Lung transplantation currently is the only treatment option for patients with end-stage lung disease. However, the finite number of suitable cadaveric organs limits the potential benefit of this treatment to an epidemiologically inconsequential minority of patients.1 Because ventricular assistance and the total artificial heart seem promising as transplant alternatives to patients with advanced heart failure, artificial lungs may provide hope for those with chronic respiratory failure.
Gas-exchange membrane technology has advanced significantly, allowing devices with improved efficiency and durability. The optimal design of an artificial lung for long-term support should be small to allow implantation, capable of providing full gas exchange, and simple in design to allow long-term and ambulatory support. An oxygenator with extremely low blood flow resistance can be perfused with venous blood by the right ventricle.2–4 This straightforward strategy would obviate the need for a mechanical pump for blood flow, therefore reducing cost, avoiding cellular blood element damage, and eliminating the risk of pump failure. Total gas exchange support can be provided if the entire right ventricular output is diverted away from the failing native lungs into the artificial lung. This application is sustainable only if total blood flow into the device does not significantly alter cardiac afterload.5,6
Whereas resistance is the opposition to steady or constant flow, impedance is the opposition to pulsatile flow, and defines cardiac afterload.7 Impedance is determined by resistance, compliance, inertance, and the amplitude and timing of pulse wave reflections. Impedance is typically expressed as a function of harmonics. Zero harmonic impedance (Z0) is analogous to resistance, or opposition to steady flow, whereas the integer harmonics represent the opposition to pulsatile flow. In physiologic circulations, first harmonic impedance (Z1) has the same frequency as the cardiac cycle, and seems to be the most important, because the majority of blood flow occurs within that harmonic.8
Our laboratory has been testing an oxygenator designed to serve as a long-term artificial lung. The device has blood flow resistance that is similar to the normal pulmonary circulation, but the original prototypes were noncompliant. We sought to characterize the right ventricular load of pulmonary replacement with this artificial lung, and compare the load using that of a newer prototype with an added compliance element.
Materials and Methods
Anesthesia and Animal Handling
Animals received humane care in compliance with the National Society for Medical Research and the Institute of Laboratory Animal Resources. The University of Michigan Committee on the Use and Care of Animals, protocol #7239, approved all experiments. Thirteen adult male Suffolk sheep (64 ± 12 kg) underwent anesthetic induction with intravenous sodium thiopental (7 mg/kg) and subcutaneous buprenorphine hydrochloride (0.6 mg). An 11 Fr cuffed tracheostomy tube was inserted, and the animals were ventilated with a Narkomed 2 anesthesia machine (North American Drager, Telford, PA), maintaining anesthesia with 0.4–4% isoflurane. Respiratory rate and tidal volumes were adjusted to maintain arterial partial pressure of CO2 (paco2) between 35 and 45 mm Hg and peak inspiratory pressure < 30 cm H2O. Positive end expiratory pressure of 5 cm H2O was applied continuously.
The femoral artery was cannulated and connected to a fluid-coupled strain-gauge pressure transducer (Abbott Critical Care Systems, Chicago, IL). A left anterolateral thoracotomy was performed. The pulmonary artery (PA) was isolated from the pulmonary valve to the bifurcation. A side-mounted pressure sensing micromanometer (Millar Instruments, Inc, Houston, TX) was inserted into the proximal PA immediately distal to the valve, and the pressure (PPA) signal displayed continuously. An ultrasonic flow probe (Transonic 24A159; Transonic Systems, Inc, Ithaca, NY) was placed around the proximal PA to measure instantaneous PA flow (QPA) and cardiac output (CO). The micromanometer and flow probe were positioned such that QPA and PPA were measured at approximately the same location.
Inflow and outflow conduits for the artificial lung consisted of short 16-mm Hemashield grafts (Boston Scientific, Natick, MA) bonded to 5/8th” polyvinylchloride tubing. The animals were anticoagulated with 100 IU/kg intravenous sodium heparin and the artificial lung inflow conduit was anastomosed end-to-side to the mid PA. The outflow conduit was anastomosed to the left atrial appendage. The artificial lung (Michigan Critical Care Consultants, Inc, Ann Arbor, MI) is an oxygenator with a rated flow of 9 l/min and a bench steady flow resistance of approximately 1.0 Woods unit in water. The device was primed with 280 ml heparinized saline, connected to the conduits, and deaired. The original noncompliant prototype was used in seven animals, and a device with a 30 ml polyurethane compliant element was used in six animals. The compliance chamber was designed to minimize volume and resistance. The modification has a compliance of between 0.6 and 0.8 ml/mm Hg up to 10 ml of volume infused. The device remained in an extracorporeal position for the extent of the data collection. Although perfusing the artificial lung, oxygen sweep gas flow rate was adjusted to maintain the paco2 between 35 and 45 mm Hg.
