Secondary Logo

Journal Logo

Effect of Artificial Lung Compliance on In Vivo Pulmonary System Hemodynamics

Sato, Hitoshi*; McGillicuddy, John W.*; Griffith, Grant W.*; Cosnowski, Amy M.*; Chambers, Sean D.; Hirschl, Ronald B.*; Bartlett, Robert H.*; Cook, Keith E.*

doi: 10.1097/01.mat.0000208961.67272.fa
Cardiac Assist
Free

This study examined the effect of artificial lung compliance (C) on pulmonary system (PS) impedance and right ventricular function during in-series attachment of the MC3 Biolung in adult sheep. Compliances, C, of 0–20 ml/mm Hg were tested at the Biolung inlet. Results indicate the PS 0th harmonic input impedance modulus was not affected by C. The PS first harmonic input impedance modulus (Z1) was 10.9 ± 3.2 mm Hg/(l/min) at C = 0 ml/mm Hg and minimized to 2.41 ± 0.79 mm Hg/(l/min) at C ≥ 0.5 ml/mm Hg. Cardiac output was 58% ± 10% of its pre-Biolung attachment, baseline value at C = 0 ml/mm Hg and was maximized to an average of 75% ± 11% at C ≥ 0.5 ml/mm Hg. The left ventricular lateral-to-anteroposterior axis length ratio, which decreases with leftward septal shift, increased with C from 0.52 ± 0.12 at C = 0 ml/mm Hg to 0.76 ± 0.06 at C = 5 ml/mm Hg (p < 0.05), but decreased slightly with C at C > 5 ml/mm Hg. Therefore, the ideal C for right ventricular function is at least 0.5 ml/mm Hg and may be as high as 5 ml/mm Hg to minimize septal shift.

From the *Departments of Surgery and Bioengineering, School of Medicine, University of Michigan, Ann Arbor, MI; and †MC3, Inc., Ann Arbor, MI.

Submitted for consideration September 2005; accepted for publication in revised form January 2006. This work was supported by a grant from the National Institutes of Health (RO1 HL69420). Robert Bartlett is part-owner of MC3, and Sean Chambers is an employee of MC3.

Reprint Requests: Dr. Keith E. Cook, PhD, University of Michigan, Dept. of Surgery & Biomedical Engineering, Medical Research Bldg, 200 Zina Pitcher Bldg, Ann Arbor, MI 48109.

Thoracic artificial lungs (TALs) are currently being developed as treatment for acute respiratory failure (ARF) or as a bridge to transplant for patients with chronic, end-stage lung disease. In the setting of ARF, TALs would be used for weeks, whereas for end-stage lung disease, TALs would be used for months. The performance of TALs can be significantly diminished by poor cardiac output (CO) during attachment. In particular, acute right ventricular (RV) dysfunction can occur during attachment of TALs in series with the pulmonary circulation. The primary cause is increased RV afterload, which has been shown to induce impaired perfusion of the RV free wall,1–3 increased myocardial oxygen demand,4 distension of the RV outflow tract,5 and septal shift.6 These effects, in turn, cause reduced RV function and cardiac output. Similar acute increases in afterload have been shown to cause RV dysfunction after heart transplantation.7,8

The primary determinant of RV afterload is the impedance of the pulmonary system. During in-series TAL attachment, blood must flow through the artificial lung and then the natural lung, and pulmonary system impedance increases. One cause of this increase is undoubtedly the increased system resistance, but insufficient TAL compliance also contributes. Adding compliance before a resistance has long been known to lower the impedance of a pulsatile flow system.9 The compliance absorbs a portion of the stroke volume during systole, lowering systolic flow rates through the artificial lung and spreading blood flow over the entire cardiac cycle. This reduces artificial lung impedance, peak pulmonary artery (PA) pressures, and the work required of the right ventricle. This effect has been proven for artificial lungs using mathematical models and in vitro and in vivo testing.10–13

No study has examined what level of compliance is ideal for RV function. This study, therefore, assessed the effect of a range of TAL compliances on RV function. The MC3 Biolung was attached in series with the pulmonary circulation of adult male sheep with an inlet variable-compliance chamber, and pulmonary circulation hemodynamics and RV function were assessed at each level of compliance. The goal was to determine the ideal compliance for the MC3 Biolung.

