Flow past mechanical heart valves (MHV) in mechanical circulatory support devices, including total artificial hearts and ventricular assist devices (VADs), is primarily implicated in thromboembolism.1 The newer valve designs, although more hemodynamically efficient, pose a significant thromboembolic risk owing to nonphysiological flow conditions at which the elevated stresses and exposure times are sufficiently high to cause platelet activation and thrombus formation.2–5 Mitigation of this risk requires lifelong anticoagulation therapy,4,6 and less thrombogenic MHV designs should therefore be developed by device manufacturers.
Design-specific thrombogenic effects of MHV with emphasis on flow fields around the hinges have been investigated with laser velocimetry7 and computational fluid dynamics (CFD).8,9 Comparison between St. Jude Medical (SJM, St. Paul, MN) and CarboMedics (CM, Austin, TX) valve hinges revealed higher velocity and shear stress magnitudes in the CM valve.7 In another study, the open pivot design (ATS Medical, Minneapolis, MN) had superior hemodynamics and lower peak stresses than the recessed hinge design (SJM).8 The effect of leaflet-housing gaps in SJM valve on peak stresses was investigated numerically10 and experimentally.11 In these studies, optimized intermediate gap clearances (∼100 μm) were concluded to be the least thrombogenic based on reduced peak stresses and the expected reduced platelet activation, while having a minimal effect on regurgitant volumes. However, these studies did not correlate numerical or flow measurements with thrombogenic markers such as platelet activation to validate the reduced thrombogenicity of the particular valve designs.
Our group has extensively investigated platelet activation in devices within a left ventricular assist device (LVAD) setup12–14 and conducted advanced numerical simulations to elucidate the effect of specific flow phases on platelet activation.2,15,16 We developed a new methodology entitled Device Thrombogenicity Emulator (DTE) that interfaces advanced numerical simulations with experimental measurements.17 We recently applied DTE to compare the thrombogenic performance of two commercial bileaflet MHV—the SJM regent and the ATS AP—in forward and regurgitant flow phases.17 The main difference between these valves is the hinge region design. The valve design-dependent resultant platelet activity was found to be higher in the SJM valve, particularly during regurgitation, in agreement with our previous results using fluid structure interaction (FSI) approach.16 Motivated by these findings and noting that the ATS valve is less thrombogenic than the widely implanted SJM MHV, we apply the DTE methodology to further optimize the original ATS design.
The DTE is used for optimizing cardiovascular device thrombogenicity by optimizing its design in the virtual numerical domain and interfacing it to an experimental apparatus that emulates the flow conditions within the device.17 Device-specific probability density function (PDF) of stress accumulation (SA) of multiple trajectories is computed in the numerical domain to establish a “thrombogenic footprint” of the device. The DTE then uses stress-loading waveforms in localized high-stress regions, or “hot-spots,” extracted from the numerical simulations and emulates these flow conditions in a programmable hemodynamic shearing device (HSD) by which the resultant platelet activity is measured.17–19
The original ATS valve has an open pivot configuration with open and closed leaflet stops protruding into the flow field (Figure 1). We use the DTE to assess the thrombogenic effects of design changes in several ATS valve features—valve-housing gap clearances, valve opening angles, and hinge modification (Figure 1). A region-specific PDF was mapped and iteratively optimized. Platelet activation measurements were performed with an innovative prothrombinase assay20 by exposing platelets in the HSD to programmed stress-loading waveforms in pertinent flow trajectories that were extracted from the numerical simulations after each design change. A description of the methodology follows.
Numerical simulations were conducted for the forward (up to 300 msec from peak systole, with the valve in fully open position17) and the regurgitant flow phases. A two-phase Newtonian fluid was used to represent the blood, with platelets considered as neutrally buoyant solid spherical particles (3 μm diameter). The simulations were discretized within a range of 5–9 million finite volume cells and 15,000–30,000 platelets seeded upstream. The simulations were performed using high mesh density approaching direct numerical simulations (DNS)21 resolution, which enables to capture flow effects in the smallest confines of complex MHV geometries while liberating from the use of turbulence models (mesh resolution within the Kolmogorov scales in MHV flows: 20–70 μm).22,23 Grid independence studies were conducted using three different grid sizes, 2, 9, and 17 × 106 cells. The difference between 9 and 17 million elements was found to be <5%. Similar optimization for time step, dt, was established at 5 × 10−4 s and was used for all simulations.
