ASAIO Journal 2004

Characterization of an Adult Mock Circulation for Testing Cardiac Support Devices

GEORGE M. PANTALOS,* STEVEN C. KOENIG,* KEVIN J. GILLARS,* GURUPRASAD A. GIRIDHARAN,*
AND

DAN L. EWERT

A need exists for a mock circulation that behaves in a physiologic manner for testing cardiac devices in normal and pathologic states. To address this need, an integrated mock cardiovascular system consisting of an atrium, ventricle, and systemic and coronary vasculature was developed specically for testing ventricular assist devices (VADs). This test conguration enables atrial or ventricular apex inow and aortic outow cannulation connections. The objective of this study was to assess the ability of the mock ventricle to mimic the FrankStarling response of normal, heart failure, and cardiac recovery conditions. The pressurevolume relationship of the mock ventricle was evaluated by varying ventricular volume over a wide range via atrial (preload) and aortic (afterload) occlusions. The input impedance of the mock vasculature was calculated using aortic pressure and ow measurements and also was used to estimate resistance, compliance, and inertial mechanical properties of the circulatory system. Results demonstrated that the mock ventricle pressurevolume loops and the end diastolic and end systolic pressurevolume relationships are representative of the Starling characteristics of the natural heart for each of the test conditions. The mock vasculature can be congured to mimic the input impedance and mechanical properties of native vasculature in the normal state. Although mock circulation testing systems cannot replace in vivo models, this conguration should be well suited for developing experimental protocols, testing device feedback control algorithms, investigating ow proles, and training surgical staff on the operational procedures of cardiovascular devices. ASAIO Journal 2004; 50:3746.

M ock circulation systems have been used extensively for in

vitro performance testing of cardiovascular devices. Although mock circulatory systems cannot replace animal and clinical trials, they form an essential and important step in the design and validation of cardiac devices. The earliest mock circulatory systems were pulse duplicators driven by stepping motors

and linear actuators1 4 and were primarily used for testing heart valves. Mock circulation systems have been used for hemodynamic, hydrodynamic, and hemolysis studies of prosthetic vessels and cardiac assist devices.510 Several physical models of the circulation have been developed to simulate and match important physiologic parameters.1116 Elaborate control schemes have been developed to mimic the physiologic behavior1719 of the cardiovascular system. Recently, Ferrari et al. developed a hybrid, numericalphysical model of the circulatory system with an electrohydraulic interface to overcome the lack of exibility in using a purely physical mock circulatory system.18 19 Mock vasculatures, such as the Pennsylvania State University mock circulation,20 have been used primarily to test, design, and validate the pumping performance of ventricular assist devices (VADs) and articial hearts. To gain insight into VAD and other cardiac device responses to a wide range of operating conditions, a mock ventricle that responds to varying preload, afterload, heart rates, and clinical conditions (normal, failing, and recovering heart) that mimic FrankStarling behavior is needed. It is also essential for a mock circulatory system to mimic the native cardiovascular system, particularly to changes in volume as well as a mock vasculature that can produce physiologically equivalent vascular input impedance. Additionally, it is desirable to be able to approximate the surgical cannulation by having ventricular, atrial, and aortic cannulation access sites for testing of different VAD designs. An in vitro circulatory system was developed to produce physiologically equivalent pressures, volumes, and ows for testing VADs. The objective of this study was to evaluate the ability of the mock circulation to faithfully reproduce normal, heart failure, and partial cardiac recovery physiologic states as a platform for testing circulatory support devices. Methods Experimental Design

Presented at the 2nd Joint EMBS-BMES Conference, October 2326, 2002, Houston, Texas. From the *Jewish Hospital Heart and Lung Institute, Department of Surgery, University of Louisville, Kentucky; and Department of Electrical and Computer Engineering, North Dakota State University, Fargo. Submitted for consideration March 2003; accepted for publication in revised form August 2003. Correspondence: George M. Pantalos, Ph.D., Jewish Hospital Cardiothoracic Surgical Research Institute, at the University of Louisville, 500 South Floyd Street, Room 102, Department of Surgery, University of Louisville, Louisville, KY 40202; E-mail gmpant02@athena.louisville.edu. DOI: 10.1097/01.MAT.0000104818.70726.E6

The objective of this study was to assess the ability of the mock circulation to mimic the FrankStarling response of the natural heart and vasculature in normal, heart failure, and partial cardiac recovery states. The pressurevolume relationship, determined by varying ventricular volume over a wide range via atrial (preload) and aortic (afterload) occlusions, was used to evaluate the response of the mock ventricle. The mock vasculature was characterized using input impedance derived from aortic pressure and ow measurements, as well as estimates of vascular resistance, compliance, and inertial properties. The mock ventricle was congured to mimic native normal, 37

38

PANTALOS ET AL.

