Secondary Logo

Journal Logo

Engineering Aspects–Novel Approaches

Multiscale Analysis of Surface Thrombosis In Vivo in a Left Ventricular Assist System

Yamanaka, Hanako*; Rosenberg, Gerson*†; Weiss, William J.*†; Snyder, Alan J.*†; Zapanta, Conrad M.*†; Siedlecki, Christopher A.*†

Author Information
doi: 10.1097/01.mat.0000181707.06225.a0
  • Free

Abstract

Cardiovascular disease remains the leading cause of death in the United States, and in 2002 was responsible for the deaths of almost 700,000 people, or 28.5% of deaths in the United States.1 In the United States in 2002, the heart transplant waiting list outnumbered the number of heart transplants performed by 1,692 people, leaving 44.5% of patients on the waiting list without heart transplants; 558 people died while waiting for a heart transplant.2 Ventricular assist devices (VADs) provide an alternative to the limited option of a heart transplant. VADs are increasingly used for bridge to transplantation, and supported almost 20% of heart transplant patients in 2003.3 More recently, the devices are being used for destination therapy, in which VADs provide patients with long-term support. A multicenter study, the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH), showed a 48% decrease in mortality by use of VADs compared with optimal medical management involving pharmacologic treatment.4 As the demand for such devices increases, improvement in performance becomes critical.

Thromboembolism is one of the major complications in VADs,5–7 occurring in 3% to 35% of bridge-to-transplant patients,8 and resulting in strokes in 16% of destination therapy VAD patients of the REMATCH study.9 A thrombus may obstruct blood flow, alter fluid dynamics causing damage to blood components, or wash off the surface as emboli and occlude vessels. The introduction of foreign materials into contact with blood, changes to fluid flow dynamics caused by the pump, and pharmacologic therapy are all insults to the biologic system involved with use of VADs that may potentially lead to thrombosis.10 A better understanding of the effects of such changes on thrombogenesis can provide criteria for the improvement of VADs.

Fluid mechanics has been closely associated with thrombosis in VADs and is an important design criterion for reducing thrombogenicity. The relationship between fluid mechanics and thrombosis in VADs is complex, with various blood components each being affected by flow differently.11–13 In animal studies comparing adult and pediatric pulsatile left ventricular assist systems (LVAS), adult LVAS were found to be thrombus-free while numerous thrombi formed on pediatric pumps.14 Fluid flow studies of the devices revealed significantly different flow dynamics between the two blood pumps of different sizes but similar geometry.15,16 Other studies have also used fluid dynamics analysis techniques to optimize pump design to reduce thrombosis.17–20 Such studies confirm the relationship between fluid flow dynamics and thrombosis in blood-contacting devices. This relationship is an important element for consideration in ongoing efforts to develop reduced-size pumps for children and smaller adult patients.

Efforts to reduce occurrences of thrombosis commonly involve pharmacologic therapy such as anticoagulation. This therapy, however, introduces the additional complication of bleeding,21,22 with the incidence of hemorrhage having been reported to range from 31% to 60%.23,24 Bleeding due to anticoagulation was reported as the cause of death in up to 11% of patients in studies of three device designs.25 Despite anticoagulation of many VAD patients, thrombosis is still a problem and new anticoagulants are constantly being developed. A closer examination of the thrombotic blood response to anticoagulation in LVAS could mitigate requirements for anticoagulation therapy.

Detection of thrombus formation on biomaterial surfaces focuses on measurement of the primary components of thrombi. Protein adsorption is widely accepted to be the initial response to blood contact with biomaterials. This is followed by platelet adhesion, activation, and aggregation, and fibrin formation culminating in formation of a thrombus.26 Thus, protein adsorption, platelet response, and fibrin formation on biomaterials are considered important measures of thrombogenicity of biomaterials.27–31

In this study, we investigated in vivo thrombus formation on the blood-contacting surface of poly(urethane urea) (PUU) blood sacs in LVAS implanted in calves for 28 to 30 days. In vivo studies allow investigation of the blood response under the complex environment found in the LVAS. Macroscopic and microscopic thrombus formation was examined throughout the various regions of the blood sacs. To the best of our knowledge, such a thorough investigation, including mapping regional differences in thrombosis in LVAS pumps, has not yet been reported. Numerous studies have investigated in vitro blood responses to flow and biomaterials,32–35 and in vivo studies discuss occurrences of thrombosis both on macroscopic and microscopic scales.30,36–38 However, these studies do not perform close examination of the location of thrombus formation within the pump. The significance of this comprehensive analysis of thrombosis throughout the blood sacs and its correlation with fluid flow dynamics in the highly complex in vivo environment of the LVAS is discussed.

