1
INTRODUCTION TO CIRCULATORY AND RESPIRATORY SYSTEM
MODELING
Gianfranco Ferrari, Marek Darowski, Tomasz Gólczewski, Krystyna Górczyńska, Maciej
Kozarski
ABSTRACT
This chapter is focused on circulatory and respiratory system modeling. It includes a brief history
of circulatory and respiratory system modeling development and a short description of the state
of art. In the chapter also basic classification of mechanical circulatory and respiratory assistance
is presented. The last part of the chapter deals with innovative approaches to modeling of both
circulatory and respiratory system which concern hybrid models and virtual organs. Hybrid
modeling consists in merging numerical and physical models or devices exchanging data among
them in real time. In this way it is possible to use the best features of numerical models
(accuracy, flexibility, low cost) without loosing the possibility of testing physical devices. Hybrid
approach is extremely useful for testing mechanical heart and lung assist devices. The idea of
virtual organs is applied to the respiratory system and is created to analyze a whole class of
problems which are not known or clearly defined before creation of a virtual organ.
1.1. INTRODUCTION
A model is a set of equations representing some aspects of a physical phenomenon. The models
representing biological systems are a special class of models that however obey to the general
principles underlying the development of any model. The difference lies in the complexity of
biological systems. In fact, for any model and especially for models of biological systems, the
key point is the amount and type of necessary simplifications. As a matter of fact any model is a
simplified representation of a real phenomenon. It is important is to know exactly the impact of
the simplification on the comprehensive system and which aspects of its behavior can be taken
into account by the model. These considerations are critical when models represent complex
biological systems where, as for example in the case of the circulatory or respiratory systems,
there are several subsystems operating concurrently. It is then evidently fundamental to separate
2
the effects of the different subsystems to achieve the necessary simplification but it is necessary
as well to develop the models preserving the possibility to take into account the interactions
among the different subsystems and the circulatory and respiratory mechanical support when it is
present.
1.2. MECHANICAL CIRCULATORY ASSISTANCE (MCA)
Krystyna Górczyńska
This paragraph does not pretend to give an exhaustive description of mechanical circulatory
assistance but rather to evidence, after some brief historical remarks, the features that may be
relevant for modeling.
The MCA can be used as a bridge to recovery when the heart inefficiency is reversible or as a
bridge to transplant when waiting for a donor’s heart. In the cases of the end-stage heart failure
only a total artificial heart can be taken into account [1],[2].
The patient’s heart condition, as well as the type of his heart disease, impose the choice of the
assist device to be used. For example, the IABP is widely used for assistance in the case of left
ventricular failure (e.g. induced by the coronary dysfunction) to unload the ventricle. On the
contrary, the main goal of the parallel LVAD assistance is to discharge pulmonary stagnation by
creating a bypass of the inefficient left ventricle to shift a volume of blood from the left atrium to
the aorta.
Thus, the MCA can be mainly used:
- in the case of heart infarction to decrease cardiac work by unloading the inefficient
ventricle during the systole and to improve its coronary perfusion during the diastole,
- to stabilize patient’s hemodynamic conditions in pre-operation stages,
- in post-operation stages (e.g. after coronary bypasses, heart valves grafts) to enable
disconnecting the patient from the cardio-pulmonary bypass,
- in the case of sudden (e.g. post-operation) ventricular insufficiency.
Historically, roller (peristaltic) pumps were the first step in MCA development. They are a
typical extra-corporeal device due to their relatively large dimensions and weight. They have
been used in heart-lung machines during surgical operations. Roller pumps were in a sense the
forerunners of MCA as it is known nowadays.
3
The next step was the development of rotary blood pumps (axial, radial, diagonal) [3],[4] that are
characterized by smaller dimensions and invasiveness. The rotary blood pumps, similarly to other
continuous-flow pumps do not need any additional heart valves. The miniature axial flow pump
developed at the Baylor School of Medicine (Houston, Texas) [5],[6],[7] has a weight of 94 g
only.
Advantages of rotary pumps are: small hemolysis and ability of appliance for a couple of days,
small dimensions and small filling volume. The main disadvantage is necessity of limiting its
rotational speed because of stress and cavitation [8].
The applied methods of assistance differ in a goal (recovery, bridge to transplant) and mode of its
usage (pulsatile, continuous flow) and connection to the circulatory system (“in series”, parallel)
as well as an extent of invasiveness of the surgical intervention.
The development of total artificial heart was parallel to MCA development. However some of the
technical solutions were adopted in both MCA and total artificial heart [8],[9].
A simple definition of mechanical circulatory assistance (MCA) can be given from strictly
engineering point of view: in the case of heart failure one or both ventricles may become no more
able to transfer to the circulatory system the energy necessary for organ perfusion. The role of the
MCA is delivering to the circulatory system the energy difference between the energy demand
and its supply produced by the failing ventricle.
Of course, the problem is not only an engineering one - the role of the MCA is more complex as
it interacts with a biological system. Nevertheless, the above simple statement leads to the
effective synthesis of one of the main goals of MCA; its importance changes in relation to the
aim of the assistance. If the MCA is used as a bridge to transplant or long term implant, the
sufficient energy transfer to the load is of a primary importance to fulfill the minimal
requirements for patient’s wellbeing. The assistance can be also applied for heart recovery. The
mechanical support consists in this case in inserting an additional pump into the circulatory
system to unload the insufficient heart and to improve hemodynamic parameters of the patient,
promoting in this way heart recovery. The energy transfer to the load is still fundamental and the
assistance should be managed to create the best conditions for heart recovery. Then, beyond the
relation between MCA and the arterial load, other variables such as coronary perfusion should be
taken into account. MCA can be inserted into the circulatory system “in series” or in parallel and
the generated flow can be pulsatile or non pulsatile [10]. This is not the place to add further
4
contributions to the debate concerning features of different types of MCA; what is important is to
be aware that different insertions and different flow types imply different solutions for modeling
the device and heart-MCA interaction as it will be shown in the examples in chapter 4. It should
be mentioned that flow pulsatility implies often but not always the presence of mechanical heart
valves that, in the simplest version, can be modeled as ideal diodes.
It is worthwhile to give a short classification of MCA in relation to its insertion place into the
circulatory system:
1. “In series” assistance:
• Left Ventricular Assist Device (LVAD).
• Rotary pumps.
• Centrifugal pumps.
• Axial flow pumps.
• Intra-Aortic Balloon Pump (IABP).
• Para-aortic counterpulsation.
• External counterpulsation - a non-invasive method based on sequential
compression of a vascular bed by cuffs placed on the legs [11],[12],[13],[14],[15].
2. Parallel assistance:
• Left Ventricular Assist Device (LVAD).
• Bi-Ventricular Assist Device (BVAD) – simultaneous assistance of the left and
right ventricle.
• Roller pumps.
• Rotary pumps.
“In series” assistance connection includes several devices, pulsatile or non pulsatile. One of the
most representative is IABP [16],[17],[18],[19],[20],[21],[22] a simple to apply and rather
effective device; it is widely used for assistance in the case of left ventricular failure induced by
e.g. coronary dysfunction. An example of IABP modeling will be presented in Chapter 4, so a
short description of this device action is needed. In the IABP, the role of the pump actuator is
played by the balloon inserted into the descending aorta through the femoral artery. The balloon,
connected to the driving unit by the catheter, is periodically filled with gas and emptied,
according to the heart rate (HR) and systole-to-diastole ratio (S/D) of the patient. The goal of the
IABP assistance is heart muscle’s recovery by its unloading, due to systolic aortic pressure
5
decrease, and better perfusion, by the diastolic aortic pressure increase - diastolic augmentation
(Figure 1). The increased coronary flow attained in this way can rise the ventricular contractility
finally resulting in the improvement of general hemodynamic conditions [16].
Figure 1 : IABP assistance. Pas – aortic pressure (from „Biocybernetyka i Inżynieria Biomedyczna 2000”, Vol. 3. Copyright by EXIT , 2001, reproduced by permission).
The IABP assistance effectiveness is dependent, besides initial circulatory and ventricular
conditions, on the balloon emptying and filling velocity as well as on time delay of the balloon
inflation and deflation (in relation to the QRS complex of the patient’s ECG) [23], on the balloon
shape, volume and position in the aorta, on the sort of gas filling the balloon and finally on the
posture [24]. Characteristic for the balloon assistance is the systolic aortic pressure and end-
diastolic pressure (EDP) decrease and the diastolic augmentation.
Another device that will be modeled in Chapters 4 and 8 is parallel, pulsatile LVAD. This device,
important for historical and practical reasons, deserves some comments.
In the case of blood stagnation in the pulmonary system, usually caused by left ventricular
inefficiency, the parallel assistance can be applied. Independently of the structure of the assist
device pumping unit, the role of such assistance is mainly to shift a volume of blood from the left
atrium to the aorta, (Figure 2), to create so called bypass of the insufficient left ventricle. When
6
the heart is assisted by the LVAD applied as a bridge to recovery, the artificial ventricle only
partially takes over the pumping function of the native ventricle [25],[26]. The LVAD action in
that case should be synchronous with the ECG signal and in general carried out on the principle
of counterpulsation; owing to such kind of assistance, harmful stagnation in the pulmonary
circulation can be discharged bringing about the arterial pulmonary and left atrial pressure drop
along with significant rise in the ventricular and arterial pressure and in total cardiac output.
Figure 2 : Block diagram of the parallel ventricular assistance (LVAD): LA – left atrium, LV – left ventricle, ALV – artificial left ventricle, ECU - electronic control system, PDU – pneumatic
drive unit, P± - control pressure (from „Biocybernetyka i Inżynieria Biomedyczna 2000”, Vol. 3. Copyright by EXIT , 2001, reproduced by permission).
Some of the assist devices available on the market for heart recovery or as a bridge to
transplant, are extracorporeal, some – implantable [8],[27],[28],[29],[30],[31].
In the case of inefficiency of both ventricles, biventricular assistance (BVAD) can be applied
using two identical LVAD and RVAD diaphragm pumps [32],[33], rotary pumps [34],[35] or e.g.
a combination of the para-aortic left heart-assist pump connected to the aorta and the IABP
inserted into the pulmonary artery [36].
Pulsatile flow assistance was the first type of assistance to be developed and clinically used. It is
essentially based on a chamber separated from the remaining part of the circulatory system by
means of artificial valves. In the assist devices, the energy can be transferred to the blood in
7
different ways: pneumatically, electromechanically, electromagnetically. The flow pulsatility is
determined, together with the artificial heart valves, by a moving piston or diaphragm that creates
the conditions for filling and ejection of the assist ventricle.
1.3. MECHANICAL RESPIRATORY ASSISTANCE
Marek Darowski, Tomasz Gólczewski
The International Standards Organization Committee is still working on standards for ventilator
mode nomenclature, but recent publications allow us to present how to distinguish basic modes of
ventilation [37],[38].
