Protocol for Relative Hydrodynamic Assessment of Tri-leaflet Polymer Valves

Heart valve disease often results from degenerative valve calcification¹, rheumatic fever², endocarditis^3,4 or congenital birth defects. When valve damage occurs, causing stenosis and/or regurgitation valve prolapse and cannot be surgically repaired, the native valve is usually replaced by a prosthetic valve. Currently available options include mechanical valves (cage-ball valves, tilting disk valves, etc.), homograft, and bioprosthetic valves (porcine and bovine valves). Mechanical valves are often recommended for younger patients based on their durability; however the patient is required to remain on anticoagulant therapy to prevent thrombotic complications⁵. Homograft and biological prosthetic valves have been effective choices to avoid blood thinner therapy; however, these valves have elevated risk for fibrosis, calcification, degeneration, and immunogenic complications leading to valve failure⁶. Tissue-engineered valves are being investigated as an emerging technology^7-9, but much still remains to be uncovered. Alternative durable, biocompatible, prosthetic valves are needed to improve the quality of life of the heart valve disease patients. Again, this valve design could replace the bioprosthesis used in transcatheter valve technology, with transcatheter approaches showing the potential for transforming the treatment of selected patients with heart valve disease¹⁰.

As stated by current standards, a successful heart valve substitute should have the following performance characteristics: "1) allows forward flow with acceptably small mean pressure difference drop; 2) prevents retrograde flow with acceptably small regurgitation; 3) resists embolization; 4) resists hemolysis; 5) resists thrombus formation; 6) is biocompatible; 7) is compatible with in vivo diagnostic techniques; 8) is deliverable and implantable in the target population; 9) remains fixed once placed; 10) has an acceptable noise level; 11) has reproducible function; 12) maintains its functionality for a reasonable lifetime, consistent with its generic class; 13) maintains its functionality and sterility for a reasonable shelf life prior to implantation."¹¹. Some of the shortcomings of existing valve prostheses may potentially be overcome by a polymer valve. Biocompatible polymers have been considered top candidates based on biostability, anti-hydrolysis, anti-oxidation, and advantageous mechanical properties such as high strength and viscoelasticity. In particular, elastomeric polymers may provide material deformation resembling native valve dynamics. Elastomers can be tailored to mimic soft tissue properties, and they may be the only artificial materials available that are bio-tolerant and that can withstand the coupled, in vivo, fluid-induced, flexural and tensile stresses, yet, move in a manner resembling healthy, native valve motion. Moreover, elastomers can be mass-produced in a variety of sizes, stored with ease, are expected to be cost-effective devices and can be structurally augmented with fibrous reinforcement.

The concept of the use of polymer materials to assemble a tri-leaflet valve is not new and has been the subject of several research investigations over the last 50 years¹², which were abandoned largely due to limited valve durability. However, with the advent of novel manufacturing methodologies^13,14, the reinforcement of polymer materials^15,16 and potentially seamless integration of polymer valve substitutes with transcatheter valve technology, there has recently been a renewed interest and activity in developing polymer valves as a potentially viable alternative to currently available commercial valves. In this light, a protocol for enabling testing of these valves to assess hydrodynamic functionality is the first step in the evaluation process; yet commercially available pulse simulator systems generally do not come equipped to accommodate tri-leaflet valve designs and contain an annular spacing to insert commercially available heart valves (e.g. tilting disc, bi-leaflet mechanical heart valves). Secondly, polymer valves are an emerging technology whose hydrodynamics can only be assessed in a relative context. Even though native heart valve pressure and flow data is available, it is important to conduct testing of native aortic porcine valves, which are biologically similar to human valves, using the same pulsatile simulator that is used to evaluate the polymer valves so as to account for measurement differences that may be system dependent. Thus, the goal of this study was to demonstrate how a commercially available pulse simulator can be fitted with an assembly to accommodate tri-leaflet valve constructs and to systematically evaluate polymer valve hydrodynamic metrics in a relative context in comparison to mechanical and native porcine heart valve counterparts. In our case, novel tri-leaflet silicone polymer valves previously developed at the University of Florida ¹³ comprised the polymer valve group.

1. Preparation

1. Design and fabricate an assembly to accommodate a tri-leaflet valve geometry. This will at minimum include a valve holder to suture-in the valve leaflets and a tube to house the valve holder and surrounding accessories to secure the assembly onto the pulse duplicator system. In our case, we utilized a commercially available pulse duplicator system available from ViVitro Labs Inc. (Victoria, BC). Valve holder design as well as pre and post assembly configurations are depicted in Figure 1.
2. The entire loop will need to be primed prior to usage. This involves two steps: i) cleansing of the entire loop system using soap solution and water, including replacement of any degraded tubes prior to use and ii) calibration of instruments connected to the loop, namely the pump being used, the flow probe, and the pressure transducers (generally measured at atrial, aortic and ventricular locations). Calibration can initially be performed using 1% saline solution and should be repeated prior to using blood-analog glycerin solution.

