Performance changes of venous valves following tissue treatment with novel in vitro system

Abstract

Objectives

The purpose of this study is to test venous valve performance and identify differences between native tissue and replacement devices developed with traditional tissue treatment methods using a new in vitro model with synchronized hemodynamic parameters and high-speed valve image acquisition.

Methods

An in vitro model mimicking the venous circulation to test valve performance was developed using hydrostatic pressure driven flow. Fresh and glutaraldehyde-treated vein segments were placed in the setup and opening/closing of the valves was captured by a high-speed camera. Hemodynamic data were obtained using synchronized hardware and virtual instrumentation.

Results

Geometric orifice area and opening/closing time of the valves was evaluated at the same hemodynamic conditions. A reduction in geometric orifice area of 27.2 ± 14.8% (p < 0.05) was observed following glutaraldehyde fixation. No significant difference in opening/closing time following chemical fixation was observed.

Conclusions

The developed in vitro model was shown to be an effective method for measuring the performance of venous valves. The observed decrease in geometric orifice area following glutaraldehyde treatment indicates a decrease in flow through the valve, demonstrating the consequences of traditional tissue treatment methods.

Keywords

Prosthetic venous valve venous system modeling performance testing chemical fixation venous valve patency

Introduction

Chronic venous diseases (CVD) are diseases of lower extremity veins mainly caused by venous valve failure, which is responsible for allowing retrograde flow of blood.^1,2 Deep veins with CVD are correlated with the development of obstructive blood clots referred to as deep vein thrombosis. Surgical treatment of venous valve disease includes direct valve repairs, vein transposition, and vein transplantation.³ Past studies have estimated that as low as 37% of transplantations and 17% of transpositions have positive long-term outcomes.^4,5 While transplantations and transpositions may be effective methods of repair in some patients, they require a high level of surgical skill. Another surgical method known as neovalve construction has been developed in which the surgeon creates a new valve by partially cutting away a layer of endothelium from the vessel wall.^6,7 This method has shown early success in patients, but no long-term study has been conducted to date. The overall high requirements and low success rate of the aforementioned procedures have led researchers to investigate alternative treatment methods such as the development of a prosthetic venous valve.⁴

Prosthetic venous valves have been developed from both synthetic materials and xenografts.^3,4,8–11 Synthetic materials have shown poor patency in previous studies, as the few studies that have been conducted did not show patency beyond a year; however, xenografts have shown patency up to two years.^12–14 Previous xenografts have shown calcification following implantation, stiffening the valves. For venous valves, this change might be disadvantageous for the performance of the valve.

In order to test venous valve performance, a need exists for a system that can obtain a detailed and comprehensive assessment of prosthetic as well as native valves. This system should allow for comparison of current designs, as well as to fully functioning native valves, in order to decide the optimal valve for implantation. The developed system should be an in vitro model optimized to provide a standardized testing environment for venous valve performance.

Adopting standards of testing of cardiac valves, the reliable and objective geometric version of the effective orifice area (EOA), called the geometric orifice area (GOA), was chosen as the primary parameter by which these valves would be evaluated.^15–18 Whereas EOA is the area of maximum flow across the valve and requires Doppler echo to measure, GOA is the measure of the projected lumen area when the valve is in the fully open position and can be measured with a camera.¹⁹ GOA is measured taking the flow and pressures across the valve into account in order to ensure normalization,²⁰ which also makes it necessary to acquire and analyze hemodynamic parameters. GOA is a good measure of valve performance as it measures to what extent flow is restricted across the open valve. GOA is also an easily repeatable measure that accurately evaluates the performance of the valve, thus being the ideal parameter for an in vitro testing system.

Multiple systems to measure the GOA have been developed in the past. One such system was developed to test the effects of glutaraldehyde on the performance of venous valves.^21–23 The system is a good model for the venous system, yet it does not provide synchronized measurements of flow, pressure, and images of the valve. Below is described a simplified system—one that can be duplicated consistently and that isolates the performance of the valves as they relate to their response flow and pressure. The isolated performance of the valves under flow would allow objective determination of which prosthetic valves would demonstrate sufficient competence, similar to healthy valves, during in vivo testing.