Back to Top | Article Outline

Materials and Methods

Thoracic Artificial Lung and Compliance Chamber

The TAL used in this study was the MC3 Biolung (MC3, Inc., Ann Arbor, MI). Blood enters the Biolung through a center channel into the middle of an elliptical fiber bundle, flows radially through the fiber bundle, and exits through dual, symmetrically placed outlets. The fiber bundle is composed of 300 μm OD, polypropylene, hollow fibers (X30 240, Membrana-Celgard, Charlotte, NC) and has a total surface area of 1.7 m2.

The variable-compliance chamber (Figure 1) used in this study was designed according to Woodruff et al.14 and consists of a rolling diaphragm fixed to a piston within a solid cylinder. An exchangeable coil spring sits upon the piston and controls the force necessary to expand the diaphragm and, thus, the chamber compliance, C. To vary the compliance level between 0.5 and 20 ml/mm Hg, the springs were changed. Methods for determining the spring required for each compliance have been published elsewhere.11 To create a situation in which C = 0, the chamber was replaced with the same length of 5/8-inch ID, 7/8-inch OD PVC tubing. The compliance chamber was attached to the inlet of the Biolung (Figure 2), and thus the circuit attached to the pulmonary circulation consisted of inlet and outlet conduits, the compliance chamber or noncompliant PVC tubing, and the Biolung.

Figure 1.

Figure 1.

Figure 2.

Figure 2.

Back to Top | Article Outline

Experimental Procedures

Seven male sheep weighing 60 ± 4.6 kg were used in this study. All sheep received humane care in compliance with the National Society for Medical Research and the Institute of Laboratory Animal Resources. Anesthesia was induced with an intravenous injection of 7 mg/kg sodium thiopental (Abbot Laboratories, North Chicago, IL) and an intramuscular injection of 0.6 mg buprenorphine hydrochloride (Buprenex Reckitt Benckiser Pharmaceuticals Inc., Richmond, VA). The sheep were intubated and mechanically ventilated using a Narkomed 2 ventilator (North American Dräger, Telford, PA), while maintaining continuous inhalational anesthesia using 0.5–5.0 vol% isoflurane (Abbot Laboratories). The ventilator was set initially at a tidal volume of ten ml/kg and a frequency of 12 breaths/min and was adjusted as needed to maintain the arterial pco2 (Paco2) between 35 and 45 mm Hg and peak inspiratory pressure < 30 cm H2O. Arterial and venous access was established by carotid and jugular catheterization using 72-inch high-pressure PVC tubing with a 0.110-inch OD and 0.065-inch ID (Abbott Critical Care Systems, North Chicago, IL). Both catheters were connected to fluid coupled pressure transducers (Abbot Critical Care Systems), and the resulting arterial and central venous pressure signals (PArt and PCV) were displayed continuously (Marquette Electronics, Milwaukee, WI). Arterial blood samples were collected as needed to ensure adequate ventilation.

A muscle-sparing, left anterolateral thoracotomy in the fourth intercostal space was performed, resecting the fourth rib for exposure. The pericardium was incised and the PA isolated between the pulmonic valve and the bifurcation. A side-mounted, pressure-sensing catheter (Millar Instruments, Houston, TX) was inserted into the proximal PA, immediately distal to the valve (Figure 3) with the resulting PA pressure (PPA) signal displayed continuously. An ultrasonic flow probe (Transonic 24AX, Ithaca, NY) was placed around the proximal PA, immediately distal to the pressure sensing catheter, in order to measure instantaneous PA flow (QPA) using a flow meter (Transonic TS420). Inflow and outflow conduits for the artificial lung consisted of 18 mm, low-porosity woven vascular grafts (Boston Scientific, Natick, MA) that were solvent bonded to 5/8-inch ID, PVC tubing. Just before graft attachment, the animal received 60 mg intravenous ketorolac (Cayman Chemical, Ann Arbor, MI) and was anticoagulated with 100 IU/kg intravenous sodium heparin (Baxter Healthcare Corporation, Deerfield, IL). The inflow conduit was anastomosed end-to-side to the proximal PA, distal to the flow probe and PA pressure catheter, and the outflow conduit was anastomosed to the distal PA (Figure 3). The compliance chamber was attached to the Biolung inlet. Both were primed with heparinized saline (1 unit/ml) and connected to the conduits such that the compliance chamber inlet was attached to the proximal PA graft, and the Biolung outlet was attached to the distal PA graft.

Figure 3.

Figure 3.

