Blood-contacting medical devices may cause thrombosis and thromboembolic related complications in patients because of the induction of nonphysiologic blood flow patterns and the exposure of blood to foreign materials. Both long term and short term use devices can cause thrombosis, resulting in patient morbidity or death.1–7 Preclinical thrombogenicity evaluation is important for ensuring the safety of blood-contacting devices and biomaterials and is generally needed for regulatory approval/clearance of new devices.8,9 One commonly used thrombogenicity evaluation method is the in vivo nonanticoagulated venous implant (NAVI) assay.10 This test involves implanting a device with a simple geometry (e.g., catheter) and a comparator device into separate veins of a large animal. The implants are left in situ for up to 4 hr before they are explanted, and then a thrombogenicity score is assigned based on the extent of thrombi that adhered to the devices’ surfaces.8,11 There are many challenges in performing a NAVI study. The implant technique, the device placement, and the size of the device as compared with the diameter of the vessel may all affect thrombus formation and lead to inconsistent thrombogenicity predictions.8,11 The inherent differences in the blood coagulability of individual animals can cause large variances in the test results, making it difficult to reproducibly discern the thrombogenicity potential of the device. In addition, this type of in vivo test is expensive and requires the animals to be sacrificed.
Several in vitro flow loop test systems have been developed to attempt to resolve some of the issues associated with the NAVI assay.11–21 However, there is currently no standardized or widely accepted in vitro test protocol to evaluate device thrombogenicity, as the test protocols of the existing flow loop systems differ greatly. For example, the duration of the test, the blood temperature, and the species of blood utilized vary substantially among different test systems. These systems are also susceptible to variability because of inherent differences between blood donors, even for the same species. We aim to reduce these discrepancies by standardizing a static control pretest (to evaluate and compensate for native blood coagulability) and a dynamic flow test loop to develop a reliable in vitro method to evaluate the thrombogenicity of medical devices and biomaterials.
Materials and Methods
Blood Preparation and Standardized Static Pretest Using Latex Tubes
There were six separate test days in total for this study. On each test day, sheep blood (Lampire Biologic Laboratories Inc, Pipersville, PA) from a live donor was drawn into containers with Anticoagulant Citrate Dextrose Solution A (ACDA, the volume ratio of ACDA to whole blood was 15:85), shipped to the testing laboratory overnight, and testing was completed within 24–36 hr of the blood draw. Immediately before starting each dynamic flow test, the blood was recalcified (by adding an appropriate amount of 2 M CaCl2 solution to the blood to obtain a targeted CaCl2 concentration of 13 mM in whole blood) and heparinized to a donor-specific concentration. Similar CaCl2 concentrations have been used in previous studies to recalcify whole blood.22,23
The heparin concentration used for each blood donor was determined using a static pretest. In the pretest, uniform latex tubes (4 cm length, 1.6 mm inner diameter (ID), and 4.8 mm outer diameter (OD), catalog number S50616A; Fisher Scientific, Hampton, NH) were incubated in recalcified blood (8 cm2/ml latex surface area to blood volume) under a series of heparin concentrations (1.0–2.4 unit/ml, 0.2 unit/ml increments) for 30 min at room temperature (Figure 1A). During incubation, the samples were gently agitated using a shaking water bath (Edvotek 10 L Digital Shaking Waterbath; Edvotek Inc., Washington, DC) set at 60 rpm. Based on preliminary data that correlated the test results from the static pretest to the dynamic test, the minimum heparin level that resulted in a thrombus surface coverage ≤ 10% on the latex during the pretest was selected as the threshold concentration. The initial concentration used to start the flow loop testing was selected to be 0.2 unit/ml less than the threshold concentration. For example, if the threshold concentration was 1.8 unit/ml, then the initial concentration would be 1.6 unit/ml for the flow loops (Figure 1B).
Activated clotting time (ACT) was also measured to determine if ACT could be used to predict donor-specific heparin concentrations. However, the large variability in the resulting clotting times (ranged from 488 to 609 s for final heparin concentrations) suggested that it would be difficult to predict appropriate heparin values using only the ACT level. More information on the ACT assessment is provided in the Supplemental Digital Content 1 (see Supplemental Digital Content, https://links.lww.com/ASAIO/A393).