Stress loading history of the seeded platelets was computed by incorporating the combined effect of stress and exposure time according to the well-established shear induced platelet activation (SIPA) concept.24 The stress tensor was extracted from the simulations along the corresponding platelet trajectories and rendered into a scalar stress value as previously described.2,25 The extremely refined DNS mesh used in this study can capture the smallest turbulent Kolmogorov scales. Thus, there is no need for turbulent modeling and inclusion of the approximated turbulent Reynolds stresses. Accordingly, only viscous stresses are present in the stress tensor and considered in the SA formulation. We further developed our formulation of the SA of a platelet (the linear product of stress, σ, and exposure time, texp) to include the stress-loading rate
26 according to the following equation:
where σi, i = 1, 2, . . ., N, is the nodal scalar value extracted from the total stress tensor described above and Δt the corresponding time step between successive nodal points.
PDF of SA
The PDF is representative of the overall thrombogenic potential of the valve17 or specific valve modification. To ensure that the SA is independent of number of seeded platelets (15,000–30,000) and spatiotemporal variations, we have used bootstrapping statistics16,17 to compare localized PDF of different valve designs. Those were calculated in respect to a sphere of interest (SOI) located at the middle of the hinge and with a radius r = 2.5 mm.17
Platelet Activation Measurement
Specific high-stress trajectories identified from the SOI were programmed into a HSD (Figure 2), a computer-controlled cone-plate-Couette viscometer capable of accurately emulating shear stress conditions found in devices, as previously described.18,19,27 Thirty milliliters of citrated blood was obtained from healthy adult volunteers in accordance with Stony Brook University Institutional Review Board regulations. Gel-filtered platelets were prepared and diluted (2 × 107/ml) in modified Tyrode's buffer.20 Viscosity was adjusted to 3.5–6.5 cP with the addition of high-molecular-weight Dextran (500 kDa, Sigma, St. Louis, MO) to achieve the desired peak shear stresses in stress-loading waveforms. Platelet activation was measured corresponding to 600 repeats of each stress-loading waveform with a modified prothrombinase assay20 at the beginning and end of the experiments (n = 5). The difference in activation was normalized to maximum activation determined by sonication17,27 and was reported for each trajectory. One-way analysis of variance was used to statistically infer activation differences between design changes and the control.
Design Changes of ATS Valve
In our previous studies, we compared the effects of the hinges of two MHV designs, ATS AP and the SJM Regent, on platelet activation.16,17 In this study, we explore design optimization changes in the ATS valve intended to achieve an improved thrombogenic performance.
Changes in Leaflet-Housing Gap Clearance.
In the original ATS valve, the gap clearance (distance between the leaflet and housing) near the hinge region is 38 μm. This distance was increased to 130 and 250 μm, correspondingly, to establish the contribution of gap clearance to the flow stresses experienced by platelets and find an optimal distance that would minimize platelet activation (Figure 1A).
Changes in Opening Angle.
According to the valve manufacturer, the maximum opening angle for the ATS valve is 85°. Experimentally measured values however are closer to 80°, indicating that the leaflets of the ATS valve do not always reach its fully opening angle.16,28 To investigate the role of the effective maximum opening angle on SA distribution and platelet activation, maximum opening angles of 80° and 85° were investigated (Figure 1B).
Hinge Region Modification.
The ATS valve has spherical convex pivots facilitating rotation of the leaflets. Three convex pivot stops on the inner circumference of the orifice ring on each side arrest the leaflets in open and closed positions. The geometry below the convex leaflet stops was modified, and a streamlined flow channel was carved along the edges of these stops (Figure 1C). Probability density function and platelet activation measurements for this modification were compared with the original valve design.
