Although many studies have sought to confirm the advantages of pulsatile flow, such as its superiority with respect to regional and global perfusion, as compared with nonpulsatile flow, its benefits remain controversial.1–3 Generally, hemodynamic energy-associated variables, such as energy equivalent pressure (EEP) and surplus hemodynamic energy (SHE), are measured to quantify pump-generated pulsatility in extracorporeal circuits.1–3 The EEP formula is based on the ratio between the area beneath the hemodynamic power curve (∫fpdt) and the area beneath the pump flow curve (∫fdt) during each pulse cycle,1,4 and SHE is calculated by multiplying the difference between EEP and mean arterial pressure by 1.332.1–4 The SHE exists only if some pulsatility is present in the pressure or flow.1 The hemodynamic energy of pulsatile flow is believed to be helpful for maintaining regional perfusion because hemodynamic energy may maintain peripheral perfusion by keeping capillary beds open and fluid moving in tissues.1
One reason for the controversy related to the superiority of pulsatile flow over nonpulsatile flow is that most hemodynamic energy is lost in extracorporeal circuits.5–7 To maximize the benefits of pulsatile flow, optimization of extracorporeal circulation (ECC) circuits is essential, which means pumps, pump settings, membrane oxygenators, arterial cannulas, and circuit designs must be carefully considered.8 Many studies have been sought to determine the pump settings and identify the oxygenator, pump, and cannula types that maximize hemodynamic energy deliver using pseudopatient models in mock test circuits.5–8 However, in these models, the aorta was represented by a straight tube line, and thus, they did not reflect real anatomical variations of the aorta. Measurements of hemodynamic energy changes in in vivo model would provide the best means of considering real anatomical variations of the aorta in studies that address the superiority of pulsatile flow over nonpulsatile flow, but such direct measures of hemodynamic energy are too invasive as massive vessel dissection is needed to calculate hemodynamic form flow and pressure measurements taken in arteries. Accordingly, direct measurements of hemodynamic energy in vivo model are difficult to perform and impractical in humans.
It is well-known that aortic cannula tip position during cardiopulmonary bypass (CPB) impacts cerebral perfusion9–11 because the flow delivered to carotid arteries is dependent on tip position.10 Thus, because hemodynamic energy delivery could be affected by the amount of flow generated by an ECC pump, aortic cannula tip position changes could influence hemodynamic energy delivery to both right and left carotid arteries. However, it is difficult to measure hemodynamic energy in carotid arteries using previously described pseudopatient models because of the use of a straight tube line to model the aorta. To overcome this limitation of pseudopatient models, a new patient model that accounts for patient-specific anatomical variations is essential. In the current study, we made and tested the feasibility of a three-dimensional (3D) printed, patient-specific, silicone, aortic model and used it to determine whether aortic cannula tip positional changes affect hemodynamic energy delivery.
Patient-Specific Silicone Model
The model systems previously used to measure hemodynamic energy have used a straight tube line to represent the aorta (Figure 1A). In the current study, we used 3D printing to produce a patient-specific aortic arch. Computer tomography (CT) images of a patient (female, 65 years old) who had undergone aortic valve replacement because of aortic regurgitation were exported in Digital Imaging and Communication in Medicine format, and a segmented 3D image of the ascending aorta was made using Mimics Base version 16 (Materialize, Leuven, Belgium; Figure 1B) and stored as a stereolithography file. A hole for aortic cannula insertion was made using 3-Matic (Materialise) on the segmented 3D image (Figure 1C). A 3D model of the ascending aorta was then printed at a scale of 1:1 scale using a 3D printer (Master EV; Carima, Seoul, South Korea; Figure 1D) using acrylonitrile-butadiene-styrene (ABS) resin. To produce a hollow, elastic model, a commercially available silicone (KE-1300T; Shin-Etsu, Tokyo, Japan) and CAT-1300 hardening agent (Shin-Etsu) were used. The silicone mixture was painted evenly (about 3 mm thick) on the surface of the ABS mold, and after drying, the silicone model was obtained by cutting the mold in half and then gluing the two halves together with silicone glue (Figure 1E).
