Secondary Logo

Journal Logo

Effect of Outflow Graft Size on Flow in the Aortic Arch and Cerebral Blood Flow in Continuous Flow Pumps: Possible Relevance to Strokes

Bhat, Sindhoor; Mathew, Jayakala; Balakrishnan, Komrakshi R.; Krishna Kumar, Ramarathnam

doi: 10.1097/MAT.0000000000000507
Adult Circulatory Support
Free

One of the most devastating complications of continuous flow left ventricular devices (CFLVADS) is stroke, with a higher incidence in HeartWare Ventricular Assist Device (HVAD) as compared with HEARTMATE II. The reason for the observed difference in stroke rates is unclear. Because outflow graft diameters are different, we hypothesized that this could contribute to the difference in stroke rates. A computational fluid–structure interaction model was created from the computed tomography (CT) scan of a patient. Pressures were used as the boundary condition and the flow through the cerebral vessels was derived as outputs. Flow into the innominate artery was very sensitive to the anastomosis angle for a 10 mm as compared with a 14 mm graft, with the net innominate flow severely compromised with a 10 mm graft at 45° angle. Aortic insufficiency seems to affect cerebral blood flow nonlinearly with an 80% decrease at certain angles of outflow graft anastomosis. Arterial return in to the arch through a narrow graft has important jet effects and results in significant flow perturbations in the aortic arch and cerebral vessels and stasis. A 10 mm graft is more sensitive to angle of insertion than a 14 mm graft. Under some conditions, serious hypoperfusion of the innominate artery is possible. Aortic incompetence results in significant decrease of cerebral blood flow. No stasis was found in the pulsatile flow compared with LVAD flow.

From the *Department of Engineering Design, Indian Institute of Technology Madras, Chennai, India; and Department of Cardiac Sciences, Fortis Malar hospital, Chennai, India.

Submitted for consideration August 2016; accepted for publication in revised form December 2016.

Disclosures: The authors have no conflicts of interest to report.

Correspondence: Ramarathnam Krishna Kumar, Department of Engineering Design, Indian Institute of Technology Madras, Chennai, India. Email: rkkumar@iitm.ac.in.

A devastating complication of continuous flow left ventricular assist devices (CFLVADS) is stroke with a higher incidence in Heartware HVAD compared with Heartmate II.1–3 There has been a paucity of studies looking at the reason for this difference in stroke rates. We hypothesized that the difference in outflow graft diameters of 10 and 14 mm could be one of the important reasons and used computational fluid dynamics (CFD) for the study.

Back to Top | Article Outline

Materials and Methods

ADINA 9.1, a commercial finite element package, was used for the fluid structure analysis.4 The process of generating a finite element model involves several steps. These steps, where a patient CT scan was converted into a three-dimensional (3D) finite element model, are given in Figure 1.

Figure 1

Figure 1

Fluid and solid were modeled separately and was solved using a two-way iterative coupling approach. Four-noded linear elastic shell elements based on large deformation and small strain with a Young’s Modulus of 0.45 MPa was used for the aortic wall. The thickness of the aortic arch and different branches were taken from literature.5

Fluid was modeled using flow-condition-based interpolation (FCBI) tetrahedral elements.4 The governing equation for blood flow follows Navier–Stokes equation with Arbitrary Lagrangian - Eulerian (ALE) formulation to account for the moving boundaries.

Blood was modeled as an incompressible Newtonian material with a laminar viscosity of 0.0035 Pa·s. This is a reasonable assumption as the blood flow in larger arteries like aorta follows the Newtonian flow.6 Mesh sensitivity study helped to fix the element size.