Table 1. Adult Mock Circulation Parameter Values for Simulating Normal, Failure, and Cardiac Recovery Physiologic Conditions Parameter Mean AoP CO LVPed DLP Normal 95 mmHg 5.0 L/m 25 mmHg 210 mmHg Failure 65 mmHg 3.0 L/m 1525 mmHg 80 mmHg Recovery 80 mmHg 4.0 L/m 1015 mmHg 150 mmHg

AoP, aortic pressure; CO, cardiac output; LVPed, left ventricular end diastolic pressure; DLP, drive line pressure.

failing heart, and partial cardiac recovery test conditions, as dened in Table 1. The fraction of the cardiac cycle for systolic duration of the mock ventricle was set to 37% for all test conditions. Heart rate (HR) was randomly selected at 80, 100, and 120 beats per minute (bpm) for each test condition. First, 30 seconds of baseline hemodynamic measurements were continuously recorded for the normal state at a randomly selected HR. Second, hemodynamic measurements were continuously recorded while gradually reducing preload by occluding atrial inow with a turn style resistor clamp up to atrial collapse and during return to baseline. Third, hemodynamic measurements were again continuously recorded while the afterload was gradually increased by occluding the aortic outow with a turn style resistor clamp up to total occlusion and following release with return to baseline. The sequence of (1) baseline, (2) reducing preload (via atrial inow occlusion), and (3) increasing afterload (via aortic outow occlusion) at each of the three randomly selected HR in the normal state was repeated for failing heart and partial cardiac recovery test conditions. Characterizing hemodynamic parameter values and ventricular pressurevolume relationships were derived for each test condition and compared with published human physiologic data.21,22 The mock vasculature and mock ventricle were appropriately adjusted (e.g., driveline pressure, peripheral resistance) to mimic normal21 and heart failure (elevated input impedance)22 test conditions. Thirty seconds of hemodynamic measurements were continuously recorded in the normal and elevated afterload states. The vascular input impedance and vascular mechanical properties (resistance, compliance, and inertance) were derived for each test condition and compared with human physiologic data. Mock Circulation The mock circulation system consists of atrium, ventricle, and systemic and coronary vasculature components (Figures 1 and 2). The articial atrium20 is made of a exible polyurethane sphere 50 mm in diameter and is connected upstream of the inow valve of the mock ventricle. The mock ventricle consists of a exing, segmented polyurethane sac inside a pressurization chamber.23 The ventricular sac is hemiellipsoid shaped and is 70 mm across at the base and 83 mm long from base to apex. The base is covered by a semirigid polymer dome 20 mm high with mounts for inow (mitral) and outow (aortic) prosthetic valves. For this particular set of experimental conditions, custom made rigid mounted trileaet polyurethane valves24 were used so that a pressurevolume catheter could

Figure 1. Illustration of adult mock circulation loop consisting of a mock ventricle with atrial or apical inow cannulation sites, mock systemic vasculature with aortic root or descending thoracic aorta outow cannulation sites, and mock coronary vasculature. Also shown are transducers for measuring hemodynamic pressures, ows, and volume.

be passed retrograde through the outow valve into the ventricular sac. Appropriately mounted tissue or mechanical valves can be used for other test conditions not requiring the retrograde passage of the pressurevolume catheter. Metered pulses of compressed air are delivered from a pneumatic controller (Utah Drive, CardioWest Technologies, Tucson, AZ) to the pressurization chamber during systole compressing the ventricular sac to form coapting quadrants, simulating contraction of the normal and dysfunctional ventricle and the delivery of the cardiac stroke volume (Figure 3). An articial aorta (polyurethane tube segment) 25 mm in diameter is connected downstream of the outow valve of the ventricular sac to the mock systemic and coronary vasculature. The mock systemic vasculature consists of four integrated chambers each 10.1 cm in diameter that represent lumped proximal resistance, systemic compliance, peripheral resistance, and venous compli-

Figure 2. Photograph of adult mock circulation loop (left) and shown with a pulsatile ventricular assist device (VAD) connected to the apex of the mock ventricle (right).

MOCK CIRCULATION

39

Figure 3. Photographs of the mock ventricular sac lled at end diastole (left) and compressed at end systole (right).