The purpose of this study is to: (1) determine changes in material properties when treating fresh valve leaflets with glutaraldehyde; (2) design a new, synchronized in vitro system for valve performance testing; and (3) evaluate valve performance in vitro for fresh and glutaraldehyde fixed valves. These objectives will aid in the development of a novel fixation method and device design to achieve optimal prosthetic valve performance relative to the natural state.

Methods

Ex vivo study

Venous valves extracted from bovine jugular veins were randomly divided into two groups: fresh (n = 6) and glutaraldehyde fixed (n = 9) valves. Valves fixed with glutaraldehyde were treated in 0.625% v/v glutaraldehyde solution for 4 h submerged under zero pressure while valves were still intact with the vein wall. Uniaxial tensile tests were performed on each valve using a 574LE Planar Biaxial machine (TestResources™, Shakopee, MN). Valves were mounted to the load cell using cardboard clamps with hooks²⁴ (Figure 1).The valves were submerged in 0.9% w/v saline at 37 degrees C and preconditioned at 10 cycles of 10% at a frequency of 1 Hz. After preconditioning, the valves underwent a tensile test in the radial direction at a strain rate of 30%/s to determine stress at a given strain. Valves were only tested as either fresh or glutaraldehyde fixed due to the necessity of having the valves still attached to the vein wall during the fixation process.

Figure 1.

Venous valves were attached to hooks on cardboard clamps and then mounted for testing. The white box surrounds the mounted venous valve. Uniaxial tensile tests were performed in the radial direction at a strain rate of 30%/s.

In vitro model

The in vitro model contains a flow circuit using the principles of gravitationally driven flow due to its simplicity and controllability. The system matches the physiological flows and pressures of the venous system and provides a low pressure, forward flow through the vein. Subsequently, irregular retrograde flow is induced originating from hydrostatic pressure against the venous valve, simulating approximate hydrostatic pressures of 60–70 mmHg that are similar to those experienced in the saphenous and popliteal veins.^25,26 Previous models utilize multiple pumps to simulate the flow elements in the venous system; however, due to the nature of venous blood flow, a system that uses a constant hydrostatic flow is an effective model of the venous system while isolating the performance of the valves as a single variable.^21–23 The system illustrated in Figure 2 consists of a flow loop with four reservoirs that are interconnected: a forward flow driving reservoir, a back reservoir, a retrograde flow reservoir used for closing venous valves, and an overflow reservoir to maintain steady-state flow throughout the system.

Figure 2.

A 2D overlay of the in vitro model. Forward flow originates from the forward flow reservoir, through the venous valve, and to the back-flow reservoir. To reverse the flow, the flow direction is changed towards the retrograde flow reservoir and the venous valve is closed under hydrostatic pressure.

The forward flow reservoir remains elevated compared to the back-flow reservoir to establish hydrostatic pressure driven flow between them. The venous valves are located between these two reservoirs. Additionally, the forward flow reservoir is connected to the retrograde flow reservoir, which sits higher than the forward reservoir. The forward reservoir is connected to both reservoirs through a solenoid controlled three-way valve. The default state of the solenoid valve is open to the back reservoir, ensuring the venous valve is in a default open position. When the venous valve needs to be activated to simulate a change in hydrostatic pressure, a relay signal is sent to the solenoid valve. This opens the solenoid valve connected to the retrograde flow reservoir, thus changing direction of the hydrostatic flow and closing the venous valve. Saline levels in each reservoir are maintained at equal levels by having all overflow directed into a single overflow reservoir, recycling all fluid back to the reservoirs. The conceptual design and completed flow circuit are shown in Figure 3.

Figure 3.