Before initiating flow through the Biolung, a baseline hemodynamic data set was taken twice with 5 minutes in between. This data set consisted of PArt, PCV, PPA, and QPA. Data were digitally acquired at a sampling frequency of 250 Hz via a 16-channel circuit board using Labview software (National Instruments, Austin, TX) and stored on a personal computer (Dell Computer Corporation, Round Rock, TX). In addition, the chest cavity was filled with heparinized saline, and a GE Vivid 5 with S3 probe was used to obtain an epicardial echocardiogram. The echocardiogram was obtained with a short axis view at papillary muscle level. End-diastolic left ventricular lateral (Ll) and anteroposterior (Lap) axis lengths (Figure 4) were measured to determine the baseline septal position.

Figure 4.

Figure 4.

Once air had been removed from the circuit, clamps on the inlet and the outlet conduits were removed and the segment of PA between the proximal and distal PA anastomoses was snared to divert the entire CO through the compliance chamber and Biolung. An ultrasonic flow probe (Transonic 14XL) was placed around the outflow conduit tubing and connected to a flow meter (Transonic TS410). A vacuum line was attached to the Biolung gas outlet in order to maintain a –20 mm Hg pressure at the gas inlet, and pure oxygen with 0.5–5% vaporized isoflurane was used as the sweep gas. Sweep gas flow rate was set equal to the Biolung blood flow rate initially and adjusted to maintain Paco2 between 35 and 45 mm Hg.

Nine different springs, providing static compliances ranging from 0.5 to 20 ml/mm Hg, were used in random order within the compliance chamber. As discussed previously, the chamber was replaced with an equivalent length of noncompliant 5/8-inch PVC tubing to provide a zero compliance control either at the beginning (n = 3) or ending (n = 4) of the experiment. Hemodynamic data and echocardiograms were acquired at each compliance level as during baseline measurements. Five minutes were allowed to pass after changing each spring type to permit for short-term hemodynamic changes. At the end of the study, the animal was killed with an intravenous injection of pentobarbital (Fatal-Plus, 90-150 mg/kg, Vortech Pharmaceuticals, Dearborn, MI).

Back to Top | Article Outline

Data Analysis

Pulsatility within the proximal PA and artificial lung outlet were calculated to determine the degree of flow pulse dampening caused by the compliance chamber. Pulsatility (P) was defined as:

in which Qmax is the maximum flow rate, Qmin is the minimum flow rate, and Qave is the average flow rate. Pulsatility change between the PA and Biolung outlet, ΔP, was thus calculated as Δ P = PPAPTAL, in which PPA and PTAL are the PA and TAL outlet pulsatilities, respectively.

Excel (Microsoft, Redmond, WA) was used to perform Fourier transforms on PPA and QPA to determine the pulmonary system input impedance. The Fourier series for PPA and QPA are represented mathematically as:

and

in which PPA,0 is the zeroth harmonic frequency modulus of PPA, equal to the average PA pressure; ωi is the ith harmonic frequency; PPA,i is the PPA amplitude or modulus at the ith harmonic frequency; φP,i is the ith harmonic frequency phase shift for PPA; QPA,0 is the zeroth harmonic frequency modulus of QPA, equal to the average PA flow rate or CO; φQ,i is the ith harmonic frequency phase shift for PA flow rate; and QPA,i is the amplitude or modulus of QPA at the ith harmonic frequency.

Only the sinusoids at the harmonic frequencies have significant amplitudes, and thus only these frequencies are important in the physical system. For physiologic flows, the harmonic frequencies are integer multiples of the heart rate, HR. Furthermore, in these flows, flow rate amplitudes decrease as the frequency of the harmonic increases.15 Therefore, our analysis focused on the zeroth harmonic frequency, or the steady component, and the first harmonic frequency, equal to HR. The zeroth and first harmonic input impedance moduli, Z0 and Z1, were calculated according to the formulas:

and

respectively. In addition, the impedance at the higher harmonics was examined by calculating the characteristic input impedance modulus, Zc, the average of the impedance moduli at harmonic frequencies between 2 and 15 Hz. Lastly, the effect of compliance on wave reflections was examined by calculating the reflection coefficient according to the equation:

Right ventricular output power, RVOP, was calculated to determine the RV workload using the formula:

in which T is the cardiac period.