Test Materials and Flow Loop Assembly
Nine materials with outer diameters of 2.1–3.2 mm were tested (Table 1): a negative control polytetrafluoroethylene (PTFE), a positive control latex,24 and seven biomaterials commonly used in blood-contacting devices. The biomaterials utilized in this study were polyvinyl chloride (PVC), two types of silicones with different formulations (Q-Sil and V-Sil), nylon, polyurethane (PU), high-density polyethylene (HDPE), and polyether block amide (PEBAX). Except for PTFE, which was available as a solid cord, all the other test materials were available as small-bore tubing. Both ends of the tubing were plugged with a PTFE cord to prevent blood flow into the lumen of the tubing. Polytetrafluoroethylene was chosen as the negative control because it has been used in many blood-contacting devices and exhibits reasonable long-term thromboresistance.25,26
The dynamic flow loop consisted of a reusable section of PVC tubing (45 cm long, 6.4 mm ID), that was used for all tests conducted with the same donor blood, inserted into a roller pump and connected to a test section via 6.4 mm OD straight polypropylene connectors. The test section (Figure 2A) of the flow loop was prepared by introducing a 12 cm long test material into the lumen of PVC tubing (32 cm long, 6.4 mm ID), through a small cut in the PVC tubing wall, and the incision site was sealed with Parafilm. One test sample was used in each blood flow loop. Using a multichannel roller pump, two loops were run simultaneously, and the loops were positioned so that the test sections remained relatively straight (Figure 2B). To maintain similar flow rates within the loops throughout the experiments, the pump tubing sections were exercised with phosphate buffered saline (PBS) for 30 min at 70 rpm and the occlusion pressure of each roller pump head was set to 150 mm Hg (at 50 rpm) before the start of the blood circulation experiments.
Whole blood (26 ml) was added to the flow loop and circulated at room temperature (21 ± 2°C) at a flow rate of 200 ml/min (65 rpm) for each of the test samples (Figure 2B). This flow rate was selected to produce a physiologic venous shear rate of approximately 130/s. After 1 hr circulation, the loops were drained and gently rinsed with PBS at a reduced pump speed of 50 rpm to avoid thrombus dislodgement from the test samples. The PVC tubing test sections were removed from the loops and dissected with surgical scissors. The test samples were then carefully removed from the tubing, photographed, and visually inspected for thrombus deposition. Before starting the next test, the reusable pump-head tubing sections of the loops were vigorously rinsed by recirculating PBS for around 3 min at the highest rpm setting of the pump (flow rate of approximately 2 L/min). The loops were then drained, and a swab was used to remove any visible thrombi that were attached to the inner lumen of the PVC tubing, before performing a final rinse with PBS.
To verify that the heparin concentration selected from the static pretest would provide appropriate anticoagulation for each blood pool, a set of control materials (negative control: PTFE and positive control: latex) were investigated first using the dynamic test loops. If the thrombus surface coverage produced on the PTFE material was ≤ 10% and on the latex was > 50% (classified as severe thrombus according to NAVI scoring scheme in the ISO 10993-4 standard8), the initial heparin concentration was used for the remainder of the experiment. Although the NAVI scoring scheme categorizes 1–25% thrombus surface coverage as minimal thrombus,8 a lower thrombus surface coverage threshold (10%)11 on the negative control was used for this study to reduce the variance of the test and to allow for a larger percentage differentiation between the positive and negative control materials. If the deposition was > 10% on the PTFE material or < 50% on the latex, the heparin concentration was increased or decreased by 0.2 unit/ml, respectively. This verification process was repeated until the thrombus surface area coverage on the control materials tested in the dynamic loops was within the acceptance criteria range (PTFE ≤ 10% and latex ≥ 50%). Once an acceptable heparin concentration was determined, the remaining test materials were investigated in a random order. Approximately 25% of the received donor blood (two out of eight blood pools in this study) had to be discarded because an acceptable heparin concentration could not be identified.
