Effect of Hematocrit and Elevated Beat Rate on the 12cc Penn State Pediatric Ventricular Assist Device

Abstract

Congenital heart disease affects approximately 40,000 infants annually in the United States with 25% requiring invasive treatment.^1,2 Due to limited number of donor hearts and treatment options available for children, pediatric ventricular assist devices (PVADs) are used as a bridge to transplant. The 12cc pneumatic Penn State PVAD is optimized to prevent platelet adhesion and thrombus formation at patient nominal conditions; however, children demonstrate variable blood hematocrit and elevated heart rates. Therefore, with pediatric patients exhibiting greater variability, particle image velocimetry is used to evaluate the PVAD with three non-Newtonian hematocrit blood analogs (20, 40, and 60%) and at two beat rates (75 and 120 bpm) to understand the device’s performance. The flow fields demonstrate a strong inlet jet that transitions to a solid body rotation during diastole. During systole, the rotation dissipates and reorganizes into an outlet jet. This flow field is consistent across all hematocrits and beat rates but at a higher velocity magnitude during 120 bpm. There are also minor differences in flow field timing and surface washing due to hematocrit. Therefore, despite patient differences in hematocrit or required pumping output, thorough surface washing can be achieved in the PVAD by altering operating conditions, thus reducing platelet adhesion potential.

Introduction

Congenital heart defects are the most common birth defect, affecting 1 out of every 125 babies born.¹ A minimum of 40,000 infants are estimated to be diagnosed with a congenital heart defect each year. Of these, approximately 25% require invasive treatment, such as heart transplantation or corrective surgery, within the first year.² However, children on the heart transplant waiting list face one of the highest mortality rates of all organ transplant patients. Each year, over 600 children are added to the list, while approximately 13% die awaiting a donor heart.^3,4

Mechanical circulatory support devices, such as ventricular assist devices (VADs), decrease the mortality rate of patients on the waiting list by augmenting cardiac output and decreasing the mechanical load on the heart.⁵ VADs have been used in adults since the late 1960’s and demonstrate an 81% and 77% survival rate in adults with a VAD at one and two years out, respectively.⁶ However, VAD use in the pediatric population is lagging due to the small number of patients, lack of available devices sized for children, and the complex geometry of congenital heart defects.⁷ Before the development of pediatric VADs (PVADs), and due to the lack of other options, full size VADs were implanted in children. Unfortunately, size disparities resulted in a high risk of neurological events and a low survival rate.⁸ Prior to PVAD development, extracorporeal membrane oxygenation (ECMO) was the primary method of pediatric mechanical circulatory support. Although ECMO was effective for short term support in children, after 10-20 days, complications such as hemorrhage and infection prevented long term use.⁹ To address the limited treatment options for small patients with severe congenital heart defects, the National, Heart, Lung, and Blood Institute established the Pediatric Circulatory Support Program in 2004 to develop assist devices for pediatric patients weighing 2 to 25 kgs.⁵

Penn State was one of five groups awarded a contract to develop a PVAD based on the pulsatile 70 cc Pierce-Donachy adult VAD.¹⁰ To prevent thrombosis, hemocompatibility of the device’s blood contacting materials and overall device hemodynamics were studied in vitro. The hemocompatibility of potential polymers for blood contacting devices were assessed using both a parallel plate flow chamber and a rotating disk system. These two experimental methods demonstrated that steady shear rates higher than 500 s⁻¹ reduced platelet deposition.^11,12 Using particle image velocimetry (PIV), hemodynamic quantification guided operating conditions, performance during weaning, and informed design criteria such as valve type and orientation.^10,13–18 Furthermore, at each operating condition and design configuration, the wall shear rates from the hemodynamic studies were determined to ensure the platelet residence time was low and the shear rates were greater than 500 s⁻¹.