Velocity contour plots, PDF, and platelet activation measurements resulting from the design modifications measured in “hot spot” trajectories are presented in Figures 3–8. The velocity contour plots show regions of flow separation and recirculation during forward flow for the valve-housing gap clearance and opening angle cases. Strong jets emanating from the hinges were observed during regurgitant flow.
The secondary flow for the original and modified valves consisted of a pair of counter-rotating vortices emanating from the jet flow that is generated in the hinge regions. The formation of these vortices seems to entrain a significant number of platelets toward elevated shear stress regions.17 Elevated shear stresses particularly during regurgitant phase, experienced by platelets because of pathological flow conditions, play a major role in SA that may lead to platelet activation.
Leaflet-Housing Gap Clearance
Simulations were performed for three hinge gap clearances: 38, 130, and 250 μm. Stress accumulation was determined for all particle trajectories in the hinge SOI (PDF: Figure 3A, forward flow; Figure 3B, regurgitant flow). A significant shift in the mean SA to lower values was observed as the gap clearance was increased. The hinge SA in forward flow for the 250 μm gap clearance (9.8 dyne × s/cm2) was lower than the 38 μm (14.97 dyne × s/cm2) and 130 μm (9.98 dyne × s/cm2) gap clearances (p < 0.01). The velocity contour plots depict stronger jets in the smallest gap clearance. Shed vortices were observed near the leaflet tip during forward flow (Figure 3C). The hinge SA in regurgitant flow for the 130 μm gap clearance (80.9 dyne × s/cm2) was lower than the 38 μm (85.6 dyne × s/cm2).
Hot-spot platelet trajectories and their corresponding stress-loading waveforms (Figure 4, A and C) were programmed into the HSD for platelet activation measurements (Figure 4, B and D). In the forward flow, the activation was significantly higher (p < 0.05) for the 38 μm case compared with the two larger gap clearances (130 and 250 μm). The mean platelet activation for the three groups was calculated as follows: a) 0.0208 ± 0.0023 (38 μm), b) 0.0079 ± 0.0007 (130 μm), and c) 0.0063 ± 0.0005 (250 μm). The lowest platelet activation value was established for the two larger gaps but was not significantly lower for the largest one. In regurgitant flow, mean activation was significantly higher (p < 0.05) for the 38 μm case (0.0438 ± 0.0035) compared with the 130 μm gap clearance (0.0186 ± 0.0008).
To estimate the effect of increased flow volumes passing through the larger clearance gaps and the effect that this may have on the platelet activation potential, we have calculated the exact amount of platelets passing through the hinges for the three different gap clearances. We found out that 1,232 platelets passed from the hinges for the control gap clearance (38 μm), 1,354 platelets (10% increase) passed from the hinges for gap clearance of 130 μm, and 1,155 platelets (6.25% decrease) passed from the hinges for gap clearance of 250 μm. The number of platelets does not change drastically when gap clearance increases and actually decreases in the larger gap (250 μm). Moreover, the peak velocity at the hinge vicinity was much higher (higher SA could result in more platelet activation) for the smallest gap clearance (u38μm=1.59m/s) and decreases for the 130 μm gap clearance (u130μm=1.46m/s). The peak velocity at the hinge vicinity increases again for the 250-μm gap clearance (u250μm=1.48m/s). Although a larger volume of blood passes through the larger pivot gaps, our simulations indicate that it does not necessarily translate into a larger number of particles passing through these regions. In addition, the PDFs indicate that the larger volumes are compensated by localized velocity reduction.
Valve Opening Angle
Large areas of flow separations were observed for the 80° opening angle case (Figure 5B). This angle produced significantly higher mean values (p < 0.01) of platelet SA (15.16 dyne × s/cm2) than the 85° opening angle (14.97 dyne × s/cm2, Figure 5A). Maximum velocities were found in the central jet at the leaflets' leading edge and are in agreement with earlier studies.16,29
Hot-spot platelet trajectories (Figure 6A) and their corresponding stresses (Figure 6B) were programmed into the HSD for platelet activation measurements (Figure 6C). Significantly higher (p < 0.05) mean platelet activation was found for the 80° opening angle case (0.0593 ± 0.0041) compared with the 85° angle case (0.0056 ± 0.0006).