ance.25 The resistance chambers contain open cell foam of different densities compressed by sealed pistons to create the required amount of resistance to ow. The compliance chambers have spring loaded pistons on roll sock diaphragms to create the required amount of arterial and venous compliance.26 Introduction ports for VAD uptake cannulae are incorporated into the atrial wall and ventricular sac apex. Introducer ports in the atrium, ventricle, aorta, and pressurization chamber enable insertion of pressure catheters. In line ow probes (Transonics, Ithaca, NY) are used for aortic, coronary, and VAD ow. Instrumentation A high delity, dual pressurevolume conductance catheter (Millar Instruments, Houston, TX) was inserted into the aortic introducer port and passed retrograde through the outow (aortic) valve and down to the ventricular apex for simultaneous mock ventricular pressure, ventricular volume, and aortic pressure measurements. Single tip, high delity catheters (Millar Instruments, Houston, TX) were inserted into introducer ports for measuring mock atrial pressure and mock ventricle driveline pressure. Aortic ow was measured with an in-line, transit time ow probe (Transonics, Ithaca, NY). Pressure and ow transducers were calibrated before and after to verify gain and offset accuracy. All instrumentation for hemodynamic measurement of pressures, ows, and volume are shown in Figure 1. Signal conditioning (amplication and offset) was accomplished using transducer ampliers (Ectron, San Diego, CA), a transit time owmeter (Transonics, Ithaca, NY), a volume conductance unit (Leycom, Sigma V, Netherlands), and other peripheral conditioners integrated in a good laboratory practice (GLP) compliant instrumentation system. Signal conditioned data were low pass ltered at 60 Hz, analog to digitally converted (AT-MIO-16E-10 and LabVIEW, National Instruments, Austin, TX) at a sampling rate of 400 Hz, and stored on a personal computer using custom data acquisition software.27 Data Analysis Characterizing hemodynamic parameter values, ventricular pressurevolume relationships, vascular input impedance, and vascular mechanical properties (resistance, compliance, and inertance) were calculated using the Hemodynamic Evaluation and Assessment Research Tool (HEART) program28 and m les developed in Matlab (MathWorks, Natick, MA). Hemodynamic Parameters Mock circulation pressures, ows, and volumes were used to calculate the following hemodynamic parameters: stroke

volume (SV) by integrating the aortic root ow, HR, and cardiac output (CO); left ventricular beginning (LVbd) and ending diastolic (LVPed), peak systolic (LVPpksys) pressure, maximum LV peak positive (dLVP/dt) and peak negative (dLVP/dt) slopes, LV end diastolic (LVVed) and end systolic (LVVes) volumes, and LV ejection fraction (LVEF LVSV LVVed); left atrial mean pressure (LAPmean); aortic peak systolic (AoPsys), end diastolic (AoPed), and mean (AoPmean) pressures; and aortic mean (AoFmean) and peak (AoFpk) ow. All hemodynamic parameters were calculated on a beat to beat basis with all beats in each data set averaged to obtain a single representative mean value for each parameter. Ventricular Pressure Volume Relationship Pressurevolume loops were created by plotting ventricular pressure against ventricular volume, where each loop represents one complete cardiac cycle (one beat). The LV volume data was calibrated by adjusting the LV stroke volume ( LVVed LVVes) to match the SV derived from aortic root ow. Baseline pressurevolume loops for normal, failing heart, and cardiac recovery test conditions were generated. The end diastolic pressurevolume relationship (EDPVR) and end systolic pressurevolume relationship (ESPVR) were determined from pressurevolume loops recorded during atrial inow and aortic outow occlusions. Vascular Input Impedance Vascular input impedance provides a rigorous way to represent the afterload of the heart and depends in part on the resistance and elastance of the vasculature and the inertia of the blood. It may be inuenced by pressure and ow wave reections. Traditionally, clinicians rely on systemic resistance [(AoPmean central venous pressure) CO] instead of the terms of the input impedance inuenced by elastance and inertial components. Any repeating waveform, including aortic pressure and ow waveforms, can be represented by a Fourier series, which is a sum of cosines and sines of differing magnitudes and phase angles at multiples of the fundamental frequency (harmonics). The input impedance of the mock vasculature was derived using aortic pressure and ow measurements during baseline recordings for each test condition using Fourier analysis algorithms and m les developed in Matlab (MathWorks, Natick, MA). The magnitudes and phases of the impedance for 10 harmonics were calculated as the ratio of the magnitudes of AoP and AoF and subtracting the phases of AoP and AoF at each harmonic. An uncertainty analysis was performed to calculate the error in the input impedance terms resulting from experimental measurement errors.29 Vascular Mechanical Properties The mechanical parameters of the mock vasculature were estimated using the four element Windkessel model,30 as shown in Figure 4, and a frequency based parameter estimation routine.31 Briey, the experimental impedance magnitudes and phases are compared with the model derived impedance magnitudes and phases, and the lumped parameter elements (resistance, compliance, inertance, and characteristic impedance) are adjusted iteratively until the error between the experimental and model derived impedance is minimized.

40

PANTALOS ET AL.

Figure 4. The four element Windkessel model used to estimate total peripheral resistance (R), systemic arterial compliance (C), inertance (L), and characteristic impedance (Z) from vascular input impedance data (AoP aortic pressure and AoF aortic ow).