However, the influence of blood constituency and its relationship to PVAD hemodynamics are not well understood. Previous PVAD fluid dynamic studies used a 40% hematocrit (Hct) non-Newtonian blood analog, but pediatric patient hematocrit can range from 20-60%.¹⁹ The percentage of red blood cells (RBCs) affects the viscoelastic behavior of blood, influencing the fluid dynamics observed. At low shear rates (<100 s⁻¹), rouleaux formation increases blood’s viscosity and elasticity exponentially. As the shear rate increases, these aggregates are broken up and begin to align with the flow. At high shear rates (>500 s⁻¹), blood viscosity and elasticity behave as an asymptotic Newtonian fluid.²⁰ Good et al. emphasized the importance of accurately representing blood’s viscoelasticity by comparing the hemodynamics within a computational pediatric aortic arch model using both Newtonian and non-Newtonian models at 20, 40, and 60% hematocrit.²¹ Hematocrit was observed to influence flow distribution to the branching arteries and velocity and wall shear stress magnitudes. Therefore, the fluid dynamics within the PVAD must be quantified at hematocrits deviating from the average 40%.

Blood constituency is not the only patient specific parameter that needs to be tested within the PVAD.²² Earlier PVAD studies only quantified the hemodynamics for weaning purposes at 86, 75 and 50 bpm.^14,16 These studies demonstrated that with a reduction in beat rate, there was an earlier loss of rotational flow, increased platelet residence time, and a reduced wall shear rate in the PVAD. However, for larger infants or infants with little to no native heart function, PVAD beat rates may need to be increased to achieve higher pump outputs. Thus, with red blood cell behavior and thrombus potential being highly dependent on shear rate and residence time, it is essential that an elevated beat rate, producing a higher pump output, be quantified to understand device performance and ensure safety.

With pediatric patients demonstrating greater variability in required pump flow and blood hematocrit, devices must demonstrate efficacy and reduced thrombus potential at more than just patient averaged conditions. Therefore, this study quantifies the fluid mechanics within the Penn State PVAD with PIV at three pediatric hematocrit analogs (20%, 40%, and 60%) and two beat rates (75 bpm and 120 bpm).

Materials and Methods

Mock Circulatory Flow Loop

An acrylic Penn State 12 cc pulsatile PVAD was manufactured to provide optical access for PIV (Fig. 1A, *pending permission). The model consisted of a pneumatically driven polyurethane diaphragm and inlet and outlet ports housing 17 mm Bjork-Shiley Monostrut (BSM) tilting disk valves, which created a more favorable surface washing than the bi-leaflet valve.^10,17 To mimic the pediatric systemic circulation, the PVAD was placed in a mock circulatory loop (Fig. 1B, *pending permission) that consisted of two compliance chambers, a resistance plate, and a reservoir. These components simulated the aortic and atrial compliance, systemic resistance, and the venous reservoir, respectively. The flow rates at the inlet and outlet were measured using ultrasonic flow probes (Transonic Systems, Inc., Ithaca, NY) and pressure transducers (Merit Medical, South Jordan, UT) measured the pressure at the pneumatic driver and outlet (Fig. 1B, *pending permission). A Wavebook data acquisition board (IOtech, Inc., Cleveland OH) recorded the flow rate and pressure waveforms at a sampling rate of 5 kHz.

Figure 1:

A. Side view schematic of the Penn State PVAD indicating the locations of the PIV (particle image velocimetry) planes of interest (7, 8.2, and 11 mm), diaphragm location, and air line [*Permission pending]. B. PVAD mock circulatory loop comprised of compliance chambers, a resistance plate, a reservoir, flow probes, and pressure transducers (¹⁴ *Permission pending). C. Sample waveform from the 75 bpm and 40% Hct operating condition defining the EDD, ejection time, and ESD. Thick black lines correspond to times PIV data were collected. D. Schematic of the user defined walls for an 8.2 mm slice in the wall shear rate code moving in a counterclockwise direction as indicated by the arrow.

Operating Conditions

The compliance and resistance of the mock circulatory loop and the pressure and vacuum applied by the pneumatic driver, matched physiological flow rates, pressures, and filling times. Each operating condition’s details are summarized in Table 1 and the standard deviation is provided for each measured parameter. Two beat rates (75 and 120 bpm) were compared. To ensure complete filling and ejection at both rates, the systolic duration, end diastolic delay (EDD), and end systolic delay (ESD) were adjusted by the pneumatic driver. A systolic duration of 340 ms was maintained at 75 bpm and 250 ms at 120 bpm. The EDD, or the time between the end of diastole and the start of systole, was defined as the time between inflow going to zero and the drive pressure beginning to increase (Fig. 1C). The EDD was kept between 0 and 10 ms to ensure complete filling.²² The ESD, or the time between the end of systole and the start of diastole, was defined as the time between outflow going to zero and the drive pressure beginning to decrease (Fig. 1C). The ESD was kept between 0 and 10 ms to ensure complete ejection and prevent cavitation.²³ The ejection time (ET), or the time between the start of systole and the point of complete ejection, was defined as the time between the start of the drive pressure increasing to the time when outflow goes to zero. Finally, the outlet pressure was maintained at an ideal 90/60 mmHg for both beat rates and was used to trigger the laser for PIV. ^10,13–18

Table 1:

Summary of operating conditions at each beat rate and hematocrit. The beat rate and ET were set on the pump and the other values were measured. The standard deviations were included for all measured parameters.