The hinge regions were modified to introduce a channel between the leaflet stops. The resulting channels are intended to increase the flow area near the hinge regions and to reduce the elevated flow stresses generated in the narrow gap clearance between the stops and the leaflets. The hinge SA (PDF: Figure 7A, forward flow; Figure 7B, regurgitant flow) for the modified valve (forward flow: 11.31 dyne × s/cm2; regurgitant flow: 76.7 dyne × s/cm2) was found to be significantly lower (p < 0.01) than the control case (forward flow: 14.97 dyne × s/cm2; regurgitant flow: 85.6 dyne × s/cm2).
Hot-spot platelet trajectories and their corresponding shear stress-loading waveforms (Figure 8A, forward flow; Figure 8C, regurgitant flow) were programmed into the HSD for platelet activation measurements (Figure 8B, forward flow; Figure 8D, regurgitant flow). The mean platelet activation for the modified valve (Figure 4B, forward: 0.0104 ± 0.0013; Figure 4D, regurgitant: 0.0438 ± 0.0035) was statistically lower (p < 0.05) than that for the original valve (Figure 8B, forward: 0.0241 ± 0.0020; Figure 8D, regurgitant: 0.0248 ± 0.0015) in both flow phases.
In this study, an optimization of the ATS AP valve thrombogenic performance was carried out with our novel DTE methodology. The ATS leaflet-housing gap clearance studies show the contribution of the hinge regions in imparting high stress to platelets. The flow fields are almost identical for the larger gap clearances of 130 and 250 μm, however with weaker jets and counter-rotating vortices in the leaflet-housing region compared with the 38 μm case in forward flow. This smallest gap clearance generates the strongest jets, increasing shear stress levels and the platelet residence time in the counter-rotating vortices emanating from the hinges, subsequently producing higher SA values. In regurgitant flow, higher stresses were also found for the smallest gap clearance. The resultant platelet activity measurements in the HSD were in complete agreement with the numerical PDF. Although the ATS valve is not recessed, optimized gap clearance is essential for reduced stresses and regurgitant flow.30 An intermediate gap clearance would be the optimal choice for minimized thrombogenic risk. A larger gap clearance does not produce significantly lower SA and may impair the valve functionality by increasing regurgitant flow volume through enhanced leakage jets formed within the hinge recess, as reported before for the SJM.7
Previous studies suggest that bileaflet MHV with larger opening angles are preferred for reduced thrombogenicity.29 Larger flow separation in the 80° opening angle accelerates the flow because of reduced cross sectional flow area, which produces higher localized velocity gradients and stresses. Exposure of platelets to increased stresses shifts the PDF to a higher mean value for the 80° opening angle. Measured platelet activation was significantly higher for the 80° opening angle trajectories, in agreement with the PDF. A larger opening angle would improve the thrombogenic performance of the ATS valve. The ATS MHV does not necessarily open fully under various operating conditions,16,28 and extra care is mandated during the design of the MHV moving parts to guarantee that the leaflets can meet their full excursion. However, individual flow patterns are influenced by valve pivot geometries and the introduction of sinuses downstream, which can dissipate stresses11 and may affect the fully open angle.
A streamlined channel was introduced in the hinge region between the leaflet and the stops intended to reduce flow induced stresses by increasing the effective flow area. A PDF of SA for the modified and the original designs shows that the modified valve has a significantly lower SA and is in complete agreement with measured platelet activation values in forward and regurgitant flow phases. Although the ATS open pivot hinge design already offers a significant improvement over the SJM MHV cavity pivot hinge design,16,17 it is clear that an additional improvement may optimize this successful design and make it less thrombogenic.
