Aortic valve stenosis is one of the most common types of acquired valvular disease1 and is conventionally treated with aortic valve replacement during cardiopulmonary bypass with cardioplegic arrest of the heart. Recent advances in minimally invasive technologies have led to the development of valve prostheses that can be implanted in the aortic position without the need to open the chest or stop the heart.2–5 In the trans-femoral approach, the prosthetic valve is introduced via a catheter inserted in the femoral artery and passed in a retrograde fashion in the ascending aorta and through the aortic annulus. In a variation of this technique, the transapical approach, the heart is entered through the apex of the left ventricle. In both cases the native valve is left in place. The stented, prosthetic valve is mounted on a balloon. When the balloon is inflated, it dilates the native stenotic valve as the prosthetic valve is placed. Calcifications on the cusps or annulus of the native valve may contribute to the onset of complications such as particulate embolization or perivalvular leaks. Also, the presence of the native cusps may lead to suboptimal prosthesis size.
During traditional, open-heart aortic valve replacement, it is routine to remove the native valve and surrounding calcifications before the placement of an artificial aortic valve.6 This prevents later embolizations and allows for a larger artificial valve to be implanted. The same benefits would likely result from native valve removal during minimally invasive valve replacement, but this is currently not done. Aortic valve removal during such an off-pump, beating-heart procedure would likely cause a sudden, massive aortic regurgitation leading to ventricular volume overload and decreased cardiac output. To overcome this, we have designed a temporary valve apparatus that limits the deleterious effects of aortic regurgitation during such a procedure. The apparatus is catheter-based for introduction into the ascending aorta via a peripheral artery and could help maintain hemodynamic stability for the duration of off-pump aortic valve removal and percutaneous implantation of a stented valve. The objective of this study was to perform an initial evaluation of the catheter-based valve apparatus in an in vitro model to test its capability to temporarily replace the aortic valve and maintain stable hemodynamics.
Materials and Methods
Several versions of the parachute valve were built and evaluated. The preliminary versions consisted of an intra-aortic balloon pump of which part of the balloon was removed and artificial chordae were added. A next generation consisted of a separate, custom-made membrane and a catheter that were assembled together. Different catheters, membrane shapes and membrane thicknesses, and chordae installation methods were tried and compared in vitro (data not shown). The valve apparatus that was used for this study (Figure 1) consisted of a 22-AWG steel guide wire (the catheter), a conical monoleaflet valve, and a series of e-PTFE chordae (strings) to prevent the blood flow from causing eversion of the valve membrane. The shaped membrane was produced by manually dipping a conical glass mandril in a viscous polyurethane (PU) solution which was obtained by dissolving PU pellets (Tecoflex SG-85 A, Thermedics, Woburn, MA) in liquid tetrahydrofurane solvent. Three consecutive dips at room temperature with a curing time of 30 minutes after each dip yielded a thin and transparent membrane with a desired combination of flexibility and strength. The membrane was then mounted on the tip of a 22-AWG wire and fixed with a suture. To form the chordae, a 5-0 Gore-Tex suture was run four times through the circular edge of the PU membrane (at 3, 6, 9, and 12 o’clock) and the ends were tied together to form four loops as chordae. The final appearance was that of a parachute mounted on a central support. Another suture was then tied through the four loops to pull them to a central point and from there run along the catheter. This provided a means to regulate the tension in the chordae and allow more or less opening of the valve when forced to do so by the aortic backflow.
To test the parachute valve, a pulse duplicator (Figure 2) was manufactured, consisting of an actuator, an atrial and ventricular assembly, and a 2-element Windkessel afterload.7 The actuator was a geared electromotor (NSH-54RL, Bodine Electric Company, Chicago, IL) connected to a pump head with rolling diaphragm (S6F air cylinder, Bellofram, Burlington, MA) via a crankshaft mechanism. The crank consisted of a screw and nut combination with a coupler for the shaft and could be adjusted continuously to select the desired stroke volume. A speed regulator for the electromotor controlled the stroke frequency (the “heart rate”). The Bellofram pump head was valveless and had only one central outlet that was connected to the ventricular chamber. Hence, during operation, fluid (water) was displaced from the ventricular chamber to the actuator cylinder as the crank moved backward (mimicking diastole), and that same fluid volume was pushed back into the ventricular chamber during forward motion of the crank (mimicking systole).
The atrial and ventricular assembly consisted of two cubic chambers (with 10 cm inner dimensions) stacked on top of each other. The top chamber was open to atmosphere and represented the left atrium, while the bottom chamber (left ventricle) was completely closed, except for the connectors described below. The horizontal wall separating the two chambers contained a 23 mm Bjork-Shiley tilting disc valve to mimic the mitral valve. One sidewall of the ventricular chamber contained a 1/2 inch pipe connection for the actuator and a 20 Fr Tuohy-Borst adaptor for insertion of an endoscope. In the opposite sidewall and aligned with this adaptor was a 3/4 inch barbed connector for attachment of the aorta.