% Hct	Beat Rate (bpm)	Average Inflow Rate (L/min)	Average Outflow Rate (L/min)	Outlet Pressure (max/min) mmHg	EDD (ms)	ESD (ms)	ET (ms)
20%	75	1.10 ± 0.05	1.12 ± 0.06	90 ± 0.24 / 60 ± 0.44	7.1 ± 0.6	9.1 ± 2.4	340
	120	1.91 ± 0.07	2.08 ± 0.05	91 ± 0.50 / 61 ± 0.53	3.9 ± 2.1	8.8 ± 3.4	250
40%	75	1.13 ± 0.01	1.20 ± 0.01	90 ± 0.20 / 61 ± 0.20	9.1 ± 1.1	7.7 ± 1.8	340
	120	1.87 ± 0.01	2.06 ± 0.01	90 ± 0.2 / 62 ± 0.08	0.8 ± 0.4	2.9 ± 2.6	250
60%	75	1.22 ± 0.01	1.16 ± 0.01	89 ± 0.22 / 64 ± 0.15	7.0 ± 3.2	7.5 ± 2.0	340
	120	1.93 ± 0.01	2.00 ± 0.02	89 ± 0.15 / 64 ± 0.42	1.7 ± 0.2	2.4 ± 1.7	250

Hematocrit Matching Fluids

To determine the influence of hematocrit on the fluid mechanics within the PVAD, three non-Newtonian blood analog fluids matched the properties of 20%, 40%, and 60% Hct pediatric blood. The fluids were composed of de-ionized water, glycerin, and Xanthan gum in various ratios (Table 2) to match patient averaged viscosity and elasticity (Fig. 2). Sodium iodide was added to match an acrylic refractive index of 1.49 and sodium thiosulfate was added in small quantities to reduce fluid cloudiness and optimize optical clarity. The fluid was seeded with 10 μm hollow glass sphere tracer particles (Potters Industries Inc., Valley Forge, PA) for PIV.

Table 2:

Composition by volume of each non-Newtonian blood analog

	Sodium Iodide	Glycerin	Water	Xanthan Gum	Sodium Thiosulfate
20% Hct	50%	18%	31%	1%	<1%
40% Hct	50%	26%	23%	1%	<1%
60% Hct	50%	36%	13%	1%	<1%

Figure 2:

A. Viscosity and B. elasticity curves of blood analogs as compared to whole blood¹⁹ at 20, 40 and 60% HCT.

Particle Image Velocimetry

Planar PIV quantified the fluid mechanics within the PVAD at each condition. The system included a dual Nd:YAG 532 nm Evergreen laser (Quantel, Les Ulis, France) which was coupled with a 500 mm spherical lens and a −25 mm cylindrical lens to illuminate an 800 μm thick laser sheet. PIV data were collected at three planes within the model (7, 8.2 and 11 mm from the edge of the inlet port) (Fig. 1A) to evaluate the flow fields at the inlet, body of the PVAD, and the outlet across conditions (Fig. 1D).^10,13–18 A PowerView Plus 4MP camera (TSI, Inc., Shoreview, MN) with a Micro-Nikkor 60 nm F1.8 lens (Nikon Corporation, Tokyo, Japan) were orthogonally mounted to the laser.

To compare data at different beat rates, time steps were calculated such that data were collected at the same percentage of diastole and systole. Data were collected at 11%, 22%, 33%, 43%, 54%, 65%, 76%, 87%, and 98% of diastole, and 12%, 26%, 41%, 56%, 71%, 85%, and 100% of systole (Fig. 1C). ^10,13–18 If the membrane entered the plane of interest and prevented the laser from illuminating particles, the time point was not collected. This interference occurred more at the end of systole and the start of diastole and in the planes closer to the diaphragm. The camera and laser were synchronized using a LaserPulse Synchronizer (TSI, Inc., Shoreview, MN) to capture 200 image pairs at each time point, triggered from the outlet pressure waveform.