To determine the pulmonic valve regurgitant fraction, the average PA flow, equal to QPA,0, and the average of all negative flow across the pulmonic valve, QPA,r, were determined. The regurgitant fraction (Fr) was calculated according to the equation:

The end-diastolic, left ventricular lateral-to-anteroposterior length ratio, Rl, was calculated to estimate the degree of septal shift according to the equation Rl= Ll/Lap, in which Ll and Lap are the left ventricular lateral and anteroposterior axis lengths, respectively. Both measurements were taken from the echocardiograms. A decrease in this ratio indicates leftward septal shift (Figure 4).

Back to Top | Article Outline

Statistical Analysis

All statistical analyses were performed using mixed model analysis within SPSS for Windows (SPSS, Chicago, IL). The experiment number was the subject variable; P, PPA, Z0, Z1, Zc, Rc, PArt, PCV, Fr, or Rl was used as the dependent variable; and compliance level was used as the repeated measure and fixed factor. Baseline data were considered to be a compliance level and included in the analysis. Pairwise, post hoc comparisons of all compliances were done using Bonferroni confidence level adjustment. Similar mixed models for CO (equal to QPA,0) and RVOP were performed using the percentage of the baseline value for both variables due to a high degree of variability in their baseline values from sheep to sheep. Thus, comparisons with baseline were excluded from the analysis. A p value < 0.05 was considered statistically significant. Data are expressed in bar graphs as mean ± SD.

Back to Top | Article Outline

Results

All of the sheep tolerated instrumentation and attachment of the artificial lung, and blood gas values were maintained within healthy ranges at all times. In one animal, impedance data were discarded because of an error in acquisition of real-time pressures; in another animal, echocardiography was not performed. Results showing the effect of compliance on hemodynamics are shown in Figures 5–13. In each Figure, the x-axis indicates the compliance proximal to the Biolung, with BL indicating baseline data before initiating flow through the artificial lung.

Figure 5.

Figure 5.

Figure 6.

Figure 6.

Figure 7.

Figure 7.

Figure 8.

Figure 8.

Figure 9.

Figure 9.

Figure 10.

Figure 10.

Figure 11.

Figure 11.

Figure 12.

Figure 12.

Figure 13.

Figure 13.

Back to Top | Article Outline

Compliance Chamber Function

Figure 5 shows the pulsatility change, ΔP, between the inlet to the compliance chamber and the Biolung outlet at all compliances. Pulsatility change during Biolung attachment increased with compliance, indicating that the compliance chamber was dampening flow as intended. Pulsatility change increased sharply from 0.29 ± 0.59 at C = 0 to 2.5 ± 0.88 at C = 2 ml/mm Hg, after which the pulsatility change increased only slightly. The change in pulsatility increased significantly with compliance up to C = 2 ml/mm Hg (p < 0.001), after which there was no significant change. Therefore, pulsatility change during Biolung use was statistically maximized at C = 2 ml/mm Hg.

Back to Top | Article Outline

Pulmonary System Hemodynamics

Figure 6 shows the mean PA pressure, PPA, at baseline and each compliance. The baseline PPA was 14.9 ± 2.4 mm Hg, significantly lower (p < 0.001) than with the Biolung at any compliance. There was no significant difference between different levels of compliance, however, with an average PPA for C = 0 to 20 ml/mm Hg of 30.8 ± 4.7 mm Hg.

Figure 7 shows results for the input impedance moduli, Z0 and Z1. The value of Z0 at all compliance levels was significantly greater than at baseline (p < 0.01), but there was no significant difference between the different compliances. Baseline Z0 was 2.87 ± 0.63 mm Hg/(l/min), whereas average Z0 from at C = 0 to C = 20 ml/mm Hg was 6.69 ± 1.14 mm Hg/(l/min). The Z1 was also significantly greater at each compliance when compared with the baseline value of 0.96 ± 0.33 mm Hg/(l/min) (p < 0.01). The Z1 decreased significantly as compliance increased to 0.5 ml/mm Hg (p < 10-6) but did not change significantly as compliance increased from 0.5 ml/mm Hg. Therefore, Z1 was minimized statistically at C = 0.5 ml/mm Hg. The maximum Z1 was 10.9 ± 3.20 mm Hg/(l/min) at C = 0, and the average Z1 for C ≥ 0.5 ml/mm Hg was 2.41 ± 0.79 mm Hg/(l/min).