The classification of ventilation modes is based on the following criteria:
1) what is the independent variable (pressure or volume) during ventilation, that can be
controlled by an anesthesiologist;
2) how breaths (mandatory or/and spontaneous) are sequenced, triggered and cycled.
According to the 1st criterion we have:
a) pressure controlled modes: artificial pressure controlled ventilation (PCV), continuous
positive airway pressure (CPAP), pressure-support ventilation (PSV), proportional-assist
ventilation (PAV) and neurally adjusted ventilatory support;
b) volume controlled modes, e.g. artificial volume controlled ventilation (VCV),
characterized by a mandatory tidal volume delivered by a ventilator to the lungs,
independent of changes in lungs mechanics and patient’s inspiratory effort as well as
chosen inspiratory flow patterns (constant, ascending/descending ramps or sinusoidal),
According to the criterion 2 we have:
a) continuous mandatory ventilation with all mandatory breath, from a ventilator,
b) continuous spontaneous ventilation, with all spontaneous breaths, when inspiratory
periods are triggered and cycled by a patients,
c) intermittent mandatory ventilation, when ventilatory assistance consists of spontaneous
and mandatory breaths.
8
During ventilatory support, the lungs and a ventilator create one mechanical system in which
interaction between variables values (airway and alveolar pressure, inspiratory and expiratory
flow, tidal volume) depends on the respiratory system mechanical properties (e.g. airways
resistance, and the lungs/thorax compliance), the mode of ventilation applied, and ventilatory
parameters settings.
VCV always gives the required minute ventilation. However, also PCV may provide a desired
ventilation in steady state conditions, i.e. when the lungs mechanisms do not change. On the other
hand, PCV is associated with a decelerating inspiratory flow, which is supposed to be responsible
for a better gas exchange1. However, in most of the modern respirators a decelerating inspiratory
flow pattern is also available during VCV [39],[40],[41],[42]. It implies that the discussion on
optimal control variables may concern not a specific mode of the artificial ventilation (PCV or
VCV) but rather the inspiratory pressure or flow pattern itself [43],[44]. The pattern may be
expected to influence both pressure and ventilation distributions in the lungs, in particular in
cases of lung pathology, which is often manifested by differences in lung mechanics between left
and right lobes.
VCV always gives a preset tidal volume VT because the airflow rate is the independent (control)
variable. However, the airway and alveolar pressures cannot be predetermined because they
depend on the lung compliance and airway resistance (Raw) (Figure 3a). In PCV, the airway
pressure is the control variable but VT changes according to the lung mechanics properties
(Figure 3b). Thus, either the pressures or VT is behind the control.
For the above reasons, in VCV there is a risk of:
• barotrauma when the pressures increase too much because of the lungs compliance smaller than
the expected,
• volutrauma of one lung in asymmetric pathologies, e.g. when airways of the other lung are
obstructed (the whole VT is loaded into the healthy lung).
In PCV, there is a risk of:
• hyperventilation (volutrauma) if the lungs compliance is greater than the expected,
• hypoventilation if the compliance is smaller or Raw is greater than the expected.
1 An example of the use of a virtual respiratory system in investigation of inspiratory pattern influence on gas exchange and blood oxygenation is presented in the Chapter 6.5
9
Figure 3 : The difference between power controlled ventilation (PoCV) and volume (VCV) or pressure (PCV) controlled ventilation from the cybernetic point of view.
R – respirator, RS – respiratory system, P – pressure, Q – airflow, VT - tidal volume, Po - power.
a) Q and VT are preset but P is uncontrolled in VCV.
b) P is preset but Q and VT are uncontrolled in PCV: their values depend on RS properties and may be dangerous for RS.
c) It depends on RS properties how quickly P increases. For example, as an increase of P causes a decrease of Q to maintain Po=Q*P preset, there is a negative feedback in PoCV. Thus, the respirator adapts own work to the RS properties during PoCV, which protects against too big values of both P and Q. The difference between power controlled ventilation (PoCV) and volume
(VCV) or pressure (PCV) controlled ventilation from the cybernetic point of view.
To avoid the above risks, Darowski [45] proposed a new method: power controlled ventilation
(PoCV). In PoCV, neither the pressure (as in PCV) nor VT (as in VCV) is the independent,
presettable variable. Instead of that, the instantaneous power, i.e. the multiplication of the
pressure and airflow rate, is the independent variable that is preset and maintained at the defined
level during inspiration. Such a connection between the pressure and airflow with their
multiplication makes that:
• on the one hand, neither the pressure (as in PCV) nor the airflow rate (as in VCV) is the
control variable, and thus both depend on the lungs state;
• on the other hand, both are under partial control because of controlled multiplication.
For example, if the current airflow value causes that the pressure starts to be too great (e.g.
because of small lungs compliance), the airflow is decreased to keep the multiplication at the
defined level. Thus, there is some kind of adaptation of the respirator work to the lungs behavior
realized by means of a negative feedback between the pressure and airflow values (Figure 3c).
10
A virtual RS seems to be the best tool for initial, comparative analysis of VCV, PCV, and PoCV
because ethical limitations would make impossible to perform such tests on real patients as the
tests require the use of all the modes in the same patient and observation which of the modes may
cause the most significant lungs injury. An example of such a comparative analysis is presented
in the Chapter 5.5.
Whether PCV, VCV or PoCV is used, the non-physiological positive pressures appear in the
lungs and thorax during the artificial ventilation. They may unfavorably influence both the lungs
and hemodynamics. This problem should be considered every time when artificial ventilation
support is applied in patients with heart or lung disease. New methods of artificial ventilation are
supposed to reduce partially these effects, maintaining the effectiveness of the breathing support.
One of such methods is PAV [46]. PAV reinforces the instantaneous patient’s breathing effort
and leaves the patient to control over many aspects of breathing (as the frequency or volume of
the inspired air).
The main assumption of the PAV mode is to produce the pressure being proportional to the
airflow, to the volume of the already inspired air, or to both. The aim of PAV is to decrease the
work of the respiratory muscles to compensate:
� muscle fatigue,
� a fall of the respiratory system compliance (C), and
� a rise of Raw.
The factors mentioned above result from different lung pathologies and the presence of the
endotracheal tube. Advantages of the PAV include:
o preservation of the patient’s own respiratory control,
o greater comfort to the patient in comparison with conventional support methods,
o smaller sedation,
o a lower risk of the patient’s over-ventilation,
o non-invasive ventilation is possible (by a face or nasal mask),
Besides advantages, PAV may has such disadvantages as:
• unfavorable dependence on patient’s respiratory activity,
• a risk of instability when the settings are incorrect or the conditions of the support change.
11
A model of RS, esp. a virtual RS, can be a good tool for analysis when potential disadvantages of
PAV might become actual. An example of such analysis as well as comparison between PAV and
PSV are presented in the Chapter 5.7.
CPAP is usually a useful method of spontaneous breathing support. If the obstruction concerns
the upper airway (e.g. as in the sleep apnea), CPAP keeps the upper airways open. Sullivan et al.
used CPAP in sleep apnea as the first [47] and nasal CPAP is a standard therapy for obstructive
sleep apnea syndrome, now. If the obstruction that concerns smallest bronchi is connected with
atelectasis, CPAP keeps bronchiole open and prevents alveoli from collapse.
Obstruction in bronchi of middle generation is of the greatest meaning from the clinical point of
view because chronic obstructive pulmonary disease and asthma are such a type of obstruction.
Despite that meaning, the cause of CPAP efficacy in such cases is not so obvious.
As CPAP means increased pressure inside bronchi, it makes those bronchi wider because of
increased transmural pressure. However, although easier inspiration through wider bronchi is
usually supposed to be the CPAP efficacy cause, the expiration period seems to be much more
critical in that case. It is because both the obstruction and a transmural pressure2 fall during
expiration decrease expiratory airflow through such bronchi3. Therefore, the smaller the
transmural pressure because of increased intrapleural pressure during the expiration, the greater
the summarized effect on the airflow. As CPAP increases the transmural pressure, it may prevent
such bronchi against collapse causing airflow limitation4, and thus it can compensate obstruction
influence on expiratory airflow. Hence, it appears that CPAP may make expiration easier.
2 When CPAP is used, the transmural pressure is equal to the CPAP value less the intrapleural pressure. 3 See the Formula (10) in the Chapter 3.2, for example. 4 See also Chapters 5.4 and 5.8 where flow limitation during forced spirometry is deliberated.
12
On the other hand, however, CPAP may make breathing harder because it shift the working point
of RS towards smaller lungs compliance (Figure 4). It can be proved that an intrathoracic
(intrapleural) pressure change (∆Pw) caused by a constant inspiratory muscles effort can be
determined with the following formula:
d
d d
Cw∆Pw=-Ps
Ca +Cw⋅ (1)
where:
Cwd, Cad – differential compliances of the chest wall and lungs, respectively;
Ps - the muscles effort expressed by the resultant pressure being an effect of that effort.
Figure 4 : Influence of the working point on the differential compliance of lungs. V – lungs volume, Pw –
the difference between the intrapleural and upper airways pressures being equal to the static
transpulmonary pressure when the air flow is equal to zero (thus, the broken line presents the lungs
compliance). The greater the lungs volume, the smaller the lungs volume increase (∆V) caused by a pressure increase (∆p), i.e. the greater the volume, the smaller the differential compliance Cad=∆V/∆p. Solid curves - the relationship between the volume and pressure for
breathing with the same frequency but different deepness of breaths.
13
Since Cad falls with lungs volume increase (Figure 4), the greater the volume, the greater the
∆Pw value caused by the same Ps (hence changes of the intrapleural pressure in Figure 5 are
bigger for the greater CPAP despite unchanged Ps). However, although ∆Pw rises with the lungs
volume, the volume increment (∆V) decreases because of the following dependence:
d
d d
CwV Ps
1 Cw / Ca∆ = ⋅
+ (2)
Hence it appears that CPAP may make breathing harder because the same respiratory muscle
effort causes less deep inspiration when CPAP is applied (Figure 5). Additionally, CPAP makes
the end-expiratory lungs volume greater, and thus a greater portion of used air is mixed with the
fresh air during inspiration, which causes that the alveolar partial pressure of O2 may be smaller.
Since, on the other hand, CPAP increasing the transmural pressure in bronchi of middle
generation should make breathing easier, it is necessary to analyze when CPAP should be applied
and when it is unprofitable (Figure 6). Thus again: investigations using a virtual RS may be
helpful. An example regarding the mechanical aspect is presented in the Chapter 5.6.
Figure 5 : Lung volume versus intrapleural pressure (from [48]) for normal airway resistance (Raw), four values of CPAP, and the breathing frequency equal to 15/min. Black points [Ppoint, Vpoint] determine the working points showed in Figure 4 (where Pw from that figure is equal to Ppoint-CPAP). Note that the greater the CPAP value, the more
horizontal the loop, i.e. the smaller the lungs volume increase.