Measurement of Patient-Specific Silicone Model Geometry
The thickness of the patient-specific silicone model was measured at 10 points on the aortic wall at 30 mm intervals and at three points on aortic arch vessels to evaluate thickness homogeneity (Figure 2A). Diameters of aortic arch branches were measured in the segmented 3D image (Figure 2B), outer diameters of aortic arch branches using the ABS mold (Figure 2C) and the inner diameters of aortic arch branches using the patient-specific silicone model using a caliper (Figure 2D). These three measurement values of aortic arch branches were compared to check whether there were differences between the segmented 3D image, ABS mold, and patient-specific silicone model. All values were measured three times.
Mechanical Properties of the Patient-Specific Silicone Model
The patient-specific silicone model and a Tygon tube (Tygon®Laboratory Tubing, AAC00037; Saint-Gobain Plastics, Akron, OH), which have been usually used to mimic the aorta, were subjected to tensile testing. All samples were cut circumferentially, and testing was conducted using a universal tensile testing device (ST-1001; SALT Co., Incheon, Korea). Sample strip widths and thicknesses were measured at same locations using a caliper (Mitutoyo Co, Kanagawa, Japan) and averaged. The width and thickness of silicone samples were 10 and 3 mm, respectively, and those of the Tygon tube samples were 10 and 2.6 mm. The gauge length of samples was 30 mm. Samples were loaded linearly at 20 mm/min, and maximum loads at failure were recorded. Testing was performed in triplicate.
In Vitro Model Experiment
The Donovan type model system was connected to a VAD pump (Korea hybrid ventricular assist device [KH-VAD]; Korea Artificial Organ Center, Seoul, Korea) that mimicked the heart (Figure 3), and the ascending aorta silicone model, and a pulsatile pump (Twin-Pulse Life Support [T-PLS], SL-1000; Newheartbio Co., Seoul, Korea) was connected to a 20 Fr aortic cannula (Aortic Perfusion Cannula; Edwards Lifesciences, Irvine, CA) through the cannula insertion hole.
The aortic cannula was inserted at three different positions, as follows; in position A, the aortic cannula tip was placed in the normal position, i.e., directed toward the middle of the transverse arch; in position B, the aortic cannula tip was rotated 90° counterclockwise from position A; and in position C, the tip was rotated 90° clockwise from position A (Figure 4).
We made our experimental setting mimic serious clinical situation when cerebral blood flow could be hindered by aortic tip positions. We thought when pump flow rate is low, the flow hindrance by aortic tip position would be maximized. Several studies that evaluated the effect of pump flow rate on cerebral blood flow and cerebral metabolism demonstrated that cerebral blood flow remained relatively constant at pump flow rates of 1.0–2.4 L/min/m2 when hypothermic bypass was used.12,13 In our study, we set the pump flow at 1.2 L/min/m2. The T-PLS generated pump flow rate of 2 L/min when its setting was 50 beats per minute (BPM). To mimic hemodilution condition during hypothermic CPB, the simulated circuit was primed with 20% glycerin solution. Viscosity is principally influenced by hematocrit and temperature.14 Viscosity remains reasonably constant if hematocrit and temperatures are the same (i.e., viscosity at 35°C with a hematocrit of 35% is about the same as when temperatures and hematocrit are 20°C and 20%.).14 Thus, we used 20°C of 20% glycerin solution.
The KH-VAD was completely stopped and did not generate any flow during T-PLS operation to mimic stoppage because of cardiac arrest. It means that there was no antegrade flow from KH-VAD to the aorta.
Flow at the outlet of the T-PLS and in the left and right carotid arteries of the silicon model was measured using an ultrasonic flow meter (T402; Transonic, Ithaca, NY), and pressures at same sites were measured using a pressure sensor (PS90030VY; Sensortechnics, Berlin, German). After starting the T-PLS pump, pressure and flow waveforms were measured simultaneously for 60 seconds using a homemade data-acquisition board at a T-PLS pump flow rate of 50 BPM. Waveforms were analyzed using MATLAB software (Mathworks, Natick, MA). Hemodynamic energy was calculated at the three aortic cannula tip positions. All experiments were repeated five times.