Materials and Methods

In Vivo Studies with Bovine Models

In vivo studies of thrombosis in a 70-cc-static stroke volume LVAS with PUU, Biospan MS/0.4 (The Polymer Technology Group, Inc., Berkley, CA) blood sacs developed and fabricated by The Pennsylvania State University39 were performed in two groups. A total of nine 70-cc LVAS were implanted in Holstein calves for 28 to 30 days. The blood flow rate was controlled to try to maintain flow rates from 5 to 6 l/min. Group 1, consisting of four calves, received postoperative anticoagulation with heparin and warfarin sodium the first 4 postoperative days and just warfarin sodium thereafter. The dosage was adjusted to maintain the prothrombin time (PT), a measure of clotting, at 1.5 to 2 times the baseline PT before surgery. Similar PT monitoring is performed for anticoagulation management in VAD patients.40 Group 2 consisted of five calves that received no postoperative anticoagulation. These calves were referred to as nonanticoagulated calves. All animals were housed and studies performed in AAALAC-accredited animal facilities. Veterinary care was administered in accordance with guidelines set forth in The Guide for Care and Use of Laboratory Animals.41

Implants were retrieved in accordance with the guidelines of the 2001 American Veterinary Medical Association Panel on Euthanasia. To avoid stagnant blood, the sacs were flushed with saline followed by 1% paraformaldehyde (PFA) fixative. This was accomplished by connecting a catheter with tubing to a 1-l bag of saline and two 1-l bags of 1% PFA. With the device still pumping, the inlet cannula was cut and the catheter connected to the bags of solution was inserted into the inlet cannula. Saline was released into the pump through the inlet cannula, clamping off the inflowing blood simultaneously. The outflow cannula was cut and the distal end of the outflow cannula was clamped off. Saline from the outflow cannula was allowed to drain. The blood sac was flushed with approximately 950 ml saline. One liter of 1% PFA was then introduced into the pump. As the first 1-l bag of 1% PFA was emptied, the pump stop command was given, and the second bag of 1% PFA was allowed to flow until the pump completely stopped. The calf was euthanized and the blood pump was retrieved. The blood sac, filled with 1% PFA, was fixed for 1 hour, replaced with saline, and stored at 4ºC.

Evaluation of PUU Blood Sacs for Thrombosis

Macroscale Evaluation

Macroscale surface evaluation was performed by mapping macroscopic thrombi on the sacs from both in vivo study groups. The sac was examined for thrombi and the size, location, and colors of each were recorded. The color of the thrombus is indicative of the presence or absence of red blood cells entrapped in a fibrin mesh. Drawings of six views of the 70-cc LVAS blood sac were created in Solid Works software (Concord, MA). Color, size, and location of thrombi were sketched onto the Solid Works sac drawings using Adobe Photoshop (San Jose, CA). Thrombus images from each implant were overlaid to identify regional patterns of thrombosis for each study group. The percentage of the sac surface covered by macroscopic thrombi was also calculated for each region.

Microscale Evaluation

Microscale surface evaluation of thrombosis was performed using scanning electron microscopy (SEM) to examine surface topography of the explanted blood sac samples. Biologic deposition of protein, platelets, and fibrin was investigated. Negative and positive in vitro controls were also prepared to help identify topographic features observed on the blood sacs by SEM. Immunofluorescent labeling and confocal microscopy were performed to validate the identity of structures observed by SEM.

In Vitro SEM Controls.

Seventeen negative control samples of PUU blood sacs that were not exposed to blood were prepared for microscopic surface evaluation by SEM. The samples were cut into approximately 1-cm2 pieces from the various regions of the blood sac. Three positive controls of fibrin clots on sac PUU were also prepared. Bovine blood drawn into citrate phosphate dextrose adenine was centrifuged for 20 minutes at 600g at room temperature to isolate the supernatant platelet-rich plasma (PRP). After transferring a portion of the PRP to a tube, the remaining blood was centrifuged at 1,500g for 20 minutes to obtain platelet poor plasma (PPP). Total volumes of 2.5, 5.0, and 6.5 μl of 1 M calcium chloride (CaCl2) were each added to 0.5 ml PPP covering a 1-cm2 sample of blood sac PUU. Samples were incubated for 2 hours at 37ºC, rinsed three times with phosphate-buffered saline (PBS), fixed with 1% PFA for 1 hour at 4ºC, and rinsed three times with PBS. A positive control of bovine platelets was also prepared by plating PRP diluted with PPP to approximately physiologic concentration on a glass coverslip. The platelets were fixed with 1% PFA for 1 hour at 4ºC. All controls were dried with an ethanol series as described in the following section for SEM.

LVAS Blood Sacs.

The blood sac analysis was divided into eight regions as illustrated in Figure 1. Two square samples approximately 0.7 cm by 0.7 cm were cut from each region. The samples were placed in the wells of 24-well polystyrene tissue culture plates (VWR, Bridgeport, NJ) and covered with PBS (pH 7.4, 10 mM) until prepared for SEM.

Figure 1.
Figure 1.:
The 70-cc blood sac. The sac was divided into eight regions: inlet, outlet, inlet side, outlet side, center front, center back, top, and bottom.
SEM Sample Preparation and Imaging.