The boundary conditions can be classified into solid and fluid boundary conditions. Fluid–structure interaction problems require realistic boundary conditions for the solid. To account for the external support to aorta from surrounding tissues and pericardium, spring elements were attached to the structural nodes externally. At peak systole, the maximum radial displacement in the ascending aorta is 1.5 mm and thoracic aorta is 0.5 mm,7 based on which the average spring stiffness was considered to be 0.8 N/m. We have used the method available in the literature8 to generate the zero-pressure geometry. Several studies, including a recent article by Osorio et al.,6 have specified a predetermined flow rate into the cerebral vessels. This boundary condition is obviously unrealistic. Farag et al.9 have specified zero pressure at the boundary as has been stated in several studies by Karmonik et al.10,11 The most sophisticated boundary conditions involve a tight coupling between 0D and 3D fluid dynamics model.12 In this study, the impedance of the rest of the arterial tree has been considered, as is the case in Kar et al.13

The flow in the cerebral vessels was, therefore, an output from the study and not predetermined, thus enhancing the validity of the findings. The complete human arterial tree was modeled using 0 D multicompartment model built in Simulink.

We defined current at the inlet points of the Zero-D model (LVAD and ascending aorta) and the voltage was measured at the required outlet branches. Mechanically, this is equivalent to prescribing the flow at the inlet and measuring the pressure at the outlet branches. The peak pressure differential between ascending aorta and thoracic aorta was 0.59 mm Hg for a healthy case and 0.2 mm Hg for LVAD-fitted case. To obtain the realistic range of pressures, we subtracted the differentials from the reference aortic pressure curve to arrive at the outlet pressure boundary conditions of the 3D model.

A CFLVAD pump giving an output of 3.6 L/min and completely closed aortic valve condition (velocity = 0 m/s) was considered for the analysis. Left ventricular assist device flow curve, ventricle, and the reference aortic pressure was taken from Travis et al.14Figure 2 shows the boundary conditions used.

Figure 2

Figure 2

We studied the blood flow in the aortic arch and its branches in six different simulations under identical conditions of LVAD flow rate and aortic pressure and they are as follows:

  1. Normal pulsatile flow
  2. Continuous flow arising out of the left ventricle through the normal aortic root
  3. Continuous flow left ventricular devices with a 10 and 14 mm outflow graft at an angle of 30° and 45° to the ascending aorta. Furthermore, simulations were done with aortic incompetence.

In each of these simulations, velocity and volume of flow, flow patterns in the aortic arch, and degree of stasis were studied.

Stasis was quantified by particle count which is a method by which particles are injected into the fluid stream numerically and its path followed with time. The ratio of the particles injected to those which stay in the system over 2 cardiac cycles is monitored and is deemed to be an objective marker for stasis. Such a technique has been used earlier by Osorio et al.6

Back to Top | Article Outline

Results

Normal Pulsatile Flow

Pulsatile flow was simulated using the pressure boundary condition calculated by Kim et al.12 The flow rates calculated in various branches closely match with that obtained by Kim et al.12 (Figure 3) and were close to the normal values seen in human subjects. The flow was completely developed by the time it reached the innominate artery and the high systolic velocities observed during systole completely fill the aorta and aid in the flow of blood into the neck vessels. The other striking feature captured in this simulation is the helical flow patterns in the aortic arch (Figure 4) The physiological benefits of helical flow in the aortic arch are now increasingly being recognized.15 The particle tracing expressed as a percentage of remaining particles in the system (Figure 7) clearly showed that only 1% of the particles injected stayed in the system beyond two cycles. Stasis was strikingly absent, even when the flow is reduced to 3.8 L/min. This is shown in Figure 7.

Figure 3

Figure 3

Figure 4

Figure 4

Back to Top | Article Outline

Continuous Flow Arising Out of the Left Ventricle Through the Normal Aortic Root

A continuous LVAD flow through the aortic valve was modeled as if the LVAD was placed inside the left ventricle. This hypothetical model brings out the difference between pulsatile and continuous flow. Figure 5 is similar to Figure 4. The flows are similar, with lower velocities.

Figure 5

Figure 5

Back to Top | Article Outline

Continuous Flow Left Ventricular Device Flow Through 10 and 14 mm Grafts at 30° and 45° Angles

The flow patterns for a 10 and a 14 mm graft at an angle of 30° and 45° are shown in Figure 6. Both flows exhibited a striking jet effect, although the peak velocities and hence the sharpness of the jet was understandably more for a 10 mm graft. Flow through the neck vessels was not affected significantly (Table 1). Importantly, there was more stasis in the arch as shown by the particle count (Figure 7).