Figure 8 shows results for Zc. Baseline Zc averaged 2.30 ± 1.07 mm Hg/(l/min). Values of Zc at C = 0, 1, 2, 3, and 4 ml/mm Hg were significantly greater than at baseline (p < 0.05), whereas Zc at C = 0.5, 5, 10, 15, and 20 ml/mm Hg were higher than baseline but not significantly different (p = 0.1 to 1.0). During TAL attachment, there were no significant differences between the Zc at any of the different compliances (p = 1.0 in all cases). Figure 9 shows results for Rc. Values of Rc did not vary significantly between any of the conditions tested.

Figure 10 and Figure 11 show CO and RVOP, respectively, as a percentage of baseline. Cardiac output was smaller at all C compared with baseline. The CO increased significantly from 58 ± 10% to 72 ± 8% as C increased from zero to 0.5 ml/mm Hg (p < 0.05). The CO was unaffected by further increases in C and averaged 75 ± 11% at C ≥ 0.5 ml/mm Hg. Unlike CO, the RVOP is larger at all C compared with baseline, but it follows a similar trend with increasing C. The RVOP increased significantly from 107 ± 37% to 151 ± 34% as C increased from 0.0 to 0.5 ml/mm Hg (p < 0.05). The RVOP was unaffected by further increases in C and averaged 150 ± 48% at C ≥ 0.5 ml/mm Hg.

Back to Top | Article Outline

Arterial System Hemodynamics

Baseline PArt was 82.0 ± 13.3 mm Hg, significantly greater (p < 0.01) than with the Biolung at any compliance. There was no significant difference between different levels of compliance, however, with an average PArt for C = 0 to 20 ml/mm Hg of 59.7 ± 13.9 mm Hg. Baseline PCV was 12.7 ± 8.4 mm Hg, significantly smaller (p < 0.01) than with the Biolung at any compliance. There was also no significant difference between different C, with a markedly larger average PCV for C = 0 to 20 ml/mm Hg of 27.6 ± 4.2 mm Hg.

Back to Top | Article Outline

Septal Shift and Pulmonic Valve Function

Figure 12 shows the left ventricular lateral-to-anteroposterior length ratio, Rl. The Rl fell significantly at all C compared with the baseline value of 0.94 ± 0.06 (p < 0.01), indicating a leftward septal shift during Biolung attachment. The Rl increased gradually with increasing C from 0.52 ± 0.12 at C = 0 ml/mm Hg to 0.76 ± 0.06 at C = 5 ml/mm Hg, but gradually decreased again at C ≥ 10 ml/mm Hg. The Rl at C = 5 ml/mm Hg was significantly larger than at C = 0, 0.5, and 15 ml/mm Hg (p < 0.05), indicating that septal shift is minimized at the intermediate compliances. Figure 13 shows an example of end-diastolic echocardiograms at baseline (a), C = 0 (b), C = 5 (c), and C = 20 ml/mm Hg (d), indicating a leftward septal shift at all compliances and a reduction in septal shift at C = 5 ml/mm Hg.

Figure 14 shows the pulmonic valve regurgitant fraction, Fr, at baseline and each compliance. Regurgitant fraction decreased from baseline at C = 0 and 0.5 ml/mm Hg and increased from baseline at all C ≥ 1 ml/mm Hg. Variability in the data was large, however, and only C = 20 shows a statistically significant difference from baseline (p < 0.001). Regurgitant fraction at C = 2 and 20 ml/mm Hg are also significantly greater than at C = 0 and 0.5 ml/mm Hg (p < 0.05), indicating a trend towards increasing regurgitation with increasing C. Figure 15 demonstrates an example of the RV outflow through the pulmonic valve at baseline (a) and when the Biolung is attached with C = 0 (b) and C = 10 ml/mm Hg (c). At baseline, flow through the pulmonic valve falls to nearly zero during diastole, indicating closure. In either case with the Biolung attached, the pulmonic valve never closes fully. When C = 0, flow through the valve is always forward, elongating ejection and reducing flow amplitude. When C = 20, ejection remained somewhat elongated but significant negative, regurgitant flow was present. In addition, the amplitude increased from those at lower compliances.

Figure 14.

Figure 14.

Figure 15.

Figure 15.