Data Collection and Statistical Analysis
To characterize thrombogenicity of the test samples, the percent thrombus surface coverage, thrombus weight, and platelet count reduction were measured. To obtain the thrombus-covered surface areas of the test samples, the lengths of the thrombi were measured using a ruler, and their widths were determined by estimating the fraction of the material’s circumference covered by the thrombi. Then the percent thrombus surface coverage was calculated by the following equation:
To estimate the thrombus weight after the dynamic flow loop testing, the test materials were dried overnight (≥ 12 hr at room temperature), and their weights were measured using an analytical balance (XS64 Analytical Balance; Mettler Toledo, Columbus, OH). The original test material weight, that was collected before insertion in the flow loop, was subtracted from the post-test measurement to calculate the dry weight of the adherent thrombi. To account for the diameter difference in the test materials, the dry thrombus weight was then divided by the blood-contacting surface area of the test material to obtain a normalized value (mg/cm2). Blood platelet counts were measured before and after the 1 hr circulation using a complete blood cell counter (Hemavet 950 FS, Drew Scientific Inc., Miami Lakes, FL) to determine the extent of platelet count reduction during the test. The data were statistically analyzed using a Friedman repeated measures analysis of variance on ranks with a Newman–Keuls post hoc test. The results were considered statistically significant if p < 0.05.
Based on the static thrombosis pretest assessment of latex tubing, the donor-specific final heparin concentrations utilized in the test flow loops ranged from 1.0 to 1.8 unit/ml. It was observed that a change in concentration of 0.4 unit/ml of heparin would have a substantial effect on the thrombus deposition (Figure 3), and a small number of donors were sensitive to an adjustment of 0.2 unit/ml of heparin. For all the dynamic loop tests in which the blood samples met the thrombus-coverage acceptance criteria of the positive and negative controls, the static pretest had accurately predicted the useable final heparin concentration within ± 0.2 unit/ml.
Representative images of thrombus deposition on the test materials after 1 hr of blood circulation with donor-specific heparinization are shown in Figure 4. The latex was significantly more thrombogenic than all the other materials, with an average thrombus surface coverage of greater than 70% (p < 0.01) (Figure 5). The PTFE negative control material and most of the biomaterials had little to no thrombus deposition with an average of less than 5% of their blood-contacting surface areas covered. The PVC and one of the silicones (Q-Sil) exhibited intermediate thrombogenicity with significantly more surface coverage than PTFE and the other biomaterials (p < 0.05), but significantly less thrombus surface coverage than latex (p < 0.01). Although the other silicone material (V-Sil) also had significantly greater thrombus surface coverage than PTFE, PU, Nylon, and HDPE (p < 0.01), we do not categorize V-Sil as an intermediate thrombogenic material because its average thrombus surface coverage was minimal (< 10%) and there was no statistical difference for the other two thrombogenicity markers, thrombus weight, and platelet count reduction.
The thrombus deposition on the latex was observed to be thicker than on the other materials. This is shown quantitatively in the thrombus weight graph (p < 0.05) (Figure 6). The thrombi on the latex had an average normalized weight of 10.8 mg/cm2, whereas thrombi on PTFE and the biomaterials (except PVC and Q-Sil) were almost unmeasurable (< 0.3 mg/cm2). The weight of the thrombus deposition on PVC and Q-Sil (0.9–1.0 mg/cm2) was much less than the latex but significantly greater than that of the other biomaterials and PTFE (p < 0.05).
The largest decrease in platelet count occurred in the loop containing the latex, with a greater than 85% platelet count reduction after 1 hr circulation (Figure 7). Polytetrafluoroethylene and most of the biomaterials had significantly lower percent platelet count decreases than latex, with an average reduction of less than 35% (p < 0.05). However, PVC and Q-Sil caused an intermediate platelet count reduction (54% and 48%, respectively) that were significantly lower than for latex (p < 0.05), but moderately higher than most of the other biomaterials and PTFE (p < 0.05). The percent platelet reduction for Q-Sil was not significantly different from that of V-Sil and nylon (p > 0.05).