Parameter estimates were made on a beat to beat basis and averaged for all beats in each data set. Results The adult mock circulation produced physiologically equivalent hemodynamic waveform magnitudes and morphology for the normal, failing heart, and cardiac recovery test conditions (Figure 5). Notably, ringing in the aortic pressure waveform coincident with the dicrotic notch was observed, which is typical of prosthetic valve closure function, particularly when rigid mounted. The aortic pressure waveform shown in Figure 5 was digitally ltered using the Matlab FILTFILT function, which is a zero phase forward and reverse digital lter. Characteristic cardiovascular parameters of the adult mock circulation for the normal, failing heart, and cardiac recovery states were comparable with physiologic parameters (Table 2). Elevated lling pressure (LVPed and LAPmean) and depressed stroke volume, peak aortic ow, contractility (dLVP/dt), external work (EW), ejection fraction (EF), and perfusion pressure (AoPmean) were successfully reproduced to mimic the failing heart test condition. The pressurevolume relationship during baseline for normal, failing heart, and partial cardiac recovery test conditions were also indicative of the native left ventricle (Figure 6). In the heart failure state, the pressurevolume loops were shifted to the right, and the area within the pressurevolume loops were reduced indicative of a larger heart with depressed function, including reduced SV, reduced stroke work, and reduced ejection fraction. In the partial cardiac recovery state, the pressure volume loops were shifted back leftward, and the area increased, mimicking a recovery was shifted back toward the normal state. The pressurevolume loops during atrial inow and aortic outow occlusions show the response of the mock ventricle to varying ventricular preload and afterload over a wide range of loading (Figure 7). The slope of the EDPVR for the heart failure state (0.32 mm Hg/ml) is nearly twice as large as the normal state (0.18 mm Hg/ml), which indicates a stiffer ventricle during lling (Figure 8). The slope of the ESPVR of the heart failure state (0.28 mm Hg/ml) is less than half that of the normal state (0.62 mm Hg/ml), which demonstrates a marked reduction in contractile function of the mock ventricle (Figure 9). The magnitude and phase characteristics of the systemic input impedance of the mock vasculature for the normal21 and

Figure 5. Adult mock circulation hemodynamic waveforms for left ventricle in (a) normal, (b) failure, and (c) partial cardiac recovery test conditions at heart rate 100 beats per minute (bpm) (AoP aortic pressure, LVP left ventricular pressure, AoF aortic ow, and LVV left ventricular volume).

heart failure22 (elevated input impedance) states are shown in Figure 10. The mechanical properties of the mock vasculature estimated using the four element Windkessel model (Figure 4) for the normal and heart failure (elevated input impedance) states are listed in Table 3. Uncertainty analysis showed less than 10% error in the magnitude and phase values for up to 10 harmonics, with less error in the lower order harmonics ( 5% error) compared with the higher order harmonics (710% error). Discussion Mock circulation systems provide an attractive platform for testing cardiac devices. Although they cannot replace in vivo models, they are well suited for developing experimental protocols, testing device performance and feedback control algorithms, investigating velocity proles, and training surgical staff with operational procedures of cardiovascular devices in normal and adverse circumstances. Some investigators connect the cardiac device directly to a mock vasculature, thereby eliminating the need for a mock atrium and mock ventricle. Additionally, many approach tuning the mock circulation to produce physiologically equivalent mean preload and afterload pressures and resistances while not considering other

Table 2. Adult Mock Circulation Hemodynamic Parameter Values for Normal, Failure, and Cardiac Recovery Test Conditions at Randomly Selected Heart Rates of 80, 100, and 120 bpm LV normal HR 100 bpm HR 120 bpm HR 80 bpm HR 100 bpm HR 120 bpm HR 80 bpm LV normal LV failure LV failure LV failure LV recovery LV recovery HR 100 bpm LV recovery HR 120 bpm

Adult Mock Circulation Parameter

LV normal

HR 80 bpm

MOCK CIRCULATION

LV SV (ml) LV HR (bpm) LV CO (L/min) LVPbd (mmHg) LVPed (mmHg) LVPpksys (mmHg) dLVP/dt (mmHg/s) dLVP/dt (mmHg/s) LV EW (mmHg-ml) LVVed (ml) LVVes (ml) LV EF (%) LAPmean (mmHg) AoPsys (mmHg) AoPdia (mmHg) AoPmean (mmHg) AoFmean (L/min) AoFpk (L/min)