14
Figure 6 : Advantages (blue) and disadvantages (red) of the support with CPAP. It seems that only experiments, also those on a virtual respiratory system, can show whether
CPAP is profitable or not in an individual case.
15
1.4. CIRCULATORY SYSTEM MODELING
Gianfranco Ferrari
The use of modeling to study, interpret and analyze circulatory phenomena is rather recent and
spread thanks to the parallel development of tools to solve and represent the equation systems at
the base of any model. The tools are of course one side of the medal, the other side is the
knowledge on cardiovascular patho-physiology that determines circulatory system modeling
evolution and complexity. From another point of view, the typically bi-directional relationship
between circulatory patho-physiology and modeling is the key point: in fact, if understanding the
circulatory phenomenon is the basis for constructing a model, the latter can suggest, reciprocally,
how to interpret or predict a circulatory phenomenon.
Another important remark is that each model is designed to represent a specific phenomenon and
that the model scope of application is consequently limited. Attention can be focused on different
aspects of circulation (fluidodynamics, blood volume distribution, interactions between different
systems as in the case of circulatory and respiratory systems) needing each of them a different
model. Finally, models can be developed to represent local or global circulatory phenomena
(taking into account the whole circulatory system). In the last case, the subject of this book,
models can be defined as comprehensive. A comprehensive model suffers from some limitations
due to the structure (mainly lumped parameters) and to the necessity to limit the model
complexity. Nevertheless, comprehensive models are valuable tools to study volume
displacements inside the circulatory network and the interaction between the heart, the arterial
system and mechanical circulatory assistance, if present.
The issue of artero-ventricular interaction deserves a brief history of its development and some
remarks. The study of the interaction between the heart and the vascular system has always been
the subject of considerable interest as it is strictly related to the most general issue of cardiac
output regulation. The possibilities offered by new analytic tools and new measurement
techniques permitted the passage from phenomenological observations to the quantitative
analysis of the heart/circulatory system interaction. These quantitative studies began in the fifties
[49],[50],[51],[52] of the last century and attained full maturity in the early seventies. It is
difficult to mention the most important studies in this field: contributions came from different
sources and, complementing each other, led to the creation of powerful instruments and
methodologies going deep into the mechanics of the heart and its interaction with the circulatory
16
system. Undoubtedly, the works of Guyton [53], Sagawa [54], Westerhof [55],[56],[57] and
Piene [58] are milestones in the study of artero-ventricular interaction.
An important achievement is the complementarity of different studies that permit, altogether, to
analyze different aspects of the same problem: the description of what happens in between the
ventricle and the circulatory system and, after all, what are the determinants of cardiac output
regulation. The first investigator to analyze the problem of cardiac output regulation and,
implicitly, of the mutual interaction between the heart and the circulatory system was Guyton
[53] who started these studies in the mid fifties of the last century: we are in debt to him of the
first analytical description of the complex interactions between different parts of the circulatory
system and, especially, of the concept of venous return and definition of the meaning and the role
of mean circulatory pressure. The studies of Sonnenblick [59], Suga [60] and Sagawa [54]
permitted to interpret ventricular mechanics and gave rise to the first models of ventricular
ejection that became more and more complete with contributions of investigators from all over
the world. In the seventies of the last century Westerhof focused his attention on arterial system
modeling, modifying the traditional windkessel [55] to fit the model to measured input
impedance of the arterial system. He was one of the first investigators to describe the heart as a
pump [56] and, on this basis, the interaction with the arterial system.
In conclusion, a set of useful tools is now available to model, analyze and interpret the behavior
of the ventricle and its interaction with the circulatory system in general and the arterial system in
particular. Ventricular modeling is well assessed and based on the variable elastance model that
from its basic definition [54] was progressively refined to include ventricular internal resistance
[61],[62] and active atria [63] . The introduction of the concepts of end-systolic ventricular
elastance and of effective arterial elastance [54] permitted to describe, on the pressure-volume
plane, the interaction between the ventricle and the arterial system by means of the working point
defined by the intersection of two lines. The approach shortly sketched here, besides being useful
in describing patho-physiological circulatory conditions, is an excellent starting point and a base
to develop methodologies facing the new challenges grown up with the quick development of
active or passive prosthetic devices developed to assist or replace the failing ventricular function
or parts of the circulatory system.
A-V interaction: graphical representation on the P-V plane
17
The representation of artero-ventricular (A-V) interaction on the pressure-volume (P-V) plane
permits some interesting considerations. Figure 7 illustrates graphically A-V interaction on the P-
V plane: the working point PW is defined by the intersection between end-systolic pressure
volume relationship (ESPVR) line and effective arterial elastance (EAE) line. Any change in
ventricular or circulatory conditions is reflected by the displacements of the working point PW.
ESPVR line reflects the contractile state of the ventricle by its slope (Ees) and intercept (V0) on
the volume axis. EAE line slope (Ea) is arterial elastance. It is defined as the ratio of total
resistance (Rt) and cardiac cycle duration (Tc). Total resistance is defined as the ratio of
ventricular end-systolic pressure (Pes) and average ventricular flow (Q). For practical
applications, Pes can be replaced by average arterial pressure [54].
The position and the slope of EAE line is modified by any change in ventricular end diastolic
volume (Ved) and in arterial circulatory parameters.
Pes
Ven
tric
ular
Pre
ssur
e
Ventricular Volume
V0 SV
PW
EAE ESPVR
P’W
SV’
EAE’
Ees
Ea E’a
VedVes
ESPVR’
Figure 7 : graphical description of artero-ventricular interaction
18
The description of A-V interaction on the P-V plane is strictly related to ventricular energetics.
As it is shown in Figure 8, each of the elements graphically represented in Figure 7 contains
information directly or indirectly related to ventricular energetics. EW is ventricular external
work, PE - ventricular potential energy. Pressure-volume area (PVA) is related to ventricular
oxygen consumption VO2 [60]. Finally, CME is cardiac mechanical efficiency. So, while
describing artero-ventricular interaction, the pressure-volume plane representation gives valuable
information about the state of the ventricle (stroke volume, ventricular volumes), ventricular
energetics and, finally, the state of the arterial network. This background, as it will be shown in
the next chapters, can be usefully applied in the analysis of different patho-physiological
circulatory conditions including the effects of prosthetic devices or of mechanical heart assist
devices. In this regard, it is worthwhile to mention two important issues:
• the number of methods and devices supporting inefficient cardiovascular system
increased rapidly during the last few years.
• Clinical studies to assess advantages and limitations of circulatory support devices and to
optimise their use are long lasting and have obvious restrictions – the health risk and
ethical problems concerning investigations on patients and animals.
Mechanical heart assistance, A-V interaction and modeling
Pes
Ven
tric
ular
Pre
ssur
e
Ventricular Volume
V0 SV
PW
EAE ESPVR
Ees
Ea
VedVes
PE
EW
⋅2 1 2
2
PVA=PE+EW
VO =b PVA+b
EWCME=
VO
Figure 8 : the pressure-volume plane representation and ventricular energetics
19
When mechanical heart assist devices exert their action, they affect directly or indirectly A-V
interaction: the development of circulatory models based on analytical tools able to represent A-
V interaction opens therefore interesting opportunities to assess and compare different
implantable devices, support devices, mechanical support methods and to improve circulatory
assistance.
Furthermore, by in vitro tests on simulators i.e. virtual patients, comprehensive models can help
to understand the complex heart/lung interaction especially during the frequent simultaneous
assistance of both respiratory and circulatory systems.
To complete this overview on circulatory assistance, it should be said that the fidelity
requirements for models or simulators are rapidly increasing both in the field of basic research
and in R&D investigations conducted mainly by Universities and Company laboratories. Finally,
another issue not yet mentioned is the increasing role of modeling and simulation in academic
education, both in technical and medical schools and in the training of medical staff to use the
new devices.
Coming back to the P-V plane representation, its potential importance lies in that the majority of
the existing active and passive prosthetic devices modify or affect ventricular volumes or
circulatory parameters or both of them. This implies a relevant direct or indirect effect on
ventricular energetics in the case of ventricular assistance or in the case of the functional
replacement of vessels sections and heart valves. Similar considerations are valid to describe the
effects of drugs or of any other therapy modifying ventricular or circulatory parameters. An
example can be paradigmatic: the IABP is an assist device that is often used to assist a failing left
ventricle. Without going into details now, its role is to unload the ventricle and improve coronary
perfusion by changing hemodynamic conditions during ventricular diastolic phase. This
mechanical action (caused by the inflation and deflation of the balloon), in spite of the limited
hemodynamic effects, changes in depth ventricular energetics. This can give rise to a virtuous
circle that improves ventricular elastance re-establishing the proper energetic relationships inside
the ventricle and permitting further recovery of the ventricle. One of the critical issues is
therefore the optimal use of the IABP to support the ventricular recovery. The analysis of the
pressure-volume loop during IABP assistance offers a powerful tool that permits to:
• merge information regarding ventricular volumes, hemodynamics and ventricular
energetics,
20
• identify criteria to optimize the control strategy of the balloon.
This framework identifies the problem of IABP assistance optimization, identifying at the same
time the requirements to develop a comprehensive circulatory model that should be aimed at
analyzing hemodynamic and energetics data predicting, to some extent, their trends during IABP
assistance.
Circulatory models, a brief history and their present
Coming back to modeling in strict sense, it should be said that the history of circulatory modeling
coincides widely with the history of bioengineering studies of the circulatory system. Circulatory
models were based from the beginning on different structures (usually numerical or physical)
reproducing the circulatory system at different levels of detail. From functional point of view,
they became progressively closer and closer to the theoretical background outlined above.
The first complex circulatory models were analogue [64] and their aim was to permit the study of
mechanical properties of the arterial beds. Only later, with the development of the first digital
computers, numerical circulatory models became fundamental and widespread [65],[66],[67].
The need for physical circulatory modeling arose and grew up with the development of the first
circulatory prosthetic devices and for research purposes [68],[69],[70],[71]. However, the growth
of circulatory modeling was to some extent independent of the methodologies developed to
analyze cardiovascular patho-physiology. Only later there was a convergence of tools (the
circulatory models) and methods (the pressure-volume plane approach to the study of artero-
ventricular interaction).
As it was said, cardiovascular research along with the development of cardiovascular
technologies stimulated the development of circulatory modeling. For example, the impulse to
the development of hydraulic modeling of the circulation came from the growth of total artificial
heart research. The development of such models is a good example of how physical models were
tailored to the specific applications. For total artificial heart in fact, the main problem was the
construction of a circuit able to simulate both systemic and pulmonary circulation, evidence the
unbalance between left and right ventricle and reproduce properly pressure-flow relationships at
the output of both ventricles. The necessary models were therefore rather simple and the many
prototypes essentially differed in the solutions chosen to realize and automate in some cases
hydraulic components [66],[68],[69],[70],[71].