The thrombus deposition and platelet count reduction data (Figures 5-7) demonstrate that the in vitro blood flow loop test system presented in the current study can repeatedly differentiate between thrombogenic (the latex positive control) and thromboresistant materials (the negative control PTFE, V-Sil, Nylon, PU, HDPE, and PEBAX). Furthermore, Q-Sil and PVC exhibited intermediate thrombogenicity potential in the tests, suggesting that this flow loop test system may be useful for comparing acute thrombogenicity of different materials (latex > PVC ≥ Q-Sil > remaining biomaterials ≥ PTFE). The use of a novel static pretest to help determine donor-specific heparin concentrations for the test loops appears to have reduced donor blood variability and therefore may provide more consistent thrombogenicity predictions. Compared to the in vivo NAVI assay, which is commonly used to evaluate acute thrombogenicity of biomaterials or devices with short-term blood contact, this in vitro test loop may offer several advantages because it: 1) utilizes donor-specific heparin concentrations to reduce donor variability; 2) eliminates surgery related confounding factors; 3) allows for a direct comparison by using the same blood pool to evaluate different materials/devices in a paired test fashion; 4) does not require blood donor animals to be sacrificed; and 5) is less expensive to perform. Thus, the in vitro blood flow loop test system described in this study may be a potential alternative or supplement to the NAVI assay.
Among the three thrombogenicity characterization endpoints used in this study, thrombus surface area coverage is a common thrombogenicity marker used in NAVI studies8 and in some dynamic in vitro assays.11 Although it is a useful tool in assessing the hemocompatibility of blood-contacting devices, it is measured manually and therefore the results may be affected by the subjective judgment of an evaluator. Also, it does not consider the thickness of the thrombi. For some medical devices, thrombi may only form on a small portion of the device’s surface; however, the thickness of those thrombi may be substantial enough to cause a safety concern. For this study, the normalized dry thrombus weight was utilized along with the thrombus surface area coverage to provide a more comprehensive thrombogenicity assessment, as the thrombus weight takes thrombus thickness into account. The reduction in circulating platelets was also included in this study as another measurement of thrombogenicity. For the specific materials tested, a larger reduction in platelet count was associated with a greater thrombus surface area coverage; however, this relationship may not be true for all devices. Previous studies have shown that some potentially thrombogenic materials are nonthromboadherent, or exhibit very little thrombus deposition.27,28 Although these materials prevent thrombus adhesion, some still promote platelet activation and aggregation, increasing the risk of thromboembolic complications.27,28 Measuring the reduction in circulating platelets within this flow system may help to detect thrombogenic devices or materials that are also nonthromboadherent.
Although it is not within the scope of this study to directly compare in vitro test results to that of in vivo animal studies, it has been shown recently that for catheters, similar thrombogenicity test results can be obtained in an in vitro flow loop when compared to a NAVI study. Grove et al.11 performed a direct comparison between an in vitro flow loop assay that used freshly harvested ovine blood with minimal heparinization (approximately 1 unit/ml) and an in vivo ovine NAVI model. Identical positive control, negative control, and test catheters were inserted into the blood flow loop and implanted into the jugular and iliac veins of seven sheep. The resulting thrombogenicity scores produced by both test methods were comparable.11 Although the in vitro flow system developed by Grove et al.11 is a potential alternative to the NAVI assay, it requires the use of fresh donor blood that is drawn within 60 min before starting the test. This is an important limitation because many testing laboratories do not have onsite access to animal blood donors or do not have the funds to maintain an animal facility. In comparison, the in vitro test system developed in this study enables the use of 24 hr postdrawn donor blood purchased from a commercial source, thus increasing the usability and reducing the overall cost of the test.
An important finding of this study was that the blood drawn 24 hr before the start of the experiment was viable for the assessments of thrombus deposition and platelet count reduction, using the current flow loop test system. In the literature, it has been recommended that blood for thrombogenicity testing be used as soon as possible and usually within 4 hr of blood draw.8,29 This study demonstrates that the use of blood older than 4 hr is possible, as long as blood usability is validated for each specific thrombogenicity test method. For the current flow loop test system, to ensure that the blood had relatively normal coagulability after overnight shipping, positive and negative control materials (PTFE and Latex) were utilized within the dynamic flow loop to qualify the blood from each donor, with the requirement to meet the thrombus surface coverage inclusion criteria (PTFE < 10%, latex > 50%) to be usable. Approximately 25% of the received donor blood could not be adjusted with heparin to meet the above inclusion criteria and was not used to complete the testing. Despite this limitation, it was more efficient and cost-effective to use a commercial source to obtain 24 hr postdraw ovine blood for the testing, in comparison to maintaining an animal facility to obtain freshly drawn blood.