Back to Top | Article Outline

Discussion

The results of this study indicate the importance of incorporating inlet compliance chambers into artificial lungs. In-series implantation with full flow to the Biolung caused the CO to drop 58% from baseline when no compliance chamber was present. Inclusion of even 0.5 ml/mm Hg of compliance was able to increase the CO to 72% of baseline. The relative improvement in CO is due primarily to a decrease in the pulsatility of blood entering the artificial lung, which in turn reduces the first harmonic input impedance modulus, Z1. The pulsatility change across the artificial lung decreased only 0.29 when compliance was zero, but improved significantly to a decrease of 1.3 at C = 0.5 ml/mm Hg. This decrease in pulsatility decreased the peak blood velocity into the Biolung, reducing systolic pressures and Z1. The magnitude of Z1 decreased from 10.9 ± 3.2 mm Hg/(l/min) at C = 0 to approximately 3.99 ± 0.86 mm Hg/(l/min) at C = 0.5 ml/mm Hg.

These results are qualitatively similar to those of an in vitro study examining the effect of compliance on MC3 Biolung impedance.12 For similar flow rates (3–5 l/min) and HR (100 beats/min), that study indicated that Biolung Z1 decreased from approximately 6.1 mm Hg/(l/min) when C = 0 to 1.3 mm Hg/(l/min) at C = 2 ml/mm Hg and a minimum of approximately 0.8 mm Hg/(l/min) at C ≥ 5 ml/mm Hg. In vivoZ1 are larger in the current study because of the cumulative impedances of the artificial lung, natural lung, and the inlet and outlet anastomoses on the PA, but the trend toward decreasing impedance with increasing compliance remains. The Z1 was minimized at a smaller compliance in vivo, most likely due to the presence of the PA compliance and a reactive RV output. The compliant PA dampens the blood flow from the RV before it reaches the artificial lung, lowering the compliance required by the artificial lung itself to minimize Z1. The compliance that minimized Z1 was likely also reduced due to changes in the RV output waveform. During in vitro studies, pump output flow rate, HR, and systolic and diastolic periods are fixed. In vivo, however, the ejection period was elongated when compliance was small, decreasing RV outflow amplitude at any flow rate. This, in turn, further reduces the compliance requirements of the TAL.

The zeroth harmonic impedance, Z0, remained unchanged despite the inclusion of compliance. This result differs from that of in vitro experiments,12 during which Z0 decreased with compliance and increased with stroke volume when C < 1 ml/mm Hg. This decrease was attributed to reduction of minor losses within the Biolung. Minor losses are flow rate–dependent energy losses caused by recirculation at fluid flow expansions, contractions, and changes in direction. The resistance added to the system by these minor losses is linearly related to the instantaneous flow rate (R = kQ), and thus instantaneous pressure drop is dependent on the square of instantaneous flow rate (Δp = kQ2). Increases in pulsatility cause increases in peak flow rates and, accordingly, lead to increases in the average pressure and Z0 in the system. All TALs possess some degree of minor losses, and thus average resistance increases with pulsatility and peak flow rates into the device. The constant Z0 during in vivo testing indicates that PA compliance and elongation of ejection is enough to dampen flow such that the effects of minor losses on Z0 are negligible.

Average Z0 during Biolung attachment, 5.75 mm Hg/(l/min), was still significantly higher than the baseline value of 2.46 mm Hg/(l/min) because of cumulative impedances of the artificial lung, compliance chamber, natural lung, and the inlet and outlet anastomoses on the PA. The increased Z0 caused a 25% reduction in CO even if compliance was large and Z1 had been minimized. At the maximum Z1, CO was reduced 42% from baseline. Therefore, minimization of Z1 would only be expected to improve CO 17%. To further improve CO during in-series attachment, one must decrease the resistance of the device, compliance chamber, or anastomoses or divert less blood flow to the TAL.

The reduction in CO is likely due to a combination of previously described effects caused by elevated PA and RV pressures. These include increased RV myocardial oxygen consumption; decreased myocardial perfusion1–3; pulmonic valve dysfunction due to distension of the RV outflow tract and PA,6,15 and contractile dysfunction due to septal shift and RV free-wall distension.5 Myocardial energetics were not studied explicitly in this study, but increased RV pressures are known to increase oxygen demand and decrease oxygen supply.1–4 End-diastolic leftward septal shift was proven in this study to occur at low compliances and decrease with increasing compliance when C < 5. This shift decreases the myocardial fiber length in the septum, decreasing the strength of the contraction. Surprisingly, this trend reversed itself when C > 5, with septal shift increasing slightly at larger compliances. Neither impedance moduli nor wave reflections explain this phenomena, as Z0, Z1, Zc, and Rc are effectively constant at C > 5 ml/mm Hg. One possibility may be that excessive compliance has negative effects on the timing of RV outflow and left ventricular inflow relative to systole, but further studies examining pulmonary vein flow would be necessary to examine this phenomenon.