21
A further stimulus came from the development of mechanical heart assistance and of several
prosthetic components that led to the construction of more sophisticated mock circulatory
systems (MCSs) tailored as well to the specific application [72],[73],[74],[75],[76]. Without
going into the details of heart assist device testing, it is interesting to point out here that, when
assistance is used for heart recovery, its role is to create the best conditions for this occurrence. In
this case, testing of the assist device should go beyond the simple evaluation of its performance
including the possibility to perform functional tests; that is to say the study of hemodynamics
conditions in relation to the control strategy of the device. More specifically, functional tests
involve the study of the mutual interaction among the heart, the circulatory network and the assist
device. That is to say, the study of A-V interaction in the presence of another pump that is the
assist device.
The situation sketched above implies the construction of circulatory models able to reproduce A-
V interactions and to react to the presence of a mechanical heart assist device.
These examples are not exhaustive as other stimuli come as well from education and from
clinical practice (surgical or intensive care procedures) where the request for hemodynamic data
analysis or trend prediction is strong and very specific. It is worthwhile to quote among the
possible applications of comprehensive circulatory modeling the study of specific environmental
conditions such as the effects of exercise, the absence or alteration of gravity, the effects of
diving.
Coming to the existing circulatory models, they are based essentially on two types of structures:
numerical and physical. The models are widely used to reproduce patho-physiological conditions
in research, education and in medical devices development and testing. Their role, beyond
education, is analyzing the experimental data and predicting the effect of care procedures
influencing, directly or indirectly, the circulatory system. Up to now, the models are computer
programs, if numerical, or specialized devices, modeling limited, strictly defined characteristics
of the circulatory system, if physical. Circulatory models may be therefore completely different
in the structure and complexity, privileging the reproduction of fluid-dynamics or volumetric
phenomena. This implies that results of simulations obtained by different investigators are often
incomparable, considerably diminishing the possible impact of the models on clinical practice.
Physical models of biological systems mechanics are generally made as physical analogues in the
form of parallel and in series connections of few fluidic lumped or distributed parameter resistors,
22
capacitances and tubes [73],[74],[75],[76]. They are of a rather simple structure and their
performance is strongly limited by the mechanical complexity, the difficult readjustment of
parameters, low long-term parameter stability and high costs. These features compromise the
evolution of models and their ability to support clinical application in a cost-effective manner.
The main limitation of a physical model is that it is always a mechanical realization of a
simplified mathematical model, while its main advantage is the possibility of its direct connection
to the devices to be tested and/or to measurement apparatus.
Numerical models, on the contrary, have a very flexible structure, with easily changeable
parameters and are much cheaper [77],[78],[79],[80]. They can have different structures (lumped
or distributed parameters, mono- and multi-dimensional) but can easily exchange data using, for
example, the multi-scale approach. However, numerical models cannot be directly connected to
devices (for assistance or measurement) working in a fluid environment. For example, a frequent
clinical interest concerns the optimization of working parameters of mechanical heart assistance
for various pathologies. The pathologies could be easily simulated in the numerical model of
cardiovascular or respiratory systems but the assist device cannot be connected to the numerical
model. A good example is again the IABP: the numerical solution of this problem implies the
necessity of using different mathematical models of the balloon in the aorta for each type and size
of the IABP. It would be difficult, if possible at all, to prove a required fidelity of these models
for different balloons and supporting devices.
Taking into account the present limitations of both physical and numerical models, a new type of
models, defined as hybrid (physical-numerical), is being developed [81],[82],[83]. Hybrid models
permit to merge numerical and physical models or devices optimizing their performance,
improving accuracy and reducing costs. In the hybrid simulators the advantages of numerical
models (flexibility, accuracy, low cost, stability) are connected with the main characteristic
feature of physical models, e.g. ability to interact directly with mechanical support devices or
measurement apparatus. It is to be remarked that hybrid models can make the model performance
independent of the structure opening the possibility to realize a modeling platform where the
structure (numerical, numerical-physical) can be easily changed without altering the model
performance and, what is more important, maintaining the possibility to compare the results in
any working condition.
23
1.5. RESPIRATORY SYSTEM MODELING
Marek Darowski, Tomasz Gólczewski
The human respiratory system like other physiological systems has a complex structure.
In order to acquaint with lungs physiology or pathology and with phenomena that takes place in
alveoli and airways we use specific tools. Measurements of flow and pressure inside the
respiratory system have a lot of limitations as they can not be invasive or because of ethical
issues. In fact, only upper airways flow and mouth pressure are usually available variables. Thus
models of the respiratory system are those specific tools that enable us to understand what is
going on inside the lungs, to help in diagnosis of lungs state or in assessment of the results of
treatment applied. There are two main research and clinical areas where models of the respiratory
system mechanics have been developed intensively during the last decades: spirometry and
mechanical ventilation of the lung. There are good reasons for this. The relatively common lung
pathology, Chronic Obstructive Pulmonary Disease (COPD) became one of the leading cause of
death recently. In order to better assess the lung patho-physiological state (COPD, asthma,
pulmonary fibrosis) a typical functional lung test, spirometry, should be supported by an
appropriate model of respiratory system mechanics.
The computer models used to simulate forced expiration during spirometry tests are morfometric,
complex models that take into account different mechanical properties of tracheal tree
generations (described by distributed or lumped parameters) and flow limitation
[84],[85],[86],[87],[88]. These models were developed assuming a symmetric [89] or an
asymmetric structure of the tracheal tree [90].
The respiratory system models used to simulate lungs mechanics during artificial ventilation were
usually based on much simpler, lumped parameter models proposed by Otis et al [91], Mead [92]
and Bates et al [93]. Otis took into account a parallel redistribution of gases in nonhomogenous
lungs, Mead included compliant airways into his model and Bates described a homogenous,
healthy lung with stress relaxation phenomenon in alveoli tissue.
The models proposed by Otis, Mead and Bates have now historical meaning. Recently,
researchers have concentrated their investigations on novel models of lung mechanics that would
better describe different patho-physiological state of lungs [88],[94],[95],[96].
In order to optimize ventilatory treatment of patients with lungs pathology new methods of
respiratory parameters estimation have been developed [97],[98],[99]. Patients with obstructive
24
airways are difficult for mechanical ventilation therapy. COPD is connected with expiratory flow
limitation, dynamic hyperinflation and rise in intrinsic end-expiratory pressure. Understanding
what are the basis of this pathology, its onset and development is of primary importance for
investigators nowadays.
Computer models concerning gas transport and exchange of gases have to take into account
different phenomena characteristic for both – respiratory and cardiovascular system. They may
help to investigate changes of gases concentrations in blood connected with some therapies [100]
or respiratory system control mechanisms [101].
Progress in both medicine and physiology requires experiments to test new methods of treatment
and support, to investigate scientific problems, verify new hypotheses, educate new staff
generation, etc. Although the final test, investigation or verification needs experiments on
animals or human beings (healthy volunteers or patients), an initial tests or verifications may be
performed with models: physical or computer (mathematical) ones. Such models might be
especially valuable in education.
The importance of the use of models increased recently because testing on patients is associated
with ethical problems.
Both mathematical (computer) and physical (e.g. mechanical) models have been developed
recently. Physical models are expensive (if they are not very simple) and difficult to adjust with
high fidelity.
Non-linear properties of the lungs are rather more difficult to be modeled by mechanical elements
than by a computer program. Moreover, the main advantage of computer models is the low cost
and easy adjustment of respiratory parameters. Indeed:
� computer models need not laboratories;
� there are no ethical problems with the use of them;
Because of the above reasons, to avoid high costs as well as ethical and economic problems
connected with experiments and investigations on animals or human beings (patients), different
computer models of RS have been developed.
Non–homogeneous mechanical properties of patient lungs are characteristic for respiratory
system pathology and create serious problems in their therapy by mechanical ventilation.
Distribution of ventilation and pressure in such complex structure is difficult to assess, also
because of difficulties with measurements. Modeling of lungs mechanics and simulation of
25
spontaneous breathing and ventilatory support may give some information and help to predict
these parameters changes according to applied ventilatory therapy. In this way, optimal
ventilatory strategy may be chosen among variety of ventilatory modes. This is only one of
potential applications of RS modeling. Medical personnel training and testing of new respirators
or new modes of ventilatory support are the next examples.
Constantly improving technology has created new types of simulators: virtual organs connected
to physical devices by means of virtual-to-real and real-to-virtual transducers. Such connection
enables researchers to utilize advantages of both computer and physical models. The above
forced us to consider a different approach that is testing physical respirators with the use of a
virtual RS. However, we have to be aware that not all models of RS can be treated as virtual RS.
26
1.6. REPRESENTATION OF CIRCULATORY-RESPIRATORY SYSTEM
INTERACTIONS
Tomasz Gólczewski, Marek Darowski
The chest contains both the respiratory system (RS) and a significant part of the cardiovascular
system (CS), i.e. the whole pulmonary circulation, left and right atriums and ventricles, great
vessels, and a part of the aorta. Therefore, cardiovascular system work is influenced by cyclic
thoracic pressure variations caused by the RS work. In particular, modeling of cardio-pulmonary
interaction has special meaning because non-physiological positive pressure, which appears in
the lungs during the artificial ventilation or ventilatory support, influences the hemodynamics
unfavorably [102],[103]. Respirators have been used for support of breathing for about 70 years.
Although many different modes of the support were tested during that time, some effects of the
positive thoracic pressure and airway pressure, being dangerous to the patient, could not be
completely eliminated. In particular, that fact concerns the adverse effect on the circulatory
system, which is manifested by a decrease of the venous return and pulmonary blood flow as well
as disturbance of filling of the heart ventricles. Thus, the existence of this mechanical interaction
suggests that modeling of CS cannot be really accurate without RS modeling unless a specific
problem is analyzed and the influence of RS may be neglected. For example, the interaction
between the respirator, RS, and CS should be considered each time when artificial ventilation or
support is applied in patients with heart disease.
On the other hand, both RS and CS realize the common task, i.e. they have to cooperate to
provide oxygen delivery to tissues and carbon dioxide removal. As the delivery and removal are
the main purpose of RS, the ventilation means both mechanical phenomena and airway gas
transfer as well as gas exchange causing blood oxygenation and decarbonation. Since the blood
oxygenation and decarbonation depend also on the pulmonary blood flow, they depend on the CS
mechanics. Certainly, oxygen and carbon dioxide transport to/from tissue with blood also depend
on the CS mechanics. For the above reasons, a comprehensive analysis of ventilation requires
models of RS and CS mechanics, airway gas transfer, gas exchange, and gas transport with blood.