The simulations were performed during the forward and regurgitant flow phases to study the effect of stress-induced platelet activation that may lead to thromboembolic complications.31 Flow disturbances caused by leaflet motion and associated transient effects during the rapid closing phase that may further add to the platelet activation potential were not included in this study. Its very short duration (10–30 msec) offers only a modest contribution to platelet activation—unless cavitation is generated during this phase.32,33 Moreover, it is only marginally related to the design optimization methodology presented in this study, which focused on the optimization of the MHV recesses and gap clearances within the valve superstructure. Physiological geometrical features such as the sinus regions were not considered. More realistic near-valve regions and valve misalignment may additionally enhance the stress levels observed in this study.2,13 Physiological, non-Newtonian blood properties13 also influence the size of the recirculation zones, which may result in increased stress-exposure time. We will address some of these issues in our future studies using the DTE methodology.
Our studies demonstrate the robustness of the DTE methodology for optimizing the thrombogenic performance of MHV. In this study, we optimized the ATS valve design and conclude that the following changes to the original valve may significantly reduce thrombogenicity: 1) increasing the effective maximum opening angle of the valve to 85°, 2) optimizing the gap clearance between the valve and the housing to be at approximately 130 μm, and 3) modifying the hinge region by carving a streamlined flow channel in the leaflet stops at the edges facing the leaflets. We note that a combination of design changes described above could additionally improve the valve thromboresistance. The excellent agreement between numerical simulations and experimental results, and the ability to interface the two, demonstrates the efficacy of DTE for estimating the effects of design parameters on device thrombogenicity. We envision that this methodology will be adopted by cardiovascular device manufacturers to optimize their device designs for achieving improved thrombogenic performance. This may reduce or even eliminate the need for anticoagulation that is mandated for most of these devices.
Supported by NIH Grant 1R01 EB008004-01 (to D.B.).
1.Bluestein D: Research approaches for studying flow induced thromboembolic complications in blood recirculating devices. Expert Rev Med Devices
1: 65–80, 2004.
2.Alemu Y, Bluestein D: Flow-induced platelet activation and damage accumulation in a mechanical heart valve: Numerical studies. Artif Organs
31: 677–688, 2007.
3.Jesty J, Yin W, Perrotta P, Bluestein D: Platelet activation in a circulating flow loop: Combined effects of shear stress and exposure time. Platelets
14: 143–149, 2003.
4.Yoganathan AP, Chandran KB, Sotiropoulos F: Flow in prosthetic heart valves: State-of-the-art and future directions. Ann Biomed Eng
33: 1689–1694, 2005.
5.Yoganathan AP, He Z, Jones SC: Fluid mechanics of heart valves. Ann Rev Biomed Eng
6: 331–362, 2004.
6.Bluestein D: Towards optimization of the thrombogenic potential of blood recirculating cardiovascular devices using modeling approaches. Expert Rev Med Devices
3: 267–270, 2006.
7.Simon HA, Leo HL, Carberry J, Yoganathan AP: Comparison of the hinge flow fields of two bileaflet mechanical heart valves under aortic and mitral conditions. Ann Biomed Eng
32: 1607–1617, 2004.
8.Govindarajan V, Udaykumar HS, Chandran KB: Flow dynamic comparison between recessed hinge and open pivot bi-leaflet heart valve designs. J Mech Med Biol
9: 161–176, 2009.
9.Govindarajan V, Udaykumar HS, Chandran KB: Two-dimensional simulation of flow and platelet dynamics in the hinge region of a mechanical heart valve. J Biomech Eng
131: 031002, 2009.
10.Leo HL, Simon HA, Dasi LP, Yoganathan AP: Effect of hinge gap width on the microflow structures in 27-mm bileaflet mechanical heart valves. J Heart Valve Dis
15: 800–808, 2006.
11.Travis BR, Marzec UM, Leo HL, et al
: Bileaflet aortic valve prosthesis pivot geometry influences platelet secretion and anionic phospholipid exposure. Ann Biomed Eng
29: 657–664, 2001.
12.Yin W, Krukenkamp IB, Saltman AE, et al
: The thrombogenic performance of a St. Jude bileaflet mechanical heart valve in a sheep model. ASAIO J
52: 28–33, 2006.
13.Yin W, Alemu Y, Affeld K, et al
: Flow-induced platelet activation in bileaflet and monoleaflet mechanical heart valves. Ann Biomed Eng
32: 1058–1066, 2004.