Calculation of Hemodynamic Energy
The EEP was defined using the following formula1,4:
where f is the pump flow rate (L/min) and p is the arterial pressure (mm Hg). Time integrals were calculated over a single pulse cycle.
SHE was defined as follows:
Statistical analysis was performed using SPSS version 21 (IBM Corporation, Armonk, NY). The Student’s t-test was used to compare tensile measurements taken for the Tygon tube and the patient-specific silicone model. One-way analysis of variance (ANOVA) was used to compare arch vessel diameters obtained using the 3D segmented image, ABS mold, and the patient-specific silicone model. ANOVA with Bonferroni multiple comparison test was used to compare variables at different aortic tip positions. The p < 0.05 was deemed significant. Results are presented as means ± standard deviations.
Feasibility of the 3D Printed Patient-Specific Aorta Model
To make ABS mold, it took 16 hours. As the patient-specific model was made by painting silicone on the ABS mold, wall thickness differences may have biased hemodynamic energy measurements. To confirm wall thickness homogeneity of the patient-specific silicon model, wall thicknesses were measured at 13 points. Wall thicknesses of 10 points on the aortic wall (from the point a to the point j, Figure 2A) ranged from 2.68 to 3.64 mm (Table 1) and of the three points on aortic arch vessels (from point
, Figure 2A) from 2.02 to 2.78 mm (Table 1).
The diameters of arch vessels in the 3D segmented model, ABS mold, and patient-specific silicone model were compared to confirm that the patient-specific silicone model well represented the anatomical geometry of the patient’s aorta. Diameters of right and left common carotid and left subclavian arteries were not significantly different between the three models (Table 2).
The 3D printed model was connected to the T-PLS pump and the Donovan type model system without difficulty and sustained the pressure generated by the T-PLS pump at a flow rate of 50 BPM without rupture or fluid leakage.
Tensile strengths, yield strengths, and elastic moduli of strips of the patient-specific silicone model and Tygon tube were measured, and all measurements obtained using the silicone model and Tygon tube were significantly different (Table 3). In fact, the tensile strength and elastic modulus values of the patient-specific model more closely resembled those of the human thoracic aorta.
Hemodynamic Energy Measurement
For the patient-specific silicone model, pressure, flow, EEP, and SHE at the right common carotid artery site (position B) were significantly greater than at positions A or C (Table 4). The %EEP at position B was not significantly different from that at position C but was significantly higher than that at position A (Table 4). At the left common carotid artery site, pressure and flow at position C were significantly higher than those at position A or B (Table 4). The EEP at position C was not significantly different from that at position B but was significantly higher than that at position A. The %EEP and SHE at position B were significantly higher than those at position A or C (Table 4).
Delivery amount of hemodynamic parameters at the right (B site of Figure 3) and left common carotid arteries (C site of Figure 3) were expressed as percentages of values at the T-PLS outflow site (A site of Figure 3) in the silicone model. It meant that the delivery percentage of hemodynamic energy at the right common carotid artery in the silicone model calculated as below:
The delivery percentage hemodynamic parameters of the right common carotid artery (%) = the hemodynamic energy amount at the right common carotid artery/the hemodynamic energy amount at T-PLS outflow site × 100
Only 44.20–45.24% of the pressure at the T-PLS outflow site was found to be delivered to the right common carotid artery, and the amount delivered was greatest when the aortic cannula tip was at position B (Table 5). Furthermore, 6.73–8.67% of flow and 29.44–30.37% of EEP at the T-PLS outflow site were delivered to the right common carotid artery, and amounts delivered were when the tip was at position B. On the other hand, 6.04–6.56% of %EEP at the T-PLS outflow site was delivered to the right common carotid artery, and the amount delivered was lowest in position A, and 2.67–2.97% of SHE at the T-PLS outflow site was delivered to the right common carotid artery, and the amount delivered was greatest in position B.