Both in vitro control and retrieved blood sac samples from in vivo studies were prepared for SEM by drying with a 50%, 60%, 70%, 80%, 90%, and 100% ethanol dehydration series. Each sample was soaked in 2 ml of each ethanol concentration for 10 minutes. The 100% ethanol was aspirated and samples were left to air dry. Samples were mounted on SEM stubs using carbon tape and coated with a 10-nm-thick gold layer or with an equivalent gold-palladium coating. Imaging was performed on a high voltage Philips XL-20 SEM (Eindhoven, Netherlands) and an AmRay 3200 EcoSEM (Bedford, MA). To obtain images representative of the sac sample’s surface, a minimum of six randomly selected areas at 500X and one 100X image were taken. Additional images at higher magnifications were taken as needed to examine specific features.

Immunofluorescence Assay.

Immunofluorescent labeling of platelets and fibrin was performed to confirm the identity of structures observed by SEM. A sample and its negative control of a bovine PRP clot were prepared. PRP clots were prepared following a protocol similar to that described previously for PPP clots. In a 24-well tissue culture polystyrene plate (VWR), 1 ml of diluted PRP was added to each of two wells. To each well, a total volume of 15 μl of 1 M CaCl2 was added. Samples were incubated for 2 hours at 37ºC, rinsed three times with PBS, fixed with 1% PFA for 1 hour at 4ºC, and rinsed three times with PBS. Retrieved sac samples were also prepared. Two samples (approximately 0.7 cm by 0.7 cm) were cut from each of three of the nonanticoagulated sacs retrieved from calves. The samples were covered with PBS in 24-well tissue culture polystyrene plates (VWR).

Samples of the PRP clots and retrieved sacs were immunofluorescently labeled for platelets and fibrinogen. Antibody solution volumes and incubation times for PRP clots differed slightly from the following given for the sac samples. PBS was replaced with 500 μl 1% rabbit serum albumin (Sigma, Milwaukee, WI) as a blocking agent. The primary antibodies, 0.5 μl goat anti-bovine fibrinogen (American Diagnostica, Stamford, CT) and 0.75 μl mouse anti-bovine αIIbβ3 (Veterinary Medical Research and Development, Pullman, WA), were added to one sample from each sac. The other sample from each sac was prepared as a negative control by excluding the primary antibody to measure nonspecific labeling of the sample. The samples were incubated overnight at 4ºC. The antibody solution was aspirated and samples were rinsed three times with approximately 2 ml of PBS. Samples were covered with 500 μl 1% rabbit serum albumin. Secondary antibodies were added to each well in the following amounts: 0.5 μl donkey anti–goat IgG-fluorescein isothiocyanate (Jackson ImmunoResearch Laboratories, West Grove, PA) and 5 μl donkey anti–mouse IgG-phycoerythrin (Jackson ImmunoResearch Laboratories). Samples were incubated for 1 hour at 37ºC while protected from light. The antibody solution was aspirated and samples were rinsed with approximately 2 ml distilled water three times.

Each sample was placed on a glass microscope slide in the middle of four small pieces of PUU glued to the slide with cyanoacrylate. A drop of Gel/Mount antifade mounting medium (Biømeda, Foster City, CA) was placed on the sample. Cyanoacrylate was applied onto the small pieces of PUU and a coverslip was placed over the calf sac sample, and held in place at all four corners with the cyanoacrylate on the PUU pieces. The small PUU pieces were used to hold the coverslip in place over the thick PUU calf sac sample that otherwise would not be held by the mounting medium alone. For the bovine PRP clots, no PUU pieces were used to mount the coverslip.

Confocal Microscopy.

Samples were imaged using the 488-nm and 543-nm lasers on a TCS SP2 AOBS confocal microscope (Leica, Deerfield, IL) with a 63X oil immersion objective. Negative controls were imaged for nonspecific and background fluorescence, followed by imaging of the samples labeled with primary and secondary antibodies.

Results

Clinical Observations

The device for each implant operated normally. The average flow rates for the nine calves were maintained between 5.13 and 6.46 l/min, with standard deviations ranging from 0.74 to 1.56 l/min. The flow rates, however, covered a wider range with values as low as 0.50 l/min and as high as 8.80 l/min (Table 1). Many of the minimum values recorded occurred as the pump was turned on the first day of implantation. Few measurements above 8.0 l/min were recorded. The average PT times of the group 1 anticoagulated calves were successfully maintained between 1.5 to 2 times the baseline values, although occasional deviations from the desired range were seen (Table 2). The implant duration was 30 days for eight of the nine implants. One implant was removed at 28 days at which time the calf was in good health and the device was operating without complications.