63 80 5.0 9.3 2.4 139.5 2146 4587 7924 152 79 48 10.5 125.5 68.0 95.3 4.9 20.1

50 100 5.0 4.7 2.4 137.4 2435 4552 6173 155 99 36 11.5 124.4 70.6 95.1 4.9 20.1

44 120 5.3 8.1 0.7 137.1 2704 4458 5609 134 87 35 11.7 135.3 73.8 95.7 4.9 21.4

37 80 3.0 0.9 13.7 93.0 1149 3214 2582 190 139 27 16.3 84.4 43.1 60.6 2.9 11.0

30 100 3.0 1.3 15.3 89.6 1356 3175 2153 188 145 23 15.0 88.8 46.9 62.6 2.9 12.7

27 120 3.2 2.1 13.5 85.5 1414 2870 1876 190 150 21 14.2 77.3 47.7 61.3 3.1 12.9

54 80 4.4 5.2 4.7 121.5 1726 3911 5586 174 117 33 11.5 103.4 55.8 79.1 4.1 16.1

41 100 4.1 2.1 4.2 114.9 2066 3600 4279 172 125 28 11.9 101.9 59.7 80.3 4.0 16.9

35 120 4.3 3.6 4.0 114.4 2200 3792 3568 171 130 24 13.2 104.4 60.5 79.5 4.0 16.4

LV, left ventricular; SV, stroke volume; HR, heart rate; CO, cardiac output; LVPbd, LV beginning diastolic pressure; LVPed, LV end diastolic pressure; LVPpksys, LV peak systolic pressure; dLVP/dt, maximum peak positive LV slope; dLVP/dt, maximum peak negative LV slope; LVEW, left ventricular external work; LVVed, LV end diastolic volume; LVVes, LV end systolic volume; LV EF, LV ejection fraction; LAPmean, left atrial mean pressure; AoPsys, aortic peak systolic pressure; AoPdia, aortic diastolic pressure; AoPmean, aortic mean pressure; AoFmean, mean aortic ow; AoFpk, peak aortic ow.

41

42

PANTALOS ET AL.

Figure 6. Adult mock ventricle pressurevolume loops for simulated normal, failure, and partial cardiac recovery baseline test conditions at heart rate 100 beats per minute (bpm).

properties such as the elastance and inertance of the mock vasculature. The mock circulation can be operated to mimic the FrankStarling response of the native ventricle for normal, failing heart, and partial cardiac recovery test conditions. Furthermore, it can be set up to test cardiac assist devices in a variety of different congurations, including the choice of atrial or apical uptake cannulation and aortic root or descending thoracic aorta return cannulation. This study demonstrated the ability of the mock circulation to produce the Frank Starling response with physiologic characteristic hemodynamic parameters and pressurevolume relationships for all test conditions. The input impedance and mechanical properties of the mock vasculature were comparable to the natural, normal vasculature, but only the resistance in the failing heart state could be faithfully reproduced. The cardiac and vascular hemodynamic waveforms of the mock circulation are comparable to natural physiology in both magnitudes (Table 2) and morphologic features3233 (Figure 5). These data demonstrate that the mock ventricle can simulate normal, failing heart, and cardiac recovery test conditions. Specically, lling pressures and ventricular volumes are elevated whereas cardiac output, stroke volume, and systolic pressures are depressed in the heart failure state compared with normal state, which are indicative of clinical ndings. The most notable departure in waveform morphologic features can be seen in the aortic pressure waveform. There is ringing at the dicrotic notch during valve closure. The ringing is associated with the limitations of a rigid mounted mechanical valve and substantial inertance in the mock vasculature introduced by the outow conduit and ow probe used to measure aortic ow. This ringing is also observed in the vicinity of a prosthetic mechanical valve that is attached to a natural vasculature.34 Despite this limitation, the characteristic hemodynamic parameter values and waveform morphology are within the physiologic envelope and are adequate for testing cardiac devices. The ventricular pressurevolume loops during baseline test conditions (Figure 6) are indicative of what has been reported clinically.35 The ventricular SV and end systolic pressure are

depressed whereas lling pressure is elevated in the heart failure state compared with normal heart state. Further, the external work of the mock ventricle in the failing heart state is nearly one-third that of the normal heart, which is also indicative of a dysfunctional heart.35 The EDPVR and ESPVR derived from the occlusion data (Figure 7) for each test condition also follow the expected physiologic response. The slope of the EDPVR for the failing heart state is approximately twice that of the normal heart state (Figure 8), which is indicative of a stiffer heart during lling. The slope of the ESPVR for the failing heart state is less than half that of the normal heart state (Figure 9), which is indicative of a less contractile heart during ejection. The combination of a stiffer heart during lling and a less contractile heart during ejection are characteristic of the natural, failing heart. Further, the pressurevolume loops, EDPVR, and ESPVR for the cardiac recovery state demonstrate an improved cardiac state with a return toward the normal heart condition. The pressurevolume relationships for each test condition are comparable with what has been reported clinically in VAD patients.36 38 The input impedance and mechanical properties of the mock vasculature are representative of the natural, normal vasculature (Figure 10). The magnitudes of the DC and 10 harmonic terms are comparable. The phase angles for the rst three harmonics show a good match between the mock and natural vasculature. However, the zero phase crossover point of the mock phase lags behind that of the native phase, and at the higher order harmonics, phase values deviate by approximately 30. These differences are likely caused by altered wave reections or inertial effects inherent in the mock vasculature because of the cross-sectional area and length of tubing, as well as areas of the mock vasculature components and the uid volumes contained therein. The mechanical properties, particularly the vascular resistance and compliance indicate a good match between the natural and mock vasculature in the normal state (Table 3). For the heart failure condition, the mock vasculature can reproduce the DC component of the input impedance, indicative of the resistive load. The phase characteristics of the input impedance of the mock vasculature show similar differences (approximately 1525) at each harmonic and a delay in the zero phase crossover point. The mechanical properties further support this nding because the resistance could be controlled but the compliance and characteristic impedance properties differed by an order of magnitude. Although the characteristic hemodynamic parameter values (mean, systolic, and diastolic) and waveform morphology can qualitatively appear to provide a good match between the natural and mock vasculature for the failing heart, differences in the input impedance and mechanical property estimates clearly demonstrate the limitations of the mock vasculature and illustrate the importance of not relying solely on waveform features, mean values, and peripheral resistance. This is a very signicant point that can be easily overlooked. It has been shown that the systemic input impedance in heart failure patients39 40 is stiffer (smaller compliance and larger resistance) compared with patients with normal ventricles.21 A stiffer vasculature can increase pulse wave propagation resulting in early return of reected wave energy increasing ventricular work load during ejection and reducing myocardial perfusion during relaxation.41