Despite such meaning of the interactions between the RS and CS mechanics as well as gas
transfer and exchange, the number of models taking into account those interactions is relatively
small in comparison with the number of models simulating separately either RS or CS. Moreover,
27
usually at least one kind of the interactions is neglected in published papers: either mechanics of
one system or gas transfer and exchange is not taken into account. For example, the mechanical
interaction between both systems during breath-hold diving was simulated in detail by Fitz-
Clarke in his interesting paper [104], but neither gas transport nor exchange was analyzed. On the
other hand, both respiratory system mechanics and airway gas transfer as well as gas exchange
were taken into account in [105], however the mechanical interaction between the systems was
neglected. Moreover, in that model, gas is transported by a liquid, whereas the gas transports
‘itself’ during normal ventilation. In some other models (e.g. [106],[107],[108]), the
cardiovascular system mechanics is either simplified or neglected at all. In particular, cyclic
blood flow is replaced with the mean, constant flow equal to the cardiac output. For that reason,
for example, the models presented in [107],[108] ignore effects of superposition of periodic heart
work and cyclic changes of the thoracic pressure caused by ventilation. If, however, the thoracic
pressure is constant, as in a problem analyzed in [106], the blood flow may be also treated as
constant. Additionally, the gas transfer is not simulated in [107],[108], which causes that the dead
space ventilation could not be simulated, and in consequence the alveolar ventilation is the same
as the minute ventilation. Therefore, simulations of influence of the tidal volume or inspiratory
pattern on blood oxygenation may give incorrect results.
It seems that the cardio-pulmonary interactions were simulated most comprehensively in [109],
because the ensemble of models consists of both CS and RS mechanics models as well as a
model of gas exchange and transfer5.
5 Some results obtained with this ensemble are presented in the Chapter 6.5
28
1.7. A NEW APPROACH TO MODELING – HYBRID MODELS
Gianfranco Ferrari, Marek Darowski, Maciej Kozarski
Hybrid models are a particular class of circulatory and respiratory models, based on numerical
and physical models merging, aimed at reducing the costs, improving accuracy and offering the
possibility to cut animal experiments. A numerical model, circulatory or respiratory, determining
the performance and possibilities of the whole model, is in general the basis of any hybrid model
as it is shown in the example of Figure 9. There two interfaces permit the merging of a numerical
circulatory model with a mechanical heart assist device. The main goal of hybrid modeling is to
test the interaction of mechanical assist devices (VADs, prosthetic devices, respirators) with the
circulatory and respiratory systems. Training of medical professionals and education are among
the possible and important applications of hybrid modeling.
The problem of circulatory and respiratory modeling
Circulatory and respiratory models are or can be applied in research, education and prosthetic
devices/components testing and development. The structure and performance of any circulatory
or respiratory model is determined by the aim it is built for. The use of such models as tools, for
example to support clinical decision, implies their exploitation in environments where velocity of
Right HeartPulmonaryCirculation
CoronaryCirculation
LVADPhysicaldevice
Hybrid model
Interface Interface
Left HeartSystemic
Circulation
Numerical model
Figure 9 : block diagram depicting a possible application of hybrid modeling
29
execution and simplicity of use are the main requirements. On the other hand, the models used as
instruments of analysis have accuracy as their main feature and imply often to measure variables
hardly available, for example, in surgical and Intensive Care Unit (ICU) environments. So, an
important issue is that the model structure is able to face the challenge of quickness of response
and simplicity to use together with the possibility to adapt easily the model to the performance
requested by specific applications. Evidently, not all types of models can meet such requirements.
A possible approach to this problem may consist in developing basic comprehensive models
characterized by a high degree of flexibility. The efforts should be therefore focused on models
simulating the whole systems considering that an appropriate design and organization of the
model could permit, together with the possibilities offered by hybrid modeling, to respond
adequately to different needs.
Hybrid models: general considerations
In the case of circulatory system, the genesis of physical models was already discussed as a
response to the development of total artificial heart and, later, of mechanical heart assist devices.
Similar considerations are valid for the respiratory system.
It has been pointed out as well that, in spite of the existence of powerful and flexible numerical
simulation tools [65],[66],[85],[110] physical models are still relevant as training tools [72], or
for specific tests [72],[73],[74],[75],[76],[85] including mechanical assist devices. However, the
structure and functions of the physical models are strictly related or even tailored to the type of
tests to be performed. Moreover, physical models represent often a compromise between the
reduction of the mechanical complexity of the model and the accuracy of the investigation.
Finally, it is often hard or even impossible to use a physical model to answer some of the
questions arising from research as in the case of mechanical circulatory assistance when it is used
for heart recovery. For example, mechanical heart assist devices can be tested in relation to their
ability to reproduce given hemodynamic conditions: in this case, the simulation circuit could be
limited to a simple lumped parameter model. If on the contrary, the investigation has to be
extended to the mutual interaction among the device, the natural ventricle and the circulatory
network, the reproduction of the ventricular function and of the artero-ventricular interaction
become fundamental. It is not trivial to point out that this is the test condition to be fulfilled
especially when mechanical heart assistance is used for heart recovery. Accurate descriptions
exist for both the reproduction of the ventricular function and of the artero-ventricular interaction
30
[54]: they are the bases of several numerical models that, however, cannot be used when it is
necessary to test a physical device. A physical device testing implies in most cases the
development of physical models that are strictly application dependent and to some extent,
tailored to the application for which they have been developed.
Limiting the attention to comprehensive lumped parameter models that have a wide field of
applications in heart assist device (HAD) and lung support testing, research and training, physical
models can hardly overcome the problems of their cost and poor flexibility together with their
limited performances. It is however important to point out that in a physical model designed for
example to test a HAD, the section of the model to be represented physically is limited to the
areas of insertion of the device (or the prosthetic device if this is the case).
With this remark in mind, the proposed solution is to merge the characteristics and the flexibility
of numerical models with the possibilities offered by physical models as it is schematized in
Figure 10 where are evidenced the possible applications of the original systems along with the
possible applications of the resulting merged “hybrid” system. The concept of “hybrid”
circulatory and respiratory modeling was developed in the last years [81],[83],[111] to overcome
Applications: Device testingTraining and
EducationResearch
Support to clinical decision
Physical models(hydraulic,electrical,
pneumatic…)
Numerical models
Hybrid model
Applications: Device testingTraining and
EducationResearch
Applications: EducationResearch
Support to clinicaldecision
Figure 10 : possible applications of a hybrid system in comparison with the possible applications of the original numerical and physical
systems
31
the traditional dichotomy between numerical and physical circulatory models preserving at the
same time the best features of both of them.
Basically, it consists in merging numerical and physical models (they can be, for example,
electrical, hydraulic or pneumatic). In this way characteristics and flexibility of numerical models
are preserved along with the possibilities offered by physical models. Their interfacing is of
course the main problem as the numerical model can be easily changed or modified if necessary
and the physical section has to be chosen in relation to the tests to be performed.
Approaches to hybrid modeling
The general concept of hybrid modeling is based on the merging of physical and numerical
models but it can be applied in different ways (Figure 11). One possibility is to start from a
physical model (hydraulic, pneumatic) and transform it into a hybrid replacing some of its parts
with a numerical model. The second possibility is to start from a basic numerical model and
modify it replacing, when and where necessary, some of its parts with a physical (hydraulic,
electrical, pneumatic) model. The third possibility is to interface a whole numerical model to a
physical device. In fact, the first application of the concept of hybrid modeling consisted in the
insertion of a numerical model into the existing physical model [81]. This solution is valid to
improve the performance of a physical model but it is not the best from the point of view of
flexibility. On the contrary, merging a physical model into the existing numerical model [82],[83]
Physical model (hydraulic,electrical,
pneumatic…)
Numerical model
Numerical model
Hybridmodel
Physical model: hydraulic, electrical,
pneumatic,…
Figure 11 : approaches to hybrid modeling
32
offers the maximal flexibility along with the possibility to minimize the physical model reduced
in this way to the barest essential.
33
1.8. A NEW APPROACH TO MODELING – VIRTUAL ORGANS AND E-
LEARNING
Tomasz Gólczewski, Marek Darowski
Virtual organs and models
Depending on model application, two kinds of models might be distinguished. Models utilized in
medical practice are of the first kind. They are used to estimate – in the case of an individual, real
patient - the values of some parameters basing on measurement data. Since precise measurement
of many factors would be impossible in everyday practice and fitting a complex model to these
measurements would be a sophisticated mathematical problem, if possible at all, models of the
first kind have to be very simple with a few parameters. Thus, simplicity of such models is an
advantage rather than an imperfection.
It has to be stressed, however, that usually parameters of such models do not correspond exactly
to their 'physical' name. For example, such a parameter as airways resistance (Raw) described
with one number (the model in Figure 12a), e.g. that is used in the Proportional Assist
Ventilation6, does not exist, in fact. In real RS, airways resistance is nonlinear and changeable: it
depends on the thoracic (intrapleural) pressure, lungs volume, alveolar pressure as well as the rate
and direction of airflow7.
Figure 12 : Fundamental models of the respiratory system. a) the simplest model characterizing the most fundamental properties of RS, i.e. the airway
resistance (R) and compliance of RS (C); Pr – respirator, Ps – spontaneous breathing) b) Otis's model, which takes into account non-homogeneity of RS (parts of different time
constant) causing gas redistribution in lungs; c) Mead's model taking into account the compliance of airways (Caw);
6 See the Chapters 1.3 and 5.7 7 See the Chapter 3.2
34
d) Mount's model (developed by Bates), which takes into account visco-elastic properties of tissues and surfactant (Rt, Ct) causing the relaxation phenomenon.
Models of the second kind simulate particular phenomena or are used in analysis of such
phenomena. Figure 12b-d present the most known simple models of RS. The simplest,
“classical” RC model (Figure 12a) is still utilized for many purposes. A bit more complicated
models take into account inertial phenomena. The most complicated of these models make
possible to reproduce lung visco-elastic properties. Now, these models have rather educational
meaning only. Present models consist of much greater number of elements; some of them may be
nonlinear, etc. Although they are much sophisticated, the general idea is usually the same: models
are used to examine or analyze a particular problem or phenomenon defined before creation of a
model.
Constantly increasing abilities of computers caused that a new, third kind of models has
appeared: virtual organs. Such artificial organs may replace animals or human patients in several
applications, such as medical education or initial scientific experiments, for example. The
difference between virtual organs and 'normal' models can be expressed with the following
statement:
A model follows a problem that is intended to be analyzed, while a virtual organ is created to
analyze a whole class of problems which are not known or clearly defined before virtual organ
creation.
To be a virtual organ, a computer model has to be sufficiently complex to be able to behave
accurately under conditions not pre-determined precisely before the model building. For example,
properties of a new respirator tested are not known before tests, and thus an observer who
manipulates this respirator should not be able to recognize whether the respirator ventilates real
lungs or virtual ones to draw accurate conclusions. The above is even more clear if a model is
intended to be used as a virtual organ in medical education. Indeed, nobody can know which
‘experiments’ a student is going to perform but results of each possible experiment should be
correct to avoid wrong education.