Data were collected at a sampling frequency of 300 Hz through a six-channel circuit board and Labview software (National Instruments, Austin, TX). Two conditions were evaluated for each animal: 1) with the artificial lung conduits clamped and the distal pulmonary artery open (native circulation), and 2) with the artificial lung conduits open and the distal pulmonary artery clamped (pulmonary replacement). Data were collected for 30 continuous seconds after equilibration had been reached as determined by stable PPA, arterial and venous oxygen saturation, paco2, and CO.
Matlab software (Mathworks) performed fast-Fourier transformation on PPA and QPA, decomposing each signal into a series of sine waves at their respective harmonics. Impedance is calculated by dividing the amplitudes of PPA and QPA at each harmonic. Resistance is the impedance at the zero harmonic (Z0). Left atrial pressure was not subtracted. Characteristic impedance (Zc) is the impedance in the absence of pressure wave reflections, and can be estimated by the average impedance between 2 and 12 Hz. Signal noise generation was determined by measuring the oscillations in vitro during steady state. Harmonics in which pressure or flow amplitudes exceeded the in vitro determined noise magnitude were excluded from calculations. Instantaneous reflected pressure waves can be determined by the following formula:
where Pb is the instantaneous reflected wave.9 Data are expressed as means and standard deviations. Statistical analysis was performed using Student two-tailed t test.
Pulmonary replacement with the noncompliant original prototype artificial lung was tolerated by all animals in the acute setting without significant change in cardiac output (6.4 ± 1.7 vs 6.0 ± 1.2 l/min, p = 0.4 (Figure 1). There was also no change in heart rate or mean arterial blood pressure. The impedance spectrum in the setting of pulmonary replacement with the original prototype is depicted in Figure 2. Z0 was similar in the native circulation and in the setting of pulmonary replacement with the noncompliant artificial lung (3.0 ± 0.8 vs 3.2 ± 0.5 Woods units, p = 0.5). There was a significant increase in Z1 with pulmonary replacement (0.4 ± 0.1 vs 5.9 ± 1.5 Woods units, p < 0.01). There was also a significant increase in Zc (0.4 ± 0.1 vs 1.4 ± 1.0 Woods units, p = 0.05). The calculated instantaneous reflected pressure wave, Pb, was of higher amplitude and peaked during systole with pulmonary replacement, as compared with reflected waves in the native circulation, which are of low amplitude (3.5 ± 0.9 vs 14.2 ± 3.6 mm Hg, p < 0.01) and occur during diastole (Figure 3). The instantaneous pulmonary artery flow pattern appeared nonpulsatile during pulmonary replacement (Figure 4).
Pulmonary replacement using the modified artificial lung with the compliance chamber did not change the animals' heart rates or mean arterial blood pressures. Addition of the compliance element to the artificial lung did not alter Z0 (3.2 ± 0.5 vs 3.5 ± 0.6 Woods units, p = 0.4 (Figure 2); however, there was a significant reduction in Z1 (5.9 ± 1.5 vs 1.9 ± 0.4 Woods units, p < 0.01). Characteristic impedance was not altered (1.4 ± 1.0 vs 1.9 ± 0.6 Woods units, p = 0.3). Pulse wave reflections were of lower amplitude (9.6 ± 3.6 vs 14.2 ± 3.6 mm Hg, p = 0.04) and much of the peak occurred during diastole (Figure 3). Pulmonary artery flow pulsatility was restored with pulmonary replacement using the compliant artificial lung (Figure 4).
To serve as a pulmonary replacement, an artificial lung with adequate surface area must not significantly alter right ventricular afterload. Although cardiac afterload has traditionally been defined by resistance, impedance more accurately and completely describes the workload of the heart, and is dependent upon resistance, inertance, compliance, and the nature of pressure wave reflections. This study examined the cardiac load of an artificial lung designed with a steady flow resistance similar to the human pulmonary circulation.
As predicted by in vitro testing, Z0, or resistance in the rigid original prototype artificial lung, was similar to the native pulmonary circulation of the anesthetized and ventilated adult sheep. However, the impedance spectrum was vastly different, with elevations seen at every harmonic, particularly the first harmonic. Characteristic impedance is the impedance in the absence of pulse wave reflections, and is inversely proportional to compliance. Zc can be estimated as the high-frequency impedance, between 2 and 12 Hz in physiologic circulations. As expected, characteristic impedance increased by more than threefold with the noncompliant artificial lung.