Many groups have developed in vitro blood flow loop systems for evaluating the thrombogenicity of blood-contacting biomaterials and devices.11–21 However, the test parameters (e.g., test duration, temperature, and flow rate) for these systems vary greatly between studies. For example, the length of the blood circulation time varies from 15 min18 to 31 hr.17 For the current study, a test duration of 1 hr was chosen because it permitted the initial donor-specific heparin concentration to be validated within the flow loop, and adjusted if needed, while still providing adequate time for thrombus deposition to occur. Also, shorter test duration allows for a greater number of materials/devices to be examined side-by-side on the same day with the same blood pool.
Although the majority of the previous referenced studies utilized a temperature of 37°C,11–15,17–21 several groups have also performed in vitro thrombogenicity evaluations at room temperature.16,30–34 When directly comparing the effect of temperature on the in vitro differentiation of material thrombogenicity, Niimi et al.33 and Braune et al.34 found that the overall results were comparable at both 37°C and room temperature. Studies have also reported that the rate of platelet aggregation and adhesion increases at room temperature compared to 37°C.30,33,34 Therefore, performing the test at room temperature may accelerate in vitro thrombus formation and thus would allow the test system to differentiate the relative thrombogenicity of different test materials after shorter test durations. In addition, testing at room temperature eliminates the need for cumbersome heating equipment and simplifies the test system, which improves testing efficiency. Although the current study demonstrated that testing at room temperature was sufficient to differentiate between materials with different thrombogenic potentials, further studies are needed to determine if testing at a physiologic temperature has a significant effect on the relative thrombogenicity comparison among different materials.
There are several limitations to the described study. Although the test is capable of differentiating thrombogenic materials and thromboresistant ones, it may not be sensitive enough to distinguish among commonly used biomaterials with subtle differences in thrombogenicity. Ovine blood was chosen for this assessment because sheep models are commonly used in the in vivo preclinical thrombogenicity evaluations of blood wetted devices.35–38 However, studies have shown that ovine blood and human blood differ in thrombogenicity potentials.39,40 Therefore, the impact of donor blood from different species (porcine, bovine, and human) will need to be investigated before standardization. Also, this testing was an acute assessment of thrombogenicity and therefore may not be appropriate or sufficient for evaluating the long-term thrombogenicity of implantable devices such as ventricular assist devices. However, this method provides a systematic comparison of biomaterials that may help improve material selection by screening out thrombogenic materials, before a chronic in vivo study.
Another limitation of the study was that the outer diameters of the test materials were not uniform and varied from 2.1 to 3.2 mm. To mitigate the possible effect of diameter on the results, the dry thrombus weight was normalized to the blood-contacting surface area of the test materials. In practice, such as in the NAVI thrombogenicity test in animals, similar sized veins are utilized to investigate medical devices of varying dimensions. Because the use of multiple-sized catheters is clinically relevant, our test system was developed to evaluate blood-contacting devices of varying sizes/diameters. However, only smooth catheter-like materials over a narrow diameter size range were evaluated in the flow loop system in the current study. Future studies will need to be performed to evaluate whether the thrombogenicity of devices of different sizes and with more complex geometries and surfaces can be effectively differentiated using this dynamic flow loop test system.
Lastly, the test methodology developed in this study requires more validation, including interlaboratory studies, before standardization. The system will be further characterized by identifying more uniform positive and negative control materials, assessing the effect of increasing the circulation time, and examining a wider range of blood flow rates.
The dynamic flow loop test system developed in this study was able to effectively compare the thrombogenicity of biomaterials in catheter-like geometries. This test method allowed the use of commercially available donor sheep blood that was used 24–36 hr postdraw, thus increasing the usability of in vitro testing and reducing the experimental costs. Most importantly, by controlling the donor-specific heparin anticoagulation levels based on a targeted static thrombogenicity assessment of latex tubing, the ability for the dynamic flow loop tests to discern the thrombogenicity potential of test materials was increased.
The authors thank Luke Herbertson, Jennifer Goode, Molly Ghosh, and Hajira Ahmad at the FDA for reviewing the article and for providing valuable input.