2. Native Aortic Valve Dissection

1. Obtain 4 fresh pig hearts with the aorta intact from a USDA approved slaughter house (Institutional Animal Care and Use Committee (IACUC) approval may be required). In our case, our dissection protocol was approved by the IACUC at Florida International University (Protocol Approval Number: 11-020). Rinse the heart with deionized water and place it in a receptacle filled with the 1% antimycotic/antibiotic and sterile phosphate buffered saline (PBS) solution and transport on ice to the hydrodynamic testing laboratory.
2. Place hearts in a dissecting pan and carefully remove the pericardium. Position the heart such that ventral side is facing you. Visually inspect and identify the four chambers of the heart and locate the aortic arch on the intact aorta.
3. Separate the heart into two halves by cutting across horizontally at approximately 0.75 in below the annulus, i.e. the junction between the aorta and the left ventricle. Carefully isolate the intact aorta still attached to the left ventricular tissue segment.
4. Examine the aortic valve located in the aortic root, the region between the ascending aorta and the lower annulus, ensuring that there is no damage or any signs of calcification.
5. Split the aorta at ~1 in above the annulus and separate the left ventricular tissue segment below the annulus to isolate the aortic valve (Figure 2a).

3. Polymer and Native Valve Suturing Process

1. Place the heart valve inside the valve holder such that the base of each valve aligns with the base of the post holder. Secure the valve in place at each post temporarily with a paper clip, but be careful not to damage the commissures or the cusps.
2. Insert the suture in the needle. Begin suturing at the bottom of the valve holder by passing the needle through the first hole, from the outside to the inside such that the needle may be easily pulled from the bottom. In a looping fashion, start suturing the valve vertically up the posts of the valve holder.
3. Progress with suturing (Figure 2b) along the circumference of the holder and secure with additional suture around the tips of the holder posts. Paper clips (Figure 2c) can be removed when the valve is completely secured using sutures to the 3 posts and at the circumference of the valve holder (Figures 2d and 2e).

4. Hydrodynamic Evaluation 

Note: Actual protocol will vary depending on specific pulse duplicator system being used. All information caontained herein used the ViVitro Pulse Duplicator Sysytem (ViVitro Labs, Inc., Vancouver, BC).

1. Bi-leaflet valve
   1. Set heart rate of pulse duplicator system to 70 beats/min.
   2. Select a flow waveform to drive the pump (in the case of the ViVitro system the S35 waveform was chosen for all hydrodynamic tests). The specific waveform utilized in our experiments is illustrated by Lim et al. (2001)¹⁷.
   3. Turn on amplifier and piston pump. Warm up for 15 min.
   4. Place bi-leaflet valve (Figure 2f) in the aortic position.
   5. Smear vacuum grease on all junctions of the device where leaks could occur.
   6. Pour glycerin/saline liquid in the atrial compartment. Note that the pulsatile duplicator system runs on 2 L of liquid with: 35%/0.7 L glycerin and 65%/1.3 L of saline solution. The saline solution is prepared using common salt well-dissolved in deionized water at a concentration of 9 mg/ml (weight/volume).
   7. Turn on the flow transducer that has been placed in the aortic position.
   8. Calibrate the pump.
   9. Proceed with the flow transducer calibration followed by the pressure transducers. Similarly to the pump, simply follow the instructions given by the ViVitest software (ViVitro Labs Inc.) for each flow and pressure under the calibrate tab.
   10. Once calibration is complete, start the pump at a low rpm until the fluid fills the aortic compartment. Check for leaks. Use additional vacuum grease if necessary.
   11. Turn the two stop-cocks (aortic and ventricular transducers) to open position.
   12. Increase the rpm of the pump until the stroke volume reaches 80 ml/beat.
   13. Permit the system to run for 10 min until flow has stabilized. Flow stabilization can be verified by observing the flow and pressure waveforms displayed in the screen. Low to none variation between cycles is a good indicator of system stabilization.
   14. In the ViVitest software select acquire mode.
   15. Click on collect 10 cycles.
   16. From the analyze mode, click on table and save the file. Also save an image of the waveforms using the photo-snap option in ViVitest.
2. Native and Polymer valves
   1. For polymer and animal valves, follow the same steps 3.1.1 - 3.1.3 from the bi-leaflet valve instructions.
   2. Place the valve holder with the sutured valve inside the glass tube from the custom made assembly. Sandwich the tube with the top and bottom pieces and secure in-place with lateral screws and nuts.
   3. Place assembly between the aortic chamber and the original aortic valve holder.
   4. Continue with steps 3.1.5 - 3.1.16 from the bi-leaflet valve instructions.