The 40.26–40.40% of pressure at the T-PLS outflow site was delivered to the left common carotid artery, and the amounts delivered at the three different tip positions were not significantly different (Table 5). The 5.20–6.39% of flow at the T-PLS outflow site was delivered to the left common carotid artery, and the amount delivered was least at position B. The 27.07–27.32% of EEP at the T-PLS outflow site was delivered to the left common carotid artery, and amounts delivered were not significantly different for the three tip positions. The 6.58–8.56% of %EEP at the T-PLS outflow site was delivered to the left common carotid artery, and the amount delivered was greatest when the tip was at position B. The 2.66–3.45% of SHE at the T-PLS outflow site was delivered to the left common carotid artery, and the amount delivered was greatest in position B.
In this study, we produced a 3D printed patient-specific aorta model to measure hemodynamic energy using an in vitro model system. To confirm wall thickness homogeneity of the patient-specific silicon model, wall thicknesses were measured at 13 points; measured values ranged from 2.02 to 2.78 mm. Arterial wall thickness depends on numerous physiologic and pathologic factors, such as age, sex, blood pressure, and smoking status.15–18 During recent years, magnetic resonance imaging has become widely used for assessing arterial wall thickness, and one such study reported mean wall thicknesses of the ascending and descending aorta of 1.89 ± 0.21 and 1.60 ± 0.22 mm, respectively.15 Wall thicknesses of the aorta differ, whereas the Tygon tube has a constant wall thickness, which presents another limitation with respect to its ability to mimic the real aorta. In the current study, we used CT images to make the 3D replica, and it is not possible to measure real aortic thicknesses by CT. We intend to remedy this shortcoming in future studies.
To determine differences between the diameters of arch vessels in the 3D segmented model, ABS mold, and patient-specific silicone model, the diameters of right and left common carotid arteries and left subclavian arteries were measured, but no significant differences were found between the three models. The diameters of arch vessels were measured because the main purpose of the current study was to obtain hemodynamic energy measurements for both common carotid arteries. It was found that the diameters of both common carotid and left subclavian arteries were not significantly different, and thus, geometric difference between the three models probably did not have undue impact on hemodynamic energy measurements. Besides the issue of patients-specific model accuracy, the time-consuming process is also a problem. In our study, it took 16 hours to make ABS mold, and more hours were needed to make a patient-specific silicone model. It means that patient-specific silicone model is difficult to be applied in urgent operation cases yet. However, the purpose of our study is to evaluate the feasibility of patient-specific silicone model for measuring the hemodynamic energy. We thought that patient-specific silicone model could be used in scheduled operations, and the time-consuming problem to print 3D models will be solved in the near future as the 3D printing technology is improved.
Several studies have shown that ECC circuit components can alter the amount of hemodynamic energy delivered by blood pumps in model circuits.8,19 Rider et al.7 compared pressure drops and SHE levels in eight 10 Fr pediatric aortic cannulas with different aortic cannula geometries in an in vitro infant model of CPB and found cannula geometries influenced both pressure drops and SHE delivery. However, in these studies, straight aortic tubing was to mimic the aorta.
It is well-known that aortic cannula position influences flow distributions in carotid arteries.11 In this previous study, a computational fluid dynamics (CFD) approach was used, which allowed consideration of patient-specific anatomical variations, but such simulations are subject to model assumptions, and thus, experimental validation is essential.20 Furthermore, the CFD method is unsuitable for measuring hemodynamic energy in ECC circuits. In the current study, a 3D printed patient-specific model was used to measure hemodynamic energy during ECC and found to provide a suitable means for measuring hemodynamic energy loss in an ECC circuit and for accessing the effects of patient-specific anatomical variations of the aorta. The variations of vessels arising from the aortic arch are numerous.21 The aortic arch type described as “normal,” i.e., the most common type, is one that gives rise to three branches: 1) the brachiocephalic trunk that branches to the right subclavian and right common carotid arteries, 2) the left common carotid artery, and 3) the left subclavian artery.21,22 In one study, eight types of aortic arch were found in 633 Caucasian Greek patients and the normal type accounted for 83% of these patients.18 Accordingly, anatomical variations of aorta are common, and these variations could substantially alter hemodynamic energy delivery.