MOCK CIRCULATION

43

Figure 7. Adult mock ventricle pressurevolume (PV) loops during a simulated vena cava (VC, dotted line) and aortic (Ao, solid line) occlusion for simulated (a) normal, (b) failure, and (c) partial cardiac recovery test conditions at heart rate 100 bpm. The pressurevolume loops follow patterns typically seen in natural ventricles.

44

PANTALOS ET AL.

Figure 8. Adult mock ventricle end diastolic pressurevolume relationship (EDPVR) for simulated normal, failure, and partial cardiac recovery test conditions at heart rate 100 beats per minute (bpm).

Figure 10. Comparison of adult mock vasculature and natural heart (native) vasculature input impedance (magnitude and phase) for (a) normal heart and (b) failing heart test conditions at heart rate 80 beats per minute (bpm).

Clinical Signicance The ability to realistically simulate normal and various abnormal conditions of cardiovascular function is a valuable asset for academic instruction and training of clinical personnel. The mock circulation described in this report accords such an opportunity with great versatility while avoiding or minimizing the need for using laboratory animals. Many principles of cardiovascular function can be demonstrated without the inclusion of a neurohumoral response. Training on the use of clinical hemodynamic instrumentation can be accomplished while students and staff are exposed to a physical rather than an electronic reproduction of cardiovascular function. Likewise, instrumentation being considered for future application can be evaluated with this system. With a mock circulatory system, students have a safe opportunity to explore the

consequences of changing parameters and can make a tangible correspondence between an increase or decrease with the pressurization of the ventricular sac and pharmacologic agents that increase or decrease ventricular contractility. Likewise, the physical correspondence can be made between a change in the peripheral vascular resistance element and pharmacologic agents that increase or decrease vascular tone and peripheral resistance. The hemodynamic consequences of hypovolemia and the appropriate response can be readily simulated. The simulation scenarios are numerous. To date, impedance data have not been used as a clinical assessment tool in the evaluation of cardiovascular patients. However, their relevance needs to be appreciated in the testing of cardiovascular devices to give an accurate evaluation of device clinical performance and reliability. Using a mock circulatory system for testing a cardiovascular device, such as a valve or an articial ventricle, that does not have an input impedance that sufciently matches the circumstances in which it will be applied can result in exposure to overloading
Table 3. Adult Mock Circulation Mechanical Property Parameter Estimates for Normal and Failing Heart Test Conditions at Heart Rates of 80 bpm Using 4-Element Windkessel Model Normal Heart Parameter Resistance - R (dyne-s/cm5) Compliance - C (ml/mmHg) Inertance - L (mmHg-s2/ml) Characteristic impedance Z (mmHg-s/ml) Native 1218 1.4 0.02 65 Mock 1335 1.3 0.01 75 Heart Failure Native 2023 0.3 0.05 1066 Mock 1902 1.3 0.02 73

Figure 9. Adult mock ventricle end systolic pressurevolume relationship (ESPVR) for simulated normal, failure, and partial cardiac recovery test conditions at heart rate 100 beats per minute (bpm).

R, arterial resistance; C, arterial compliance; L, arterial inductance; Z, arterial characteristic impatience.