This study also indicated that increasing compliance increased pulmonic valve regurgitation during Biolung attachment. Recent results of artificial lung implantation in pigs have shown that increased pulmonary resistance causes the pulmonic valve to become incompetent, potentially due to RV outflow tract distension.14 These results also indicated that the degree of regurgitation through the incompetent valve is dependent on the magnitude of end-ejection inertia change in the system.

Any change in inertia in a fluid mechanical system leads to an associated pressure change between the inlet and outlet: ΔPI= I(dQ/dt), in which I is the inertance and Q is the instantaneous flow rate. The inertance of a vessel is equal to ρL/A, in which ρ is the density of blood and L and A are the length and cross-sectional area of the vessel, respectively. In both the natural and artificial lungs, this pressure change is largest in the large, conducting channels where blood flow velocity, Q/A, is the largest. The sheep PA is approximately 3 inches long with a diameter of about 1 inch, and thus I = 1.6 g/cm4. The artificial lung tubing in this experiment was fairly long to accommodate attachment of the compliance chamber. Inflow and outflow grafts were composed of approximately 4 inches of 18 mm ID vascular graft with 4 inches of 5/8-inch ID tubing. An additional 4 inches of 5/8-inch ID tubing was used between the compliance chamber and artificial lung and 8 inches of 5/8-inch ID tubing between the artificial lung outlet and the outflow graft. The inertance before and after the compliance chamber were thus 9.6 g/cm4 and 20.4 g/cm4, respectively.

Normal pulmonary system inertance is thus low, and inertia plays only a minor role in determining pulmonary pressures during normal pulmonary flow. For example, if peak flow is approximately 20 l/min, minimum flow is zero, and the period of declining ejection is approximately 0.16 seconds, the average ΔPI= –2.5 mm Hg. However, during artificial lung attachment, the inertial pressure change across the pulmonary system is much larger and can significantly impact the pulmonary system. During artificial lung attachment without compliance, total I = 30 g/cm4. If cardiac output is decreased 40% due to high Z0, peak flow = 12 l/min and ΔPI= –28 mm Hg. In this case, there is a significant pressure gradient pulling blood from the RV and through the artificial lung. During end-ejection, flow rates through the pulmonary system are small and resistance has a reduced effect on PA pressure. Under these conditions, the inertial pressure drop in the PA dominates, eliminating any regurgitation through the incompetent pulmonic valve. If compliance is increased, however, dQ/dt decreases downstream from the compliance chamber and thus ΔPI becomes less negative. If compliance is large, dQ/dt downstream from the compliance chamber is negligible. In this case, only the upstream I, 9.6 g/cm4, effects the system. If peak flow is now 75% of baseline, 15 l/min, then ΔPI= –11 mm Hg. This pressure gradient remains significant, but is markedly less than when C = 0.

Thus, when the pulmonic valve is incompetent, increasing compliance in the system decreases the pressure gradient pulling fluid from the RV during end-ejection. This leads to increased regurgitation. This effect can be seen if one compares Figures 15a, b, and c. In both TAL attachment cases, the pulmonic valve never closes due to the high pulmonary system resistance. At C = 0, ΔPI is highly negative, fluid is pulled from the RV, the ejection period is elongated vs. baseline, and fluid flow through the pulmonic valve is always positive. At C = 20 ml/mm Hg, ΔPI is less negative, and there is marked regurgitation.

Clinically, the Biolung will possess less tubing than in the current study. The presence of the relatively large compliance chamber at the inlet requires outlet graft tubing to be approximately 4 inches longer, and the tubing flow probes and deep thorax of the sheep adds a few additional inches to the grafts. Clinical use, therefore, will have less inertia and greater regurgitation. The inefficiency from regurgitation that is added by increased compliance is small, however, in comparison to compliances beneficial effect on impedance, pulmonary system pressures, and CO.