Wave reflections in pulsatile systems occur with changes in outflow geometry, such as constrictions or bifurcations. The speed with which pressure waves travel is directly proportional to the stiffness of the vasculature.9 In the normal pulmonary circulation, pulse wave reflections are minimized with its large surface area and gentle tapering of blood vessels. In addition, the compliant vascular bed slows the speed of wave transmission. Reflection peaks occur during the diastolic phase of the cardiac cycle, avoiding direct opposition to forward flow. The geometry within the noncompliant original prototype artificial lung generated a large pulse wave reflection, and the stiff conduits resulted in rapid return during the systolic phase of the cardiac cycle, directly opposing cardiac ejection. As such, the ventricle is unable to completely eject. Without recoil of the elastic large vessels of the lungs, diastolic pressure rapidly falls, and the full ventricle continues to empty during diastole. This results in the relatively nonpulsatile flow pattern seen with pulmonary replacement using the noncompliant original prototype and the persistence of diastolic flow. Echocardiograms will be performed to confirm these physiologic findings.
The addition of the compliance element to the artificial lung slightly increased its resistance, but not to any considerable degree. The modified device did not provide enough compliance to the system to lower characteristic impedance; however, there was an important reduction in first harmonic impedance. The compliance chamber dampens the reflected pressure wave and delays its return to the heart until later in the cardiac cycle. The physiologic benefit can be seen with the restoration of pulsatile flow generated by the right ventricle.
The consequence of impedance elevations is unknown. There are few studies in which cardiac performance is evaluated as a function of impedance, while maintaining normal resistance. Piene and Sund,10 using a Langendorf preparation, demonstrated alterations in right ventricular output with changing compliance; however, these adaptations had not been reproduced in vivo. Although there was no evidence of right ventricular failure in these short-term experiments with open-chested animals, cardiac decompensation may occur gradually over time, or acutely with elevation of afterload associated with chest closure. Assessment of myocardial oxygen consumption may validate concerns of potential heart failure, and long-term implantations will be helpful in characterizing the nature and timing of cardiac dysfunction. Furthermore, this construct makes possible the in vivo characterization of right ventricular alterations with variable changes in compliance.
In conclusion, pulmonary replacement with a low-resistance artificial lung is tolerated by the right ventricle in this short-term model. However, the lack of compliance and large pulse wave reflections change the impedance spectrum and alter cardiac performance. The addition of a compliance element reduces the amplitude of pressure wave reflections, and to a certain degree normalizes ventricular function. Further refinements in artificial lung compliance will be instrumental in quantifying how much compliance is necessary to avoid late ventricular failure.
1. Arcasoy SM, Kotloff RM: Medical progress: Lung transplantation. N Engl J Med
340: 1081–1091, 1999.
2. Fazzalari FL, Montoya PJ, Bonnell MR, et al.: The development of an implantable artificial lung. ASAIO J
40: M728–M731, 1994.
3. Lynch WR, Montoya PJ, Brandt DO, et al.: Hemodynamic effect of a low resistance artificial lung in series with the native lungs of sheep. Ann Thor Surg
69: 351–356, 2000.
4. Haft JW, Montoya JP, Alnajjar O, et al.: An artificial lung reduces pulmonary input impedance and improves right ventricular efficiency in acute pulmonary hypertension. J Thorac Cardiovasc Surg
122: 1094–1100, 2001.
5. 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.
6. Haft JW, Bull JL, Rose R, et al.: Design of an Artificial lung compliance chamber for pulmonary replacement. ASAIO J
49: 35–40, 2003.
7. Milnor WR.: Pulsatile blood flow. N Engl J Med
287: 27–34, 1972.
8. Kussmaul WG, Altschuler JA, Matthai WH, Laskey WK: Heart failure: Right ventricular-vascular interaction in congestive heart failure. Importance of low frequency impedance. Circulation
88: 1010–1015, 1993.
9. Westerhof N, Sipkema P, Van Den Bos GC, Elzinga G: Forward and backward waves in the arterial system. Cardiovasc Res
6: 648–56, 1972.
Copyright © 2005 by the American Society for Artificial Internal Organs
10. Piene H, Sund T: Flow and power output of right ventricle facing load with variable input impedance. Am J Physiol
237: H125–H130, 1979.