5. Post Processing

1. Flow and Pressure Waveforms
   1. Average the data collected for each of the waveforms collected, i.e. aortic pressure (AP), ventricular pressure (VP), and flow rate (Q).
   2. For each group of valve (polymer, porcine native aortic valve and bi-leaflet), plot the corresponding AP, VP and Q versus time relationships on the same plot.
   3. For the AP, superimpose normal, native aortic valve¹⁸, and bi-leaflet prosthetic valve¹⁹ plots from the literature for validation purposes.
2. Hydrodynamic metrics
   1. For each valve tested, the following hydrodynamic metrics should be computed: a) Forward flow pressure drop and maximum transvalvular pressure (TVP), b) the aortic root mean square (RMS) forward flow rate, c) aortic forward flow, closing, leakage and total regurgitant volume, d) valve end orifice area (EOA), e) transaortic forward flow, closing, leakage and total energy losses.
      1. Forward flow pressure drop is computed from TVP readings and can be categorized into 3 time intervals, P: interval that starts and ends with 0 TVP, F: interval with forward flow and H: interval starting with 0 TVP and ending with 0 flow. Maximum TVP is the maximum pressure gradient recorded across the valve from the aortic and ventricular pressure readings.
      2. The RMS forward flow rate (Q_rms) provides a useful metric for quantifying the magnitude of forward flow rate as follows:
         
         Where 'n' is the total number of time points collected, 'Q_i' is the instantaneous flow rate measurement collected in order 'i'.
      3. The aortic forward, closing and leakage volumes are computed based on the following time intervals, Forward: beginning of forward flow through the valve (t_o), to the end of forward flow (t₁); Closing: from t₁ till the instance of valve closure (t₂); Leakage: from t₂ till the end of the cardiac cycle (t₃). Total regurgitant volume is simply the sum of closing and leakage volumes.
      4. The EOA based on blood properties can be computed for the 3 intervals, P, F and H from the mean TVP during each of these periods as²⁰:
         
      5. Energy losses are defined as follows²¹:
         

Representative flow and pressure waveforms are shown in Figures 3, 4 and 5. The plots were averaged over the sample size of valves tested for each group, which was, n = 5, 4, and 2 valves for polymer, native porcine and bi-leaflet groups, respectively. The mean hydrodynamic metrics and the standard error of the mean for these sample sizes are presented in Table 1.

Figure 1. (a) Schematic of the ViVitro pulse duplicator system showing the primary components that implement a Windkessel model for physiologically relevant flows (figure presented here with permission from ViVitro Systems, Inc, BC, Canada). (b) Rapid prototyped valve holder configuration to suture and secure silicone or native porcine valves in-place. (c) Modification of the ViVitro pulsatile loop to accommodate tri-leaflet valve constructs. Click here to view larger figure.

Figure 2. (a) Native porcine valve. (b) Top view of polymer valve leaflets. (c) Side view of polymer valve after suturing and securing in-place within valve-holder. (d) Saint Jude bi-leaflet mechanical valve. Click here to view larger figure.

Figure 3. Mean instantaneous flow rates of the 3 valves tested (n = 5, 4, and 2 valves for polymer, native porcine and bi-leaflet, respectively). Flow rate was measured using an electromagnetic flow meter connected to a noninvasive flow probe placed at the interface location of the ventricle and aortic chambers (see Figure 1a). Click here to view larger figure.

Figure 4. Mean instantaneous ventricular pressure of the 3 valves tested (n = 5, 4, and 2 valves for polymer, native porcine and bi-leaflet, respectively). Ventricular pressure was measured in the ventricle chamber using a micro-tip pressure transducer. Superimposed literature ventricular pressure values for native and bi-leaflet valves (Diameter: 29 mm) were obtained from¹⁸ and¹⁹, respectively. Click here to view larger figure.

Figure 5. Mean instantaneous aortic pressure of the 3 valves tested (n = 5, 4, and 2 valves for polymer, native porcine and bi-leaflet, respectively). Aortic pressure was measured just downstream from the aortic valve position using a micro-tip pressure transducer. Superimposed literature aortic pressure literature values for native and bi-leaflet (Diameter: 29 mm) valves were obtained from¹⁸ and¹⁹, respectively. Click here to view larger figure.

Table 1. Mean and Standard Error of the Mean (SEM) Hydrodynamic metrics computed for the heart valves tested (n = 5, 4, and 2 valves for polymer, native porcine and bi-leaflet, respectively). The following intervals should be noted: P: interval that starts and ends with 0 TVP, F: interval with forward flow and H: interval starting with 0 TVP and ending with 0 flow. Mean diameters of the valves were as follows: Polymer valve (n=5): 22 mm; Native porcine valve (n=4): 20 mm; bi-leaflet (n=2): 23 mm. Small sample size for bi-leaflet valve was due to limited samples available for research use; the two bi-leaflet valves tested were previously donated to the Biomedical Engineering Department at Florida International University by Saint Jude Medical (Saint Paul, MN).