Table 1

Table 1

Figure 6

Figure 6

Figure 7

Figure 7

Figure 6 clearly shows that no fluid enters the brachiocephalic for a 10 mm graft, while the flow is present for a 14 mm graft. The diffused jet of a 14 mm graft makes it possible for the fluid to enter all the vessels, making it less vulnerable for graft angle. On the other hand, 10 mm graft, because of the angle, directs the fluid toward the subclavian artery, resulting in a loss of fluid entry into the brachiocephalic. A small negative flow was observed with a 10 mm graft. Similar negative flow was observed in an earlier study for cardiopulmonary bypass (CPB).16 About 9% of particles stayed after two cycles for a 14 mm graft at 45° (Figure 7). One may conclude from these studies that angle of entry is important for a smaller graft because of the jet effect. Larger grafts because of the diffused nature of the jet distribute the fluid more evenly.

Back to Top | Article Outline

Effect of Aortic Incompetence on Cerebral Blood Flow in Continuous Flow Left Ventricular Devices

The effect of aortic regurgitation on arch flow was interesting. Because of a large pressure differential across the valve, a small 2 mm diameter opening in the aortic valve created 20% regurgitation. Normally there is a stationary column of blood above the aortic valve and below the out flow graft to “support” the jet of flow from the VAD.

This standing column below the jet is disturbed by aortic regurgitation and the jet is deflected downwards causing dramatic changes in the blood flow pattern in the arch (Figure 8) resulting in decreased cerebral perfusion and this is more pronounced as the angle of graft insertion approaches 90°. This has obvious clinical implications.

Figure 8

Figure 8

Back to Top | Article Outline

Discussion

One of the obvious differences between Heartmate II and HVAD is the diameter of the outflow graft and there has been no study looking at its role in the causation of strokes. Although the jet effect of the outflow graft has been noted in several studies,16,17 the relationship between graft diameter and stasis and eddies in the arch and cerebral blood flow have not been investigated in detail.

Under normal pulsatile conditions or continuous flow through the aortic valve as if LVAD is placed in the ventricle, blood flow in the aortic arch and cerebral vessels is fully developed with a striking absence of stasis in the aortic arch. On the basis of our simulations, it is beneficial for future VAD designs, to use this concept of flow originating close to the aortic valve through a large orifice.

The flow into the innominate and carotids depend on the direction and the width of the jet from the graft. A 14 mm graft of Heartmate II, while sharing the same flow characteristics as HVAD, has a more diffused jet and is not as sensitive to graft attack angle. Flow in the innominate is significantly reduced at certain angles with a 10 mm graft, leading to the hypothesis that hypoperfusion may be one of the important contributors to the difference in the stroke rates observed.

One of the surprising findings in our study was the impact of aortic incompetence on cerebral blood flow prompting the possibility of it being a risk factor. Depending on the graft angle, aortic incompetency may result in a disproportionate drop in cerebral blood flow. This finding may have relevance in regular CPB as well, as aortic incompetence happens quite often under bypass conditions if the aorta is not clamped.

Progressive drop in cognitive function has been observed in LVADS in almost one in four patients but has been attributed to old age.18 As per our study, cerebral hypoperfusion may well be an important cause.

The “normal” circle of Willis shows a lot of variations and is complete in less than 30% of individuals19 and an absence of the posterior communicating artery observed commonly by middle age.20 An absent posterior communicating artery is an important contributor to stroke if the innominate is occluded.21 Our finding of a drop in blood flow in the innominate artery assumes significance in the light of these findings.

Several publications have highlighted the preponderance of strokes involving the right cerebral hemisphere, both in LVADS and in CPB.22 Our findings suggest that hypoperfusion and stasis in the innominate could equally be responsible apart from other factors.