In model experiments, mechanical property differences between Tygon tubes and real aortas are problematic because it affect hemodynamic energy measurements. The tensile strength, yield strength, and elastic modulus of human thoracic aorta have been reported to be 0.3–0.8,23 3.8,24 and 0.4–1 N/mm2,25 respectively, and in the current study, the tensile strength and elastic modulus values of the patient-specific model more closely resembled those of the human thoracic aorta than the Tygon tube.
Importantly, when we measured hemodynamic energy using the 3D printed silicone model, we found that EEP and SHE of both right and left carotid arteries were affected by aortic cannula tip position.
Neurological dysfunction is one of the most common complications of CPB and may lead to comorbidities and increase hospital stays,9 and thus, proper cerebral perfusion is essential to minimize the risks of perioperative neurological morbidities.9–11 Several authors have suggested that pulsatile flow is better than nonpulsatile flow in terms of reducing these risks, especially in pediatric CPB cases,26 because hemodynamic energy reflects blood flow to the brain.27 However, the superiority of pulsatile flow in the context of organ perfusion remains a topic of debate. One reason for this is that hemodynamic energy is dependent on ECC circuit components; in fact, most hemodynamic energy is absorbed by ECC circuits.
By CFD method, one study showed that nonpulsatile perfusion resulted in moderately enlarged maximum wall shear stress values (2.4-fold), whereas pulsatile perfusion considerably affected the maximum shear forces (51-fold).28 Because maximum turbulent kinetic energy and wall shear stress at the tip of the cannulas are primarily dependent on the canula diameter, opting for large-sized cannulas is important.28 It is well-known that proper large arterial cannula with short tips are most suitable for pulsatile perfusion which have the greatest influence on the pressure drop over the cannula and the SHE delivery.29 Besides aortic cannula geometries, the findings of the current study suggest that even changes in aortic cannula tip position can alter hemodynamic energy delivery to right and left common carotid arteries and that hemodynamic energy delivery to carotid arteries is only a small fraction of that at available at pump outlets. Furthermore, these results could not have been obtained using a straight tube line to replicate the aorta. To measure hemodynamic energy accurately at different aortic cannula tip positions in a model circuit, we suggest that the devised 3D printed patient-specific aorta model provides a feasible option. In the near future, we can simulate and decide which the aortic tip cannula position or geometry would be best to deliver the hemodynamic energy to the carotid artery with 3D printed patient-specific aorta model before surgery.
Our study has several limitations. In the current study, we stopped the KH-VAD that mimic the patient heart in the mock system to replicate the cardiac arrest. If the patient heart did eject the blood flow to the aorta, those flow could mix with ECC flow and make turbulent flow to the carotid arteries. The turbulent flow could be different by the percentage of assisted circulation. In this study, we did not consider the delivery amount of hemodynamic energies to carotid arteries could affect by the percentage of assisted circulation.
The pump flow and aortic pressure in our experiment were lower than usual ECC condition, and hemodynamic energy changes by aortic tip cannula position could be different in the higher pump flow or aortic pressure conditions. In the future study, we should evaluate whether the hemodynamic energy change by aortic cannula tip position would be affected by the aortic pressure or pump flow difference.
We did not consider arterial stiffness changes by atherosclerosis on the aortic wall. Arterial stiffness reflect the arterial compliance and distensibility.30 Atherosclerosis is a pathological condition of the intima, which is characterized by lipid accumulation, inflammatory cells, vascular smooth muscle cell migration, foam cell development, connective tissue fibers, and calcium deposits.30 It is hard to mimic atherosclerotic plaque by 3D printing technology because the atherosclerotic plaque has complex composition. In case of calcium deposits, there are lots of ongoing studies to mimic calcified atheroma in the aorta by a 3D printers that produce multimaterial photopolymer.31 For example, a rigid and a flexible photopolymer can be combined to print a model with flexible material for soft tissues, such as arteries, and hard material for a calcification or bone.31 In the near future, it will be possible to make patient-specific aorta model that even consider calcifications in the aorta wall.
Beside atherosclerosis, aging decreases elasticity of the common carotid artery.32 It means changed elasticity of the carotid artery might impact the flow changes in the carotid arteries. We could not mimic the elasticity changes by aging in our patient-specific model. For imitating real carotid artery elasticity, the new materials that can mimic various elasticity of the carotid artery by aging should be developed.