To simulate cases, which had been not known before the virtual organ creation, a model has to be
both complex and specific. 'Normal' models may describe mathematically different physiological
phenomena, while the virtual organ has to reflect physical and physiological properties that are
35
the cause of those phenomena. For example, both the model of RS that is presented by Schuessler
et al. [112] and the virtual RS elaborated by Golczewski and Darowski [85] simulate airflow
limitation during forced expiration. However, in Schuessler’s model, Raw and the flow limitation
are described with two different formulas, which is some kind of consistency lack. In the virtual
RS, Raw is described in such a way that Raw means ‘normal’ resistance during normal breathing
and reflects flow limitation during forced expiration. Moreover, the formula that describes the
flow limitation in [112] is not derived from physiological properties; it is an assumed formula
with such values of parameters which give agreement between the phenomenon and the formula
results. In the virtual RS, Raw is derived from a physiological property that is the commonly
known experimental dependence of Raw on the lungs volume8.
Fitting parameters of a simpler model to measurement data is usually a goal of its use. In
particular, such fitting is the task of models of the first kind. In the case of virtual organs, such
fitting parameters to measurement data seems to be rather impossible. Indeed, such fitting in a
case of real organs is called diagnosing and is one of the two most fundamental, sophisticated
problems of medicine (treatment is the second one). Since virtual organs should be enough
complex to imitate real organs, diagnosing rather than fitting could be regarded in the case of
such models. Therefore, any application assuming fitting parameters of a virtual organ to
measurement data is rather pointless; direct diagnosis of real patients would be more rationale.
The meaning of virtual organs increases in the modern science. For example, Virtual
Physiological Human is one of three long-term strategic priorities recommended to the attention
of the European Commission by European Alliance for Medical and Biological Engineering and
Science (EAMBES). In the connection with the 7th Framework Program of the European
Community for research, technological development and demonstration activities (2007 to 2013
yr), EAMBES proposed to create virtual organs, which will be connected in the one virtual
human in the future [113].
Anticipating such global tendency, the virtual RS mentioned above was developed as a kind of
artificial organ for respirators and support methods testing [114],[115],[116] as well as a medical
education tool [117]. However, a stand-alone virtual RS would be too simple tool to replace the
real RS in more advanced research or education. Indeed, oxygen delivery to tissues and carbon
8 The derivation is presented in the Chapter 3.2
36
dioxide removal are the final goal of RS. Both the delivery and removal depend on work of the
cardiovascular system. On the other hand, the work of the cardiovascular system depends on
activity of RS because of influence of the thoracic and alveolar pressures on the pulmonary
circulation, heart work, and venous return. Additionally, the work of both system depends on gas
tensions in blood. Hence it appears that more advanced studies have to regard mechanical activity
of both systems as well as gas transport and exchange. Figure 13 illustrates a framework of such
studies.
Figure 13 : A framework of cardio-respiratory studies. As oxygen delivery and carbon dioxide removal are the fundamental goal of the cardio-
respiratory system, the model of gas transfer and exchange is the central point (it consists of modules of: AGT - airways gas transfer, GE - gas exchange, BGT - blood gas transport).
Respiratory system mechanics influences AGT, GE, and circulation. Cardiovascular system mechanics influences BGT and GE. A support of respiration and circulation can be both
simulated and realized with physical devices.
Virtual organs verification
A model can be treated as a virtual organ if it behaves as the real organ under different
conditions. In particular, all or almost all known fundamental physiological phenomena should be
observed. For that reason, accurate simulations of several such phenomena can be treated as
37
virtual organ verification. Such a kind of verification seems to be the most proper since if known
phenomena would be simulated inaccurately, a virtual organ could not be reliable when a new
unknown phenomenon is studied. Certainly, even the most accurate simulations of known
phenomena cannot be treated as a proof that results of unknown phenomenon study will be true
without any doubts. However, it is the property of all scientific theories: such a theory has to
describe the known knowledge correctly but its predictions should be always confirmed by means
of real, direct experiments. Considering the above, a virtual organ has to be verified like each
theory, i.e. it has to be checked
whether a theory called “the virtual organ” corresponds to all known facts or not.
The main feature of virtual reality is more or less impossibility of recognition of differences
between real and virtual objects. In the case of a virtual patient, the above means that it should be
difficult to recognize which results of a diagnostic examination concern real patients and which
have been obtained for a virtual patient. For example, as spirometry is the fundamental diagnostic
method in the case of RS, comparison between results of spirometry of the virtual RS and real
patients has been treated as a part of virtual RS verification9.
Virtual organs in e-learning
Virtual organs, as the other models, are useful in research, however, they may be of special
significance in medical education.
Education consists in learning facts, rules, and how to manipulate them. Learning may be
performed with Internet, which is the base of distal-learning being a kind of e-learning,
i.e. learning utilizing multimedial methods, especially those using computers. Recently,
the meaning of such methods increases very quickly. However, since medical education consists
also in practice with patients, it might seem that e-learning cannot be useful in medical education.
Nevertheless, since virtual patients can replace real patients partially, e-learning may be yet
possible in medical education, at least in an initial period.
9 Such virtual spirometry of the virtual RS is presented in the Chapter 5.4.
38
Application of virtual patients in education is of special significance because there are no ethical,
financial or legal problems when a student makes a mistake, even if he/she ‘kills’ such a patient.
Additionally:
• such a patient is easy to duplicate, and thus each student may have own patient,
• a student can choose a convenient place and time for learning,
• an instructor can simulate at each moment a disease that he want to discuss, while a real
patient suffering from a chosen disease may be not available in the lecture day.
In general, the virtual patient enables students either to ‘introduce’ a disease by themselves or to
manipulate the patient with a disease ‘introduced’ by an instructor. For example, a virtual organ
enables students to learn how a disease and its severity influence results of patient examination.
The Chapter 5.8 presents the Tgol.e-spirometry system [117] being such an application of
the virtual RS presented in [85]. That system is the virtual RS supplemented with a user interface
enabling medical students or instructors to ‘introduce’ obstructive and/or restrictive lungs disease
of various severity and observe results of the forced spirometry.
Figure 14 : An example of virtual patient use in medical education.
The following example of intensive care of a virtual patient may be another illustration of virtual
organs usefulness in professional advancement (Figure 14). In this example, a student should
manage long-term artificial volume controlled ventilation of an old patient who has to be turned
into the lateral position to avoid bed sore. The ventilation of such a patient may vary after his/her
position change, which leads to a fall of arterial blood oxygenation (Figure 15). A student should
find a method of profitable change of ventilation leading to blood oxygenation improvement.
39
Figure 15 : Arterial blood oxygenation falls after patient’s position change. The position of a virtual patient has been changed from the supine to the left lateral one. The
oxygenation is expressed with the percentage of oxygenated hemoglobin.
A student can observe at the computer screen that when position is changed, arterial blood
oxygenation falls significantly in two minutes. Student’s aim is to improve the oxygenation with:
• either an increase of the fraction of oxygen in the inspired air (FIO2)
• or an increase of the minute ventilation caused by an increase of the ventilation frequency or
tidal volume
• or change of ventilation method.
Probably, the student increases FIO2 as the first. In general, it would be justified since the minute
ventilation was correct before position change, and thus a problem with gas transfer or exchange
rather than with the volume controlled ventilation could be supposed. However, even significant
increase of FIO2 causes insufficient increase of the oxygenation (Figure 16). Further increasing
of toxic oxygen concentration seems to be not a good solution. Therefore, an increase of the
minute ventilation may be proposed by the student. However, if he/she tried to improve the
oxygenation with an increase of the ventilation frequency, it would not attain the goal. If he/she
tried to improve the oxygenation with an increase of the tidal volume, it would attain the goal not
before he/she increased the tidal volume two times (Figure 17).
40
Both too high FIO2 and too big VT are dangerous for lungs. Therefore, the student has to look for
better solution, and thus he/she should change the method of ventilation. If he/she chose the
independent ventilation, the oxygenation would be improved sufficiently with increase of neither
FIO2 nor VT.
The above exercise could be supplemented with some explanations or the student would have
possibility to perform some other simulations to find explanations by him/herself.
Figure 16 : Arterial blood oxygenation alteration after FiO2 increase.
Figure 17 : Arterial blood oxygenation alteration after tidal volume (VT) increase.
41
1.9. CONCLUSION
Circulatory and respiratory modeling have a long lasting history that was not always linear,
depending on the experimental and research demand. However the potential of modeling
approach to clinical and experimental needs has not been fully explored yet. The realization of
models that can be regarded as “virtual humans” is an ambitious scheme. However, having in
mind this goal, it is possible to pursue it in steps constructing models characterized by a high
degree of flexibility and modularity. The advantage lies in using the same models for different
applications raising the comparability and repeatability of results. The next chapters will address
this issue and will illustrate the model design and realization along with their possible
applications.
42
REFERENCES
[1] Samuels L, Entwistle J, Holmes E, Fitzpatrick J, Wechsler A. Use of the AbioCor
replacement heart as destination therapy for end-stage heart failure with irreversible
pulmonary hypertension. J Thorac Cardiovasc Surg 2004; 128(4):643-45.
[2] Delgado RM, Nawar M, Loghin C, Myers TJ, Gregoric ID, Pool T et al. Catheterization
of the AbioCor™ Implantable Replacement Heart. Evaluation of the Unique Physiology
Created by the Device. Tex Heart Inst J. 2006;33(3):359–360.
[3] Schistek R. Various Design of Nonpulsatile Blood Pumps. In: Unger F, Assisted
circulation 3. Part II: Cardiac Assistance. London Paris Tokyo Hong Kong Berlin
Heidelberg New York: Springer – Verlag: 1989;20:235-42.
[4] Pagani FD. Continuous-Flow Rotary Left Ventricular Assist Device with “3rd
Generation Design”. Elsevier: Seminars in Thoracic and Cardiovasc Surg. 2008;
20(3):255-63.
[5] Wieselthaler GM, Schima H, Hiesmayr M, Pacher R, Laufer G, Noon GP et al. First
Clinical Experience With the DeBakey VAD Continuous-Axial-Flow Pump for Bridge
to Transplantation. Circulation. 2000;101: 356-59.
[6] Goldstein DJ, Zucker M, Arroyo L, Baran D, McCarthy PM, Loebe M et al. Safety and
feasibility trial of the MicroMed DeBakey ventricular assist device as a bridge to
transplantation. Am Coll Cardiol. 2005;45:962-63.
[7] Palino MA, Ohye RC, Chang AC, Gajarski RJ, Bove EL, Devaney EJ. Bridge to
Transplant Using the MicroMed DeBakey Ventricular Assist Device in a Child with
Idiopatic Dilated Cardiomyopathy. Annals Thorac Surg. 2006; 81(3):1118-121.
[8] Reul H, Kaufmann R, Siess Th. Cardiac Assist devices. In: Verdonck P, Intra and
Extracorporeal Cardiovascular Fluid Dynamics. Nova Scotia. Canada: WITpress:
1998;1:233-57.