Interestingly, strokes during CPB have received a lot of attention using CFDs studies.23 Kaufmann et al.24 have shown that jet effect through a narrow aortic cannula, 24 or 26 F,(8–8.5 mm) has an impact on the flow through neck vessels. Although we agree with the views expressed by Kaufmann et al.25 that cerebral autoregulation plays an important role in rotary blood pump and LVADs, it should be borne in mind that about 20% of the patients even in CPB have impaired autoregulation.26

The site, angle of insertion, and the degree of beveling of an outflow graft are of course highly variable and are never measured precisely during surgery. Also, the findings of this study are valid for this particular arch anatomy which is very variable as is the branching pattern of the neck vessels. Nevertheless and precisely for this reason, this study serves to highlight the importance of these factors in flow into cerebral vessels when arterial return is through a narrow graft. Also the choice of 30° and 45° was merely to highlight the importance of attack angle.

We are aware of the shortcomings of this study.

  1. This is a theoretical simulation and needs to be validated. We are currently looking to validate this hypothesis with particle image velocimetry (PIV) studies. The experiments should be carried out on patient-specific models.
  2. We have not taken cerebral autoregulation into consideration. That is a subject of an ongoing study.
  3. Because these findings are valid for this particular arch anatomy, a patient-specific simulation may be the answer.
Back to Top | Article Outline

Conclusion

Arterial return through a narrow graft has important jet effects and results in significant flow perturbations in the aortic arch and cerebral vessels and stasis. A 10 mm graft is more sensitive to the angle of insertion than a 14 mm graft. Under some conditions, serious hypoperfusion of the innominate artery is possible. Aortic incompetence results in a significant decrease of cerebral blood flow. In some vulnerable situations, anatomic or physiological, these factors can result in a stroke. This study offers some insights into the possible reasons for the observed differences in stroke rate between Heartmate II and HVAD.