Completed flow loop design. The reservoirs have an overflow on sections on the back that maintain a uniform saline level in each reservoir. The overflow drains into the overflow reservoir where the fluid is recycled back to the main and upper reservoirs. Final modifications included both scissor lifts and multiple outlets on reservoirs. A fiberscope is connected to the camera.

The flows and pressures in the system can be calculated by using a combination of the continuity equation (Eq.1) and Bernoulli’s equation (Eq.2).^27,28 Bernoulli’s equation balances energy along streamlines of a fluid assuming no viscous loss. The flow due to hydrostatic pressure can be related to the height difference between the two reservoirs using Bernoulli’s equation in combination with the continuity equation. This flow can then be adjusted by raising or lowering the heights of each reservoir, each of which sits on a scissor lift for quick adjustment. Considering Poiseuille’s Law of Resistance (Eq.2), the flow rate of the system is calculated by multiplying the pressure drop, or change in reservoir height, by the hydrostatic resistance from tube diameters and fluid viscosity across the system.²⁷

Q = v A

(1)

P_{t} = P + \frac{1}{2} ρ v^{2} + ρ g h

(2)

Δ P = Q R

(3)

Reducing the total resistance across the system lowers the reservoir height differential required to generate a specific flow. In order to minimize resistance, the tubing diameter was maximized without making it larger than the tested vein. Also, the total tubing length was minimized while maintaining the necessary entrance length to ensure laminar flow. Diameter changes throughout the system were also minimized.

Within the circuit, each vein is placed in an acrylic clamp containing a gel mold that mimics in vivo soft tissue (Clear Ballistics, Ft. Smith, AR). Different-sized gel molds are created for use with different sized veins. The system uses an electromagnetic flow meter (Carolina Medical Electronics, King, NC) to measure the flow upstream of the venous valve. Physiologic flows for deep human veins are between 0.3 and 0.5 L/min², while the system has a range from 0.2 to 2.5 L/min with an accuracy and resolution of 0.01 L/min. Laminar flow is established with sufficient entrance length from a long, straight section of pipe following a non-straight section. Maintaining laminar flow eliminates any variability that may arise in valve performance due to the inconsistency of fluid dynamics caused by disturbed or turbulent flow. To achieve the necessary entrance length for laminar flow, the length of the piping can be generally estimated to be ten times the diameter of the tubing.

Pressure transducers (ADInstruments, Dunedin, NZ) are used to measure pressure data upstream and downstream of the venous valve. This allows the GOA to be quantified as a variable against the pressure differential across the valve. The pressure drop can then be used as another method to quantify the performance of the valves. The change in GOA of the valves is measured through the use of a high-speed Basler Ace 1280-60 camera (Basler, Ahrensburg, DE) that is placed downstream of the venous valve and records the opening and closing of the valve leaflets in concurrence with the pressure and flow readings. The camera is attached to a custom-designed rigid fiberscope (Advanced Inspection Technologies, Melbourne, FL). Laminar flow is undisturbed since the scope is placed downstream of the valve. The distance between the scope and the valve can be adjusted by a syringe-inspired design in which the scope replaces the plunger of the syringe and can move back and forth along the imaging axis.

Flow, input and output pressures, and the high-speed video data are transmitted to a NI-cDaq 9188 CompactDAQ Ethernet chassis (National Instruments, Austin, TX) and then transferred into LabVIEW System Design Software (National Instruments), where it is synchronized and processed. The data points are sampled at 625 Hz, which was chosen from a list of preselected data acquisition rates in order to maximize the number of data points acquired without overflowing the buffer system, which is to be described. These data points remain in sync for transfer into LabVIEW during acquisition of the flow and pressure data and are recorded by specific data acquisition modules (National Instruments). The data points are stored within a buffer to later be sampled and logged. Before logging the data, each point must be aligned with the video captured from the high-speed camera that does not have an NI module to directly feed into the NI-cDaq 9188. To accomplish this, the camera trigger is hard-wired using a Programmable Functional Interface line from the NI-cDaq 9188. The camera is controlled with a trigger line from the NI-cDaq 9188. Each time the camera is triggered to capture a single frame, the data buffer of the associated time points of pressure and flow data is emptied. In contrast to the acquisition rate of the data points, the camera is triggered at 50 Hz giving a frame rate of 50 fps. For proper synchronization, the triggering of each video frame grabs the 12 previous data points in the data buffer, associating the data points with that particular frame. Additionally, this ensures there are enough data points stored to account for all data acquired during a single frame. The synchronized data are then passed through a custom low-pass filter to erase any signal noise affecting the data.