In summary, these results suggest that a compliance of 0.5 ml/mm Hg is sufficient for in-series attachment in order to maximize CO but that a compliance of 5.0 ml/mm Hg allows for the most normal RV geometry. Current polyurethane compliance chambers used by MC3 with the Biolung have a pressure dependent compliance that is 0.80 ml/mm Hg at pressures ≤ 10 mm Hg, but decreases with increasing pressure to 0.45 ml/mm Hg at a pressure of 40 mm Hg.12 This compliance should be capable of maximizing CO under conditions similar to this study. Clinically, the amount of compliance required to maximize RV function will depend on patient hemodynamics (HR and CO) and attachment mode. Patients with low heart rates and high cardiac outputs will have higher stroke volumes and require more compliance. Lastly, patients implanted in parallel will require less or no compliance. Series, proximal PA to distal PA attachment was chosen for this study because it is the most strenuous on the heart and typically requires the most compliance. Parallel, PA to LA, attachment lowers the resistance of the pulmonary system and more effectively utilizes the compliance of the pulmonary circulation. In this setting, RV dysfunction is less likely, and less compliance should be required. In conclusion, a compliance > 0.5 ml/mm Hg is ideal for the MC3 Biolung in order to maximize CO. However, decreases in device Z0should be pursued in order to further maximize cardiac output.

Back to Top | Article Outline

Acknowledgments

The authors thank the Vascular Surgery Franchise at Boston Scientific for their generous donation of the Cooley Low Porosity Vascular Grafts and Peng Li, MD, PhD, for his assistance with the echocardiograms.

Back to Top | Article Outline

References

1.Brooks H, Kirk ES, Vokonas PS, et al: Performance of the right ventricle under stress: Relation to right coronary flow. J Clin Invest 50: 2176–2183, 1971.
2.Vlahakes GJ, Turley K, Hoffman JIE: The pathophysiology of failure in acute right ventricular hypertension: Hemodynamic and biochemical corrections. Circulation 63: 87–94, 1981.
3.Schulman DS, Biondi JW, Zohgbi S, et al: Coronary flow limits right ventricular performance during positive end-expiratory pressure. Am Rev Respir Dis 141: 1531–1537, 1990.
4.Elberry JR, Lucke JC, Feneley MP, et al: Mechanical determinants of myocardial oxygen consumption in conscious dogs. Am J Physiol 269: H609–H620, 1995.
5.Chuong CJ, Sacks MS, Templeton G, et al: Regional deformation and contractile function in canine right ventricular free wall. Am J Physiol 260: H1224–H1235, 1991.
6.Dong S, Smith ER, Tyberg JV: Changes in the radius of curvature of the ventricular septum at end diastole during pulmonary arterial and aortic constriction in the dog. Circulation 86: 1280–1290, 1992.
7.Kirklin JK, Naftel DC, Kirklin JW, et al: Pulmonary vascular resistance and the risk of heart transplantation. J Heart Transplant 7: 331–336, 1988.
8.Delgado JF, Gomez-Sanchez MA, Saenz de la Calzada C, et al: Impact of mild pulmonary hypertension on mortality and pulmonary artery pressure profile after heart transplantation. J Heart Lung Transplant 20: 942–948, 2001.
9.Piene H, Sund T: Flow and power output of right ventricle facing load with variable input impedance. Am J Physiol 237: H125–H130, 1979.
10.Boschetti F, Perlman CE, Cook KE, Mockros LF: Hemodynamic effects of attachment modes and device design of a thoracic artificial lung. ASAIO J 46: 42–48, 2000.
11.Haft JW, Bull JL, Rose R, et al: Design of an artificial lung compliance chamber for pulmonary replacement. ASAIO J 49: 35–40, 2003.
12.McGillicuddy JW, Chambers SD, Galligan DT, et al: In vitro, fluid mechanical effects of thoracic artificial lung compliance. ASAIO J 52: 789–794, 2005.
13.Lick SD, Zwischenberger JB, Wang D, et al: Improved right heart function with a compliant inflow artificial lung in series with the pulmonary circulation. Ann Thorac Surg 72: 899–904, 2001.
14.Woodruff SJ, Sharp MK, Pantalos GM: Compact compliance chamber design for the study of cardiac performance in microgravity. ASAIO J 43: 316–320, 1997.
15.Cook KE, Perlman CE, Seipelt R, et al: Pulmonic valve function during artificial lung attachment. In: BMES Annual Conference Proceedings, 2003. Landover, MD: BMES.
Copyright © 2006 by the American Society for Artificial Internal Organs