Table 1
Table 1:
Average Blood Flow Rates, Standard Deviations, and Minimum and Maximum Values Recorded for Anticoagulated and Nonanticoagulated Calves
Table 2
Table 2:
Average Prothrombin Times (PTs), Standard Deviations, and Minimum and Maximum Values Recorded for Anticoagulated and Nonanticoagulated Calves

Macroscopic Thrombosis Observations

Overlay images of macroscopic thrombosis mappings for each group of anticoagulated and nonanticoagulated calf blood sacs are shown in Figure 2. Little macroscopic thrombus formation was seen in the sacs with the exception of one sac from each group. Thrombosis in the inlet and outlet ports was minimal, occurring in only two of the nine blood sacs. In the anticoagulated blood sacs, one sac had a relatively large number of thrombi and the remaining three sacs had few or no thrombi (Figure 2A). Similar results were obtained for the five nonanticoagulated calf blood sacs of group 2 (Figure 2B). One sac had a relatively large number of thrombi and the remaining sacs had few or no thrombi. Detailed observations for each sac are listed in Table 3. No major difference in macroscopic thrombus formation was observed between blood sacs from groups 1 and 2. In all blood sacs, no red thrombus was observed. The total sac surface coverage by macroscopic thrombi was 0.23% and 0.16% for anticoagulated and nonanticoagulated sacs, respectively (Table 4).

Figure 2.
Figure 2.:
Macroscopic mapping of thrombi. (A) Overlay of the four anticoagulated blood sacs. Thrombi are numbered with the sac numbers. Thrombi on sac 4 with the most thrombi are unlabeled to provide greater visibility of thrombus distribution. (B) Overlay of four nonanticoagulated blood sacs (mapping of thrombi for first sac not included, see text). The unlabeled thrombi were observed on sac 3.
Table 3
Table 3:
Observations of Macroscopic Thrombosis On Blood Sacs from Groups 1 and 2
Table 4
Table 4:
Percentage of Sac Surface Coverage by Macroscopic Thrombi

Microscopic Thrombosis Observations

Confocal Microscopy

To demonstrate the presence of biologic structures on the retrieved sacs, immunofluorescent labeling and confocal microscopy were used. Images showing bovine fibrin and platelets are shown in Figure 3. All negative controls showed minimal nonspecific labeling. Figures 3A and 3B show the structures of bovine fibrin and platelets, respectively, in PRP clots, and confirm successful labeling of the structures. Similarly, fibrin and platelet structures were observed on the nonanticoagulated calf sac samples (Figures 3C and 3D).

Figure 3.
Figure 3.:
Confocal microscopy images of fibrin and platelets. (A) Fibrin in a bovine PRP clot. (B) Bovine platelets in a PRP clot. (C) Fibrin on the surface of a blood sac retrieved from a nonanticoagulated calf. (D) A layer of fibrin and a platelet covered with fibrinogen on the surface of a blood sac retrieved from a nonanticoagulated calf. Solid arrows: fibrin. Dotted arrow lines: platelets.

In Vitro SEM Controls.

The negative control PUU material without biologic contact was relatively smooth with few microscopic features visible by SEM (Figure 4). On one sample, however, in an area approximately 240 μm by 72 μm, a single group of branching strand-like features was observed. This feature was considered to be an anomaly observed on a single image of 102 images taken at 500X on 17 negative control samples. Positive controls of PPP clots on PUU consisted of well-formed fibrin meshes with ridges, folds, and branches (Figure 4B). The positive control of PRP plated on glass showed platelets adhered and spreading (Figure 4C). Similar images from the controls were used to identify structures observed on the blood sacs retrieved from calves.

Figure 4.
Figure 4.:
SEM control images. (A) Representative negative control of PUU with no biologic contact consists of a primarily smooth surface (imaged at 500X). (B) Positive control of a PPP clot with fibrin mesh illustrates the topography of fibrin observed in several of the blood sacs retrieved from calves (500X). (C) Positive control of PRP plated on glass. Structures resembling the bovine platelets shown here were observed on most blood sacs (5,000X). Solid arrows: fibrin. Dotted arrow lines: platelets.

LVAS Blood Sacs.

Microscopic evaluation by SEM provided information on the topography of the different blood sac regions after 30 days implantation in calves. An array of structures such as circular masses from 1 μm up to 10 μm in diameter and branching fibrils was observed on the surfaces of the anticoagulated and nonanticoagulated blood sacs (Figure 5). Rough topography as shown in Figure 5A was frequently observed throughout the sacs. The small size of these structures and their uniformity across the surface suggested such topography resulted from the formation of a protein layer. Other features resembled unspread and fully spread platelets (Figures 5B and 5C) and fibrin (Figure 5D) seen in the positive controls. These structures also closely resemble the platelets and fibrin imaged by confocal microscopy, and will be referred to as platelets and fibrin from hereon. The dark circles in Figure 5C were observed frequently and are believed to be pits in the PUU possibly due to polymer composition and fabrication process, although degradation is a possibility as suggested in studies investigating the stability of similar polymers.42,43 Features resembling a protein layer and unspread platelets were most abundant and observed in several of the anticoagulated and nonanticoagulated sacs.