Back to Top | Article Outline

References

1. Pagani FD, Milano CA, Tatooles AJ, et al. HeartWare HVAD for the treatment of patients with advanced heart failure ineligible for cardiac transplantation: Results of the ENDURANCE destination therapy trial. J Hear Lung Transplant. 2015.34: S9.
2. Stulak JM, Davis ME, Haglund N, et al. Adverse events in contemporary continuous-flow left ventricular assist devices: A multi-institutional comparison shows significant differences. J Thorac Cardiovasc Surg. 2016.151: 177–189.
3. Lalonde SD, Alba AC, Rigobon A, et al. Clinical differences between continuous flow ventricular assist devices: A comparison between HeartMate II and HeartWare HVAD. J Card Surg. 2013.28: 604–610.
4. Theory and Modeling Guide Volume III : ADINA CFD & FSI. 2012.III.
5. Reymond P, Merenda F, Perren F, Rüfenacht D, Stergiopulos N. Validation of a one-dimensional model of the systemic arterial tree. Am J Physiol Heart Circ Physiol. 2009.297: H208–H222.
6. Osorio AF, Osorio R, Ceballos A, et al. Computational fluid dynamics analysis of surgical adjustment of left ventricular assist device implantation to minimise stroke risk. Comput Methods Biomech Biomed Engin. 2013.16: 622–638.
7. Weber TF, Müller T, Biesdorf A, et al. True four-dimensional analysis of thoracic aortic displacement and distension using model-based segmentation of computed tomography angiography. Int J Cardiovasc Imaging. 2014.30: 185–194.
8. Raghavan ML, Ma B, Fillinger MF. Non-invasive determination of zero-pressure geometry of arterial aneurysms. Ann Biomed Eng. 2006.34: 1414–1419.
9. Farag MB, Karmonik C, Rengier F, et al. Review of recent results using computational fluid dynamics simulations in patients receiving mechanical assist devices for end-stage heart failure. Methodist Debakey Cardiovasc J. 2014.10: 185–189.
10. Karmonik C, Bismuth J, Shah DJ, Davies MG, Purdy D, Lumsden AB. Computational study of haemodynamic effects of entry- and exit-tear coverage in a DeBakey type III aortic dissection: technical report. Eur J Vasc Endovasc Surg. 2011.42: 172–177.
11. Karmonik C, Anderson J, Partovi S, et al. Quantitative description of hemodynamics alterations in the aorta after LVAD implantation: A computational fluid dynamics study. Radiat Oncol. 2007.69: 8–9.
12. Kim HJ, Vignon-Clementel IE, Figueroa CA, et al. On coupling a lumped parameter heart model and a three-dimensional finite element aorta model. Ann Biomed Eng. 2009.37: 2153–2169.
13. Kar B, Delgado RM 3rd, Frazier OH, et al. The effect of LVAD aortic outflow-graft placement on hemodynamics and flow: Implantation technique and computer flow modeling. Tex Heart Inst J. 2005.32: 294–298.
14. Travis AR, Giridharan GA, Pantalos GM, et al. Vascular pulsatility in patients with a pulsatile- or continuous-flow ventricular assist device. J Thorac Cardiovasc Surg. 2007.133: 517–524.
15. Liu X, Pu F, Fan Y, Deng X, Li D, Li S. A numerical study on the flow of blood and the transport of LDL in the human aorta: The physiological significance of the helical flow in the aortic arch. Am J Physiol Heart Circ Physiol. 2009.297: H163–H170.
16. Kaufmann TA, Hormes M, Laumen M, et al. Flow distribution during cardiopulmonary bypass in dependency on the outflow cannula positioning. Artif Organs. 2009.33: 988–992.
17. Inci G, Sorgüven E. Effect of LVAD outlet graft anastomosis angle on the aortic valve, wall, and flow. ASAIO J. 2012.58: 373–381.
18. Fendler TJ, Spertus JA, Gosch KL, et al. Incidence and predictors of cognitive decline in patients with left ventricular assist devices. Circ Cardiovasc Qual Outcomes. 2015.8: 285–291.
19. Saikia B, Handique A, Phukan P, Lynser D, Jamil M. Study of anomalies in the circle of Willis using magnetic resonance angiography in north eastern India. J Anat Soc India. 2014.63: 67–73.
20. Suemoto C, Grinberg L, Aparecida MSS, Jacob Filho W, Pasqualucci C. Anatomical variations of circle of Willis in an autopsy study in the city of São Paulo. Braz J Morphol Sci. 2008.25: 157–214.
21. Hendrikse J, Hartkamp MJ, Hillen B, Mali WP, van der Grond J. Collateral ability of the circle of Willis in patients with unilateral internal carotid artery occlusion: Border zone infarcts and clinical symptoms. Stroke. 2001.32: 2768–2773.
22. Kaufmann TA, Schmitz-Rode T, Steinseifer U. Implementation of cerebral autoregulation into computational fluid dynamics studies of cardiopulmonary bypass. Artif Organs. 2012.36: 754–758.
23. Kato TS, Ota T, Schulze PC, et al. Asymmetric pattern of cerebrovascular lesions in patients after left ventricular assist device implantation. Stroke. 2012.43: 872–874.
24. Kaufmann TA, Neidlin M, Büsen M, Sonntag SJ, Steinseifer U. Implementation of intrinsic lumped parameter modeling into computational fluid dynamics studies of cardiopulmonary bypass. J Biomech. 2014.47: 729–735.
25. Kaufmann T, Schmitz-Rode T, Moritz A, Steinseifer U. Effect of outflow cannula placement and pulsatility of blood pumps on cerebral blood flow and wall shear stress during cardiac assist. Blucher Mech Eng Proc. 2012.1: 822–835.
26. Ono M, Joshi B, Brady K, et al. Risks for impaired cerebral autoregulation during cardiopulmonary bypass and postoperative stroke. Br J Anaesth. 2012.109: 391–398.
Keywords:

left ventricular assist device; stroke; mechanical circulatory support

Copyright © 2017 by the American Society for Artificial Internal Organs