Fresh vs. glutaraldehyde treated valve performance testing

Five fresh veins were tested under the same hemodynamic conditions using the in vitro model, at both 0.5 L/min and 1.0 L/min while maintaining pressures from 60 to 70 mmHg both before and after the valve, and subsequently underwent chemical fixation submerged under zero-pressure static fixation with 0.17%v/v glutaraldehyde solution for 4 h. Following glutaraldehyde fixation, veins were again tested using the in vitro model under the same conditions as previously stated. Changes in valve performance were measured by calculating GOA using a custom-built script in MATLAB™. This script uses edge detection to automatically distinguish between valve and non-valve area. The edge detection can be done in one of two ways. If there is sufficient contrast between the pixel intensity of the valve and the background, a Sobel operator can be used at multiple points. However, this edge detection method can have errors due to venous valves being thin and often semi-transparent. In order to eliminate such errors, manual tracing of the valve is used to outline the GOA when the valve is fully open. The output obtained is GOA measurements in pixels, which is then converted to area in centimeters using reference measurements. These reference measurements were made by controlling the lumen diameter with the vein clamp and using ImageJ to calculate a pixel to distance ratio. GOA calculations were made by evaluating three consecutive frames at three separate time points. The nine frames, which are all at equivalent flow rates, are averaged together. This is done to ensure that fluttering of the valve leaflets does not impact the GOA measurement, something that was not accounted for in previous studies. The pressure measurements obtained are used to standardize the GOA measurements for each valve. Opening and closing times were calculated by counting the number of frames between the initiation of valve activation and when the valve becomes either fully open or closed.

Results

Ex vivo study

Following the completion of tensile testing, fresh valves were compared to glutaraldehyde fixed valves by calculating the average stress at each strain for both groups (Figure 4). The average stress/strain characteristics for fresh and glutaraldehyde fixed valves can be observed in Figure (4(a)). Using an unpaired t-test, a statistically significant increase in stress was found for glutaraldehyde treated valves at 15%, 20%, and 25% strain (p < 0.01, p < 0.05, p < 0.05, respectively) indicated by * in Figure 4(b).

Figure 4.

Uniaxial testing was performed for fresh (n = 6) and glutaraldehyde (n = 9) valves to determine stress at a given percent strain. (a) Average stress/strain characteristics for fresh and glutaraldehyde (Glut) fixed valves. (b) A significant increase in stress was found for glutaraldehyde treated valves at all three percent strains of interest (p < 0.05).

As venous valves must be highly elastic to maintain function, the observed increase in stress can have a detrimental impact on valve performance. This observation indicates that chemical fixation using glutaraldehyde as a fixative will result in a change in venous valve functionality. This hypothesis was tested by measuring the GOA of the valves before and after glutaraldehyde fixation, using the in vitro performance-testing mode.

In vitro study

The synchronization of pressures and flows of the in vitro model can be observed in Figure 5(a). The pressure differential at valve close was observed to be between 4 and 12 mmHg for all valves. The changes in GOA due to glutaraldehyde fixation are shown in Figure 5(b) with * indicating a significant different (p < 0.05).

Figure 5.

(a) Synchronization of pressure, flow, and video data. (b) GOA was calculated for fresh valves at 0.5 L/min and 1.0 L/min, which then underwent glutaraldehyde fixation (n = 5). A significant difference was observed at both flow rates (p < 0.05). (c) Opening and closing time was calculated for both fresh and glutaraldehyde fixed valves. No significant difference in opening or closing times was observed at either flow rate for either group. GOA: geometric orifice area.