1. Wiper A, Hashmi I, Srivastava V, et al. Guide wire thrombus formation during trans-femoral TAVI. Cardiovasc Revasc Med 2014.15: 360–361.
2. Chin T, Priyesh P, Islam AM. A novel approach to extraction of a large thrombus on the intraventricular guide-wire during transcatheter aortic valve replacement. Catheter Cardiovasc Interv 2017.89: 495–498.
3. Scalone G, Brugaletta S, Garcia-Garcia HM, et al. Frequency and predictors of thrombus inside the guiding catheter during interventional procedures: An optical coherence tomography study. Int J Cardiovasc Imaging 2015.31: 239–246.
4. Gobeil F, Juneau C, Plante S. Thrombus formation on guide wires during routine PTCA procedures: A scanning electron microscopic evaluation. Can J Cardiol 2002.18: 263–269.
5. Eckman PM, John R. Bleeding and thrombosis in patients with continuous-flow ventricular assist devices. Circulation 2012.125: 3038–3047.
6. de Mel A, Cousins BG, Seifalian AM. Surface modification of biomaterials: A quest for blood compatibility. Int J Biomater 2012.2012: 707863.
7. Li S, Henry JJ. Nonthrombogenic approaches to cardiovascular bioengineering. Annu Rev Biomed Eng 2011.13: 451–475.
8. ISO 10993–4: 2017, Biological evaluation of medical devices —Part 4: Selection of tests for interactions with blood. 2017.Arlington, VA, International Organization for Standardization (ISO).
9. Guidance for Industry and Food and Drug Administration Staff: Use of International Standard ISO 10993-1, “Biological evaluation of medical devices - Part 1: Evaluation and testing within a risk management process.” 2016.Maryland, US, US Food and Drug Administration, Silver Spring.
10. Wolf MD, Anderson JM. Boutrand J. Practical approach to blood compatibility assessments: General considerations and standards. in Biocompatibility and Performance of Medical Devices. 2012, Cambridge, United Kingdom, Woodhead Publishing Ltd, pp. 159–200.
11. Grove K, Deline SM, Schatz TF, Howard SE, Porter D, Smith ME. Thrombogenicity testing of medical devices in a minimally heparinized ovine blood loop. J Med Devices 2017.11: 0210080210088.
12. Krajewski S, Neumann B, Kurz J, et al. Preclinical evaluation of the thrombogenicity and endothelialization of bare metal and surface-coated neurovascular stents. AJNR Am J Neuroradiol 2015.36: 133–139.
13. Stang K, Krajewski S, Neumann B, et al. Hemocompatibility testing according to ISO 10993-4: Discrimination between pyrogen- and device-induced hemostatic activation. Mater Sci Eng C Mater Biol Appl 2014.42: 422–428.
14. van Oeveren W, Tielliu IF, de Hart J. Comparison of modified chandler, roller pump, and ball valve circulation models for in vitro
testing in high blood flow conditions: Application in thrombogenicity testing of different materials for vascular applications. Int J Biomater 2012.2012: 673163.
15. Walker EK, Nauman EA, Allain JP, Stanciu LA. An in vitro
model for preclinical testing of thrombogenicity of resorbable metallic stents. J Biomed Mater Res A 2015.103: 2118–2125.
16. Weaver JD, Ku DN. Biomaterial testing for covered stent membranes: Evaluating thrombosis and restenosis potential. J Biomed Mater Res B Appl Biomater 2012.100: 103–110.
17. Böswald M, Lugauer S, Bechert T, Greil J, Regenfus A, Guggenbichler JP. Thrombogenicity testing of central venous catheters in vitro. Infection 1999.27(suppl 1): S30–S33.
18. Nguyen KT, Su SH, Sheng A, et al. In vitro
hemocompatibility studies of drug-loaded poly-(L-lactic acid) fibers. Biomaterials 2003.24: 5191–5201.
19. Walter T, Rey KS, Wendel HP, et al. Thrombogenicity of sirolimus-eluting stents and bare metal stents: Evaluation in the early phase after stent implantation. In Vivo 2010.24: 635–639.