Figure 5.
Figure 5.:
Array of topography observed on surfaces of explanted PUU blood sacs. Topography resembling proteinaceous layers (A), unspread platelets (B), spread platelets (C), and fibrin (D), as seen in positive controls, were observed in the explanted blood sacs. Dotted arrow lines show platelets.

In sacs 1, 2, and 4 of the anticoagulated group, rough proteinaceous topography similar to Figure 5A was observed in most regions. In sac 1, a low to medium density of platelets as shown in Figure 5B was found in all regions except the outlet. The inlets and outlets were densely covered with the fibrillar features shown in Figure 5D. In sac 2, a few platelets were present at the top and center front, respectively. Similar to sac 1, the inlets and outlets were densely covered with fibrin. On sac 3, few features were observed. Samples from the outlet and center front had low densities of proteinaceous topography. No other relevant structures were observed. Sac 4 had a low density of small proteinaceous structures seen in all regions except the center front. The inlet side had a sparse distribution of features resembling spread platelets. No fibrillar structures were observed.

All five nonanticoagulated group 2 sacs had low to high densities of rough topography, suggesting formation of protein layers in all or most regions. Sac 1 had features resembling unspread platelets in all regions except the inlet. The ports were covered with structures resembling fibrin. In sac 2, the majority of such proteinaceous topography was observed in the inlet, outlet, center back, and top. Platelets about 5 μm in diameter were observed in the centers, sides, and bottom of the sac in medium to high densities. Fibrillar structures were observed mainly in the inlet and to a lesser extent in the outlet side, top, and bottom. In sac 3, the surface appeared to be coated with a protein layer throughout the sac. Unspread platelets were observed at a low density in the inlet and outlet sides. However, structures resembling fully spread platelets were observed in all regions except the inlet and outlet. No fibrin was observed. In sac 4, the highest density of proteinaceous topography was observed in the inlet and outlet. Unspread platelets were observed on some samples from the ports, center front, outlet side, and top. No spread platelets were observed. Fibrillar structures were observed on one sample from the top of the sac. Sac 5 had low to high densities of proteinaceous topography throughout the sac. Platelets were observed on one sample from each of the inlet, center back, and inlet side, and on both samples from the top and bottom regions at low densities. Fibrillar structures were observed in all samples from the inlet and outlet but not in any other region.

Anticoagulation of calves from group 1 resulted in differences in microscopic observations between groups 1 and 2. Figure 6 compares representative SEM images showing the surfaces from different regions of anticoagulated and nonanticoagulated blood sacs. The inlets and outlets of both groups had fibrin deposition (Figures 6A–D). Fibrin was observed in one or both ports of three of four anticoagulated and three of five nonanticoagulated blood sacs. In all other regions of the sac, this structure resembling fibrin was rarely seen. Only one of the nine sacs had fibrin in regions other than the ports. Deposition of unspread and spread platelets differed between the two groups. The platelets were observed more frequently in group 2 nonanticoagulated blood sacs compared to group 1 anticoagulated sacs (Figures 6E–P). Only one out of four anticoagulated sacs showed platelets in most regions, whereas all five nonanticoagulated sacs had at least moderate coverage by either spread or unspread platelets. Anticoagulation given to group 1 calves appeared to hinder platelet adhesion but not fibrin formation, whereas in group 2 (nonanticoagulated calves), platelet adhesion was abundant.

Figure 6.
Figure 6.:
Effects of anticoagulation on surface topography of blood sacs. Surface topographies of the eight different regions examined are shown. Surface topography of the anticoagulated sacs (images on left) differed from that of the nonanticoagulated sacs (images on right). Platelet-like features were scarce in all regions of the anticoagulated blood sacs. In contrast, in nonanticoagulated sacs such platelet-like structures were present in abundance in all regions except for the inlets and outlets. However, in both groups of sacs the primary structures observed in the inlets and outlets were branching features similar in structure to fibrin clots. Solid arrows: fibrin. Dotted arrow lines: unspread and spread platelets.

The location within the blood sac also appeared to affect the observed surface topography. Figure 7 shows representative SEM images from the eight sampled regions of the anticoagulated calf blood sac 2, demonstrating the location-dependent surface topography. Fibrin was observed primarily on the inlets and outlets of the blood sacs, as seen in Figures 7A and 7C. Such structures were absent in all other regions (Figures 7B and 7D–H). Similar location-dependent topography was observed in the nonanticoagulated calf sacs as shown in Figure 8 for sac 5. Fibrin was the main feature that appeared to be location-dependent in all sacs, with these structures being observed almost exclusively in the inlets and outlets of four of the nine sacs. Unspread and spread platelets were least abundant in the inlets and outlets. The differences observed in surface topography with sac location may be a result of differing fluid dynamics between regions.