14.Yin W, Gallocher S, Pinchuk L, et al
: Flow induced platelet activation in a St. Jude mechanical heart valve, a trileaflet polymeric heart valve and a St. Jude tissue valve. Artif Organs
29: 826–831, 2005.
15.Bluestein D, Yin W, Affeld K, Jesty J: Flow-induced platelet activation in a mechanical heart valve. J Heart Valve Dis
13: 501–508, 2004.
16.Dumont K, Vierendeels J, Kaminsky R, et al
: Comparison of the hemodynamic and thrombogenic performance of two bileaflet mechanical heart valves using a CFD/FSI model. J Biomech Eng
129: 558–565, 2007.
17.Xenos M, Girdhar G, Alemu Y, et al
: Device thrombogenicty emulator (DTE)—design optimization methodology for cardiovascular devices: A study in two bileaflet MHV designs. J Biomech
2010 (in press).
18.Girdhar G, Bluestein D: Biological effects of dynamic shear stress in cardiovascular pathologies and devices. Expert Rev Med Devices
5: 167–181, 2008.
19.Girdhar G, Xu S, Jesty J, Bluestein D: In vitro model of platelet-endothelial activation due to cigarette smoke under cardiovascular circulation conditions. Ann Biomed Eng
36: 1142–1151, 2008.
20.Jesty J, Bluestein D: Acetylated prothrombin as a substrate in the measurement of the procoagulant activity of platelets: Elimination of the feedback activation of platelets by thrombin. Anal Biochem
272: 64–70, 1999.
21.Moin P, Mahesh K: Direct numerical simulation: A tool in turbulence research. Annu Rev Fluid Mech
30: 539–578, 1998.
22.Ellis JT, Healy TM, Fontaine AA, et al
: Velocity measurements and flow patterns within the hinge region of a Medtronic Parallel bileaflet mechanical valve with clear housing. J Heart Valve Dis
5: 591–599, 1996.
23.Ge L, Dasi LP, Sotiropoulos F, Yoganathan AP: Characterization of hemodynamic forces induced by mechanical heart valves: Reynolds vs. viscous stresses. Ann Biomed Eng
36: 276–297, 2008.
24.Hellums JD: 1993 Whitaker lecture: Biorheology in thrombosis research. Ann Biomed Eng
22: 445–455, 1994.
25.Apel J, Neudel F, Reul H: Computational fluid dynamics and experimental validation of a microaxial blood pump. ASAIO J
47: 552–558, 2001.
26.Yeleswarapu KK, Antaki JF, Kameneva MV, Rajagopal KR: A mathematical model for shear-induced hemolysis. Artif Organs
19: 576–582, 1995.
27.Nobili M, Sheriff J, Morbiducci U, et al
: Platelet activation due to hemodynamic shear stresses: Damage accumulation model and comparison to in vitro measurements. ASAIO J
54: 64–72, 2008.
28.Aoyagi S, Arinaga K, Fukunaga S, et al
: Leaflet movement of the ATS valve in the aortic position: Unique behavior observed in 19-mm valves. Ann Thorac Surg
82: 853–857, 2006.
29.King MJ, David T, Fisher J: Three-dimensional study of the effect of two leaflet opening angles on the time-dependent flow through a bileaflet mechanical heart valve. Med Eng Phys
19: 235–241, 1997.
30.Emery RW, Van Nooten GJ, Tesar PJ: The initial experience with the ATS Medical mechanical cardiac valve prosthesis. Ann Thorac Surg
75: 444–452, 2003.
31.Bluestein D, Li YM, Krukenkamp IB: Free emboli formation in the wake of bi-leaflet mechanical heart valves and the effects of implantation techniques. J Biomech
35: 1533–1540, 2002.
32.Bluestein D, Menon S, Wu ZJ, et al
: Closing behavior of a new bileaflet mechanical heart valve. ASAIO J
39: M398–M402, 1993.
33.Bluestein D, Einav S, Hwang NH: A squeeze flow phenomenon at the closing of a bileaflet mechanical heart valve prosthesis. J Biomech
27: 1369–1378, 1994.