Processing

The PIV images were processed to compute the velocity magnitude and wall shear rate (WSR) at each time point and operating condition to assess the potential for platelet adhesion. First, raw images were masked to remove non-fluid background or reflections within the PIV image. This was done by manually identifying the acrylic walls of the PVAD model using an in-house MATLAB script.^17,24 The PVAD was separated into four walls based on the curvature of the PVAD (Fig. 1D). Starting at wall one, and moving in a counterclockwise manner, each wall was manually identified and the program scanned the intensity gradient of the image to fit the user selected curve to the PVAD wall. The background was then removed from the image.

The masked images were processed in Insight4G using a Hart cross-correlation algorithm²⁵ to quantify the particle displacement within each interrogation region. The processed images were then averaged in Tecplot Focus (Tecplot Inc., Bellevue, WA) to visualize the velocity contours. To directly compare the flow fields between the different operating conditions, the velocity fields were normalized by their respective average inlet velocities, determined from the average inlet flow rate recorded at each condition (Table 1) and the BSM valve’s effective orifice area (1.1 cm²).

Lastly, to quantify WSR within the PVAD, the masked images were processed in a custom MATLAB script developed by Hochereon et al.²⁴ The initial wall masking code output the (x,y) coordinate location of each PVAD wall in the counterclockwise direction that the walls were selected in. The WSR code used the wall locations to identify the interrogation region near each wall (Fig. 1D), ensured the interrogation region was greater than 10% of its actual size for accuracy, and calculated the centroid. Next, to obtain WSR, the mean tangential velocity gradient was calculated between the centroid of the interrogation region and the PVAD wall. The average WSR along the length of each wall (x-axis) over time (y-axis) was visualized with a WSR contour map to assess if the WSRs were greater than 500 s⁻¹ at any point in the cycle, ensuring sufficient surface washing to prevent platelet adhesion.¹² The contour maps depicted both the magnitude of the WSR and direction the walls were being washed in, with positive WSR demonstrating flow moving in the counterclockwise direction along the user selected points and negative WSR demonstrating flow moving clockwise along the wall. If the surfaces are not appropriately washed, they will be depicted by zero shear in green.

5. Results

To summarize the PVAD flow fields, the normalized velocities during three key time points were presented at each operating condition: 54% diastole or mid-diastole (Fig. 3), 84% diastole or late diastole (Fig. 4), and 43% systole or mid-systole (Fig. 5). These time points were selected to best display the flow features of interest observed within the PVAD. Additionally, the WSR at 75 (Fig. 6) and 120 bpm (Fig. 7) for all three Hct were presented over the entire cycle when the membrane was not obstructing image capture.

Figure 3:

PVAD fluid mechanics at the 54% diastole, 8.2 mm plane for A. 75 bpm and 20% Hct, B. 75 bpm and 40% Hct, C. 75 bpm and 60% Hct, D. 120 bpm and 20% Hct, E. 120 bpm and 40% Hct, and F. 120 bpm and 60% Hct.

Figure 4:

PVAD fluid mechanics at the 87% diastole, 8.2 mm plane for A. 75 bpm and 20% Hct, B. 75 bpm and 40% Hct, C. 75 bpm and 60% Hct, D. 120 bpm and 20% Hct, E. 120 bpm and 40% Hct, and F. 120 bpm and 60% Hct.

Figure 5:

PVAD fluid mechanics at the 43% systole, 8.2 mm plane for A. 75 bpm and 20% Hct, B. 75 bpm and 40% Hct, C. 75 bpm and 60% Hct, D. 120 bpm and 20% Hct, E. 120 bpm and 40% Hct, and F. 120 bpm and 60% Hct.

Figure 6:

WSR at 75 bpm and at A. 20%, B. 40%, and C. 60% Hct conditions for walls 1-4. Note the difference of scales between 75 and 120 bpm due to the elevated WSR at 120 bpm. The dashed line in each frame indicates the transition from diastole to systole. Green contour regions describe surface washing less than 500 s⁻¹.