Figure 7.
Figure 7.:
Location-dependent surface topography in blood sac 2 after 30 days in calves with anticoagulation. The main difference was observed in the inlet and outlet ports of several blood sacs. In these regions, fibrin-like structures were seen most abundantly and few platelet-like structures were present.
Figure 8.
Figure 8.:
Location-dependent surface topography in nonanticoagulated blood sac 5 after 30 days in calves. In the inlet and outlet ports, fibrin-like structures were the dominating topography. Platelet-like structures were observed in other regions such as the top of the sac between the ports (B).

Discussion

Precise control of the flow rates in the pumps was difficult because of the presence of a healthy native left ventricle of the calf as well as changes in flow through the pump with movement of the animal from standing to lying down. The effect that the low and high transient blood flow rates may have on thrombogenesis in vivo is unclear. It has not been established whether the transient high flow rates are enough to affect thrombosis at the surface. It is unknown if the higher flows can wash biologic deposition that may occur during periods of lower flow rates. Similarly, the effect of variation in prothrombin times of the anticoagulated calves on thrombosis is unclear. Although the average prothrombin times were maintained well above the baseline times, it is unknown if the few abnormal prothrombin times negate the effectiveness of the anticoagulation therapy. The consequences of low transient flow rates and prothrombin times on chronic thrombosis are topics for further investigation.

Evaluation of the anticoagulated and nonanticoagulated sacs showed variations in thrombus formation with both anticoagulation and location within the sac. At the macroscopic level, no difference in the number and size of macroscopic thrombi was observed between anticoagulated and nonanticoagulated blood sacs. However, at the microscopic level, features resembling platelets were observed less in anticoagulated sacs. In the high shear stress ports, observations were similar for both anticoagulated and nonanticoagulated groups at both macroscale and microscale. Anticoagulation administered in these studies appeared to hinder platelet adhesion but did not reduce the formation of the fibrillar structures. The anticoagulation of group 2 calves with heparin and warfarin sodium should have reduced fibrin formation and perhaps indirectly, reduced platelet adhesion. Heparin binds to antithrombin III, which then binds to thrombin, and inactivates thrombin, a key factor in the common pathway of the coagulation cascade. Warfarin sodium inhibits synthesis of the vitamin K-dependent coagulation factors VII, IX, X, and prothrombin, which taken together are involved in all three pathways of the coagulation cascade. Heparin and warfarin sodium appeared to have no effect on formation of fibrillar structures in the high shear stress ports but did reduce platelet adhesion in the sacs. Regional differences were observed at both the macroscopic and microscopic scales, with the high shear stress inlet and outlet ports differing from all other regions. At the macroscopic level, fewer thrombi were recorded in the ports. At the microscopic level, features resembling fibrin were observed almost exclusively in the ports and platelet-like features were common in the lower shear stress regions. Such differences in observations between study groups and with sac location may be related to the fluid dynamics in the sacs.

The microscopic observations by SEM indicating region-dependent topography can be compared to region-dependent shear stress measurements reported for the adult 70-cc sac16 and a 50-cc blood sac.20 The shear stress measurements in the 70-cc blood sac published by Baldwin et al.16 under similar operating conditions to our in vivo studies concluded there was a difference in Reynold’s shear stress magnitudes in the inlet and outlet compared to the rest of the sac. They showed that the shear stresses in the inlet and outlet regions by the valves were 1,000 dynes/cm2 or less, and even higher during regurgitant jets. All other regions within the blood sac had shear stresses below 500 dynes/cm2. Our study showed regional differences in macroscopic thrombi and microscopic surface topography corresponding with differences in flow found by Baldwin and colleagues, with fibrin observed primarily in regions of shear stresses as high as 1,000 dynes/cm2 and platelets adhered to surfaces with shear stresses below 500 dynes/cm2. As mentioned previously, adhesion of platelets to the surface was reduced by anticoagulation in the lower shear stress regions of < 500 dynes/cm2. However, formation of fibrin in the higher stress regions of the inlets and outlets remained unaffected by anticoagulation. These results suggest that the differences in platelet and fibrin adhesion may be caused by varied fluid dynamics in the blood sacs.

The shear stress reported by Baldwin and colleagues to occur in the adult LVAS are considerably above the values at which thrombus deposition normally occurs. A range of critical shear rates from 1,000 to 2,000 s–1, corresponding to shear stresses from 40 to 80 dynes/cm2 in blood, has been reported above which minimal platelet adhesion occurs.11,12,26 Maximum platelet surface coverage has been reported to occur at shear rates between 50 to 500 s–1, corresponding to shear stresses from 2 to 20 dynes/cm2 in blood, and similarly fibrin deposition is thought to decrease with increased shear rate.12,26 The shear stresses in the adult 70-cc blood sac are believed to be high enough to provide effective washing of the surface thereby limiting thrombus deposition.16 However, the in vivo results presented here suggest differently, with thrombus formation at the surface occurring even at shear stresses at least six times greater than the shear rate literature reports it being inhibited.