[9] Dowling RD, Etoch SW, Stevens KA, Johnson AC, Gray LA Jr et al. Current Status of
the AbioCor Implantable Replacement Heart. Annals of Thoracic Surg 2001;71:S147-
S149.
[10] Unger F, Assisted circulation 3. Part II: Cardiac Assistance. London Paris Tokyo Hong
Kong Berlin Heidelberg New York: Springer-Verlag, 1989.
43
[11] Zheng ZS, Li TM, Kambic H, Chen GH, Yu LQ, Cai SR et al. Sequential external
counterpulsation (SECP) in China. Trans Am Soc Artif Intern Organs 1983;29:599-603.
[12] Applebaum RM, Kasliwal R, Tunick PA, Konecky N, Katz ES, Trehan N, Kronzon I.
Sequential external counterpulsation increases cerebral and renal blood flow. Am Heart
J 1997;133(6):611-15.
[13] Ochoa AB, deJong A, Grayson D, Franklin B, McCullough P. Effect of enhanced
external counterpulsation on resting oxygen uptake in patients having previous coronary
revascularization and in healthy volunteers. Am J Cardiol 2006; 98:613-615.
[14] Nichols WW, Estrada JC, Braith RW, Owens K, Conti CR. Enhanced external
counterpulsation treatment improves arterial wall properties and wave reflection
characteristics in patients with refractory angina. J Am Coll Cardiol 2006; 48:1208-
1214.
[15] Erdling A, Bondesson S, Pettersson T, Edvinsson L. Enhanced external counter
pulsation in treatment of refractory angina pectoris: two year outcome and baseline
factors associated with treatment failure. BMC Cardiovasc Disord 2008; 8 (39): 1-7.
[16] Balooki H, editor. Clinical Application of Intra-Aortic Balloon Pump. New York:
Futura Publ. Comp Inc; 1977.
[17] Kawaguchi O, Pue WE, Daily BB, Pierce WS. Ventriculoarterial coupling with intra-
aortic balloon pump in acute ischemic heart failure. J Thorac Cardiovasc Surg
1999;117(1):164-71.
[18] Sauren LD, Reesink KD, Selder JL, Beghi C, van der Veen FH, Maessen JG. The acute
effect of intra-aortic balloon counterpulsation during extracorporeal life support: an
experimental study. Artif Organs 2007;31(1):31-8.
[19] Bian X, Downy HS. Enchanced Intra-Aortic Balloon Pump: Markedly Improved
Systemic Hemodynamics and Cardiac Function in Canines with Severe, Acute Left
Ventricular Failure. Artif Organs 2002;26(8):727-33.
[20] Ntalianis AS, Drakos SG, Charitos C, Dolou P, Pierrakos CN, Terrovitis JV et al.
Effects of intra-aortic balloon pump versus centrifugal pump on myocardial energetics
and systemic circulation in a porcine model of rapidly worsening acute heart failure.
ASAIO J 2008; 54(6):600-5.
44
[21] Olasińska-Wiśniewska A, Mularek-Kubzdela T, Grajek S, Breborowicz P, Seniuk W,
Podzerek T. Indications, results of therapy and factors which influence survival in
patients treated with intra-aortic balloon counterpulsation. Kardiol Pol. 2008;66(9):950-
5; disc. 956-7.
[22] Onorati F, Santarpino G, Rubino AS, Caroleo S, Dardano A, Scalas C, Gulletta E,
Santangelo E, Renzulli A. Body perfusion during adult cardiopulmonary bypass is
improved by pulsatile flow with intra-aortic balloon pump. Int J Artif Organs. 2009
Jan;32(1):50-61.
[23] Jaron D, Moore TW, He P. Theoretical considerations regarding the optimisation of
cardiac assistance by intra-aortic balloon pumping. IEEE Trans Biomed Eng
1983;30(3):177-86.
[24] Khir AW, Price S, Hale C, Young DA, Parker KH, Pepper JR. Intra-aortic balloon
pumping: does posture matter? Artif Organs 2005;29(1):36-40.
[25] Pae WE Jr, Rosenberg G, Pierce WS. Ventricular Assistance: The Pennsylvania State
University Experience. In: Unger F, Assisted circulation 3. Part II: Cardiac Assistance.
London Paris Tokyo Hong Kong Berlin Heidelberg New York: Springer – Verlag:
1989;10:115-31.
[26] Górczyńska K, Darowski M, Kozarski M. Cardiovascular, Respiratory and Veno-
Lymphatic Assistance. Modelling and Simulation. Biocybernetics and Biomedical
Engineering 2002; 22(2-3):145-75.
[27] Swartz MT, Pennington DG, McBride LR et al. Temporary Mechanical Circulatory
Support: Clinical Experience with 148 Patients. In: Unger F, Assisted circulation 3. Part
II: Cardiac Assistance. London Paris Tokyo Hong Kong Berlin Heidelberg New York:
Springer – Verlag: 1989;11:132-51.
[28] Farrar DJ, Hill JD, Pennington DG, McBride LR, Holman WL, Kormos RL et al.
Preoperative an Postoperative comparison of patients with univentricular and
biventricular support with the Thoratec ventricular assist device as a bridge to cardiac
transplantation. J Thorac Cardiovasc Surg 1997;113: 202-209.
[29] Portner PM. A totally implantable heart assist system: the Novacor program. In: Akutsu
T, Koyanagi H, Heart replacement. Tokyo: Springer-Verlag: 1993; 71-80.
45
[30] Samuels LE, Holmes EC, Hagan H, Gopalan R, Droogan C, Ferdinand F. The Thoratec
Implantable Ventricular Assist Device (TVAD): initial clinical experience. Heart Surg
Forum 2006;9(4):E690-2.
[31] Farkas EA, Elefteriades JA. Assisted circulation: experience with the Novacor Left
Ventricular Assist System. Expert Rev Med Devices 2007; 4(6): 769-74.
[32] Ferrari G, Górczyńska K, Mimmo R, De Lazzari C, Clemente F, Tosti G et al. Mono
and biventricular assistance: their effect on ventricular energetics. Int J Artif Organs
2001;6(24):380-91.
[33] Ono M. Heart Assist Devices: statistics and numerical data. Masui 2009;58(3)327-36.
[34] Endo G, Araki K, Oshikawa M, Kojima K,Saitoh T, Nsakamura K et al. Control
Strategy for Biventricular Assistance with Mixed-Flow Pumps. Artif Organs
2000;8(24):594-9.
[35] Forresster MD, Myers TJ, Gregoric ID, Frazier OH. Saphenous Vein Graft Flow during
Left Ventricular Assistance with an Axial Flow Pump. Tex Heart Inst J 2006;
33(2):222-24.
[36] Nanas JN, Moulopoulos SD. Counterpulsation: Historical Background, Technical
Improvements, Hemodynamic and Metabolic Effects. Cardiology 1994;84(3):156-67.
[37] Chatburn RL. Classification of ventilator modes: update and proposal for
implementation. Respir Care 2007;52(3):301-23.
[38] Chatburn RL. Which ventilators and modes can be used to deliver noninvasive
ventilation? Respir Care 2009;54(1):85-101.
[39] Al-Saady N, Bennett ED. Decelerating inspiratory flow waveform improves lung
mechanics and gas exchange in patients on intermittent positive-pressure ventilation. Int
Care Med 1985;11(2):68-75.
[40] Raul JL Jr. Inspiratory flow patterns: the „shape“ of ventilation. Respir Care
1993;38(1):132-40.
[41] Slutsky AS. Mechanical ventilation. Chest 1993;104(6):1833-59
[42] Munoz J, Guerrero JE, Palomino R, De La Calle B. Pressure-controlled ventilation
versus controlled mechanical ventilation with decelerating inspiratory flow. Crit Care
Med 1993;21:1143-48.
46
[43] Ravenscraft SA, Burke WC, Marini JJ. Volume-cycled decelerating flow: an alternative
form of mechanical ventilation. Chest 1992;101:1342-51
[44] Alvarez A, Subirana M, Benito S. Decelerating flow ventilation effects in acute
respiratory failure. J Crit Care 1998;13(1):21-25.
[45] Darowski M. Volume controlled with self adapting inspiratory flow pattern- a new
approach to mechanical ventilation of the lungs. Biocybernetics and Biomedical
Engineering 1995;15:5-15.
[46] Younes M. Proportional Assist Ventilation, a New Approach to Ventilatory Support.
Am Rev Respir Dis 1992;145:114–20.
[47] Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversal of obstructive sleep apnea by
continuous positive airway pressure applied through the nares. Lancet 1981;1:862-5.
[48] Golczewski T, Darowski M. Virtual respiratory system in investigation of CPAP
influence on optimal breathing frequency in obstructive lungs disease. Nonlinear
Biomedical Physics 2007;1:item 6.
[49] Case RB, Berglund E, Sarnoff SJ. Ventricular functions. II. Quantitative relationship
between coronary flow and ventricular function with studies on unilateral failure. Circ
Res 1954;2(4):319-25.
[50] Guyton AC. Determination of cardiac output by equating venous return curves with
cardiac response curves. Physiol Rev 1955;35(1):123-9.
[51] Sagawa K. Analysis of ventricular pumping capacity as a function of input and output
pressure loads. In Reeve EB, Guyton AC, editors. Physical bases of circulatory
transport: regulation and exchange. Philapelphia: W.B. Saunders Company; 1967.
p.141.
[52] Sarnoff SJ. Myocardial contractility as described by ventricular function curves. Physiol
Rev 1955;35(1):107-22.
[53] Guyton AC, Jones CE, Coleman TG. Circulatory physiology: Cardiac output and its
regulation. Philapelphia: W.B. Saunders Company, 1973.
[54] Sagawa K, Maughan L, Suga H, Sunagawa K. Cardiac contraction and the Pressure-
Volume Relationship. New York: Oxford University Press,1988.
[55] Westerhof N, Elzinga G, Sipkema P. An artificial arterial system for pumping hearts. J
Appl Physiol 1971;31(5):776-81.
47
[56] van den Horn CJ, Westerhof N, Elzinga G. Interaction of heart and arterial system. Ann
Biomed Eng 1984;12:151-62.
[57] Westerhof N, Sipkema P, Elzinga G, Murgo JP, Giolma JP. Arterial Impedance. In:
Hwang NHC, Gross DR, Patel DJ, editors. Quantitative Cardiovascular Studies.
Baltimore: University Park Press; 1979. p.111-50.
[58] Piene H. Impedance matching between ventricle and load. Ann Biomed Eng
1984;12:191-207.
[59] Sonnenblick EH. Force-velocity relations in mammalian heart muscle. Am J Physiol
1962;202:931-9.
[60] Suga H. Total mechanical energy of a ventricle model and cardiac oxygen consumption.