The aorta has elastic properties and is capable of storing approximately 50% of the stroke volume in the systolic phase by extending the aortic walls.33,34 When the aortic valve is closed in early diastole, the stored volume is forwarded to the peripheral circulation, which is because of the so-called Windkessel effect.33,34 In the current study, the patient-specific aorta model might not guarantee the Windkessel effect in the carotid artery because the elasticity of silicone is different with the real aorta or carotid arteries. Although our patient-specific aorta model had many limitation to mimic real patient aorta or carotid arteries because of material elasticity difference between the real aorta and silicone, the anatomical resemblance between our model and patient is a meaningful improvement to solve the limitations of the present pseudopatient mock systems.
The described 3D printed patient-specific aorta model was found to be feasible for measuring hemodynamic energy during ECC at three different aortic cannula positions. Furthermore, we believe that the described model provides a suitable means of determining the effects of patient-specific anatomical variations of the aorta in model systems.
1. Ji B, Undar AComparison of perfusion modes on microcirculation during acute and chronic cardiac support: Is there a difference? Perfusion 2007.22: 115–119,
2. Undar AMyths and truths of pulsatile and non-pulsatile perfusion during acute and chronic cardiac support. Artif Organs 2004.28: 439–443,
3. Undar A, Ji B, Lukic B, et alQuantification of perfusion modes in terms of surplus hemodynamic energy
levels in a simulated pediatric CPB model. ASAIO J 2006.52: 712–717,
4. Shepard RB, Simpson DC, Sharp JFEnergy equivalent pressure. Arch Surg 1966.93: 730–740,
5. Undar A, Owens WR, McGarry MC, et alComparison of hollow-fiber membrane oxygenators in terms of pressure drop of the membranes during normothermic and hypothermic cardiopulmonary bypass in neonates. Perfusion 2005.20: 135–138,
6. Undar A, Eichstaedt HC, Masai T, et alComparison of six pediatric cardiopulmonary bypass pumps during pulsatile and nonpulsatile perfusion. J Thorac Cardiovasc Surg 2001.122: 827–829,
7. Rider AR, Ji B, Kunselman AR, et alA performance evaluation of eight geometrically different 10 Fr pediatric arterial cannulae under pulsatile and nonpulsatile perfusion conditions in an infant cardiopulmonary bypass model. ASAIO J. 2008.54: 306–315,
8. Lim CH, Yang S, Choi JW, Sun KOptimizing the circuit of a pulsatile extracorporeal life support system in terms of energy equivalent pressure and surplus hemodynamic energy
. Artif Organs 2009.33: 1015–1020,
9. Neidlin M, Steinseifer U, Kaufmann TAA multiscale 0-D/3-D approach to patient-specific adaptation of a cerebral autoregulation model for computational fluid dynamics studies of cardiopulmonary bypass. J Biomech 2014.47: 1777–1783,
10. Kaufmann TA, Hormes M, Laumen M, et alFlow distribution during cardiopulmonary bypass in dependency on the outflow cannula positioning. Artif Organs 2009.33: 988–992,
11. Kaufmann TA, Hormes M, Laumen M, et alThe impact of aortic/subclavian outflow cannulation for cardiopulmonary bypass and cardiac support: A computational fluid dynamics study. Artif Organs 2009.33: 727–732,
12. Cook DJ, Proper JA, Orszulak TA, Daly RC, Oliver WCEffect of pump flow rate on cerebral blood flow during hypothermic cardiopulmonary bypass in adults. J Cardiothorac Vasc Anesth 1997.11: 415–419,
13. Rogers AT, Prough DS, Roy RC, et alCerebrovascular and cerebral metabolic effects of alterations in perfusion flow rate during hypothermic cardiopulmonary bypass in man. J Thorac Cardiovasc Surg 1992.103: 363–368,