A compliant aorta was fabricated from the same material as the valve by dipping an aluminum rod in the PU resin. This cylindrical aorta (2.5 cm diameter, 20 cm long) connected the ventricular chamber to the compliance chamber. The compliance chamber was a closed cylinder (9 cm inner diameter, 20 cm high) in which the entrapped volume of air (combined with the elasticity of the aorta) determined the compliance of the afterload. The compliance chamber also contained a 20 F Tuohy-Borst adaptor for insertion of an endoscope from the distal viewpoint. This was aligned with the aortic connector. The outflow connector of the compliance chamber was orthogonal to the endoscope insertion port and was connected to the atrial chamber via 3/4 inch polyvinyl chloride (PVC) tubing and a ball valve. The ball valve mimicked the total vascular resistance and formed the resistor element of the 2-element Windkessel afterload. Such an afterload setup was previously tested and characterized and is capable of simulating a wide range of hemodynamic conditions and impedances that cover the pathophysiological range.8
Aortic flow was measured with a 20A Transonic flow probe and a T106 flow meter (Transonic Systems Inc., Ithaca, NY) around the aorta, proximally to the parachute valve. Luer connectors were installed in the ventricular and compliance chambers and TruWave pressure transducers (Edwards Lifesciences LLC., Irvine, CA) were used with a Spacelabs 510 patient monitor (Spacelabs Health care, Issaquah, WA) to acquire pressures at those locations. All data were digitized at 50 Hz with a DI-158U data acquisition system and Windaq software (DataQ Instruments, Akron, OH) A 5 mm diameter, rigid endoscope (model 1011, SST, Chicago, IL) was connected to a 3-chip endoscopic camera system (Model 888, Stryker Endoscopy, San Jose, CA) for continuous imaging of the valve function. Illumination was provided by a Q-5,000 light source (Stryker) and images were recorded continuously on DVD.
The “parachute” valve was first positioned in the compliance chamber where it did not influence hemodynamics. The actuator stroke volume was set at 100 mL and a heart rate of 60 beats/min. Pressures in the left ventricle and compliance chamber (arterial pressure) and flow through the PU aorta were acquired and are displayed in the left hand panels of Figure 3. Next, the balloon was moved retrograde into the PU aorta and the tension in the string holding the chordae was adjusted to obtain maximum flow and diastolic pressure. The hemodynamic data with the valve in position are shown in the right hand panels of Figure 3.
Ten consecutive beats of the control data (no valve) and the data with the parachute valve were analyzed and each beat was processed individually. Mean flows and pressures were calculated, and maximum and minimum pressures determined. For each beat, the positive and negative portions of the flow data were integrated over time to obtain the forward and regurgitant volumes, respectively. Control data and valve data were compared with a Wilcoxon non-parametric test with a significance level of 0.05.
Ventricular pressure, arterial pressure (in the compliance chamber) and flow through the PU aorta without the parachute valve are displayed in the left panels of Figure 3. The right panels of Figure 3 show the same measurements with the parachute valve placed in the PU aorta. Ten consecutive beats in each condition were further analyzed and each beat was processed individually. The results are presented in Table 1. Inserting the catheter-based parachute valve induced a marked improvement of simulated hemodynamics. Without the valve, the ventricular and arterial pressure waveforms were virtually the same. (Figure 3; can also be derived from the systolic and diastolic pressure values in Table 1). With the parachute valve in place, diastolic ventricular pressure was approximately 7 mm Hg, which was a consequence of the noncompliant ventricular assembly. The fluid in the ventricular and atrial chamber above the pressure transducer kept generating pressure even when there was no pumping action. Without an aortic valve, the diastolic arterial pressure dropped to the base level of the ventricular pressure. Since coronary blood flow is mostly diastolic, a very low diastolic pressure like that seen here, would rapidly cause the onset of myocardial ischemia in a clinical setting. Insertion of the parachute valve augmented diastolic arterial pressure to 38.3 mm Hg. Even though there was a significant change in diastolic pressure by inserting the parachute valve, these values were influenced by the properties of the afterload in accordance with cardiac-vascular coupling mechanisms. Total arterial compliance as determined from the parachute valve data and the decay time method9 approximated 0.9 mL/mm Hg. The total vascular resistance for the given data was 0.85 mm Hg·s/mL (14.2 wood units). A more compliant system would reduce the pulse pressure and could further augment the diastolic pressure.
For both conditions, the forward volume generated by ventricular ejection approximated the set piston stroke volume of 100 mL (Figure 3 and Table 1). Without an aortic valve, however, two-thirds of that volume was aspirated again in the ventricle, which means that only one-third of ventricular filling came from the atrium. This massive regurgitation was reflected in the low cardiac output of 2.08 L/min. The parachute valve significantly reduced the regurgitation volume, leading to an instantaneous increase in flow to 4.66 L/min. Since the forward volume was equal for both conditions, the pressure buildup upstream of the ball valve resistor should be equal during systole under both conditions. However, as a consequence of the parachute valve’s ability to reduce retrograde flow, the fluid remained under pressure in the compliance chamber during diastole. The pressure started at a higher level following systole; therefore, the same forward volume resulted in a higher systolic pressure in subsequent cycles. The systolic pressure was increased by 25 mm Hg by inserting the parachute valve (from 75.4 mm Hg to 100.2 mm Hg) and this also resulted in a higher mean arterial pressure (65.3 mm Hg with the parachute valve vs. 31.0 mm Hg without a valve).