Am J Physiol 1979;236(3):H498-H505.
[61] Shroff SG, Janicki JS, Weber KT. Evidence and quantitation of left ventricular systolic
elastance. Am J Physiol 1985;249 (Heart Circ Physiol 18): H358-70.
[62] Campbell KB, Kirkpatrick RD, Knowlen GG, Ringo JA. Late-systolic pumping
properties of the left ventricle. Deviation from elastance-resistance behaviour. Circ Res
1990;,66:218-33.
[63] Korakianitis T, Shi Y. Numerical comparison of hemodynamics with atrium to aorta
and ventricular apex to aorta VAD support. ASAIO J 2007;53(5):537-48.
[64] Noordergraf A, Verdouw PD, van Brummelen AGW, Wiegel FW. Analog of the arterial
bed. In Attinger EO, editor. Pulsatile blood flow. New York: Mc Graw Hill; 1964.
p.373-87.
[65] Avolio AP. Multi-branched model of the human arterial system. Med Biol Eng Comput
1980;18(6):709-18.
[66] Dagan J. Pulsatile mechanical and mathematical model of the cardiovascular
system. Med Biol Eng Comput 1982;20(5):601-7.
[67] Burattini R, Gnudi G. Computer identification of models for the arterial tree input
impedance: comparison between two new simple models and first experimental results.
Med Biol Eng Comput 1982;20(2):134-44.
[68] Donovan FM jr. Design of a hydraulic analog of the circulatory system for evaluating
artificial hearts. Biomater Med Devices Artif Organs 1975;3:439-449.
48
[69] Rosenberg G, Phillips WM, Landis DL, Pierce WS. Design and evaluation of the
Pennsylvania State University mock circulatory system. ASAIO J 1981;4:41-49.
[70] Swanson WM, Clark RE. A simple cardiovascular system simulator: design and
performance. J Bioeng 1977;1:135-145.
[71] Arabia M, Akutsu T. A new test circulatory system for research in cardiovascular
engineering. Ann Biomed Eng 1984;12:29-48.
[72] Ferrari G, De Lazzari C, Mimmo R, Tosti G, Ambrosi D, Górczyńska K. A computer
controlled mock circulatory system for mono and biventricular assist device testing. Int
J Artif Organs 1998;21(1):26-36.
[73] Bowles CT, Shah SS, Nishimura K, Clark C, Cumming DV, Pattison CW et al.
Development of mock circulation models for the assessment of counter pulsation
systems. Cardiovasc Res 1991;25(11):901-08.
[74] Schima H, Baumgartner H, Spitaler F, Kuhn P, Wolner E. A modular mock circulation
for hydromechanical studies on valves, stenoses, vascular grafts and cardiac assist
devices. Int J Artif Organs 1992;15(2):417-21.
[75] Knierbein B, Reul H, Eilers R, Lange M, Kaufmann R, Rau C. Compact mock loops of
the systemic and pulmonary circulation for blood pump testing. Int J Artif Organs 1992;
15(1):40-8.
[76] Woodard JC, Rock SM, Portner PM. A sophisticated electromechanical ventricular
simulator for ventricular assist system testing. ASAIO Trans 1991;37(3):M210-11.
[77] Zhou J, Armstrong GP, Medvedev AL, Smith WA, Golding LA, Thomas JD. Numeric
modeling of the cardiovascular system with a left ventricular assist device. ASAIO J
1999;45(1):83-9.
[78] Vollkron M, Schima H, Huber L, Wieselthaler G. Interaction of the cardiovascular
system with an implanted rotary assist device: simulation study with a refined computer
model. Artif Organs 2002;26(4):349-59.
[79] Geertsema AA, Rakhorst G, Mihaylov D, Blanksma PK, Verkerke GJ. Development of
a numerical simulation model of the cardiovascular system. Artif Organs
1997;21(12):1297-301.
[80] Pontrelli G. A multiscale approach for modelling wave propagation in an arterial
segment. Comput Methods Biomech Biomed Engin 2004;7(2):79-89.
49
[81] Ferrari G, Kozarski M, De Lazzari C, Clemente F, Merolli M, Tosti G et al. A hybrid
(numerical-physical) model of the left ventricle. Int J Artif Organs 2001;24(7):456-62.
[82] Ferrari G, Kozarski M, De Lazzari C, Górczyńska K, Tosti G, Darowski M.
Development of a hybrid (numerical-hydraulic) circulatory model: prototype testing and
its response to IABP assistance. Int J Artif Organs 2005;28(7):750-759.
[83] Kozarski M, Ferrari G, Zieliński K, Górczyńska K, Pałko KJ, Tokarz A et al. A New
Hybrid Electro-Numerical Model of the Left Ventricle. Comput Biol Med
2008;38(9):979-89.
[84] Olender MF, Clark JW, Stevens PM. Analog computer simulation of maximum
expiratory flow limitation. IEEE Trans Biomed Eng 1976:23:445–52.
[85] Golczewski T, Darowski M. Virtual respiratory system for education and research:
simulation of expiratory flow limitation for spirometry, Int J Artif Organs 2006;29:961-
72.
[86] Lambert RK, Wilson TA, Hyatt RE, Rodarte JR. A computational model for expiratory
flow. J Appl Physiol: Respirat Environ 1982;52:44-56.
[87] Polak AG. A forward model for maximum expiration. Comput Biol Med 1998;28:613-
25.
[88] Polak AG, Lutchen KR. Computational model for forced expiration from asymmetric
normal lungs. Ann Biomed Eng 2003;31:891–907.
[89] Weibel ER. Morphometry of the human lung, New York, Academic Press, 1963.
[90] Horsfield K, Dart G, Olson DE, Filley GF, Cumming G. Models of the human bronchial
tree. J Appl Physiol 1971;31:207-17.
[91] Otis AB, McKerrow CB, Bartlett RA, Mead J, McIlroy MB, Selver-Stone NJ, Radford
Jr EP., Mechanical factors in distribution of pulmonary ventilation. J Appl Physiol
1956;8:427–43.
[92] Mead J. Contribution of compliance of airways to frequency-dependent behavior of
lungs. J Appl Physiol 1969;26:670-673.
[93] Bates JH, Brown KA, Kochi T. Respiratory mechanics in the normal dog determined by
expiratory flow interruption. J Appl Physiol 1989;67:2276-85.
50
[94] Avanzolini G, Barbini P, Bernardi F, Cevenini G, Gnudi G. Role of the mechanical
properties of tracheobronchial airways in determining the respiratory resistance time
course. Ann Biomed Eng 2001;29(7):575-86.
[95] Khirani S, Biot L, Eberhard A, Baconnier P. Positive end expiratory pressure and
expiratory flow limitation: a model study. Acta Biotheor 2001;49(4):277-90.
[96] Lambert RK. Simulation of the effects of mechanical nonhomogeneities on expiratory
flow from human lungs. J Appl Physiol 1990;68(6):2550-63.
[97] Avanzolini G, Barbini P. A versatile identification method applied to analysis of
respiratory mechanics. IEEE Trans Biomed Eng 1984;31(7):520-6.
[98] Lutchen KR, Kaczka DW, Suki B, Barnas G, Cevenini G, Barbini P. Low-frequency
respiratory mechanics using ventilator-driven forced oscillations. J Appl Physiol
1993;75(6):2549-60.
[99] Kaczka DW, Ingenito EP, Lutchen KR. Technique to determine inspiratory impedance
during mechanical ventilation: implications for flow limited patients. Ann Biomed Eng
1999;27(3):340-55.
[100] Ursino M, Colí L, Brighenti C, Chiari L, de Pascalis A, Avanzolini G. Prediction of
solute kinetics, acid-base status, and blood volume changes during profiled
hemodialysis. Ann Biomed Eng 2000;28(2):204-16.
[101] Ursino M, Magosso E, Avanzolini G. An integrated model of the human ventilatory
control system: the response to hypercapnia. Clin Physiol 2001;21(4):447-64.
[102] Beyar R, Goldstein Y. Model studies of the effects of the thoracic pressure on the
circulation. Ann Biomed Eng 1987;15:373-83.
[103] Abraham E, Yoshihara G. Cardiorespiratory effects of pressure controlled ventilation in
severe respiratory failure. Chest 1990;98:1445-49.
[104] Fitz-Clarke JR. Computer simulation of human breath-hold diving: cardiovascular
adjustments. Eur J Appl Physiol 2007;100:207-24.
[105] Corno C; Fiore GB, Costantino ML. A mathematical model of neonatal tidal liquid
ventilation integrating airway mechanics and gas transfer phenomena. IEEE Trans
Biomed Eng 2004;51:604-11.
[106] Golczewski T. Gas exchange in virtual respiratory system – simulation of ventilation
without lungs movement. Int J Artif Organs 2007;30:1047-56.
51
[107] Niranjan SC, Bidani A, Ghorbel F, et al. Theoretical Study of Inspiratory Flow
Waveforms during Mechanical Ventilation on Pulmonary Blood Flow and Gas
Exchange. Comp Biomed Res 1999;32:355-90.
[108] Winkler T, Krause A, Kaiser S. Simulation of mechanical respiration using a
multicompartment model for ventilation mechanics and gas exchange. Int J Clin Mon
Comp 1995;12:231-9.
[109] Golczewski T, Zielinski K, Ferrari G, Palko KJ, Darowski M. Influence of ventilation
mode on blood oxygenation - investigation with Polish virtual lungs and Italian model
of circulation. Biocybernetics and Biomedical Engineering 2009 [in press].
[110] Heldt T, Shim EB, Kamm RD, Mark RG. Computational modeling of cardiovascular
response to orthostatic stress. J Appl Physiol 2002;92(3):1239-54.
[111] Darowski M, Kozarski M, Gólczewski T. Model studies on respiratory parameters for
different lung structures. Biocybernetics and Biomedical Engineering 2000;20(2):67-77.
[112] Schuessler TF, Gottfried SB, Bates JH. A model of the spontaneously breathing patient:
application to intrinsic PEEP and work of breathing. J Appl Physiol 1997;82:1694-703.
[113] Recommendations of the European Alliance for Medical and Biological Engineering
and Science for the 7th Framework Program of the European Commission. EAMBES
Documents, March 2006.
[114] Golczewski T, Kozarski M, Darowski M. The respirator as a user of virtual lungs.
Biocybernetics and Biomedical Engineering 2003;23(2):57-66.
[115] Gólczewski T, Darowski M. Influence of ventilatory mode on respiration parameters –
investigation on virtual lungs. Biocybernetics and Biomedical Engineering
2003;23(3):63-72
[116] Gólczewski T. Gas exchange in virtual respiratory system – simulation of ventilation
without lungs movement. Int J Artif Organs 2007;30:1047-56.
[117] Tomalak W, Golczewski T, Michnikowski M, Darowski M. Virtual respiratory system
for interactive e-learning of spirometry Eur Resp Rev 2008;17:36-8.