MOCK CIRCULATION

45

or underloading conditions that could yield misleading test results. The response to these unrealistic results may be an unnecessary redesign of a device with corresponding compromises in device performance. Hence, the use of a mock circulation system with a valid input impedance is an essential tool in the prediction of the clinical performance of a device from laboratory testing. Limitations The performance of the mock circulation during normal, failing heart, and cardiac recovery test conditions is representative of clinical observations from a purely hemodynamic/ hydrodynamic viewpoint. Clearly, a mock circulation is not intended to replace in vivo models and is incapable of replicating all expected clinical responses. For instance, the mock circulation cannot be used to study neurohumoral responses, tissue remodeling, activation of regulatory proteins, or changes in genetic phenotype. Furthermore, a number of mock circulation and numerical models have been proposed for investigating the performance of cardiac devices. The authors do not intend to suggest that the mock circulation system presented in this article is superior to these other systems but rather to demonstrate that it is able to reproduce the FrankStarling response reported clinically and that it would be of interest for testing cardiac devices and for approximating the input impedance of the natural vasculature. A key current limitation of the mock circulation used in this investigation is that it only represents the systemic circulation restricting testing to left ventricular assist devices. A comparable pulmonary circulation has been developed25 and will be incorporated into future investigations to allow testing during the simulation of right ventricular and biventricular failure. The current mock circulation also lacks authenticity in certain details of the arterial waveform (ringing) and impedance phase responses (i.e., inductance), particularly in the failing heart vasculature. The instrumentation used to record hemodynamic variables has inherent measurement errors associated with each technique (i.e., pressure error 1 mm Hg, ow error 0.5 L/min), which the authors attempted to minimize by using GLP compliant test equipment, calibration procedures, and documentation practices. The methods used to analyze these data can also introduce error, particularly the calculation of the EDPVR, ESPVR, and lumped-parameter model for assessing vascular mechanical properties. Despite these limitations, the mock circulation presented here should be a valuable research tool for assessing the performance of cardiac devices, developing experimental protocols in a controlled environment in preparation for in vivo studies, and training personnel on the operation and maintenance of cardiac devices. Acknowledgment
This project was funded by a grant from the Jewish Hospital Research Foundation (Louisville, KY). The authors would like to thank CardioWest Technologies, Tucson, AZ, for providing the pneumatic heart controller essential for the conduct of this investigation.

References
1. Wieting DW: Design and evaluation of a system suitable for analyzing ow behavior of prosthetic human heart valves (masters thesis). University of Texas, Austin, 1966.

2. Cornhill JF: An aorticleft ventricular pulse duplicator used in testing prosthetic aortic heart valves. J Thorac Cardiovasc Surg 73: 550 558, 1977. 3. Scotten LN, Dipl T, Walker DK, Brownlee RT: Construction and evaluation of a hydromechanical simulation facility for the assessment of mitral valve prostheses. Journal of Medical Engineering and Technology 3: 1118, 1979. 4. Walker DK, Modi VJ, Brownlee RT: In vitro assessment of mitral valve prosthesis. J Thorac Cardiovasc Surg 79: 680 688, 1980. 5. Schima H, Tsangaris S, Zilla P, Kadletz M, Wolner E: Mechanical simulation of shear stresses on the walls of peripheral arteries. J Biomech 23: 845 851, 1990. 6. Sipkema P, Latham RD, Westerhof N, Rubal BJ, Slife DM: Isolated aorta setup for hemodynamic studies. Ann Biomed Eng 18: 491503, 1990. 7. Schima H, Baumgartner H, Spitaler F, Kuhn P, Wolner E: A modular mock circulation for hydromechanical studies on valves, stenosis, vascular grafts and cardiac assist devices. Int J Artif Organs 15: 417 421, 1992. 8. Garrison LA, Frangos JA, Geselowitz DB, Lamson TC, Tarbell JM: A new mock circulatory loop and its application to the study of chemical additive and aortic pressure effects on hemolysis in the Penn State electric ventricular assist device. Artif Organs 18: 397 407, 1994. 9. Helal MA, Watts KC, Marble AE: Hydrodynamic simulation of arterial networks which include compliant and rigid bypass grafts. J Biomech 27: 277287, 1994. 10. Vandenberghe S, Segers P, Meyns B, Verdonck P: Hydrodynamic characterization of ventricular assist devices. Int J Artif Organs 24: 470 477, 2001. 11. Knierbein B, Reul H, Eilers R, Lange M, Kaufmann R, Rau G: Compact loops of the systemic and pulmonary circulation for blood pump testing. Int J Artif Organs 15: 40 48, 1992. 12. Ferrari G, Delazzari C, Mimmo R, Ambrosi D, Tosto G: Mock circulatory system for in vitro reproduction of the left ventricle, the arterial tree and their interaction with a left-ventricular assist device. J Med Eng Technol 18: 8795, 1994. 13. Vermette P, Thibault J, Laroche G: A continuous and pulsatile ow circulation system for evaluation of cardiovascular devices. Artif Organs 22: 746 752, 1998. 14. Sharp MK, Dharmalingam RK: Development of a hydraulic model of the human systemic circulation. ASAIO J 45: 535540, 1999. 15. Schaub JD, Koenig SC, Schroeder MJ, Ewert DL, Drew GA, Swope RD: Development of a ow feedback pulse duplicator system with rhesus monkey arterial input impedance characteristics. ASAIO J 45: 334 338, 1999. 16. Ferrari G, Nicoleti A, DeLazzari C, et al: A physical model of the human systemic arterial tree. Int J Artif Organs 23: 647 657, 2000. 17. Baloa LA, Boston JR, Antaki JF: Elastance based control of a mock circulatory system. Ann Biomed Eng 29: 244 251, 2001. 18. Ferrari G, Kozarski M, DeLazzari C, et al: A hybrid (numericalphysical) model of the left ventricle. Int J Artif Organs 24: 456 462, 2001. 19. Ferrari G, DeLazzari CD, Kozarski M, et al: A hybrid mock circulatory system: Testing a prototype under physiologic and pathological conditions. ASAIO J 48: 487 494, 2002. 20. Rosenberg G, Phillips WM, Landis DL, Pierce WS: Design and evaluation of the Pennsylvania State University mock circulatory system. ASAIO J 4: 41 49, 1981. 21. Nichols WW, Conti CR, Walker WE, Milnor WR: Input impedance of the systemic circulation in man. Circ Res 40: 451 458, 1977. 22. Koenig SC, Shae M, Pearson A, et al: Intraoperative assessment of acute hemodynamic changes after partial left ventriculectomy. J Card Surg 14: 152156, 2000. 23. Pantalos GM, Hayes J, Khanwilkar P, Bearnson G, Olsen D: Left ventricular simulator for cardiovascular device testing. ASAIO J 42: 46, 1996. 24. Pantalos GM, Chiang B, Bishop D, et al: Development of smaller articial ventricles and valves made by vacuum forming. Int J Artif Organs 11: 373380, 1988. 25. Gillars KJ: Design and validation of a four-region hydraulic model