A decrease in the measured GOA was observed at both 0.5 L/min and 1.0 L/min. The decrease in GOA at each flow rate was 27% and 28%, respectively. These changes were both found to be significant (p < 0.05) after performing paired t-tests between the fresh and glutaraldehyde treated valves at each flow rate. Opening time increased by 2.1% while closing time decreased by 14.6% (Figure 5(c)), neither of which were found to be statistically significant (p > 0.05).

Discussion

The in vitro results provided by the model can be used to determine which valve is the most effective in a controlled environment and parameters to influence this efficiency can be isolated and individually adjusted. The performance testing demonstrated a significant decrease in GOA at both 0.5 L/min and 1.0 L/min due to the increased stiffness and decreased elasticity of the valve leaflets. The results obtained aligned with those from previous in vitro models, proving its validity in providing an accurate and efficient performance evaluation of venous valves.^21–23 These results were consistent with the uniaxial testing of the valves, which showed that glutaraldehyde significantly increases the stress at a given strain for venous valve tissue.

The developed in vitro model was demonstrated as the first synchronized evaluation of the performance of isolated venous valves through the use of GOA measurements. As the acquisition rate of the high-speed camera differs from that of the pressure and flow data, the synchronization is imperative. This aspect was achieved through the use of hardware and virtual instrumentation. The system effectively measured the GOA, taking into account both pressure and flow, and also measured the closing time of the valves normalized to the pressure drop across the valve. Through synchronization, every video frame has an associated pressure and flow that can be used to ensure that conditions of each valve are similar, thus allowing objective comparison of the GOA data. The opening and closing time demonstrated no significant change following glutaraldehyde fixation. These results were obtained at 50 Hz; however, if a difference is found at a higher frequency, the difference in fluid amount compared to the GOA difference is negligible.

As xenografts have shown better long-term competence than synthetic materials, treating animal tissue is a necessity in prosthetic development. However, the results of this study demonstrate that glutaraldehyde is not likely to be a viable method. Thus, there is a vital need for a novel fixation method for xenografts that will not increase tissue stiffness to better maintain native tissue properties. By creating a new fixation method that does not alter the tissue’s mechanical properties and valve performance, a more successful xenograft may be developed that offers longer patency than past prosthetic venous valves. Future studies will focus on developing such a method and evaluating treated venous valves using the in vitro model presented.

Footnotes

Acknowledgements

The authors acknowledge Cockrum’s Meat Processing and Taxidermy (Rudy, AR) for their professional help and generosity in providing bovine jugular veins for this study. This work was presented at the 2017 American College of Phlebology Conference.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Marc Girardot is co-owner and CEO of BioMed Design LLC. Morten Jensen and Marc Girardot are co-owners of VeinTek LLC.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the Arkansas Research Alliance.

Ethical Approval

N/A

Guarantor

Contributorship

GE developed testing system, performed data collection, and created figures. MW wrote first draft of manuscript. HJ and KH provided clinical insight and background information for study basis. MG was involved with chemical fixation process. MJ conceived study and aided in testing system development. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

References

Eberhardt

Chronic venous insufficiency.

Circulation 2005; 111: 2398–2409.

Eklöf

et al . Updated terminology of chronic venous disorders: the VEIN-TERM transatlantic interdisciplinary consensus document. J Vasc Surg 2009; 49: 498–501.

Jan de Borst

Moll

Percutaneous venous valve designs for treatment of deep venous insufficiency.

J Endovasc Ther 2012; 19: 291–302.

Pavcnik

Chronic venous insufficiency and bioprosthetic bicuspid square stent based venous valve for transcatheter placement. Acta Clinica Croatia 2002; 41: 93–97.