Figure 7:

WSR at 120 bpm and at A. 20%, B. 40% and C. 60% Hct conditions for walls 1-4. Note the difference of scales between 75 and 120 bpm due to the elevated WSR at 120 bpm. The dashed line in each panel indicates the transition point from diastole to systole. Green contour regions describe surface washing less than 500 s⁻¹.

Across all operating conditions, fluid entered the PVAD as a strong inlet jet. By 54% diastole, there were two inlet jets, one stronger than the other, due to the major and minor orifice of the BSM valve (Fig. 3D) that set up the start of a rotational field at the base of the PVAD. As diastole continued and with the outlet valve closed, these two jets merged into a single jet that fed into the solid body rotation. The fluid did not enter as early in diastole or penetrate as deep into the PVAD for the 60% Hct case compared to the 20% Hct case at both beat rates (Fig. 3D, 3F). This was corroborated by the WSR contour plots that demonstrated a reduced WSR on surface 3 (S3) at the outlet side base of the PVAD (Fig. 6C, 7C). For an increased beat rate, the fluid had more momentum, thus a greater velocity magnitude and WSR which allowed the fluid to penetrate deeper into the PVAD earlier in the cardiac cycle, washing the surface (Fig. 3A, 3D, Fig. 6, Fig. 7). Additionally, there was little to no difference between the WSR or flow field development due to the blood analog Hcts at the elevated beat rate (Fig. 7).

By late diastole (Fig. 4), both the inlet and outlet valves were closed which further allowed for a strong rotational field. The WSR during the solid body rotation was critical for washing the blood contacting surfaces to prevent platelet adhesion and stagnation regions (Fig. 6, Fig. 7). Other than at 75 bpm, 20% and 60% Hct, every condition had a WSR greater than 500 s⁻¹ at some point in the cycle. The velocity magnitude and WSR increased during late diastole at an elevated beat rate (Fig. 4D–F, Fig. 6, Fig. 7), providing greater surface washing.

At the end of diastole and the start of systole, once the outlet valve opened, the flow started to align with the outlet valve. This rapidly decreased the velocity magnitude of the rotational flow field. By mid-systole, two outlet jets formed, similar to the two inlet jets, due to the major and minor orifice of the BSM valve (Fig. 5). The major orifice was oriented towards the outer wall to promote greater surface washing along wall 4 (Fig. 5B, Fig. 6, Fig. 7). With the increase in beat rate, there was little to no difference between 75 and 120 bpm normalized velocity in the outlet jet region. There was however, a greater WSR at the outer wall of the PVAD (Surface 4, S4) at the elevated beat rate due to the higher velocity magnitude prior (Fig. 6, Fig. 7). Lastly, the area the jet occupied at the outlet was larger, extending further into the sac at 20% Hct than 60% Hct which increased the WSR along the entire length of the outlet (Fig. 5, Fig. 6, Fig. 7).

Discussion

To address the lack of pediatric circulatory assist devices for bridge to transplant in congenital heart defect patients, Penn State University is developing a pneumatically driven pulsatile PVAD. To optimize the PVAD design, hemodynamics and blood contacting materials were first evaluated within an in vitro environment to prevent platelet adhesion. Since these initial PVAD studies, the PVAD has undergone minor design improvements and is currently in pre-clinical testing. The pump has been tested extensively in lambs and has demonstrated excellent biocompatibility even with heparin restricted to the first two post-operative weeks, to maintain normal coagulation during the acute phase response in which fibrinogen and platelet activity increase²⁶. However, these prior PIV and animal studies assessed the device at 75 bpm or lower or with a single patient averaged 40% Hct blood analog. Pediatric patients exhibit a wider range of hematocrit and require variable pump outputs depending on their size or native heart functionality. In vivo testing at the higher beat rate of 120 bpm will commence this year and thus, the fluid mechanics within the Penn State PVAD must be quantified at a wider range of pediatric conditions, including two beat rates (75 and 120 bpm) and three blood analog hematocrits (20, 40, and 60%).