In this study, we have demonstrated the utility of modifying a commercially available pulsatile duplicator unit to accommodate tri-leaflet valve geometries so that hydrodynamic testing of polymer and native porcine valves can be performed. Specifically in our case, the system modified was a ViVitro left heart and systemic simulator system (Figure 1a) controlled via the ViViTest data acquisition system (ViVitro Systems, Inc, Victoria, BC, Canada). However, the system is not unlike several in vitro, pulsatile flow loops which all utilize a two-component Windkessel model to mimic flow and pressure waveforms of relevance to the human circulation^22-25. These two-component Windkessel systems typically consist of a pulsatile pump, a compliance chamber that mimics the distensibility of the arteries, and a peripheral resistance controller that can be used to regulate the vascular resistance. The equation that describes the two-component Windkessel model is:

where C is the compliance, R the resistance, Q(t) is the volumetric flow rate as a function of time and P is the arterial pressure (i.e. either in the pulmonary artery or aorta). In this context, we believe that a similar modification can be made to accommodate tri-leaflet valves in other pulsatile simulators as well. Specifically in our case, to house a tri-leaflet valve structure in the aortic valve location, an assembly primarily of acrylic plastic (Plexiglass) casing that contained a rapid prototype valve holder and sutured tri-leaflet valve (Figures 1b and 1c) could be easily integrated and removed from the primary ViVitro system. Hydrodynamic testing was subsequently conducted similar to other studies performed by Baldwin et al.²⁶ and Wang et al.²⁵ Instantaneous flow rate was measured using an electromagnetic flowmeter system (Figure 3). Real-time measurement of pressure was recorded at the ventricular and conduit location using microtip transducers at a set heart rate of 70 beats/min (Figures 4 and 5). The testing fluid was a blood-analog liquid, comprising deionized water to glycerin in a 65% to 35% ratio and 9 g/L of NaCl, mimicking blood viscosity (~3.3 cP).

We initially tested a mechanical bi-leaflet valve and the obtained mean pressure wave forms were compared to literature values¹⁹. Some ventricular pressure variability was observed possibly owing to different pump mechanisms in place to drive fluid flow as well as geometry and specific settings of different pulse duplicator systems such as size of the ventricular chamber, specific valve mimicking the mitral valve location, heart rate chosen, physiological flow waveform selected, etc. On the other hand, the aortic waveforms were found to be very similar and system-independent. This exercise was repeated for native porcine valves and again, larger variability in ventricular pressure was observed when comparing our results to the literature¹⁸. However, it is important to note that within our system, instantaneous flow rates as well as both ventricular and aortic pressures were similar regardless of the valve that was tested, i.e. polymer and native with assembly or bi-leaflet without assembly. This exercise is important to perform because one needs to ensure that modifications to the duplicator system with an assembly do not considerably alter local flow and/or pressure conditions. Secondly, these results indicate that as a means of system validation, at minimum, comparable aortic pressures need to be derived across pulse duplicator platforms or the valve being tested. The interpretation of the hydrodynamic variables themselves is a matter of individual polymer valve design specifics. Standards such as ISO (International Organization for Standardization) 5840 used in the evaluation of cardiac valve prostheses can serve as a guide to assess various parameters associated with the polymer valve geometry, manufacturing and material properties. These parameters can be further optimized and hydrodynamic testing subsequently revisited to ensure that the standards needed for FDA submission are met.

For example, in our polymer valves, comparable energy losses and lower regurgitant volumes versus native and bi-leaflet valves suggested acceptable workloads on the left ventricle²¹ and efficient valve closure (Table 1). However, the closing dynamics resulted in a relatively higher polymer valve maximum TVP gradient (versus bi-leaflet valves), which in our case, warrants further mechanical evaluation of silicone material being used to fabricate the valves to ensure that the higher stress does not cause leaflet rupture, and that a sufficient factor of safety can be put in place. In conclusion, we have demonstrated that an assembly consisting of a housing unit, glass tube and a valve holder can be fabricated to accommodate tri-leaflet structures such as polymer valves which can be sutured in-position. Comparative flow and pressure waveforms across native, prosthetic and polymer valves that are being developed need to be obtained. Second, the pressure waveforms need to be validated with literature values. A limitation of our approach is that ventricular waveforms are pulse duplicator system specific and are likely to show differences; however aortic pressure waveforms should be comparable across platforms or valve being tested if sufficient valve functionality exists. A future direction of this work is to further optimize the polymer valve material, manufacturing process and geometry. Hydrodynamics tests will subsequently be repeated under identical conditions so as to determine if functional improvements are quantitatively observed by comparing the current and previous hydrodynamic metrics computed.