Lehtola

Oinonen

Sugano

et al . Deep venous reconstructions: long-term outcome in patients with primary or post-thrombotic deep venous incompetence. Eur J Vasc Surg 2007; 35: 487–493.

Maleti

Lugli

Neovalve construction in postthrombotic syndrome.

J Vasc Surg 2006; 43: 794–799.

Lugli

Guerzoni

Garofalo

et al . Neovalve construction in deep venous incompetence. J Vasc Surg 2009; 49: 156–162.

Pavcnik

Uchida

Kaufman

et al . Percutaneous management of chronic deep venous reflux: review of experimental work and early clinical experience with bioprosthetic valve. Vasc Med 2008; 13: 75–84.

Rosenbloom

Schuler

Bishara

et al . Early experimental experience with a surgically created, totally autogenous venous valve: a preliminary report. J Vasc Surg 1988; 7: 642–646.

10.

Hill

Schmidt

Evancho

et al . Development of a prosthetic venous valve. J Biomed Mater Res 1985; 19: 827–832.

11.

Midha

Long-term patency of a polymer vein valve. Atlanta, GA: School of Bioengineering, Georgia Institute of Technology, 2009.

12.

Uflacker

Percutaneously introduced artificial venous valve: experimental use in pigs. In: The 1993 Annual Meeting of the Western Angiographic & Interventional Society, Portland, OR, 29 September–3 October, 1993.

13.

Midha

Design of a vein valve replacement.

J Vasc Surg 2012; 55: 303.

14.

Gomez-Jorge

Percutaneous deployment of a valved bovine jugular vein in the swine venous system: a potential treatment for venous insufficiency.

J Vasc Interv Radiol 2000; 11: 931–936.

15.

Tien

Chen

Berwick

et al . Role of sinus in prosthetic venous valve. Eur J Vasc Surg 2014; 48: 98–104.

16.

Cho

Prosthesis-patient mismatch after mitral valve replacement: comparison of different methods of effective orifice area calculation.

Yonsei Med J 2016; 57: 328–336.

17.

Gharaie

Morsi

A novel design of a polymeric aortic valve.

Int J Artif Organs 2015; 38: 259–270.

18.

Marquez

Comparative hydrodynamic evaluation of bioprosthetic heart valves. J Heart Valve Dis 2001; 10: 802–811.

19.

Garcia

Kadem

What do you mean by aortic valve area: geometric orifice area, effective orifice area, or Gorlin area?

J Heart Valve Dis 2006; 15: 601–608.

20.

Rittgers

Oberdiet

Pottala

Physiologically-based testing system for the mechanical characterization of prosthetic vein valves.

BioMed Eng OnLine 2007; 6: 1–12.

21.

Lee

Abolfathi

DeLaria

et al . In vitro testing of venous valves. Trans Am Soc Artif Intern Organs 1991; 37: 266–268.

22.

DeLaria

Phifer

Roy

et al . Hemodynamic evaluation of a bioprosthetic venous prosthesis. J Vasc Surg 1993; 18: 577–586.

23.

Nguyen

Nashef

Abolfathi

et al . Method for testing venous valves. United States: Baxter International Inc., assignee, 1994.

24.

Khoiy

KAR.

On the biaxial mechanical response of porcine tricuspid valve leaflets.

J Biomech Eng 2016; 138: 104504-1–104504-6.

25.

Arnoldi

Venous pressure in the leg of healthy human subjects at rest and during muscular exercise in the nearly erect position.

Acta Chir Scand 1965; 130: 570–583.

26.

Stick

Hiedl

Witzleb

Venous pressure in the saphenous vein near the ankle during changes in posture and exercise at different ambient temperatures.

Eur J Appl Physiol 1993; 66: 434.

27.

Currie

IG.

Fundamental mechanics of fluids. 4th ed. New York: CRC Press, 2013.

28.

Wilkes

JO.

Fluid mechanics for chemical engineers. 2nd ed. New Jersey, USA: Pearson Education Inc., 2006.