Microscale evaluation in this work consisted of SEM imaging, which provides topographic information. It does not conclusively identify platelets and fibrin and, therefore, should be supplemented with techniques such as immunolabeling to confirm the identity of thrombus components. However, the use of extensive controls both positive and negative, preliminary confocal images, two in vivo groups with several animals each, and intense surface sampling, should provide ample information for identification of the structures observed by SEM.

Conclusion

This study illustrates the need for thorough in vivo evaluation of thrombosis in LVAS beyond the macroscopic into the microscale for comparison of regions with varied flow dynamics. Differences between animals that received anticoagulation and those that did not were seen primarily in increased numbers of platelets in the nonanticoagulated group. Results showed strong location dependence within the blood sac at both macroscale and microscale. There appear to be correlations between fluid dynamics and macroscopic/microscopic thrombi within the sac even at these higher-than-physiologic shear stresses. Further investigation of the relationship between thrombosis and levels of shear stress may allow us to establish new parameters for reduction of thrombosis in LVAS.

Acknowledgments

The authors thank Dr. Walter Pae, Dr. Ben Sun, Dr. Tigran Khalapyan, Allen Prophet, George Felder, John Reibson, Bryan Fritz, Karen Bussard, Barbara Bush, and Brad Doxtater for their assistance with animal implant, care, and device retrieval. Financial support was provided by NIH (R01HL60276-01A1and N01-HV-58156).