The opening and closing of the valve is illustrated in Figure 4. When opened by a negative pressure gradient and the resulting retrograde flow, the valve assumed its original molded conical shape and resembled a deployed parachute that self-centered in the vessel. Strong tension in the chordae arose and could be monitored and regulated manually in the distal suture along the catheter. Regulating the tension in this suture gave control over the opening of the parachute and allowed for the parachute to adapt to a range of aortic diameters. During systole, the valve collapsed and the tension in the chordae fell to zero. Since the collapsed valve apparatus had free radial movement and was not forced in a central position, the valve membrane was pushed flat against the wall as seen in Figure 4. However, the data in Figure 3 suggest that this was a negligible obstruction because the systolic pressure gradient was comparable with and without the valve.
Currently there is no technology available to remove the native aortic valve and clean a calcified annulus during minimally invasive valve replacement surgery. However, such devices are under development (US Patent 7201761) and their clinical use can be expected in the future. These in vitro data support earlier in vivo data, demonstrating that valve removal in a beating heart causes massive regurgitation and drastically reduces cardiac output.10,11 Therefore, the application of a minimally invasive valve removal device needs to be paired with a means to prevent regurgitation while the native valve is being removed and a stented replacement valve is being placed. To our knowledge, our prototype parachute valve is the first device designed for this purpose and can function as an easy-to-place, temporary aortic valve.
In vivo evaluation of the parachute valve is currently ongoing in a series of off-pump porcine models. Initial results are promising and show agreement with the in vitro data described above. For optimal clinical use, two or three sizes of the parachute valve may be needed, combined with an echo based selection method. Therefore, future in vitro evaluation is planned to assess the relationship between valve effectiveness and the parachute to aorta diameter ratio by inserting porcine, ovine, and bovine aortas in the pulse duplicator. Further improvement in the mounting technique of the parachute membrane on the catheter is currently being investigated, as is a regulator for more accurately controlling the tension in the chordae.
A catheter-based parachute valve was tested in an in vitro cardiovascular simulator and its function as a temporary aortic valve was evaluated. We found that the parachute valve significantly improved hemodynamics compared with the situation where there is no aortic valve. We foresee that technological advances will make off pump removal of the native aortic valve possible and desired for improved valve replacement surgery. In such a scenario, insertion of a temporary valve in the ascending aorta can contribute to maintaining stable hemodynamics during surgery.
The authors thank the personnel of the Medical Devices Laboratory of the University of Pittsburgh for their assistance with the material selection and fabrication of the parachute valve. The authors are also grateful to Shannon Wyszomierski, PhD, for proof reading and editing the manuscript.
1. Iung B, Baron G, Butchart EG, et al
: A prospective survey of patients with valvular heart disease in Europe: The Euro Heart Survey on Valvular Heart Disease. Eur Heart J
24: 1231–1243, 2003.
2. Andersen HR, Knudsen LL, Hasenkam JM: Transluminal implantation of artificial heart valves. Description of a new expandable aortic valve and initial results with implantation by catheter technique in closed chest pigs. Eur Heart J
13: 704–708, 1992.
3. Kuklinski D, Attmann T, Weigang E, et al
: Future horizons in surgical aortic valve replacement: Lessons learned during the early stages of developing a transluminal implantation technique. ASAIO J
50: 364–368, 2004.
4. Moazami N, Bessler M, Argenziano M, et al
: Transluminal aortic valve placement. A feasibility study with a newly designed collapsible aortic valve. ASAIO J
42: M381–M385, 1996.
5. Lutter G, Kuklinski D, Berg G, et al
: Percutaneous aortic valve replacement: An experimental study. I. Studies on implantation. J Thorac Cardiovasc Surg
123: 768–776, 2002.
6. Carr JA, Savage EB: Aortic valve repair for aortic insufficiency in adults: A contemporary review and comparison with replacement techniques. Eur J Cardiothorac Surg
25: 6–15, 2004.
7. Westerhof N, Stergiopulos N: Models of the arterial tree. Stud Health Technol Inform
71: 65–77, 2000.
8. Vandenberghe S, Segers P, Meyns B, Verdonck P: Hydrodynamic characterisation of ventricular assist devices. Int J Artif Organs
24: 470–477, 2001.
9. Simon AC, Safar ME, Levenson JA, et al
: An evaluation of large arteries compliance in man. Am J Physiol
237: H550–H554, 1979.
10. Folts J, Rowe C: Coronary and hemodynamic effects of temporary acute aortic insufficiency in intact anesthetized dogs. Circ Res
35: 238–246, 1974.
11. Nakao S, Nagatomo T, Kiyonaga K, et al
: Influences of localized aortic valve damage on coronary artery blood flow in acute aortic regurgitation: An experimental study. Circulation
76: 201–207, 1987.