To compare the flow fields across operating conditions, the velocity fields were normalized using their respective average velocities entering the PVAD. The PVAD demonstrated similar flow fields across all conditions and all three imaging planes. Generally, the PVAD had a strong inlet jet that resulted a solid body rotation during diastole. This rotation continued into early systole until the outlet valve opened and the flow realigned into two strong outlet jets, with the major orifice along the outer wall. These flow fields were consistent with prior studies that evaluated the Penn State PVAD at beat rates of 50, 75, and 86 bpm at 40% Hct.^{10,13,15,16,18,22,27}

By increasing the beat rate to 120 bpm, there were stronger inlet and outlet jets, a stronger recirculation region, and increased WSR, all promoting better surface washing (Fig. 6, Fig. 7). At 120 bpm the fluid had more momentum, producing a recirculation region that developed earlier in diastole and lasted longer into systole than the 75 bpm condition. This trend was consistent with Rozelle et al. who evaluated the PVAD with a 40% Hct analog at 75 bpm and 50 bpm and found that with a reduced beat rate, there was a reduced velocity magnitude, an earlier loss of the solid body rotation, and an increased blood residence time due to decreased WSR.¹⁶ Previously, bovine animal studies with the 50 cc adult VAD at 75 bpm developed microthrombi on the outlet side, which correlated with regions of WSR less than 500 s⁻¹.²⁸ By increasing the pump output, as was done in this study, the stronger jets and rotational flows produced WSRs greater than 500 s⁻¹ along all walls, reducing the risk of platelet deposition on the PVAD PUU.^11,12

In addition to assessing the PVAD at two beat rates, the effect of blood viscoelasticity was evaluated using three hematocrit blood analogs (20, 40, and 60%). Although the viscoelasticity did not affect the general flow patterns observed in the PVAD, the choice of blood analog influenced the overall velocity magnitude, penetration of the jet into the pump body, and area occupied by the outlet jet. At 75 bpm, the inlet jets penetrated deeper into the sac, developed a rotational field earlier in diastole, and produced outlet jets that occupied a larger area at 20% Hct (Fig. 3A, 4A, 5A). Also, with increasing hematocrit, the development of each of the flow structures were delayed. The differences between flow field evolution and amount of surface washing due to the depth of fluid penetration to the base of the PVAD were attributed to the fluid having a decreased momentum with increasing Hct. At 60% Hct, there was likely an increased viscous dissipation due to the increased viscous effects, which reduced the momentum of the fluid. Next, a Xanthan gum blood analog mimicked the shear thinning behavior of blood by providing nonuniform viscosity and elasticity dependent on shear rate.²⁰ By increasing the shear rate at an elevated beat rate (Fig. 6, Fig. 7), the blood analogs produced a more Newtonian behavior, combatting the effect of hematocrit and viscous energy dissipation. Therefore, at 120 bpm, with the increased shear rate and Newtonian behavior, there was little to no difference with hematocrit. Regardless of the time to develop the rotational flow or the degree of surface washing at each hematocrit, pump operating conditions could be adjusted for each patient to ensure the WSR exceeds 500 s⁻¹ at some point in the cardiac cycle.^11,12 Thus, the PVAD could achieve sufficient washing to reduce the risk of platelet adhesion.

Limitations

For this study, the primary limitation was the wall region used for calculating the WSR. PIV is a 2D measurement that inherently produces greater uncertainty near the wall region²⁹ Additionally, the near wall region identification occurs by a user defined mask where the user defines a wall as close to the flow field as possible²⁴. However, regardless of the user defined mask, Hochareon et al. obtained an error of less than 20% for each interrogation window by thoroughly validating the shear rate calculation methodology against four theoretical solutions. They found this error could be reduced by optimizing PIV resolution and data collection settings.

Conclusions

With children exhibiting variability in required pump beat rate and blood hematocrit, the Penn State PVAD was evaluated at two beat rates (75 and 120 bpm) and three hematocrits (20, 40, and 60%). Under all conditions, during diastole there was a strong inlet jet that transitioned to a solid body rotation. During systole, the solid body rotation dissipated as the flow realigned to form an outlet jet. The increase in beat rate increased the velocity magnitude and the rotational period, leading to greater surface washing. Hematocrit had a greater impact at 75 bpm where there was reduced surface washing with increased %Hct. Thus, by increasing the beat rate, there would be little to no effect of hematocrit and improved surface washing. Moving forward, there should be a further analysis on the temporal impact of WSR to determine if exceeding 500 s⁻¹ at some point in the cardiac cycle is sufficient or if this needs to be sustained over a specific time frame to prevent platelet adhesion. Regardless, this study demonstrated reduced platelet adhesion potential at most patient specific operating conditions and indicated the PVAD operating condition could be safely optimized for each patient’s needs.