References

1. Kochanek KD, Murphy SL, Anderson RN, Scott C: Deaths: Final data for 2002. Natl Vital Stat Rep 53: 1–116, 2004.
2. 2003 Annual Report of the U.S. Organ Procurement and Transplantation Network and the Scientific Registry of Transplant Recipients: Transplant Data 1993-2002, 2003.
3. Dembitsky WP: The influence of ASAIO-based cardiovascular technology on patients and society. ASAIO J 50: xxiiii–xxviii, 2004.
4. Stevenson LW, Miller LW, Desvigne-Nickens P, et al: Left ventricular assist device as destination for patients undergoing intravenous inotropic therapy: A subset analysis from REMATCH (Randomized Evaluation of Mechanical Assistance in Treatment of Chronic Heart Failure). Circulation 110: 975-981, 2004.
5. Piccione Jr W: Left ventricular assist device implantation: Short and long-term surgical complications. J Heart Lung Transplant 19: S89–94, 2000.
6. Hampton CR, Verrier ED: Systemic consequences of ventricular assist devices: alterations of coagulation, immune function, inflammation, and the neuroendocrine system. Artif Organs 26: 902–908, 2002.
7. Portner PM, Jansen PG, Oyer PE, et al: Improved outcomes with an implantable left ventricular assist system: A multicenter study. Ann Thorac Surg 71: 205–209, 2001.
8. Pennington DG, Oaks TE, Lohmann DP: Permanent ventricular assist device support versus cardiac transplantation. Ann Thorac Surg 68: 729–733, 1999.
9. Lazar RM, Shapiro PA, Jaski BE, et al: Neurological events during long-term mechanical circulatory support for heart failure: The Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) experience. Circulation 109: 2423–2427, 2004.
10. Wagner WR, Schaub RD, Sorensen EN, et al: Blood biocompatibility analysis in the setting of ventricular assist devices. J Biomater Sci Polym Ed 11: 1239–1259, 2000.
11. Balasubramanian V, Slack SM: The effect of fluid shear and co-adsorbed proteins on the stability of immobilized fibrinogen and subsequent platelet interactions. J Biomater Sci Polym Ed 13: 543–561, 2002.
12. Savage B, Saldivar E, Ruggeri ZM: Initiation of platelet adhesion by arrest onto fibrinogen or translocation on von Willebrand factor. Cell 84: 289–297, 1996.
13. Horbett TA: Principles underlying the role of adsorbed plasma-proteins in blood interactions with foreign materials. Cardiovascular Pathology 2: S137–S148, 1993.
14. Daily BB, Pettitt TW, Sutera SP, Pierce WS: Pierce-Donachy pediatric VAD: progress in development. Ann Thorac Surg 61: 437–443, 1996.
15. Bachmann C, Hugo G, Rosenberg G, et al: Fluid dynamics of a pediatric ventricular assist device. Artif Organs 24: 362–372, 2000.
16. Baldwin JT, Deutsch S, Geselowitz DB, Tarbell JM: LDA measurements of mean velocity and Reynolds stress fields within an artificial heart ventricle. J Biomech Eng 116: 190–200, 1994.
17. Okamoto E, Hashimoto T, Inoue T, Mitamura Y: Blood compatible design of a pulsatile blood pump using computational fluid dynamics and computer-aided design and manufacturing technology. Artif Organs 27: 61–67, 2003.
18. Burgreen GW, Antaki JF, Wu ZJ, Holmes AJ: Computational fluid dynamics as a development tool for rotary blood pumps. Artif Organs 25: 336–340, 2001.
19. Sorensen EN, Burgreen GW, Wagner WR, Antaki JF: Computational simulation of platelet deposition and activation: I. Model development and properties. Ann Biomed Eng 27: 436–448, 1999.
20. Hochareon P, Manning KB, Fontaine AA, et al: Wall shear-rate estimation within the 50cc Penn State artificial heart using particle image velocimetry. J Biomech Eng 126: 430–437, 2004.
21. Peppas NA, Langer R: New challenges in biomaterials. Science 263: 1715–1720, 1994.
22. Slater JP, Rose EA, Levin HR, et al: Low thromboembolic risk without anticoagulation using advanced-design left ventricular assist devices. Ann Thorac Surg 62: 1321–1328, 1996.
23. Hunt SA, Frazier OH: Mechanical circulatory support and cardiac transplantation. Circulation 97: 2079–2090, 1998.
24. McBride LR, Naunheim KS, Fiore AC, et al: Clinical experience with 111 Thoratec ventricular assist devices. Ann Thorac Surg 67: 1233–1239, 1999.
25. Minami K, El-Banayosy A, Sezai A, et al: Morbidity and outcome after mechanical ventricular support using Thoratec, Novacor, and HeartMate for bridging to heart transplantation. Artif Organs 24: 421–426, 2000.
26. Colman RW, Hirsh J, Marder VJ, Clowes AW, George JN (eds): Hemostasis and Thrombosis: Basic principles and clinical practice. Philadelphia, Lippincott Williams & Wilkins, 2000.
27. Nojiri C, Okano T, Jacobs HA, et al: Blood compatibility of PEO grafted polyurethane and HEMA/styrene block copolymer surfaces. J Biomed Mater Res 24: 1151–1171, 1990.
28. Park K, Mao FW, Park H: Morphological characterization of surface-induced platelet activation. Biomaterials 11: 24–31, 1990.
29. Jahangir AR, McClung WG, Cornelius RM, et al: Fluorinated surface-modifying macromolecules: Modulating adhesive protein and platelet interactions on a polyether-urethane. J Biomed Mater Res 60: 135–147, 2002.
30. Park K, Shim HS, Dewanjee MK, Eigler NL: In vitro and in vivo studies of PEO-grafted blood-contacting cardiovascular prostheses. J Biomater Sci Polym Ed 11: 1121–1134, 2000.
31. Schaub RD, Kameneva MV, Borovetz HS, Wagner WR: Assessing acute platelet adhesion on opaque metallic and polymeric biomaterials with fiber optic microscopy. J Biomed Mater Res 49: 460–468, 2000.
32. Slack SM, Turitto VT: Fluid dynamics and hemorheologic considerations. Cardiovasc Pathol 2: 11S–21S, 1993.
33. Aldenhoff YB, Blezer R, Lindhout T, Koole LH: Photo-immobilization of dipyridamole (Persantin) at the surface of polyurethane biomaterials: Reduction of in-vitro thrombogenicity. Biomaterials 18: 167–172, 1997.
34. Frank RD, Dresbach H, Thelen H, Sieberth HG: Glutardialdehyde induced fluorescence technique (GIFT): A new method for the imaging of platelet adhesion on biomaterials. J Biomed Mater Res 52: 374–381, 2000.
35. Alanazi A, Nojiri C, Noguchi T, et al: Improved blood compatibility of DLC coated polymeric material. ASAIO J 46: 440–443, 2000.
36. Schenk S, Weber S, Luangphakdy V, et al: In vivo performance and biocompatibility of the MagScrew ventricular assist device. ASAIO J 49: 594–598, 2003.
37. Ichikawa S, Linneweber J, Motomura T, et al: In vivo evaluation of the NEDO biventricular assist device with an RPM dynamic impeller suspension system. ASAIO J 49: 578–582, 2003.
38. Keogh JR, Wolf MF, Overend ME, et al: Biocompatibility of sulphonated polyurethane surfaces. Biomaterials 17: 1987–1994, 1996.
39. Pierce WS, Snyder AJ, Rosenberg G, et al: A long-term ventricular assist system. J Thorac Cardiovasc Surg 105: 520–524, 1993.
40. Thoratec Laboratories Corporation. Anticoagulation Protocol 15579 (Rev. A). Pleasanton, CA, Thoratec Laboratories Corporation, 2004.
41. National Institutes of Health. Guide for the Care and Use of Laboratory Animals. Bethesda, U.S. Government Printing Office (NIH publication No. 86-23), 1985.
42. Schubert MA, Wiggins MJ, Schaefer MP, et al: Oxidative biodegradation mechanisms of biaxially strained poly(etherurethane urea) elastomers. J Biomed Mater Res 29: 337–347, 1995.
43. Wu L, Weisberg DM, Runt J, et al: An investigation of the in vivo stability of poly(ether urethaneurea) blood sacs. J Biomed Mater Res 44: 371–380, 1999.
Copyright © 2005 by the American Society for Artificial Internal Organs