20. Mrowietz C, Franke RP, Seyfert UT, Park JW, Jung F. Haemocompatibility of polymer-coated stainless steel stents as compared to uncoated stents. Clin Hemorheol Microcirc 2005.32: 89–103.
21. Paul R, Marseille O, Hintze E, et al. In vitro
thrombogenicity testing of artificial organs. Int J Artif Organs 1998.21: 548–552.
22. Ogawa S, Ohnishi T, Hosokawa K, Szlam F, Chen EP, Tanaka KA. Haemodilution-induced changes in coagulation and effects of haemostatic components under flow conditions. Br J Anaesth 2013.111: 1013–1023.
23. Pretorius E, Oberholzer HM, van der Spuy WJ, Franz RC. Comparing techniques: The use of recalcified plasma in comparison with citrated plasma alone and in combination with thrombin in ultrastructural studies. Hematology 2011.16: 337–340.
24. ASTM F2888-19: Standard practice for platelet leukocyte count—An in-vitro measure for hemocompatibility assessment of cardiovascular materials. 2019.West Conshohocken, PA, ASTM International (ASTM).
25. Roll S, Müller-Nordhorn J, Keil T, et al. Dacron vs. PTFE as bypass materials in peripheral vascular surgery–systematic review and meta-analysis. BMC Surg 2008.8: 22.
26. Mazzaccaro D, Febis ED, Settembrini A, Tassinari L, Carmo M, Settembrini P. Long-term results of PTFE trilaminate graft versus venous graft and composite graft for below-the-knee revascularization. J Cardiovasc Surg (Torino) 2014.55: 685–691.
27. Ratner BD. Blood compatibility--a perspective. J Biomater Sci Polym Ed 2000.11: 1107–1119.
28. Llanos GR, Sefton MV. Immobilization of poly(ethylene glycol) onto a poly(vinyl alcohol) hydrogel: 2. Evaluation of thrombogenicity. J Biomed Mater Res 1993.27: 1383–1391.
29. Braune S, Walter M, Schulze F, Lendlein A, Jung F. Changes in platelet morphology and function during 24 hours of storage. Clin Hemorheol Microcirc 2014.58: 159–170.
30. Zhang JN, Wood J, Bergeron AL, et al. Effects of low temperature on shear-induced platelet aggregation and activation. J Trauma 2004.57: 216–223.
31. Lu Q, Malinauskas RA. Comparison of two platelet activation markers using flow cytometry after in vitro
shear stress exposure of whole human blood. Artif Organs 2011.35: 137–144.
32. Jamiolkowski MA, Woolley JR, Kameneva MV, Antaki JF, Wagner WR. Real time visualization and characterization of platelet deposition under flow onto clinically relevant opaque surfaces. J Biomed Mater Res A 2015.103: 1303–1311.
33. Niimi Y, Ishiguro Y, Nakata Y, Goto T, Morita S, Yamane S. Platelet adhesion to heparin coated oxygenator fibers under in vitro
static conditions: Impact of temperature. ASAIO J 2001.47: 361–364.
34. Braune S, Fröhlich GM, Lendlein A, Jung F. Effect of temperature on platelet adherence. Clin Hemorheol Microcirc 2016.61: 681–688.
35. Tuzun E, Harms K, Liu D, et al. Preclinical testing of the Levitronix Ultramag pediatric cardiac assist device in a lamb model. ASAIO J 2007.53: 392–396.
36. Tuzun E, Roberts K, Cohn WE, et al. In vivo
evaluation of the HeartWare centrifugal ventricular assist device. Tex Heart Inst J 2007.34: 406–411.
37. Teigen C, Stanley JR, Johnsont P, Gross C. Preclinical evaluation of the InCraft” aortic endograft in a sheep model. Vascular 2014.22: 13–19.
38. Zhou K, Niu S, Bianchi G, et al. Biocompatibility assessment of a long-term wearable artificial pump-lung in sheep. Artif Organs 2013.37: 678–688.
39. Goodman SL. Sheep, pig, and human platelet-material interactions with model cardiovascular biomaterials. J Biomed Mater Res 1999.45: 240–250.
40. Pelagalli A, Belisario MA, Tafuri S, et al. Adhesive properties of platelets from different animal species. J Comp Pathol 2003.128: 127–131.