46

PANTALOS ET AL. of the human cardiovascular system (masters thesis). University of Utah, Salt Lake City, 2003. Woodruff SJ, Sharp MK, Pantalos GM: Improved compliance chamber design for the study of cardiac performance in microgravity. ASAIO J 43: 316 320, 1997. Drew GA, Koenig SC: Biomedical patient monitoring, data acquisition, and playback with LabVIEW. Chapter 2, in Martin HT, Martin ML (eds), LabVIEW for Automative, Telecommunications, Semiconductor, Biomedical, and Other Applications. Upper Saddle River, NJ: Prentice Hall PTR, pp. 9298, 2000. Schroeder MJ, Perrault B, Ewert DL, Koenig SC: HEART: an automated beat-to-beat cardiovascular analysis package using Matlab. Comput Biol Med (in press). Koenig SC, Schroeder MJ, Wheeler B, Jorgenson J, Ewert DL: Uncertainty analysis of measurement error in calculating vascular input impedance: How many harmonics can be resolved? 2nd Joint EMBSBMES Conference, Houston, TX, October 23 26, 2002. Stergiopulos N, Westerhof BE, Westerhof N: Total arterial inertance as the fourth element of the Windkessel model. Am J Physiol 276: H81H88, 1999. Essler S, Schroeder MJ, Phaniraj V, Koenig SC, Latham RD, Ewert DL: Single beat estimation of arterial vascular parameters for transient and steady beats with application to hemodynamic state under variant gravitational conditions. Ann Biomed Eng 27: 486 97, 1999. Nichols WW, ORourke MF: McDonalds Blood Flow in Arteries: Theoretical, Experimental and Clinical Principles, Fourth edition. Oxford: Oxford University Press, 1998. Guyton AC: Textbook of Medical Physiology, 10th ed. Philadelphia: WB Saunders, 2000. Pantalos GM, Kim C, Robison P, Everett S, Olsen DB: Characterization of natural and total articial heart acceleration. ASAIO Trans 35: 235238, 1989. Sagawa K, Maughan WL, Suga H, Sunagawa K: Cardiac Contraction and the PressureVolume Relationship. Oxford: Oxford University Press, 1998. Levin HR, Oz MC, Chen JM, Packer M, Rose EA, Burkhoff D: Reversal of chronic ventricular dilation in patients with endstage cardiomyopathy by prolonged mechanical unloading. Circulation 91: 27172710, 1995. Burkhoff D, Holmes JW, Madigan J, Barbone A, Oz MC: Left ventricular assist device-induced reverse ventricular remodeling. Prog Cardiovasc Dis 43: 19 26, 2000. Madigan JD, Barbone A, Choudhri AF, et al: Time course of reverse modeling of the left ventricle during support with a left ventricular assist device. J Thorac Cardiovasc Surg 121: 902 908, 2001. Pepine CJ, Nichols WW, Conti CR: Aortic input impedance in heart failure. Circulation 58: 460 465, 1978. Nichols WW, Pepine CJ, Conti CR: Vascular load dened by the aortic input impedance spectrum. Federation Proceedings 39: 196 201, 1980. Nichols WW, Pepine CJ: Ventricular/vascular interaction in health and heart failure. Compr Ther 18: 1219, 1992.

26.

33. 34. 35. 36.

27.

28.

29.

37. 38.

30.

31.

39. 40. 41.

32.