14. Gravlee GPA Practical Approach to Cardiac Anesthesia. 2013.Philadelphia, Lippincott Williams & Wilkins,
15. Mensel B, Kühn JP, Schneider T, Quadrat A, Hegenscheid KMean thoracic aortic wall thickness determination by cine MRI with steady-state free precession: Validation with dark blood imaging. Acad Radiol 2013.20: 1004–1008,
16. Rosero EB, Peshock RM, Khera A, Clagett P, Lo H, Timaran CHSex, race, and age distributions of mean aortic wall thickness in a multiethnic population-based sample. J Vasc Surg 2011.53: 950–957,
17. Zanchetti A, Bond MG, Hennig M, et alRisk factors associated with alterations in carotid intima-media thickness in hypertension: Baseline data from the European Lacidipine Study on Atherosclerosis. J Hypertens 1998.16: 949–961,
18. Bourque JM, Schietinger BJ, Kennedy JL, et alUsefulness of cardiovascular magnetic resonance imaging of the superficial femoral artery for screening patients with diabetes mellitus for atherosclerosis. Am J Cardiol 2012.110: 50–56,
19. Wang S, Kunselman AR, Clark JB, Undar AIn vitro hemodynamic evaluation of a novel pulsatile extracorporeal life support system: Impact of perfusion modes and circuit components on energy loss. Artificial Organs 2015.39: 59–66,
20. Büsen M, Kaufmann TA, Neidlin M, Steinseifer U, Sonntag SJIn vitro
flow investigations in the aortic arch during cardiopulmonary bypass with stereo-PIV. J Biomech 2015.48: 2005–2011,
21. Natsis KI, Tsitouridis IA, Didagelos MV, Fillipidis AA, Vlasis KG, Tsikaras PDAnatomical variations in the branches of the human aortic arch in 633 angiographies: Clinical significance and literature review. Surg Radiol Anat 2009.31: 319–323,
22. Liechty JD, Shields TW, Anson BJVariations pertaining to the aortic arches and their branches; with comments on surgically important types. Q Bull Northwest Univ Med Sch 1957.31: 136–143,
23. Silver FH, Christiansen DL, Buntin CMMechanical properties of the aorta: A review. Crit Rev Biomed Eng 1989.17: 323–358,
24. Kinley CE, Marble AECompliance: A continuing problem with vascular grafts. J Cardiovasc Surg (Torino) 1980.21: 163–170,
25. Learoyd BM, Taylor MGAlterations with age in the viscoelastic properties of human arterial walls. Circ Res 1966.18: 278–292,
26. Undar APulsatile versus nonpulsatile cardiopulmonary bypass procedures in neonates and infants: From bench to clinical practice. ASAIO J 2005.51: vi–x,
27. Su XW, Guan Y, Barnes M, Clark JB, Myers JL, Undar AImproved cerebral oxygen saturation and blood flow pulsatility with pulsatile perfusion during pediatric cardiopulmonary bypass. Pediatr Res 2011.70: 181–185,
28. Assmann A, Benim AC, Gül F, et alPulsatile extracorporeal circulation during on-pump cardiac surgery enhances aortic wall shear stress. J Biomech 2012.45: 156–163,
29. Wang S, Haines N, Undar AQuantification of pressure-flow waveforms and selection of components for the pulsatile extracorporeal circuit. J Extra Corpor Technol 2009.41: P20–P25,
30. Cecelja M, Chowienczyk PRole of arterial stiffness in cardiovascular disease. J R Soc Med Cardiovasc Dis. DOI 10.1258/cvd.2012.012016.
31. Giannopoulos AA, Steigner ML, George E, et alCardiothoracic applications of 3-dimensional printing. J Thorac Imaging 2016.31: 253–272,
32. Reneman RS, van Merode T, Hick P, Hoeks APFlow velocity patterns in and distensibility of the carotid artery bulb in subjects of various ages. Circulation 1985.71: 500–509,
33. Gülan U, Lüthi B, Holzner M, Liberzon A, Tsinober A, Kinzelbach WExperimental investigation of the influence of the aortic stiffness on hemodynamics in the ascending aorta. IEEE J Biomed Health Inform 2014.18: 1775–1780,
34. Belz GGElastic properties and Windkessel function of the human aorta. Cardiovasc Drugs Ther 1995.9: 73–83,